WEBVTT
00:00:06.240 --> 00:00:07.200
Next curve.
00:00:10.439 --> 00:00:11.205
Hey everyone.
00:00:11.205 --> 00:00:15.794
Welcome to this episode of the Nextcar Rethink Podcast.
00:00:15.794 --> 00:00:22.335
And of course, this is our special Silicon Future Series, which I do in partnership.
00:00:22.335 --> 00:00:30.074
In collaboration, wonderful collaboration with the illustrious Carl Fre of.
00:00:30.675 --> 00:00:34.755
Cambrian AI and soon to be quantum research.
00:00:34.755 --> 00:00:34.765
Right.
00:00:34.765 --> 00:00:35.630
It became even quantum,
00:00:35.865 --> 00:00:36.255
I think.
00:00:36.344 --> 00:00:37.395
Ah, yeah.
00:00:37.395 --> 00:00:37.454
The
00:00:37.454 --> 00:00:38.174
next big thing.
00:00:38.564 --> 00:00:38.984
Yeah.
00:00:38.984 --> 00:00:40.155
And then of course.
00:00:41.490 --> 00:00:43.820
The, quantum dotie.
00:00:46.340 --> 00:00:46.880
Are you ever gonna
00:00:46.880 --> 00:00:47.984
run Jim McGregor for
00:00:47.984 --> 00:00:48.064
me?
00:00:50.445 --> 00:00:54.875
Jim McGregor of serious research Clan mcg.
00:00:55.850 --> 00:00:56.179
Yeah.
00:00:56.390 --> 00:00:56.929
Okay.
00:00:57.020 --> 00:00:57.829
Clan McGregor.
00:00:58.310 --> 00:00:58.924
Yeah.
00:00:58.924 --> 00:01:03.119
Who has, some amazing, amazing kids?
00:01:03.490 --> 00:01:05.950
who are just remarkable programmers.
00:01:05.950 --> 00:01:13.620
I I saw Damien, by the way, at, Katie Cadence live, 2026 and, yeah, you have really cool kids, man.
00:01:14.204 --> 00:01:14.594
Thank you.
00:01:14.814 --> 00:01:17.965
anyway, welcome everyone.
00:01:19.254 --> 00:01:28.665
And of course we cover, all the exciting stuff that you need to know in the world, the semiconductors, ai, and apparently now quantum.
00:01:28.665 --> 00:01:29.144
Quantum.
00:01:29.665 --> 00:01:42.515
Jim and Carl have been going deep into all this quantum stuff that everyone is starting to get pretty excited about, and, they're gonna basically break down the landscape.
00:01:43.260 --> 00:01:45.540
and so this is gonna be a great education for me.
00:01:45.599 --> 00:01:49.900
and, these two gentlemen are going to basically lay it out.
00:01:49.900 --> 00:01:52.000
This is what the landscape looks right now.
00:01:52.000 --> 00:02:03.219
And this looks, this is how you need to structure your thinking around what the market will look like and how the technol, the competing technologies actually, right.
00:02:03.219 --> 00:02:08.090
There are many different approaches, how those will, likely, and may be.
00:02:08.525 --> 00:02:10.444
Play out in the future.
00:02:10.444 --> 00:02:12.724
And so it is a really interesting area.
00:02:12.754 --> 00:02:14.525
It's formative, I believe.
00:02:14.594 --> 00:02:23.335
I don't know if you both agree, but I'm really excited, to have this, presentation, by the both of you on this, emerging.
00:02:23.784 --> 00:02:29.875
Well, do you wanna call it emerging topic or is this just something that is hot but people need
00:02:30.504 --> 00:02:33.050
to It's pretty one, it's one of the most important topics for the next decade.
00:02:34.060 --> 00:02:34.930
Yeah, absolutely.
00:02:35.080 --> 00:02:35.770
Totally agree.
00:02:36.039 --> 00:02:36.639
Totally agree.
00:02:38.229 --> 00:02:40.180
And then we get into neuromorphic computing.
00:02:40.180 --> 00:02:41.439
So we can do even more there.
00:02:41.439 --> 00:02:41.500
Yeah.
00:02:42.099 --> 00:02:42.370
More.
00:02:42.375 --> 00:02:42.645
Okay.
00:02:42.789 --> 00:02:43.060
Yeah.
00:02:43.060 --> 00:02:44.680
We'll have another, talk on that.
00:02:44.680 --> 00:02:44.949
Right?
00:02:45.009 --> 00:02:47.800
we're gonna make his head explode, Carl, if we keep doing this.
00:02:47.900 --> 00:02:48.560
But you know what?
00:02:48.560 --> 00:02:52.810
For those of you who are listening on audio, you really want to go to YouTube and watch this episode.
00:02:52.810 --> 00:03:00.650
'cause we're gonna have some graphics and, if you follow us on LinkedIn, when we post, on LinkedIn or, other, social.
00:03:01.055 --> 00:03:04.729
Outlets will, provide the presentation material.
00:03:05.030 --> 00:03:18.139
But, for those of you who are just listening in, you really want to, check out the YouTube, episode in the video so that you can see, the charts that these two gentlemen are prepared and they're quite pretty.
00:03:18.830 --> 00:03:20.479
You guys did a very good job.
00:03:20.479 --> 00:03:20.719
Pretty job.
00:03:20.719 --> 00:03:20.990
Yeah.
00:03:20.990 --> 00:03:21.344
Well, thank you.
00:03:21.365 --> 00:03:22.250
You are pretty too.
00:03:23.840 --> 00:03:26.509
You guys are very gorgeous in that regard.
00:03:26.750 --> 00:03:27.020
So,
00:03:27.889 --> 00:03:31.590
well, when you have a face for radio, you have to make your slides look good.
00:03:36.030 --> 00:03:36.360
That's a good one.
00:03:37.240 --> 00:03:37.569
Right.
00:03:37.599 --> 00:03:38.050
Okay.
00:03:38.050 --> 00:03:45.759
So on that note, before we get started, please remember to like, share and react and comment on this episode and also subscribe here on YouTube and Buzzsprout.
00:03:46.330 --> 00:03:48.849
Listen to us on your favorite podcast platform.
00:03:48.879 --> 00:03:52.569
Opinions and statements by my guests are their own and don't reflect mine.
00:03:52.960 --> 00:03:54.310
Are those the next curve?
00:03:54.340 --> 00:03:58.205
Although I think I'm, we're gonna be in very good alignment since you guys are really.
00:03:58.960 --> 00:04:01.870
Going to be giving a masterclass here.
00:04:01.870 --> 00:04:06.129
So I'm all ears and I think the audience will be as well.
00:04:06.400 --> 00:04:15.219
And, opinions and statements, like I said, are their own, but we're doing this for informational purposes and in this case it's educational, it's gonna be great.
00:04:15.250 --> 00:04:22.709
we do this to provide an open forum for, discussion and debate on all things, ai.
00:04:23.449 --> 00:04:27.319
Semiconductor industry and now Quantum.
00:04:27.319 --> 00:04:32.069
And so with that, gentlemen, Where should we start this conversation about Quantum?
00:04:32.230 --> 00:04:34.240
let me start, start with my, if that's okay.
00:04:34.240 --> 00:04:37.180
They're a little higher level and Jim gets into more detail.
00:04:37.839 --> 00:04:41.884
It's actually been a big week this week, or last week, I guess since we posted next week.
00:04:42.345 --> 00:04:46.824
because, this week was, quantum World, world Quantum Computing Day.
00:04:47.500 --> 00:04:48.819
On, Tuesday, right, Jim?
00:04:49.089 --> 00:04:49.300
Yeah.
00:04:49.300 --> 00:04:50.350
Supported by Nvidia.
00:04:50.350 --> 00:04:52.269
It was NVIDIA's Quantum computing
00:04:52.269 --> 00:04:57.740
Nvidia had their second quantum day, not associated with GTC this time.
00:04:57.740 --> 00:05:05.970
It was a standalone event with Jensen Wong speaking, and some of the leaders in Quantum, that are, at least those that are friendly with Nvidia, they were there.
00:05:05.970 --> 00:05:09.689
Those that are not as friendly weren't there, but, it was really great.
00:05:09.730 --> 00:05:11.889
Nvidia made some important announcements.
00:05:12.279 --> 00:05:12.430
Yeah.
00:05:12.509 --> 00:05:21.540
With their new, quantum model for basically improving the error detection and error mitigation correction.
00:05:21.814 --> 00:05:24.295
The problem when an error happens, where did it happen?
00:05:24.894 --> 00:05:27.795
And that's where GPUs can help quantum a lot.
00:05:28.214 --> 00:05:33.615
So there's a really emerging view that GPUs and Quantum are not gonna be competing technologies.
00:05:33.944 --> 00:05:41.305
They're really gonna be, collaborative, technologies that help bring quantum to the useful quantum computing era.
00:05:41.605 --> 00:05:54.375
So what I've done here is just created a slide,'cause this gets very confusing for all of us, of what is going on in this marketplace and how can you have this many billions of dollars pouring into a market that's currently producing.
00:05:55.115 --> 00:05:56.345
Call it zero revenue.
00:05:56.975 --> 00:06:03.545
but, as Jensen Wong says, the most interesting markets of this, the markets of$0 that will grow into billions of dollars.
00:06:04.024 --> 00:06:09.279
And many of us are firmly of the belief that is exactly what will happen with quantum computing.
00:06:09.790 --> 00:06:11.620
It's not gonna replace ai.
00:06:11.620 --> 00:06:15.009
It will facilitate ai, and AI will facilitate quantum.
00:06:15.009 --> 00:06:18.360
It's a reinforcing, cycle here, it's virtuous cycle.
00:06:18.629 --> 00:06:20.579
So on the left you have the modalities.
00:06:20.639 --> 00:06:26.699
By modality, you have basically mean what kind of physics are these guys using to create these quantum states?
00:06:27.120 --> 00:06:29.939
And the next, column is basically in plain English.
00:06:30.000 --> 00:06:30.689
How's it work?
00:06:30.790 --> 00:06:31.495
what's this stuff all about?
00:06:32.579 --> 00:06:34.139
Next column is key companies.
00:06:34.500 --> 00:06:36.389
And then finally, the maturity level.
00:06:36.689 --> 00:06:40.560
This is saying how close were they to producing quantum advantage.
00:06:40.560 --> 00:06:54.370
By quantum advantage, we mean that you could solve meaningful problems in a cost-effective way that you cannot solve, in, digital computing technologies regardless of how big your supercomputer is.
00:06:55.225 --> 00:06:56.064
so let's start this.
00:06:56.095 --> 00:06:57.235
I'll try to do this quickly.
00:06:57.235 --> 00:06:58.915
You have superconducting qubits.
00:06:58.915 --> 00:07:00.625
This is the one that's been around the longest.
00:07:00.625 --> 00:07:06.725
This is where IBM and Google ti, as well as IQM and OQC.
00:07:07.074 --> 00:07:12.555
this is where you see those huge chandeliers of gold that are used as.
00:07:12.699 --> 00:07:20.139
Final stage of cooling that creates a, sta a a temperature of millikelvin.
00:07:20.139 --> 00:07:22.490
So this is colder than outer space.
00:07:22.850 --> 00:07:29.519
And that is where the actual supercomputing chip, superconducting chip, excuse me, can operate.
00:07:29.819 --> 00:07:29.910
Mm-hmm.
00:07:29.910 --> 00:07:36.120
And so you have these back rooms full of these huge refrigerators that eventually chill down through these chandeliers.
00:07:36.704 --> 00:07:39.014
To the bottom where this chip is gonna operate.
00:07:39.345 --> 00:07:42.644
The advantage of this technology is it's really fast.
00:07:43.045 --> 00:07:44.334
the disadvantages.
00:07:45.149 --> 00:07:50.339
It has to operate ultra low temperatures, which is very expensive in space, consumes a lot of space.
00:07:50.920 --> 00:07:53.389
but, it's also, unfortunately very noisy.
00:07:53.779 --> 00:07:59.990
so if you think about a normal semiconductor chip, it's okay that you're, you know, 10 10 to the minus seven.
00:08:01.060 --> 00:08:08.949
But here we're talking about 10 to minus two or three, so that means one, one in a hundred or one in a thousand results are bad.
00:08:09.480 --> 00:08:12.689
And considering you're producing hundreds of millions of results, that's not good.
00:08:12.990 --> 00:08:17.910
So you have to identify these errors and you have to correct mitigate these errors, which means reduce their impact.
00:08:18.209 --> 00:08:20.939
But you also eventually have to correct these errors.
00:08:21.600 --> 00:08:27.660
This is often done by teaming together physical cubits to create a logical cubit.
00:08:27.959 --> 00:08:33.970
The ratio of the, of which is super important to both speed and economics.
00:08:34.070 --> 00:08:34.820
if you can do.
00:08:35.309 --> 00:08:38.730
Two or three physical cubit to produce one logical cubit.
00:08:39.179 --> 00:08:42.330
You're gonna be three times more powerful than one.
00:08:42.335 --> 00:08:49.475
an architect that maybe takes 10 cu logical qubits or a hundred logical cubits to produce one, excuse me, physical cubits.
00:08:49.970 --> 00:08:52.159
To produce one logical cubit.
00:08:52.190 --> 00:08:54.559
Logical cubits are what matter.
00:08:54.990 --> 00:09:04.740
that's the, if you can try to distill it down and across all these modalities, logical cubits and their performance and costs are really what matter matters here.
00:09:05.129 --> 00:09:06.480
It's what you're using.
00:09:06.480 --> 00:09:11.909
It's really what you're using this form, the gates, the circuits, the those, the circuits, random circuits that you're actually developing.
00:09:12.059 --> 00:09:16.169
And by the way, if you've never seen pictures of the Poughkeepsie Supercomputing.
00:09:16.659 --> 00:09:21.789
center in New York that IBM has, it kinda looks like an invasion of the body snatchers.
00:09:22.840 --> 00:09:24.610
But all these hanging refrigerators.
00:09:24.610 --> 00:09:26.870
All these hanging refrigerators, they're huge.
00:09:27.620 --> 00:09:29.245
They're huge as well.
00:09:29.245 --> 00:09:33.715
As well as at the TJ Watson Research Center where they have a major demo center now.
00:09:33.924 --> 00:09:34.014
Mm-hmm.
00:09:34.254 --> 00:09:38.034
A demon demonstrating the, IBM Quantum system two.
00:09:38.754 --> 00:09:42.174
now if you go down the list, trap s kind of nest, these are kind of.
00:09:43.565 --> 00:09:44.914
They're not random orders here.
00:09:44.914 --> 00:09:52.620
These are in decreasing state of maturity, potentially increasing potential to be disruptive.
00:09:52.620 --> 00:10:04.830
Tract ions are much more, high, much higher fidelity, which means that they're not gonna be as error prone, but it's hard to scale that's in progress and it's hard to get the performance.
00:10:04.830 --> 00:10:06.330
They don't have the same performance.
00:10:06.720 --> 00:10:13.750
As a supercomputing super, conducting circuit does, they, these ions, they're exotic ions.
00:10:13.750 --> 00:10:15.009
I can't even remember their names.
00:10:15.009 --> 00:10:17.350
They're not normal elements you could think of.
00:10:17.750 --> 00:10:27.684
they're trapped in control with lasers and walking around the computer show floor at, let's say TTC last month, you could actually see models of what these look like.
00:10:28.044 --> 00:10:29.184
Very high fidelity.
00:10:29.884 --> 00:10:31.445
There's gotta work on performance.
00:10:31.445 --> 00:10:43.524
And you can see the companies that are listed there, neutral Atoms, kind of a dark horse here, is late, late in development, but is showing tremendous potential.
00:10:44.095 --> 00:10:44.605
Tremendous.
00:10:44.980 --> 00:10:47.769
Promise because they operate at room temperature.
00:10:47.769 --> 00:10:54.360
You don't need these massive refrigeration systems to get them down to sub, sub one degree Kelvin.
00:10:54.929 --> 00:10:57.899
so that means they're a lot more portable, they're more affordable.
00:10:58.210 --> 00:10:59.259
They're more flexible.
00:10:59.590 --> 00:11:00.399
and they're fast.
00:11:00.960 --> 00:11:10.450
They're not as fast as superconducting qubits or even as trapped ion cu qubits, but they can be as fast when you combine them, because of the, cause.
00:11:10.450 --> 00:11:18.450
You combine several physical, neutral atom qubits to create a, logical, qubit and get pretty much the same performance.
00:11:18.975 --> 00:11:31.745
bot is, using light particles and you can see some of the companies there that have been quite vocal about their ability, or again, kind of like neutral atoms, early stage, early days, but very promising technology.
00:11:32.899 --> 00:11:35.570
The next one is spin qubits or quantum dots.
00:11:35.570 --> 00:11:42.269
These are electron spin in a semiconductor structure, and you can see the companies they're the most notable are probably Intel.
00:11:42.730 --> 00:11:44.379
again, this is early stage.
00:11:44.379 --> 00:11:49.750
It is not something you're gonna, you, or maybe even your drill from will be using, but it does have a lot of potential.
00:11:50.029 --> 00:11:52.220
Intel actually started on Supercomputing.
00:11:52.610 --> 00:11:52.820
Yeah.
00:11:53.164 --> 00:11:56.139
And they switched and then switched to the, the spin bits.
00:11:56.139 --> 00:11:56.860
So That's right.
00:11:57.490 --> 00:11:58.330
It, it did.
00:11:58.330 --> 00:12:02.289
So a lot of the, you have to realize that this has all been in development stage.
00:12:02.289 --> 00:12:09.679
So some of these companies have actually evaluated or worked on multiple different, multiple of these different, qubit solutions or modality.
00:12:09.679 --> 00:12:09.740
Yeah.
00:12:10.580 --> 00:12:11.450
Quantium is like that.
00:12:11.450 --> 00:12:13.500
Quantium used to be known as, Honeywell.
00:12:13.779 --> 00:12:14.429
They used to be part of
00:12:14.429 --> 00:12:19.715
They used to be part of Honeywell and they spun'em out and they really decided to focus at that time on trapped ions.
00:12:19.715 --> 00:12:21.965
But before they were more of an application layer.
00:12:22.294 --> 00:12:28.184
And, really trying to develop, independent of the modality, they decided to just focus on trapped ions.
00:12:28.184 --> 00:12:29.504
They think that's the sweet spot.
00:12:29.715 --> 00:12:36.034
So, Carl, some of the audience might be wondering, as you go down this list, at what point are you.
00:12:36.230 --> 00:12:40.399
Departing from superconducting and super cool, like this, near mm-hmm.
00:12:40.889 --> 00:12:42.960
zero Kelvin cooling.
00:12:43.240 --> 00:12:45.100
so what does that look like here?
00:12:45.100 --> 00:12:48.309
Because you mentioned neutral atom is room temperature, right?
00:12:48.309 --> 00:12:48.909
And so, yeah.
00:12:49.370 --> 00:12:57.700
what are we looking at in terms of the environment that these different modalities, outside of the supercomputing, operate in?
00:12:58.610 --> 00:13:04.570
Well, it's actually, supercomputing and the quantum and kneeling that use the, ultra low temperatures.
00:13:04.620 --> 00:13:12.730
everything else is either a higher temperature or, at room temperature, especially when you start talking about using photons or ions and stuff like that.
00:13:12.820 --> 00:13:13.360
Right, right.
00:13:13.419 --> 00:13:13.750
Okay.
00:13:13.990 --> 00:13:14.169
Yeah.
00:13:14.169 --> 00:13:14.440
Cool.
00:13:14.590 --> 00:13:15.129
Exactly.
00:13:15.565 --> 00:13:15.715
All,
00:13:16.355 --> 00:13:18.125
quantum and kneeling is kind of a niche.
00:13:18.215 --> 00:13:22.625
It has commercial users today, and there's really two players that are Dwa and Fujitsu.
00:13:23.554 --> 00:13:29.674
and then there's topological qubits, which created a tremendous roar in the market when mi, when Microsoft disclosed it.
00:13:29.975 --> 00:13:31.355
They're the only ones doing this.
00:13:31.654 --> 00:13:42.044
If they can make it work, which is a big, if, if they can make it work, it could be the winner, but it's very far from production, probably the least developed research platform.
00:13:42.539 --> 00:13:47.399
Yeah, but it was pretty amazing at the time, how much, how much buzz it created.
00:13:47.610 --> 00:13:50.669
But, earlier you mentioned, you said super computing.
00:13:50.669 --> 00:13:51.750
You meant superconducting,
00:13:51.750 --> 00:13:53.490
I meant superconducting conducting.
00:13:53.490 --> 00:13:54.809
Just wanted to clarify that.
00:13:55.210 --> 00:13:55.539
Thank you, Leonard.
00:13:55.809 --> 00:13:56.110
Yeah.
00:13:57.174 --> 00:13:58.644
Some pros and cons here.
00:13:59.075 --> 00:14:01.355
I've probably covered quite a bit of this already.
00:14:01.360 --> 00:14:01.519
Yeah,
00:14:01.524 --> 00:14:03.565
Pros, superconducting, fast gates.
00:14:03.625 --> 00:14:07.164
Very strong ecosystem, by the way, because they've been around a long time.
00:14:07.404 --> 00:14:15.125
And IBM's willing to spend a lot of money to, to, facilitate the collaboration of research institutions around the world.
00:14:15.644 --> 00:14:16.514
it's the near term.
00:14:16.514 --> 00:14:16.634
Later.
00:14:17.220 --> 00:14:17.549
Right.
00:14:17.600 --> 00:14:20.365
big, big, strong, force in the industry.
00:14:20.815 --> 00:14:26.965
Trapped ions, they're really the contender right now have fault tolerance systems, neutral atoms.
00:14:27.014 --> 00:14:28.154
They're the dark horse.
00:14:28.184 --> 00:14:34.705
But there is one company, for example, which is inflection, did a SPAC in February.
00:14:35.049 --> 00:14:40.960
And so they're now publicly traded and, they're seeing a lot of interest from investors in neutral atoms.
00:14:41.500 --> 00:14:46.899
and then they also have an existing business, which is called quantum sensing.
00:14:46.899 --> 00:14:50.259
We're not talking quantum sensing here, we're talking quantum computing.
00:14:50.840 --> 00:14:55.580
so they can use that to fund and finance the development of quantum computing.
00:14:56.330 --> 00:15:00.559
FO photonic, very strong quantum networks and distributed quantum computing.
00:15:01.100 --> 00:15:05.779
quantum dots promising long-term pathway for building these on silicon.
00:15:05.779 --> 00:15:10.750
On CM OS compatible, chips, topological, potentially a breakthrough.
00:15:10.950 --> 00:15:12.750
very high risk, very high reward.
00:15:13.110 --> 00:15:14.309
There's quantum and kneeling.
00:15:14.370 --> 00:15:15.779
It's commercially viable today.
00:15:16.105 --> 00:15:18.995
DD Wade's doing quite well, in their niche.
00:15:19.044 --> 00:15:19.884
but it is a niche.
00:15:19.884 --> 00:15:20.965
It's not gonna be a general purpose.
00:15:20.965 --> 00:15:21.414
Quantum computing.
00:15:22.929 --> 00:15:26.409
With that, I'll turn it over to Jim and you can, dive into your analytics.
00:15:26.409 --> 00:15:26.649
Jim,
00:15:26.889 --> 00:15:35.340
I took a, another, a similar view to what, Carl did and I have to kind of preface this and say that, all these numbers are relative.
00:15:35.490 --> 00:15:41.799
In other words, things that, they're how I kind of, positioned the different factors of these different technologies.
00:15:41.830 --> 00:15:45.789
And quite honestly, you have to remember that all these technologies are still developing.
00:15:45.789 --> 00:15:48.100
So they are changing and they will change.
00:15:48.100 --> 00:15:54.250
So even looking at the spider radar graph that you're looking at, all these different technologies will change over time.
00:15:54.700 --> 00:15:55.750
And if you.
00:15:56.269 --> 00:15:57.860
Click down one slide here.
00:15:58.080 --> 00:15:58.799
we'll see.
00:15:58.980 --> 00:16:04.440
When you start looking at the numbers and you start looking at how they're positioned, supercomputing provides that best balance.
00:16:04.769 --> 00:16:10.490
Trapped on trapped ion provides that, high performance solution or the best performing solution.
00:16:10.940 --> 00:16:17.139
The photonic and quantum dots, which are very similar, have the best manufacturability capabilities today.
00:16:17.450 --> 00:16:21.289
although the topological, that was the one that, Microsoft is working on is.
00:16:21.580 --> 00:16:26.350
Theoretically the best from an error correction standpoint or error mitigation standpoint.
00:16:26.350 --> 00:16:26.679
Mitigation.
00:16:26.740 --> 00:16:28.570
So it will be very interesting.
00:16:28.659 --> 00:16:32.679
And then once again that a kneeling is really a specialized niche.
00:16:32.769 --> 00:16:36.909
I don't ever see that necessarily being a general market capability.
00:16:37.330 --> 00:16:45.600
and, one more slide., I outlined some of those, the same thing that, Carl did, but you have to remember one thing you have to really remember, and let's go to the next slide.
00:16:46.120 --> 00:16:46.509
Leon.
00:16:47.710 --> 00:16:48.190
Leonard
00:16:49.360 --> 00:16:52.750
d Neil is, is that, is that
00:16:52.750 --> 00:16:53.649
my basic Leon?
00:16:53.649 --> 00:16:55.809
But I, you know, I, I, is that my Scott team?
00:16:56.620 --> 00:16:57.610
Yeah, exactly.
00:16:59.110 --> 00:17:08.009
I, no, you have to remember where we are and the fact that, a lot of people said, well, think about where AI was in 2015.
00:17:08.190 --> 00:17:09.599
Before, everything exploded.
00:17:10.019 --> 00:17:12.539
We're kind of there in what we call the Nisc era.
00:17:12.539 --> 00:17:17.690
It's kind of noisy, and unstable qubits where we're really still learning about it.
00:17:17.930 --> 00:17:23.180
But you have to understand that the industry has actually been working on this since the mid nineties.
00:17:23.390 --> 00:17:25.250
I mean, this is a technology that's taken.
00:17:25.549 --> 00:17:29.210
So when, when everyone says, oh, it's gonna be decades away, it's been decades.
00:17:29.240 --> 00:17:31.220
this has been in development for a long time.
00:17:31.220 --> 00:17:33.200
we went through a long period of just.
00:17:34.019 --> 00:17:36.420
Understanding is this even feasible?
00:17:36.420 --> 00:17:39.960
And then kind of proof of concept era through the early two thousands.
00:17:39.960 --> 00:17:43.720
And now we're in an era where we understand it a little bit better.
00:17:43.880 --> 00:17:47.569
we're having to do a lot of error mitigation and suppression just.
00:17:47.569 --> 00:17:51.255
To be able to get rid of some of that noise that's in these different types of qubits.
00:17:51.255 --> 00:17:55.825
And obviously there are many different ways to create a qubit, and read a qubit.
00:17:56.174 --> 00:17:59.625
but we've created quantum utility, and this is probably the most important thing.
00:18:00.134 --> 00:18:06.194
Why are there billions and billions of dollars going in this, not just in terms of the technology, but also in terms of the use of it?
00:18:06.464 --> 00:18:10.065
I mean, you have banks, you have materials companies, you have.
00:18:10.914 --> 00:18:18.265
Scientific research organizations all pouring billions into this because they see this as a very significant leap in terms of computing.
00:18:18.295 --> 00:18:20.035
This is an accelerator technology.
00:18:20.244 --> 00:18:34.075
It's not going to displace classical computing, and this is why AI and standard computing and basically the accelerated computing environment we're in now is gonna be so beneficial to this because basically it's gonna help us on the front end.
00:18:34.454 --> 00:18:47.690
Basically create, the quantum circuits to organize the data and everything else that feeds into this quantum accelerator, and then understand that information and do any error correction as it comes out of that quantum computer.
00:18:48.140 --> 00:18:53.210
So basically this quantum computer or QPU quantum computing unit, it's gonna be another.
00:18:53.589 --> 00:19:02.400
Type of rack that sits in the data center or even maybe remotely that's tied to that standard supercomputer or ai, powerhouse.
00:19:02.730 --> 00:19:07.440
and right now we're finding that, you know, even when you get to like a hundred qubits and that's.
00:19:07.575 --> 00:19:08.714
physical qubits.
00:19:09.105 --> 00:19:18.275
You're getting into an area where it's starting to be interesting, where you can understand, how to use the technology, and even run some, interesting simulations.
00:19:18.305 --> 00:19:29.375
And when you get to a thousand physical qubits, then you're getting into an area where you're creating, a significant number of logical com, qubits, maybe a hundred or 200 logical compute qubits.
00:19:29.744 --> 00:19:45.535
Creating gates, creating, circuits and actually exceeding that capability, as Carl said, creating quantum advantage where you're exceeding that capability of even what a supercomputer or any supercomputer on the planet can do to do today.
00:19:45.894 --> 00:19:52.224
Now you have to remember that there's always a probability, risk here in terms of, you know, how close to an actual.
00:19:52.259 --> 00:20:06.619
Answer it is, but you have to remember that what it's trying to do is beyond what we've been able to do before, analyzing the human genome, modeling the, the earth and the weather patterns, actually discovering new material technologies.
00:20:06.625 --> 00:20:21.005
it's very, very significant and we're already getting to that quantum utility, being able to use this quantum technology and going forward, we're really looking at this here that we're calling fault tolerant era, where basically we understand the qubits a little bit more.
00:20:21.005 --> 00:20:23.585
We can control the fidelity of the qubits a little bit more.
00:20:23.944 --> 00:20:33.815
And once again, this could be in any of those different modalities, but we get to a point where the accuracy is fairly high and we're getting to a point where we can scale.
00:20:34.005 --> 00:20:35.505
And that's probably the most important thing.
00:20:36.015 --> 00:20:43.420
So when you see these rapid things, we're basically doubling to quadrupling the number of physical qubits that we can do every year.
00:20:44.005 --> 00:20:56.255
depending on the different modalities, that's significant because we're gonna be able to go from, thousands of cubits today, or hundreds or to thousands of qubits today to millions of cubits, shortly after the, into the next decade.
00:20:56.704 --> 00:20:58.345
someplace in between, 20.
00:20:58.589 --> 00:21:00.690
32 to 2035.
00:21:01.180 --> 00:21:25.839
we place it come somewhere around 2034, but that's very significant and once again, as Carl said, once you get into that range and you're only doing anywhere from, it could be a couple physical cubits to a hundred cubits per logical qubit, you're still getting into a point with a million physical qubits to where you can do some very, significant things that you couldn't do anywhere else or any place else.
00:21:26.855 --> 00:21:39.654
And in fact the, all the companies we've mentioned all have roadmaps that explain how they will get to a million, qubits, which is sort of the milestone for quantum advantage.
00:21:39.865 --> 00:21:44.845
Probably some areas that can actually get to quantum advantage, far lower numbers of cubits.
00:21:45.115 --> 00:21:47.365
But in terms of a more general.
00:21:48.055 --> 00:21:56.315
Platform for doing scientific work and, and research in areas that we don't touch today because today's computer simply can't do it.
00:21:56.855 --> 00:22:04.484
So it's gonna open up all kinds of areas of computational, applications that do not exist at all today.
00:22:04.755 --> 00:22:07.125
So it is funny, some people talk about how.
00:22:07.549 --> 00:22:15.390
You know, there's an article about three or four months ago, some guy wrote, I think it was in Wall Journal, that, computing's all gonna come down to the sides of the shoebox.
00:22:15.390 --> 00:22:15.720
Right.
00:22:15.720 --> 00:22:17.519
And it's all gonna be quantum and stuff like this.
00:22:17.519 --> 00:22:20.630
And I'm like, well, I don't know what this guy's smoking, but I'd like some.
00:22:21.134 --> 00:22:21.674
It's crazy.
00:22:21.884 --> 00:22:24.075
This is not gonna replace the traditional.
00:22:24.210 --> 00:22:24.289
Yeah.
00:22:24.585 --> 00:22:25.430
And, and
00:22:26.174 --> 00:22:27.825
Carl's address is no.
00:22:30.339 --> 00:22:31.664
Yeah, and you know what?
00:22:31.664 --> 00:22:51.305
I think that is one of the biggest clarifications that needs to be made at the moment because I think there's, so many misconceptions about what a quantum application is and one of the things that was very clear from Quantum Day and I managed to get.
00:22:51.720 --> 00:23:04.329
You know, listening on some of it, and I'd love to give both of your takes on this, but for the most part, the best way of, thinking about what these applications are the ones that are traditionally, you know, for supercomputing, right?
00:23:04.329 --> 00:23:06.069
Whether it's for, scientific, very much
00:23:06.069 --> 00:23:06.079
so.
00:23:07.089 --> 00:23:07.480
Right.
00:23:07.480 --> 00:23:08.079
Certain,
00:23:08.085 --> 00:23:12.164
Doing that analysis, doing the scientific research, doing those types of applications.
00:23:12.164 --> 00:23:16.875
And quite honestly, I don't know how you feel, Carl, but I think that, it's very much like AI modeling.
00:23:16.904 --> 00:23:22.095
I think we're gonna end up with different types of quantum modalities for different types of applications.
00:23:22.335 --> 00:23:28.394
I think we're gonna find as we get farther down the road that, you know, and some of this is definitely AI enabled.
00:23:28.394 --> 00:23:35.769
On the front end and backend, that we're gonna be able to do, cer, certain modalities are gonna fit certain applications better than others, and.
00:23:36.644 --> 00:23:37.065
Yeah.
00:23:37.795 --> 00:23:46.325
in looking at this chart specifically, that growth chart, it gets to a point, you have to remember that, this is very much like the AI was, over a decade ago.
00:23:46.664 --> 00:23:48.105
we're gonna see changes in the industry.
00:23:48.115 --> 00:23:50.305
So we're probably gonna see a consolidation.
00:23:50.545 --> 00:23:53.714
Some of these companies will get bought for their IP or bought all out.
00:23:54.015 --> 00:24:00.444
We may have, larger companies doing multiple modalities by the time we get to this fault tolerant era.
00:24:00.775 --> 00:24:08.259
So there's going to be a lot of change and that's why everyone really needs to pay attention because, we don't know what the end solution is going to be yet.
00:24:09.174 --> 00:24:09.684
Yeah, that's,
00:24:09.684 --> 00:24:12.355
and are you saying there's gonna be kind of a Cambrian explosion?
00:24:12.805 --> 00:24:13.684
Couldn't resist.
00:24:13.684 --> 00:24:15.480
couldn't resist my bad.
00:24:16.559 --> 00:24:17.430
Lots of explosion,
00:24:17.430 --> 00:24:18.839
self-promotion, my apologies.
00:24:19.319 --> 00:24:21.539
No, I agree with Jim.
00:24:21.539 --> 00:24:24.170
I think we'll see specialization start to occur.
00:24:24.170 --> 00:24:31.829
Right now we're just trying to get to the point where we can demonstrate quantum advantage across a reasonable number of, universal spaces.
00:24:32.339 --> 00:24:40.049
but as that becomes more general purpose, you'll then see it specialized and you'll see specific quantum modalities.
00:24:40.734 --> 00:24:45.234
Being used for certain problems and others being used for other problems.
00:24:45.595 --> 00:24:47.934
So I think there's room for all here to succeed.
00:24:48.444 --> 00:25:02.640
And, quite honestly, there's plenty of money to be made by all of the players, Most of the players that we're seeing today have a significant research and results that will turn into competitive advantage in specific areas.
00:25:02.730 --> 00:25:03.269
Yeah, and
00:25:03.269 --> 00:25:03.569
I,
00:25:03.569 --> 00:25:03.720
and
00:25:03.720 --> 00:25:12.779
there's one other area that, since Leonard, since you brought up the misconceptions, the other misconception we have to kind of address is that quantum is gonna break everything.
00:25:14.289 --> 00:25:14.579
Yeah.
00:25:14.904 --> 00:25:21.265
It is a, it is definitely a threat because,, traditional security algorithms can be broken by a quantum computer in record times.
00:25:21.315 --> 00:25:24.765
However, you have to remember that, IBM especially, that's been list, yeah.
00:25:24.765 --> 00:25:27.015
Kind of leading the charge on these nist.
00:25:27.525 --> 00:25:34.734
Algorithms that basically to actually, you know, create algorithms that quantum computers can't solve for new security algorithms.
00:25:35.035 --> 00:25:42.625
They've been working on this for over a decade and they continue, and these new NIST algorithms keep coming out every year, every couple of years.
00:25:42.954 --> 00:25:48.414
So, quantum computing is not designed to be an encryption, to be a hacker's gold mine.
00:25:48.865 --> 00:25:58.515
It is something that every company today has to be aware of and has to start thinking about, integrating these NIST algorithms because they need to be quantum safe today.
00:25:58.575 --> 00:26:10.444
Because there is the theory that, and I completely believe this, you can go and scrape, information today and possibly use quantum computers tomorrow to actually decipher the security algorithms.
00:26:10.714 --> 00:26:11.724
So that is.
00:26:12.259 --> 00:26:14.180
A part of the quantum discussion.
00:26:14.180 --> 00:26:16.400
It's not the part of the quantum discussion.
00:26:17.069 --> 00:26:33.269
And I, I think that maybe that's a topic that we talk about, separately because, one of the risks going forward is that, especially around this, quantum era security problem, that.
00:26:33.704 --> 00:26:40.045
Some of the thinking it is based on, fundamental misconceptions about quantum.
00:26:41.230 --> 00:26:53.140
And I think, you guys are doing a lot of, good work that will be a basis of clarifying a lot of those misconceptions because I think quantum computing is in general not well understood.
00:26:53.140 --> 00:26:54.759
They think that, for instance.
00:26:55.904 --> 00:27:01.724
Quantum computing will displace and replace accelerated computing, right?
00:27:02.244 --> 00:27:05.025
Traditional computing as you both, cited before.
00:27:05.335 --> 00:27:09.134
One of the things that I think might be, a good clarification.
00:27:09.759 --> 00:27:17.769
Education for the audience might be helping them understand what is the role of accelerated computing vis-a-vis quantum computing.
00:27:17.769 --> 00:27:28.630
Jim, you mentioned, that AI will be important in enabling and, scaling, helping, quantum computing to scale.
00:27:28.630 --> 00:27:34.690
There was a lot of talk at Quantum Day actually last year about how.
00:27:35.220 --> 00:27:45.549
AI and accelerated computing, could, enable or be a foundation for these control planes for these different modalities of, quantum, computing runtime.
00:27:45.549 --> 00:27:49.059
So, how does this all kind of interrelate with each other?
00:27:49.059 --> 00:27:52.345
Because, you do have Nvidia obviously making big investments.
00:27:53.279 --> 00:28:00.769
In the quantum ecosystem, but also positioning their technologies for quantum computing enablement.
00:28:02.059 --> 00:28:07.519
Maybe both of you can take some time to stitch that picture together for the audience.
00:28:08.134 --> 00:28:15.714
Well, you have to remember that we're dealing with quantum mechanics, so we're actually dealing with physics and physical systems here in understanding, quantum computing.
00:28:16.015 --> 00:28:26.244
You have to remember that classical computing has always been a part of that, but also has some manual tasks, like actually trying to calibrate the qubits, actually trying to make sure that they're going.
00:28:26.460 --> 00:28:33.339
Perform the way you hope they're going to perform, because they're always in a state of flux, and that's a huge, huge challenge.
00:28:33.619 --> 00:28:35.750
Nvidia and I credit Nvidia for this.
00:28:35.750 --> 00:28:41.000
Instead of trying to jump into the quantum computing fray, they've said, listen, we're going to enable it.
00:28:41.000 --> 00:28:42.619
We're gonna be the key enabler here.
00:28:42.829 --> 00:28:48.920
So they've created their icing, their Nvidia icing models for to help with that calibration on the front end.
00:28:49.470 --> 00:29:07.619
Basically, hopefully automate that to where you don't have, it's not a manual process, it's an automated process and a reiterative process for each run that you're running, but also on the backend for doing the error correction, which, as we've mentioned before, just getting the right information, ensuring that you have accurate information is a challenge.
00:29:07.619 --> 00:29:12.690
Today we're using a combination of suppression, mitigation, and error correction today.
00:29:12.930 --> 00:29:15.630
We hope we get to, and that error correction really isn't.
00:29:15.849 --> 00:29:17.230
That great at this point in time.
00:29:17.440 --> 00:29:30.500
So we're hoping we're gonna get to a point where, A, we have more steady states of the quantum, of the qubits, and B so we have higher fidelity, and B, we get to a point where we can actually do quantum error correction.
00:29:30.819 --> 00:29:34.390
With a very good understanding and very high speeds to be able to do that.
00:29:34.690 --> 00:29:52.819
So I mean, Nvidia, and this is a perfect application for it, trying, taking these mult multi parallel solutions, basically their GPUs and saying, listen, just like we power AI with this parallelism, we can power these tasks in the front end and back end of quantum computing.
00:29:52.829 --> 00:30:00.150
I credit them because I think it's going to be, what they're doing is not important for one specific modality or one specific partner.
00:30:00.390 --> 00:30:03.839
It's gonna be important for the entire industry in any modality that's out there.
00:30:05.130 --> 00:30:08.369
Yeah, I hats off to Nvidia for really.
00:30:08.950 --> 00:30:13.930
Putting themselves at the center, the conversation of quantum computing without having a quantum computer.
00:30:14.480 --> 00:30:17.059
that, that's pretty magical, accomplishment.
00:30:17.279 --> 00:30:20.130
I do wonder how they're gonna monetize that going forward.
00:30:20.539 --> 00:30:22.190
but they'll sell enough GPUs.
00:30:22.190 --> 00:30:23.000
it will definitely help them.
00:30:23.000 --> 00:30:24.049
But more importantly.
00:30:24.414 --> 00:30:30.654
It puts them at the center of the ecosystem for most quantum computing because everybody has to air correction.
00:30:30.684 --> 00:30:34.285
Everybody has to do calibration, q calibration.
00:30:34.674 --> 00:30:36.505
And so, that's an important first step.
00:30:36.505 --> 00:30:37.914
I don't think it's the last step.
00:30:38.585 --> 00:30:42.605
I'd like both of your reactions to the comment that I'm gonna make right now.
00:30:43.059 --> 00:30:48.660
the way I view it is, Nvidia is doing a really good job of positioning themselves in practical terms, right?
00:30:48.660 --> 00:30:48.750
Mm-hmm.
00:30:48.990 --> 00:30:56.329
Not just fluffy marketing terms as the picks and shovels for, quantum computing.
00:30:56.329 --> 00:31:01.519
So this hockey stick that you have here, Jim, there, the.
00:31:02.630 --> 00:31:18.174
The concept is that they will be able to, provide that, control plane, and enable these, execution environments, the design of these execution environments so that quantum computing can scale.
00:31:18.859 --> 00:31:32.529
The other thing that I also noticed, from conversations and presentations coming out of Quantum Day is that, accelerated computing is great for, discovering applications, right?
00:31:32.529 --> 00:31:34.480
Because some of these applications are not known.
00:31:35.329 --> 00:31:38.809
And it's not that you're going to use quantum computing to figure those out.
00:31:38.809 --> 00:31:45.779
You actually need to have, accelerated computing or AI to support that discovery process.
00:31:46.099 --> 00:31:48.980
much like what we see with, drug discovery, right?
00:31:48.980 --> 00:31:55.099
So there's a role there and, your question, Carl, how are the likes of Nvidia?
00:31:55.099 --> 00:31:58.670
Because I assume that they're gonna be more players than just Nvidia in this game.
00:31:59.210 --> 00:32:00.470
How will they monetize?
00:32:00.470 --> 00:32:06.380
Well, they would monetize by supporting these important activities in not only scaling.
00:32:07.025 --> 00:32:16.724
Quantum computing itself from in terms of progressing all these different modalities, but also the discovery of.
00:32:17.275 --> 00:32:22.605
the applications, but then also the design of the systems, right?
00:32:22.605 --> 00:32:23.505
And the circuits.
00:32:23.565 --> 00:32:26.025
That was the other thing that I took away as well.
00:32:26.025 --> 00:32:27.255
I think in general, you're right.
00:32:27.355 --> 00:32:29.694
it was really interesting article.
00:32:29.694 --> 00:32:31.464
I believe it's Wall Street Journal Suite.
00:32:31.974 --> 00:32:36.714
But it was talking about how Jenssen's investment strategy is not to pick winners.
00:32:36.785 --> 00:32:37.714
Don't pick winners.
00:32:37.755 --> 00:32:42.789
Invest in everybody who has a shot and eventually you'll hit one or two that really.
00:32:43.244 --> 00:32:44.714
Give you a good return on investment.
00:32:45.305 --> 00:32:47.345
and I think that's what he's doing right now with Quantum.
00:32:47.375 --> 00:32:50.315
He's saying, look, I don't know which one's gonna be best.
00:32:50.644 --> 00:32:55.444
In fact, I don't know which ones are gonna be best at which problems to Jim's point earlier.
00:32:55.934 --> 00:33:02.615
I'm gonna invest in all of them by giving them the tools they need to accelerate, fault tolerant computing era.
00:33:02.920 --> 00:33:08.230
Get there sooner, get there with mo with higher quality qubits and lower costs.
00:33:08.599 --> 00:33:10.849
right now cost is not in the equation, right?
00:33:10.849 --> 00:33:10.970
It is.
00:33:10.970 --> 00:33:12.019
Just can you get there?
00:33:12.019 --> 00:33:12.079
Yeah.
00:33:12.440 --> 00:33:17.890
But eventually, if you can get there, well then it'll, we'll determine the winner is well, how much it costs.
00:33:18.490 --> 00:33:22.359
And how manufacturable is it and how high quality is it?
00:33:22.359 --> 00:33:23.440
How dependable is it?
00:33:23.440 --> 00:33:24.609
All that stuff, right?
00:33:25.069 --> 00:33:29.819
all the Illities will come into play, but right now capability, is king.
00:33:29.880 --> 00:33:30.865
And so he's investing in everybody.
00:33:31.230 --> 00:33:31.289
Yeah.
00:33:31.779 --> 00:33:33.940
And expects to be able to make money along the way.
00:33:34.660 --> 00:33:35.815
Carl, You're getting soft.
00:33:36.039 --> 00:33:37.240
I'm out in the open.
00:33:37.240 --> 00:33:38.410
You can take a shot.
00:33:38.410 --> 00:33:38.769
Come on.
00:33:39.710 --> 00:33:44.640
one of the interesting things is that, with any new technology, there usually comes new business models.
00:33:44.670 --> 00:33:47.519
We don't know what those business models look like at this point in time.
00:33:47.910 --> 00:33:50.279
That's going to be, have to evolve over time.
00:33:50.309 --> 00:33:54.880
And, and like Nvidia, we saw them do an end-to-end solution for ai.
00:33:54.910 --> 00:33:55.295
We see.
00:33:55.730 --> 00:33:58.700
IBM doing the same thing for quantum computing.
00:33:59.029 --> 00:34:03.200
They're the only company I know of that has, well, D-Wave and others.
00:34:03.200 --> 00:34:09.500
But really, most of the modalities are relying on kind of a ecosystem partnerships to go through there.
00:34:09.500 --> 00:34:10.670
And I'm not knocking that.
00:34:10.670 --> 00:34:14.744
I think that's great too, but IBM puts so much in end to end from.
00:34:15.150 --> 00:34:25.360
The refrigeration solutions to the systems, to the wiring, to, they have invested billions and continue to in the entire thing, and now they've got a partnership with a MD.
00:34:25.644 --> 00:34:25.704
Yeah.
00:34:26.364 --> 00:34:29.425
Now, if you don't know, these two companies have worked together for decades.
00:34:29.844 --> 00:34:29.905
Yeah.
00:34:29.905 --> 00:34:33.204
On copper interconnects and silicon on insulator technology.
00:34:33.204 --> 00:34:42.324
They have helped pioneer, these two of these companies have helped pioneer some of the major breakthroughs we've had in semiconductor manufacturing and design over the past three or four decades.
00:34:42.625 --> 00:34:47.684
So having those two work together, I think is gonna be even, is gonna be a big opportunity.
00:34:48.135 --> 00:34:53.135
And trying to see what, because once again, now you've got another ai company that's.
00:34:54.320 --> 00:34:55.070
Fray there.
00:34:55.429 --> 00:34:59.929
And it'll be interesting to see how everything kind of meshes together as we get farther on down the road.
00:34:59.980 --> 00:35:05.460
and it's funny'cause some of the quantum company, quantum computer companies are promoting on-prem systems.
00:35:05.969 --> 00:35:09.719
I don't know if that's going to be the preferred business model.
00:35:09.719 --> 00:35:19.929
I still think that for the vast majority, cloud-based, kind of rental solutions are probably going to be the most likely scenario for quantum computing, especially for the next decade.
00:35:21.139 --> 00:35:23.965
I think the exception to that's probably in two areas.
00:35:23.965 --> 00:35:26.934
One is in financial services and the other is, government.
00:35:27.414 --> 00:35:30.985
you look at a lot of these companies right now, they're, they have two primary.
00:35:31.929 --> 00:35:33.550
Markets, customer segments.
00:35:33.695 --> 00:35:39.190
One, one is, is researchers and the other is national defense.
00:35:39.579 --> 00:35:39.670
Mm-hmm.
00:35:40.059 --> 00:35:49.695
and so national def both those areas, I think there is a role for on-prem quantum computing, even though it may be an internal self-hosted cloud.
00:35:50.175 --> 00:35:51.644
so kind of a hybrid model.
00:35:52.394 --> 00:35:52.635
Yeah.
00:35:52.635 --> 00:35:58.425
And you, it's important to note that, there's obviously a lot of investment from the research community too, matter of fact.
00:35:58.425 --> 00:35:58.514
Mm-hmm.
00:35:58.875 --> 00:36:07.184
IBM has been putting some of their systems out there in the research community, like, uh, believe the University of Cleveland or something like that, or mm-hmm.
00:36:07.425 --> 00:36:10.635
They've gotta be universities and research organizations around the world now.
00:36:10.905 --> 00:36:13.545
So it's gonna be interesting to see how this all plays out.
00:36:13.545 --> 00:36:17.054
And matter of fact, two of the companies are spin outs from Oxford University.
00:36:17.380 --> 00:36:29.760
So, it's, I I think we're probably, maybe not to the point where we saw transformer models and we saw LLMs with, ai, but we're getting there very quickly and by the end of the decade it's gonna get exciting.
00:36:30.210 --> 00:36:30.570
Yeah.
00:36:31.110 --> 00:36:32.639
Sounds a lot like super computing.
00:36:34.295 --> 00:36:34.585
It's
00:36:34.804 --> 00:36:35.034
how it's,
00:36:35.110 --> 00:36:36.730
super computing on steroids.
00:36:36.730 --> 00:36:37.630
Steroids, yeah.
00:36:37.630 --> 00:36:38.920
It's quantum computing.
00:36:38.920 --> 00:36:41.440
So, Carl, do you think we're gonna have quantum factories.
00:36:43.110 --> 00:36:44.039
Quantum factories.
00:36:44.099 --> 00:36:47.429
It sounds like Jim described, quantum infrastructure.
00:36:47.429 --> 00:36:47.670
Right.
00:36:47.670 --> 00:36:52.900
The whole supply chain and the build out large, data centers and or quantum centers.
00:36:52.949 --> 00:36:53.280
yeah.
00:36:53.324 --> 00:36:59.014
I think there will be quantum factories, because the cost of these systems is still quite high.
00:36:59.465 --> 00:36:59.554
Yeah.
00:36:59.554 --> 00:37:02.085
And so you'd amortize those costs over as many.
00:37:02.880 --> 00:37:05.159
Use cases and customers as as you can.
00:37:05.699 --> 00:37:19.824
And so building quantum factories where you have CPUs, GPUs, xus, and Q all basically providing services for applications we can barely imagine today.
00:37:20.525 --> 00:37:21.425
This has been great.
00:37:21.425 --> 00:37:23.485
This has really been, informative.
00:37:23.704 --> 00:37:30.480
I know I've learned a lot from both of you in this 30 minutes or so that we've had this chat.
00:37:30.480 --> 00:37:33.780
And, it's great to have an opportunity also to talk about Quantum Day.
00:37:34.199 --> 00:37:37.320
I, I thought, that it was really well done program.
00:37:38.144 --> 00:37:41.494
even some of the detail they went into Was incredible.
00:37:41.525 --> 00:37:41.795
I,
00:37:42.034 --> 00:37:42.335
yeah,
00:37:42.364 --> 00:37:47.764
I need to go back and watch it probably a dozen more times because some of the mathematics, especially in the last presentation,
00:37:48.695 --> 00:37:48.905
yeah.
00:37:48.905 --> 00:37:49.894
I was in transit.
00:37:49.894 --> 00:37:52.625
I was trying to listen to this stuff as I was on a flight.
00:37:52.625 --> 00:37:53.284
So,
00:37:53.344 --> 00:37:56.255
yeah, the professor from Israel was just, oh my god.
00:37:56.255 --> 00:37:58.715
He, his mathematics depth was incredible.
00:38:00.815 --> 00:38:02.434
So, gentlemen, Thank you so much.
00:38:02.434 --> 00:38:03.300
It was really our pleasure.
00:38:03.485 --> 00:38:05.224
Really, really great, great stuff.
00:38:05.224 --> 00:38:18.364
And, I think the, audience will undoubtedly find it incredibly grounding and, a good orientation as to how they should, view the aqua, the Cambrian explosion.
00:38:19.034 --> 00:38:20.445
That is quantum explosion.
00:38:20.954 --> 00:38:21.764
The what?
00:38:21.974 --> 00:38:23.144
The quantum explosion.
00:38:23.144 --> 00:38:24.855
The quantum explosion.
00:38:25.215 --> 00:38:27.315
Nah, I gotta make it Cambrian.
00:38:27.510 --> 00:38:28.000
Cambrian.
00:38:29.585 --> 00:38:30.144
Cambrian.
00:38:30.489 --> 00:38:32.170
Cambrian Quantum explosion.
00:38:32.559 --> 00:38:32.860
I've got
00:38:32.860 --> 00:38:33.219
all these qu
00:38:33.550 --> 00:38:34.329
the qc,
00:38:34.960 --> 00:38:35.289
I've got all
00:38:35.289 --> 00:38:37.570
these Cambrian fossils.
00:38:37.570 --> 00:38:39.519
I can't change'em all to quantum fossils.
00:38:43.514 --> 00:38:56.534
So, thank you gentlemen and everyone, if you're interested in tapping into, Jim's research calls research, reach out to both of these gentlemen and their respective firms.
00:38:56.534 --> 00:39:01.445
Of course, Jim McGregor is with tedious research.
00:39:02.409 --> 00:39:02.829
How's that?
00:39:02.829 --> 00:39:02.980
It's
00:39:03.144 --> 00:39:03.465
pretty good.
00:39:03.905 --> 00:39:03.925
Pretty good.
00:39:04.449 --> 00:39:04.659
Great.
00:39:04.659 --> 00:39:04.719
Yeah.
00:39:05.320 --> 00:39:05.739
I mean, sad.
00:39:06.280 --> 00:39:12.309
Hey, I'm an honorary Scotsman now because my nickname, or at least my Scottish name is, Liam, right?
00:39:12.400 --> 00:39:13.030
Yes, exactly.
00:39:14.650 --> 00:39:17.320
I gotta live up to that honorary designation.
00:39:17.320 --> 00:39:19.059
So yes, reach out to him at cheer.
00:39:19.280 --> 00:39:22.940
research@www.teusresearch.com.
00:39:22.940 --> 00:39:28.440
He's also on Forbes and he's quite active He Times and other publications.
00:39:28.519 --> 00:39:36.565
And then of course we have Carl Fre of Cambrian Quantum AI Research.
00:39:37.309 --> 00:39:39.530
Put Hy keep putting hyphens on the company name.
00:39:39.530 --> 00:39:42.949
Cambrian AI hyphen Quantum hyphen Research.
00:39:43.000 --> 00:39:49.264
So catch him at www.cambrianaium.com.
00:39:50.349 --> 00:39:51.159
Did I get that right?
00:39:51.545 --> 00:39:52.085
That's right.
00:39:52.420 --> 00:39:52.900
Yeah.
00:39:52.900 --> 00:39:54.579
And then, he has a substack.
00:39:54.610 --> 00:40:00.079
He has, a, he's on Forbes and he's on LinkedIn.
00:40:00.139 --> 00:40:04.860
He's also public, well published, across a number of publications and outlets.
00:40:05.309 --> 00:40:08.840
And of course, please subscribe to our podcast here.
00:40:08.889 --> 00:40:18.880
We do a, A, an episode every month to cover everything that's happening that you should know about, in the world of.
00:40:20.034 --> 00:40:23.335
AI, semiconductors and quantum.
00:40:23.335 --> 00:40:43.994
So, make sure to check us out on YouTube, subscribe, like, share, and then check out the audio version on Buzz Sprouts and listen to us on your favorite podcast platform and subscribe to the next curve research portal@wwwnextcurve.com for the tech and industry insights that matter.
00:40:44.235 --> 00:40:45.315
Gentlemen, thank you so much.
00:40:45.315 --> 00:40:45.639
This is awesome.
00:40:46.500 --> 00:40:46.949
Thank you.
00:40:47.309 --> 00:40:48.900
We will see you by the end of the month.
00:40:49.179 --> 00:40:49.900
talk to you soon.
00:40:49.900 --> 00:40:50.139
Yep.
00:40:50.235 --> 00:40:50.525
Yeah.
00:40:51.184 --> 00:40:51.605
Cheers.
00:40:52.105 --> 00:40:52.525
Cheers.
00:00:06.240 --> 00:00:07.200
Next curve.
00:00:10.439 --> 00:00:11.205
Hey everyone.
00:00:11.205 --> 00:00:15.794
Welcome to this episode of the Nextcar Rethink Podcast.
00:00:15.794 --> 00:00:22.335
And of course, this is our special Silicon Future Series, which I do in partnership.
00:00:22.335 --> 00:00:30.074
In collaboration, wonderful collaboration with the illustrious Carl Fre of.
00:00:30.675 --> 00:00:34.755
Cambrian AI and soon to be quantum research.
00:00:34.755 --> 00:00:34.765
Right.
00:00:34.765 --> 00:00:35.630
It became even quantum,
00:00:35.865 --> 00:00:36.255
I think.
00:00:36.344 --> 00:00:37.395
Ah, yeah.
00:00:37.395 --> 00:00:37.454
The
00:00:37.454 --> 00:00:38.174
next big thing.
00:00:38.564 --> 00:00:38.984
Yeah.
00:00:38.984 --> 00:00:40.155
And then of course.
00:00:41.490 --> 00:00:43.820
The, quantum dotie.
00:00:46.340 --> 00:00:46.880
Are you ever gonna
00:00:46.880 --> 00:00:47.984
run Jim McGregor for
00:00:47.984 --> 00:00:48.064
me?
00:00:50.445 --> 00:00:54.875
Jim McGregor of serious research Clan mcg.
00:00:55.850 --> 00:00:56.179
Yeah.
00:00:56.390 --> 00:00:56.929
Okay.
00:00:57.020 --> 00:00:57.829
Clan McGregor.
00:00:58.310 --> 00:00:58.924
Yeah.
00:00:58.924 --> 00:01:03.119
Who has, some amazing, amazing kids?
00:01:03.490 --> 00:01:05.950
who are just remarkable programmers.
00:01:05.950 --> 00:01:13.620
I I saw Damien, by the way, at, Katie Cadence live, 2026 and, yeah, you have really cool kids, man.
00:01:14.204 --> 00:01:14.594
Thank you.
00:01:14.814 --> 00:01:17.965
anyway, welcome everyone.
00:01:19.254 --> 00:01:28.665
And of course we cover, all the exciting stuff that you need to know in the world, the semiconductors, ai, and apparently now quantum.
00:01:28.665 --> 00:01:29.144
Quantum.
00:01:29.665 --> 00:01:42.515
Jim and Carl have been going deep into all this quantum stuff that everyone is starting to get pretty excited about, and, they're gonna basically break down the landscape.
00:01:43.260 --> 00:01:45.540
and so this is gonna be a great education for me.
00:01:45.599 --> 00:01:49.900
and, these two gentlemen are going to basically lay it out.
00:01:49.900 --> 00:01:52.000
This is what the landscape looks right now.
00:01:52.000 --> 00:02:03.219
And this looks, this is how you need to structure your thinking around what the market will look like and how the technol, the competing technologies actually, right.
00:02:03.219 --> 00:02:08.090
There are many different approaches, how those will, likely, and may be.
00:02:08.525 --> 00:02:10.444
Play out in the future.
00:02:10.444 --> 00:02:12.724
And so it is a really interesting area.
00:02:12.754 --> 00:02:14.525
It's formative, I believe.
00:02:14.594 --> 00:02:23.335
I don't know if you both agree, but I'm really excited, to have this, presentation, by the both of you on this, emerging.
00:02:23.784 --> 00:02:29.875
Well, do you wanna call it emerging topic or is this just something that is hot but people need
00:02:30.504 --> 00:02:33.050
to It's pretty one, it's one of the most important topics for the next decade.
00:02:34.060 --> 00:02:34.930
Yeah, absolutely.
00:02:35.080 --> 00:02:35.770
Totally agree.
00:02:36.039 --> 00:02:36.639
Totally agree.
00:02:38.229 --> 00:02:40.180
And then we get into neuromorphic computing.
00:02:40.180 --> 00:02:41.439
So we can do even more there.
00:02:41.439 --> 00:02:41.500
Yeah.
00:02:42.099 --> 00:02:42.370
More.
00:02:42.375 --> 00:02:42.645
Okay.
00:02:42.789 --> 00:02:43.060
Yeah.
00:02:43.060 --> 00:02:44.680
We'll have another, talk on that.
00:02:44.680 --> 00:02:44.949
Right?
00:02:45.009 --> 00:02:47.800
we're gonna make his head explode, Carl, if we keep doing this.
00:02:47.900 --> 00:02:48.560
But you know what?
00:02:48.560 --> 00:02:52.810
For those of you who are listening on audio, you really want to go to YouTube and watch this episode.
00:02:52.810 --> 00:03:00.650
'cause we're gonna have some graphics and, if you follow us on LinkedIn, when we post, on LinkedIn or, other, social.
00:03:01.055 --> 00:03:04.729
Outlets will, provide the presentation material.
00:03:05.030 --> 00:03:18.139
But, for those of you who are just listening in, you really want to, check out the YouTube, episode in the video so that you can see, the charts that these two gentlemen are prepared and they're quite pretty.
00:03:18.830 --> 00:03:20.479
You guys did a very good job.
00:03:20.479 --> 00:03:20.719
Pretty job.
00:03:20.719 --> 00:03:20.990
Yeah.
00:03:20.990 --> 00:03:21.344
Well, thank you.
00:03:21.365 --> 00:03:22.250
You are pretty too.
00:03:23.840 --> 00:03:26.509
You guys are very gorgeous in that regard.
00:03:26.750 --> 00:03:27.020
So,
00:03:27.889 --> 00:03:31.590
well, when you have a face for radio, you have to make your slides look good.
00:03:36.030 --> 00:03:36.360
That's a good one.
00:03:37.240 --> 00:03:37.569
Right.
00:03:37.599 --> 00:03:38.050
Okay.
00:03:38.050 --> 00:03:45.759
So on that note, before we get started, please remember to like, share and react and comment on this episode and also subscribe here on YouTube and Buzzsprout.
00:03:46.330 --> 00:03:48.849
Listen to us on your favorite podcast platform.
00:03:48.879 --> 00:03:52.569
Opinions and statements by my guests are their own and don't reflect mine.
00:03:52.960 --> 00:03:54.310
Are those the next curve?
00:03:54.340 --> 00:03:58.205
Although I think I'm, we're gonna be in very good alignment since you guys are really.
00:03:58.960 --> 00:04:01.870
Going to be giving a masterclass here.
00:04:01.870 --> 00:04:06.129
So I'm all ears and I think the audience will be as well.
00:04:06.400 --> 00:04:15.219
And, opinions and statements, like I said, are their own, but we're doing this for informational purposes and in this case it's educational, it's gonna be great.
00:04:15.250 --> 00:04:22.709
we do this to provide an open forum for, discussion and debate on all things, ai.
00:04:23.449 --> 00:04:27.319
Semiconductor industry and now Quantum.
00:04:27.319 --> 00:04:32.069
And so with that, gentlemen, Where should we start this conversation about Quantum?
00:04:32.230 --> 00:04:34.240
let me start, start with my, if that's okay.
00:04:34.240 --> 00:04:37.180
They're a little higher level and Jim gets into more detail.
00:04:37.839 --> 00:04:41.884
It's actually been a big week this week, or last week, I guess since we posted next week.
00:04:42.345 --> 00:04:46.824
because, this week was, quantum World, world Quantum Computing Day.
00:04:47.500 --> 00:04:48.819
On, Tuesday, right, Jim?
00:04:49.089 --> 00:04:49.300
Yeah.
00:04:49.300 --> 00:04:50.350
Supported by Nvidia.
00:04:50.350 --> 00:04:52.269
It was NVIDIA's Quantum computing
00:04:52.269 --> 00:04:57.740
Nvidia had their second quantum day, not associated with GTC this time.
00:04:57.740 --> 00:05:05.970
It was a standalone event with Jensen Wong speaking, and some of the leaders in Quantum, that are, at least those that are friendly with Nvidia, they were there.
00:05:05.970 --> 00:05:09.689
Those that are not as friendly weren't there, but, it was really great.
00:05:09.730 --> 00:05:11.889
Nvidia made some important announcements.
00:05:12.279 --> 00:05:12.430
Yeah.
00:05:12.509 --> 00:05:21.540
With their new, quantum model for basically improving the error detection and error mitigation correction.
00:05:21.814 --> 00:05:24.295
The problem when an error happens, where did it happen?
00:05:24.894 --> 00:05:27.795
And that's where GPUs can help quantum a lot.
00:05:28.214 --> 00:05:33.615
So there's a really emerging view that GPUs and Quantum are not gonna be competing technologies.
00:05:33.944 --> 00:05:41.305
They're really gonna be, collaborative, technologies that help bring quantum to the useful quantum computing era.
00:05:41.605 --> 00:05:54.375
So what I've done here is just created a slide,'cause this gets very confusing for all of us, of what is going on in this marketplace and how can you have this many billions of dollars pouring into a market that's currently producing.
00:05:55.115 --> 00:05:56.345
Call it zero revenue.
00:05:56.975 --> 00:06:03.545
but, as Jensen Wong says, the most interesting markets of this, the markets of$0 that will grow into billions of dollars.
00:06:04.024 --> 00:06:09.279
And many of us are firmly of the belief that is exactly what will happen with quantum computing.
00:06:09.790 --> 00:06:11.620
It's not gonna replace ai.
00:06:11.620 --> 00:06:15.009
It will facilitate ai, and AI will facilitate quantum.
00:06:15.009 --> 00:06:18.360
It's a reinforcing, cycle here, it's virtuous cycle.
00:06:18.629 --> 00:06:20.579
So on the left you have the modalities.
00:06:20.639 --> 00:06:26.699
By modality, you have basically mean what kind of physics are these guys using to create these quantum states?
00:06:27.120 --> 00:06:29.939
And the next, column is basically in plain English.
00:06:30.000 --> 00:06:30.689
How's it work?
00:06:30.790 --> 00:06:31.495
what's this stuff all about?
00:06:32.579 --> 00:06:34.139
Next column is key companies.
00:06:34.500 --> 00:06:36.389
And then finally, the maturity level.
00:06:36.689 --> 00:06:40.560
This is saying how close were they to producing quantum advantage.
00:06:40.560 --> 00:06:54.370
By quantum advantage, we mean that you could solve meaningful problems in a cost-effective way that you cannot solve, in, digital computing technologies regardless of how big your supercomputer is.
00:06:55.225 --> 00:06:56.064
so let's start this.
00:06:56.095 --> 00:06:57.235
I'll try to do this quickly.
00:06:57.235 --> 00:06:58.915
You have superconducting qubits.
00:06:58.915 --> 00:07:00.625
This is the one that's been around the longest.
00:07:00.625 --> 00:07:06.725
This is where IBM and Google ti, as well as IQM and OQC.
00:07:07.074 --> 00:07:12.555
this is where you see those huge chandeliers of gold that are used as.
00:07:12.699 --> 00:07:20.139
Final stage of cooling that creates a, sta a a temperature of millikelvin.
00:07:20.139 --> 00:07:22.490
So this is colder than outer space.
00:07:22.850 --> 00:07:29.519
And that is where the actual supercomputing chip, superconducting chip, excuse me, can operate.
00:07:29.819 --> 00:07:29.910
Mm-hmm.
00:07:29.910 --> 00:07:36.120
And so you have these back rooms full of these huge refrigerators that eventually chill down through these chandeliers.
00:07:36.704 --> 00:07:39.014
To the bottom where this chip is gonna operate.
00:07:39.345 --> 00:07:42.644
The advantage of this technology is it's really fast.
00:07:43.045 --> 00:07:44.334
the disadvantages.
00:07:45.149 --> 00:07:50.339
It has to operate ultra low temperatures, which is very expensive in space, consumes a lot of space.
00:07:50.920 --> 00:07:53.389
but, it's also, unfortunately very noisy.
00:07:53.779 --> 00:07:59.990
so if you think about a normal semiconductor chip, it's okay that you're, you know, 10 10 to the minus seven.
00:08:01.060 --> 00:08:08.949
But here we're talking about 10 to minus two or three, so that means one, one in a hundred or one in a thousand results are bad.
00:08:09.480 --> 00:08:12.689
And considering you're producing hundreds of millions of results, that's not good.
00:08:12.990 --> 00:08:17.910
So you have to identify these errors and you have to correct mitigate these errors, which means reduce their impact.
00:08:18.209 --> 00:08:20.939
But you also eventually have to correct these errors.
00:08:21.600 --> 00:08:27.660
This is often done by teaming together physical cubits to create a logical cubit.
00:08:27.959 --> 00:08:33.970
The ratio of the, of which is super important to both speed and economics.
00:08:34.070 --> 00:08:34.820
if you can do.
00:08:35.309 --> 00:08:38.730
Two or three physical cubit to produce one logical cubit.
00:08:39.179 --> 00:08:42.330
You're gonna be three times more powerful than one.
00:08:42.335 --> 00:08:49.475
an architect that maybe takes 10 cu logical qubits or a hundred logical cubits to produce one, excuse me, physical cubits.
00:08:49.970 --> 00:08:52.159
To produce one logical cubit.
00:08:52.190 --> 00:08:54.559
Logical cubits are what matter.
00:08:54.990 --> 00:09:04.740
that's the, if you can try to distill it down and across all these modalities, logical cubits and their performance and costs are really what matter matters here.
00:09:05.129 --> 00:09:06.480
It's what you're using.
00:09:06.480 --> 00:09:11.909
It's really what you're using this form, the gates, the circuits, the those, the circuits, random circuits that you're actually developing.
00:09:12.059 --> 00:09:16.169
And by the way, if you've never seen pictures of the Poughkeepsie Supercomputing.
00:09:16.659 --> 00:09:21.789
center in New York that IBM has, it kinda looks like an invasion of the body snatchers.
00:09:22.840 --> 00:09:24.610
But all these hanging refrigerators.
00:09:24.610 --> 00:09:26.870
All these hanging refrigerators, they're huge.
00:09:27.620 --> 00:09:29.245
They're huge as well.
00:09:29.245 --> 00:09:33.715
As well as at the TJ Watson Research Center where they have a major demo center now.
00:09:33.924 --> 00:09:34.014
Mm-hmm.
00:09:34.254 --> 00:09:38.034
A demon demonstrating the, IBM Quantum system two.
00:09:38.754 --> 00:09:42.174
now if you go down the list, trap s kind of nest, these are kind of.
00:09:43.565 --> 00:09:44.914
They're not random orders here.
00:09:44.914 --> 00:09:52.620
These are in decreasing state of maturity, potentially increasing potential to be disruptive.
00:09:52.620 --> 00:10:04.830
Tract ions are much more, high, much higher fidelity, which means that they're not gonna be as error prone, but it's hard to scale that's in progress and it's hard to get the performance.
00:10:04.830 --> 00:10:06.330
They don't have the same performance.
00:10:06.720 --> 00:10:13.750
As a supercomputing super, conducting circuit does, they, these ions, they're exotic ions.
00:10:13.750 --> 00:10:15.009
I can't even remember their names.
00:10:15.009 --> 00:10:17.350
They're not normal elements you could think of.
00:10:17.750 --> 00:10:27.684
they're trapped in control with lasers and walking around the computer show floor at, let's say TTC last month, you could actually see models of what these look like.
00:10:28.044 --> 00:10:29.184
Very high fidelity.
00:10:29.884 --> 00:10:31.445
There's gotta work on performance.
00:10:31.445 --> 00:10:43.524
And you can see the companies that are listed there, neutral Atoms, kind of a dark horse here, is late, late in development, but is showing tremendous potential.
00:10:44.095 --> 00:10:44.605
Tremendous.
00:10:44.980 --> 00:10:47.769
Promise because they operate at room temperature.
00:10:47.769 --> 00:10:54.360
You don't need these massive refrigeration systems to get them down to sub, sub one degree Kelvin.
00:10:54.929 --> 00:10:57.899
so that means they're a lot more portable, they're more affordable.
00:10:58.210 --> 00:10:59.259
They're more flexible.
00:10:59.590 --> 00:11:00.399
and they're fast.
00:11:00.960 --> 00:11:10.450
They're not as fast as superconducting qubits or even as trapped ion cu qubits, but they can be as fast when you combine them, because of the, cause.
00:11:10.450 --> 00:11:18.450
You combine several physical, neutral atom qubits to create a, logical, qubit and get pretty much the same performance.
00:11:18.975 --> 00:11:31.745
bot is, using light particles and you can see some of the companies there that have been quite vocal about their ability, or again, kind of like neutral atoms, early stage, early days, but very promising technology.
00:11:32.899 --> 00:11:35.570
The next one is spin qubits or quantum dots.
00:11:35.570 --> 00:11:42.269
These are electron spin in a semiconductor structure, and you can see the companies they're the most notable are probably Intel.
00:11:42.730 --> 00:11:44.379
again, this is early stage.
00:11:44.379 --> 00:11:49.750
It is not something you're gonna, you, or maybe even your drill from will be using, but it does have a lot of potential.
00:11:50.029 --> 00:11:52.220
Intel actually started on Supercomputing.
00:11:52.610 --> 00:11:52.820
Yeah.
00:11:53.164 --> 00:11:56.139
And they switched and then switched to the, the spin bits.
00:11:56.139 --> 00:11:56.860
So That's right.
00:11:57.490 --> 00:11:58.330
It, it did.
00:11:58.330 --> 00:12:02.289
So a lot of the, you have to realize that this has all been in development stage.
00:12:02.289 --> 00:12:09.679
So some of these companies have actually evaluated or worked on multiple different, multiple of these different, qubit solutions or modality.
00:12:09.679 --> 00:12:09.740
Yeah.
00:12:10.580 --> 00:12:11.450
Quantium is like that.
00:12:11.450 --> 00:12:13.500
Quantium used to be known as, Honeywell.
00:12:13.779 --> 00:12:14.429
They used to be part of
00:12:14.429 --> 00:12:19.715
They used to be part of Honeywell and they spun'em out and they really decided to focus at that time on trapped ions.
00:12:19.715 --> 00:12:21.965
But before they were more of an application layer.
00:12:22.294 --> 00:12:28.184
And, really trying to develop, independent of the modality, they decided to just focus on trapped ions.
00:12:28.184 --> 00:12:29.504
They think that's the sweet spot.
00:12:29.715 --> 00:12:36.034
So, Carl, some of the audience might be wondering, as you go down this list, at what point are you.
00:12:36.230 --> 00:12:40.399
Departing from superconducting and super cool, like this, near mm-hmm.
00:12:40.889 --> 00:12:42.960
zero Kelvin cooling.
00:12:43.240 --> 00:12:45.100
so what does that look like here?
00:12:45.100 --> 00:12:48.309
Because you mentioned neutral atom is room temperature, right?
00:12:48.309 --> 00:12:48.909
And so, yeah.
00:12:49.370 --> 00:12:57.700
what are we looking at in terms of the environment that these different modalities, outside of the supercomputing, operate in?
00:12:58.610 --> 00:13:04.570
Well, it's actually, supercomputing and the quantum and kneeling that use the, ultra low temperatures.
00:13:04.620 --> 00:13:12.730
everything else is either a higher temperature or, at room temperature, especially when you start talking about using photons or ions and stuff like that.
00:13:12.820 --> 00:13:13.360
Right, right.
00:13:13.419 --> 00:13:13.750
Okay.
00:13:13.990 --> 00:13:14.169
Yeah.
00:13:14.169 --> 00:13:14.440
Cool.
00:13:14.590 --> 00:13:15.129
Exactly.
00:13:15.565 --> 00:13:15.715
All,
00:13:16.355 --> 00:13:18.125
quantum and kneeling is kind of a niche.
00:13:18.215 --> 00:13:22.625
It has commercial users today, and there's really two players that are Dwa and Fujitsu.
00:13:23.554 --> 00:13:29.674
and then there's topological qubits, which created a tremendous roar in the market when mi, when Microsoft disclosed it.
00:13:29.975 --> 00:13:31.355
They're the only ones doing this.
00:13:31.654 --> 00:13:42.044
If they can make it work, which is a big, if, if they can make it work, it could be the winner, but it's very far from production, probably the least developed research platform.
00:13:42.539 --> 00:13:47.399
Yeah, but it was pretty amazing at the time, how much, how much buzz it created.
00:13:47.610 --> 00:13:50.669
But, earlier you mentioned, you said super computing.
00:13:50.669 --> 00:13:51.750
You meant superconducting,
00:13:51.750 --> 00:13:53.490
I meant superconducting conducting.
00:13:53.490 --> 00:13:54.809
Just wanted to clarify that.
00:13:55.210 --> 00:13:55.539
Thank you, Leonard.
00:13:55.809 --> 00:13:56.110
Yeah.
00:13:57.174 --> 00:13:58.644
Some pros and cons here.
00:13:59.075 --> 00:14:01.355
I've probably covered quite a bit of this already.
00:14:01.360 --> 00:14:01.519
Yeah,
00:14:01.524 --> 00:14:03.565
Pros, superconducting, fast gates.
00:14:03.625 --> 00:14:07.164
Very strong ecosystem, by the way, because they've been around a long time.
00:14:07.404 --> 00:14:15.125
And IBM's willing to spend a lot of money to, to, facilitate the collaboration of research institutions around the world.
00:14:15.644 --> 00:14:16.514
it's the near term.
00:14:16.514 --> 00:14:16.634
Later.
00:14:17.220 --> 00:14:17.549
Right.
00:14:17.600 --> 00:14:20.365
big, big, strong, force in the industry.
00:14:20.815 --> 00:14:26.965
Trapped ions, they're really the contender right now have fault tolerance systems, neutral atoms.
00:14:27.014 --> 00:14:28.154
They're the dark horse.
00:14:28.184 --> 00:14:34.705
But there is one company, for example, which is inflection, did a SPAC in February.
00:14:35.049 --> 00:14:40.960
And so they're now publicly traded and, they're seeing a lot of interest from investors in neutral atoms.
00:14:41.500 --> 00:14:46.899
and then they also have an existing business, which is called quantum sensing.
00:14:46.899 --> 00:14:50.259
We're not talking quantum sensing here, we're talking quantum computing.
00:14:50.840 --> 00:14:55.580
so they can use that to fund and finance the development of quantum computing.
00:14:56.330 --> 00:15:00.559
FO photonic, very strong quantum networks and distributed quantum computing.
00:15:01.100 --> 00:15:05.779
quantum dots promising long-term pathway for building these on silicon.
00:15:05.779 --> 00:15:10.750
On CM OS compatible, chips, topological, potentially a breakthrough.
00:15:10.950 --> 00:15:12.750
very high risk, very high reward.
00:15:13.110 --> 00:15:14.309
There's quantum and kneeling.
00:15:14.370 --> 00:15:15.779
It's commercially viable today.
00:15:16.105 --> 00:15:18.995
DD Wade's doing quite well, in their niche.
00:15:19.044 --> 00:15:19.884
but it is a niche.
00:15:19.884 --> 00:15:20.965
It's not gonna be a general purpose.
00:15:20.965 --> 00:15:21.414
Quantum computing.
00:15:22.929 --> 00:15:26.409
With that, I'll turn it over to Jim and you can, dive into your analytics.
00:15:26.409 --> 00:15:26.649
Jim,
00:15:26.889 --> 00:15:35.340
I took a, another, a similar view to what, Carl did and I have to kind of preface this and say that, all these numbers are relative.
00:15:35.490 --> 00:15:41.799
In other words, things that, they're how I kind of, positioned the different factors of these different technologies.
00:15:41.830 --> 00:15:45.789
And quite honestly, you have to remember that all these technologies are still developing.
00:15:45.789 --> 00:15:48.100
So they are changing and they will change.
00:15:48.100 --> 00:15:54.250
So even looking at the spider radar graph that you're looking at, all these different technologies will change over time.
00:15:54.700 --> 00:15:55.750
And if you.
00:15:56.269 --> 00:15:57.860
Click down one slide here.
00:15:58.080 --> 00:15:58.799
we'll see.
00:15:58.980 --> 00:16:04.440
When you start looking at the numbers and you start looking at how they're positioned, supercomputing provides that best balance.
00:16:04.769 --> 00:16:10.490
Trapped on trapped ion provides that, high performance solution or the best performing solution.
00:16:10.940 --> 00:16:17.139
The photonic and quantum dots, which are very similar, have the best manufacturability capabilities today.
00:16:17.450 --> 00:16:21.289
although the topological, that was the one that, Microsoft is working on is.
00:16:21.580 --> 00:16:26.350
Theoretically the best from an error correction standpoint or error mitigation standpoint.
00:16:26.350 --> 00:16:26.679
Mitigation.
00:16:26.740 --> 00:16:28.570
So it will be very interesting.
00:16:28.659 --> 00:16:32.679
And then once again that a kneeling is really a specialized niche.
00:16:32.769 --> 00:16:36.909
I don't ever see that necessarily being a general market capability.
00:16:37.330 --> 00:16:45.600
and, one more slide., I outlined some of those, the same thing that, Carl did, but you have to remember one thing you have to really remember, and let's go to the next slide.
00:16:46.120 --> 00:16:46.509
Leon.
00:16:47.710 --> 00:16:48.190
Leonard
00:16:49.360 --> 00:16:52.750
d Neil is, is that, is that
00:16:52.750 --> 00:16:53.649
my basic Leon?
00:16:53.649 --> 00:16:55.809
But I, you know, I, I, is that my Scott team?
00:16:56.620 --> 00:16:57.610
Yeah, exactly.
00:16:59.110 --> 00:17:08.009
I, no, you have to remember where we are and the fact that, a lot of people said, well, think about where AI was in 2015.
00:17:08.190 --> 00:17:09.599
Before, everything exploded.
00:17:10.019 --> 00:17:12.539
We're kind of there in what we call the Nisc era.
00:17:12.539 --> 00:17:17.690
It's kind of noisy, and unstable qubits where we're really still learning about it.
00:17:17.930 --> 00:17:23.180
But you have to understand that the industry has actually been working on this since the mid nineties.
00:17:23.390 --> 00:17:25.250
I mean, this is a technology that's taken.
00:17:25.549 --> 00:17:29.210
So when, when everyone says, oh, it's gonna be decades away, it's been decades.
00:17:29.240 --> 00:17:31.220
this has been in development for a long time.
00:17:31.220 --> 00:17:33.200
we went through a long period of just.
00:17:34.019 --> 00:17:36.420
Understanding is this even feasible?
00:17:36.420 --> 00:17:39.960
And then kind of proof of concept era through the early two thousands.
00:17:39.960 --> 00:17:43.720
And now we're in an era where we understand it a little bit better.
00:17:43.880 --> 00:17:47.569
we're having to do a lot of error mitigation and suppression just.
00:17:47.569 --> 00:17:51.255
To be able to get rid of some of that noise that's in these different types of qubits.
00:17:51.255 --> 00:17:55.825
And obviously there are many different ways to create a qubit, and read a qubit.
00:17:56.174 --> 00:17:59.625
but we've created quantum utility, and this is probably the most important thing.
00:18:00.134 --> 00:18:06.194
Why are there billions and billions of dollars going in this, not just in terms of the technology, but also in terms of the use of it?
00:18:06.464 --> 00:18:10.065
I mean, you have banks, you have materials companies, you have.
00:18:10.914 --> 00:18:18.265
Scientific research organizations all pouring billions into this because they see this as a very significant leap in terms of computing.
00:18:18.295 --> 00:18:20.035
This is an accelerator technology.
00:18:20.244 --> 00:18:34.075
It's not going to displace classical computing, and this is why AI and standard computing and basically the accelerated computing environment we're in now is gonna be so beneficial to this because basically it's gonna help us on the front end.
00:18:34.454 --> 00:18:47.690
Basically create, the quantum circuits to organize the data and everything else that feeds into this quantum accelerator, and then understand that information and do any error correction as it comes out of that quantum computer.
00:18:48.140 --> 00:18:53.210
So basically this quantum computer or QPU quantum computing unit, it's gonna be another.
00:18:53.589 --> 00:19:02.400
Type of rack that sits in the data center or even maybe remotely that's tied to that standard supercomputer or ai, powerhouse.
00:19:02.730 --> 00:19:07.440
and right now we're finding that, you know, even when you get to like a hundred qubits and that's.
00:19:07.575 --> 00:19:08.714
physical qubits.
00:19:09.105 --> 00:19:18.275
You're getting into an area where it's starting to be interesting, where you can understand, how to use the technology, and even run some, interesting simulations.
00:19:18.305 --> 00:19:29.375
And when you get to a thousand physical qubits, then you're getting into an area where you're creating, a significant number of logical com, qubits, maybe a hundred or 200 logical compute qubits.
00:19:29.744 --> 00:19:45.535
Creating gates, creating, circuits and actually exceeding that capability, as Carl said, creating quantum advantage where you're exceeding that capability of even what a supercomputer or any supercomputer on the planet can do to do today.
00:19:45.894 --> 00:19:52.224
Now you have to remember that there's always a probability, risk here in terms of, you know, how close to an actual.
00:19:52.259 --> 00:20:06.619
Answer it is, but you have to remember that what it's trying to do is beyond what we've been able to do before, analyzing the human genome, modeling the, the earth and the weather patterns, actually discovering new material technologies.
00:20:06.625 --> 00:20:21.005
it's very, very significant and we're already getting to that quantum utility, being able to use this quantum technology and going forward, we're really looking at this here that we're calling fault tolerant era, where basically we understand the qubits a little bit more.
00:20:21.005 --> 00:20:23.585
We can control the fidelity of the qubits a little bit more.
00:20:23.944 --> 00:20:33.815
And once again, this could be in any of those different modalities, but we get to a point where the accuracy is fairly high and we're getting to a point where we can scale.
00:20:34.005 --> 00:20:35.505
And that's probably the most important thing.
00:20:36.015 --> 00:20:43.420
So when you see these rapid things, we're basically doubling to quadrupling the number of physical qubits that we can do every year.
00:20:44.005 --> 00:20:56.255
depending on the different modalities, that's significant because we're gonna be able to go from, thousands of cubits today, or hundreds or to thousands of qubits today to millions of cubits, shortly after the, into the next decade.
00:20:56.704 --> 00:20:58.345
someplace in between, 20.
00:20:58.589 --> 00:21:00.690
32 to 2035.
00:21:01.180 --> 00:21:25.839
we place it come somewhere around 2034, but that's very significant and once again, as Carl said, once you get into that range and you're only doing anywhere from, it could be a couple physical cubits to a hundred cubits per logical qubit, you're still getting into a point with a million physical qubits to where you can do some very, significant things that you couldn't do anywhere else or any place else.
00:21:26.855 --> 00:21:39.654
And in fact the, all the companies we've mentioned all have roadmaps that explain how they will get to a million, qubits, which is sort of the milestone for quantum advantage.
00:21:39.865 --> 00:21:44.845
Probably some areas that can actually get to quantum advantage, far lower numbers of cubits.
00:21:45.115 --> 00:21:47.365
But in terms of a more general.
00:21:48.055 --> 00:21:56.315
Platform for doing scientific work and, and research in areas that we don't touch today because today's computer simply can't do it.
00:21:56.855 --> 00:22:04.484
So it's gonna open up all kinds of areas of computational, applications that do not exist at all today.
00:22:04.755 --> 00:22:07.125
So it is funny, some people talk about how.
00:22:07.549 --> 00:22:15.390
You know, there's an article about three or four months ago, some guy wrote, I think it was in Wall Journal, that, computing's all gonna come down to the sides of the shoebox.
00:22:15.390 --> 00:22:15.720
Right.
00:22:15.720 --> 00:22:17.519
And it's all gonna be quantum and stuff like this.
00:22:17.519 --> 00:22:20.630
And I'm like, well, I don't know what this guy's smoking, but I'd like some.
00:22:21.134 --> 00:22:21.674
It's crazy.
00:22:21.884 --> 00:22:24.075
This is not gonna replace the traditional.
00:22:24.210 --> 00:22:24.289
Yeah.
00:22:24.585 --> 00:22:25.430
And, and
00:22:26.174 --> 00:22:27.825
Carl's address is no.
00:22:30.339 --> 00:22:31.664
Yeah, and you know what?
00:22:31.664 --> 00:22:51.305
I think that is one of the biggest clarifications that needs to be made at the moment because I think there's, so many misconceptions about what a quantum application is and one of the things that was very clear from Quantum Day and I managed to get.
00:22:51.720 --> 00:23:04.329
You know, listening on some of it, and I'd love to give both of your takes on this, but for the most part, the best way of, thinking about what these applications are the ones that are traditionally, you know, for supercomputing, right?
00:23:04.329 --> 00:23:06.069
Whether it's for, scientific, very much
00:23:06.069 --> 00:23:06.079
so.
00:23:07.089 --> 00:23:07.480
Right.
00:23:07.480 --> 00:23:08.079
Certain,
00:23:08.085 --> 00:23:12.164
Doing that analysis, doing the scientific research, doing those types of applications.
00:23:12.164 --> 00:23:16.875
And quite honestly, I don't know how you feel, Carl, but I think that, it's very much like AI modeling.
00:23:16.904 --> 00:23:22.095
I think we're gonna end up with different types of quantum modalities for different types of applications.
00:23:22.335 --> 00:23:28.394
I think we're gonna find as we get farther down the road that, you know, and some of this is definitely AI enabled.
00:23:28.394 --> 00:23:35.769
On the front end and backend, that we're gonna be able to do, cer, certain modalities are gonna fit certain applications better than others, and.
00:23:36.644 --> 00:23:37.065
Yeah.
00:23:37.795 --> 00:23:46.325
in looking at this chart specifically, that growth chart, it gets to a point, you have to remember that, this is very much like the AI was, over a decade ago.
00:23:46.664 --> 00:23:48.105
we're gonna see changes in the industry.
00:23:48.115 --> 00:23:50.305
So we're probably gonna see a consolidation.
00:23:50.545 --> 00:23:53.714
Some of these companies will get bought for their IP or bought all out.
00:23:54.015 --> 00:24:00.444
We may have, larger companies doing multiple modalities by the time we get to this fault tolerant era.
00:24:00.775 --> 00:24:08.259
So there's going to be a lot of change and that's why everyone really needs to pay attention because, we don't know what the end solution is going to be yet.
00:24:09.174 --> 00:24:09.684
Yeah, that's,
00:24:09.684 --> 00:24:12.355
and are you saying there's gonna be kind of a Cambrian explosion?
00:24:12.805 --> 00:24:13.684
Couldn't resist.
00:24:13.684 --> 00:24:15.480
couldn't resist my bad.
00:24:16.559 --> 00:24:17.430
Lots of explosion,
00:24:17.430 --> 00:24:18.839
self-promotion, my apologies.
00:24:19.319 --> 00:24:21.539
No, I agree with Jim.
00:24:21.539 --> 00:24:24.170
I think we'll see specialization start to occur.
00:24:24.170 --> 00:24:31.829
Right now we're just trying to get to the point where we can demonstrate quantum advantage across a reasonable number of, universal spaces.
00:24:32.339 --> 00:24:40.049
but as that becomes more general purpose, you'll then see it specialized and you'll see specific quantum modalities.
00:24:40.734 --> 00:24:45.234
Being used for certain problems and others being used for other problems.
00:24:45.595 --> 00:24:47.934
So I think there's room for all here to succeed.
00:24:48.444 --> 00:25:02.640
And, quite honestly, there's plenty of money to be made by all of the players, Most of the players that we're seeing today have a significant research and results that will turn into competitive advantage in specific areas.
00:25:02.730 --> 00:25:03.269
Yeah, and
00:25:03.269 --> 00:25:03.569
I,
00:25:03.569 --> 00:25:03.720
and
00:25:03.720 --> 00:25:12.779
there's one other area that, since Leonard, since you brought up the misconceptions, the other misconception we have to kind of address is that quantum is gonna break everything.
00:25:14.289 --> 00:25:14.579
Yeah.
00:25:14.904 --> 00:25:21.265
It is a, it is definitely a threat because,, traditional security algorithms can be broken by a quantum computer in record times.
00:25:21.315 --> 00:25:24.765
However, you have to remember that, IBM especially, that's been list, yeah.
00:25:24.765 --> 00:25:27.015
Kind of leading the charge on these nist.
00:25:27.525 --> 00:25:34.734
Algorithms that basically to actually, you know, create algorithms that quantum computers can't solve for new security algorithms.
00:25:35.035 --> 00:25:42.625
They've been working on this for over a decade and they continue, and these new NIST algorithms keep coming out every year, every couple of years.
00:25:42.954 --> 00:25:48.414
So, quantum computing is not designed to be an encryption, to be a hacker's gold mine.
00:25:48.865 --> 00:25:58.515
It is something that every company today has to be aware of and has to start thinking about, integrating these NIST algorithms because they need to be quantum safe today.
00:25:58.575 --> 00:26:10.444
Because there is the theory that, and I completely believe this, you can go and scrape, information today and possibly use quantum computers tomorrow to actually decipher the security algorithms.
00:26:10.714 --> 00:26:11.724
So that is.
00:26:12.259 --> 00:26:14.180
A part of the quantum discussion.
00:26:14.180 --> 00:26:16.400
It's not the part of the quantum discussion.
00:26:17.069 --> 00:26:33.269
And I, I think that maybe that's a topic that we talk about, separately because, one of the risks going forward is that, especially around this, quantum era security problem, that.
00:26:33.704 --> 00:26:40.045
Some of the thinking it is based on, fundamental misconceptions about quantum.
00:26:41.230 --> 00:26:53.140
And I think, you guys are doing a lot of, good work that will be a basis of clarifying a lot of those misconceptions because I think quantum computing is in general not well understood.
00:26:53.140 --> 00:26:54.759
They think that, for instance.
00:26:55.904 --> 00:27:01.724
Quantum computing will displace and replace accelerated computing, right?
00:27:02.244 --> 00:27:05.025
Traditional computing as you both, cited before.
00:27:05.335 --> 00:27:09.134
One of the things that I think might be, a good clarification.
00:27:09.759 --> 00:27:17.769
Education for the audience might be helping them understand what is the role of accelerated computing vis-a-vis quantum computing.
00:27:17.769 --> 00:27:28.630
Jim, you mentioned, that AI will be important in enabling and, scaling, helping, quantum computing to scale.
00:27:28.630 --> 00:27:34.690
There was a lot of talk at Quantum Day actually last year about how.
00:27:35.220 --> 00:27:45.549
AI and accelerated computing, could, enable or be a foundation for these control planes for these different modalities of, quantum, computing runtime.
00:27:45.549 --> 00:27:49.059
So, how does this all kind of interrelate with each other?
00:27:49.059 --> 00:27:52.345
Because, you do have Nvidia obviously making big investments.
00:27:53.279 --> 00:28:00.769
In the quantum ecosystem, but also positioning their technologies for quantum computing enablement.
00:28:02.059 --> 00:28:07.519
Maybe both of you can take some time to stitch that picture together for the audience.
00:28:08.134 --> 00:28:15.714
Well, you have to remember that we're dealing with quantum mechanics, so we're actually dealing with physics and physical systems here in understanding, quantum computing.
00:28:16.015 --> 00:28:26.244
You have to remember that classical computing has always been a part of that, but also has some manual tasks, like actually trying to calibrate the qubits, actually trying to make sure that they're going.
00:28:26.460 --> 00:28:33.339
Perform the way you hope they're going to perform, because they're always in a state of flux, and that's a huge, huge challenge.
00:28:33.619 --> 00:28:35.750
Nvidia and I credit Nvidia for this.
00:28:35.750 --> 00:28:41.000
Instead of trying to jump into the quantum computing fray, they've said, listen, we're going to enable it.
00:28:41.000 --> 00:28:42.619
We're gonna be the key enabler here.
00:28:42.829 --> 00:28:48.920
So they've created their icing, their Nvidia icing models for to help with that calibration on the front end.
00:28:49.470 --> 00:29:07.619
Basically, hopefully automate that to where you don't have, it's not a manual process, it's an automated process and a reiterative process for each run that you're running, but also on the backend for doing the error correction, which, as we've mentioned before, just getting the right information, ensuring that you have accurate information is a challenge.
00:29:07.619 --> 00:29:12.690
Today we're using a combination of suppression, mitigation, and error correction today.
00:29:12.930 --> 00:29:15.630
We hope we get to, and that error correction really isn't.
00:29:15.849 --> 00:29:17.230
That great at this point in time.
00:29:17.440 --> 00:29:30.500
So we're hoping we're gonna get to a point where, A, we have more steady states of the quantum, of the qubits, and B so we have higher fidelity, and B, we get to a point where we can actually do quantum error correction.
00:29:30.819 --> 00:29:34.390
With a very good understanding and very high speeds to be able to do that.
00:29:34.690 --> 00:29:52.819
So I mean, Nvidia, and this is a perfect application for it, trying, taking these mult multi parallel solutions, basically their GPUs and saying, listen, just like we power AI with this parallelism, we can power these tasks in the front end and back end of quantum computing.
00:29:52.829 --> 00:30:00.150
I credit them because I think it's going to be, what they're doing is not important for one specific modality or one specific partner.
00:30:00.390 --> 00:30:03.839
It's gonna be important for the entire industry in any modality that's out there.
00:30:05.130 --> 00:30:08.369
Yeah, I hats off to Nvidia for really.
00:30:08.950 --> 00:30:13.930
Putting themselves at the center, the conversation of quantum computing without having a quantum computer.
00:30:14.480 --> 00:30:17.059
that, that's pretty magical, accomplishment.
00:30:17.279 --> 00:30:20.130
I do wonder how they're gonna monetize that going forward.
00:30:20.539 --> 00:30:22.190
but they'll sell enough GPUs.
00:30:22.190 --> 00:30:23.000
it will definitely help them.
00:30:23.000 --> 00:30:24.049
But more importantly.
00:30:24.414 --> 00:30:30.654
It puts them at the center of the ecosystem for most quantum computing because everybody has to air correction.
00:30:30.684 --> 00:30:34.285
Everybody has to do calibration, q calibration.
00:30:34.674 --> 00:30:36.505
And so, that's an important first step.
00:30:36.505 --> 00:30:37.914
I don't think it's the last step.
00:30:38.585 --> 00:30:42.605
I'd like both of your reactions to the comment that I'm gonna make right now.
00:30:43.059 --> 00:30:48.660
the way I view it is, Nvidia is doing a really good job of positioning themselves in practical terms, right?
00:30:48.660 --> 00:30:48.750
Mm-hmm.
00:30:48.990 --> 00:30:56.329
Not just fluffy marketing terms as the picks and shovels for, quantum computing.
00:30:56.329 --> 00:31:01.519
So this hockey stick that you have here, Jim, there, the.
00:31:02.630 --> 00:31:18.174
The concept is that they will be able to, provide that, control plane, and enable these, execution environments, the design of these execution environments so that quantum computing can scale.
00:31:18.859 --> 00:31:32.529
The other thing that I also noticed, from conversations and presentations coming out of Quantum Day is that, accelerated computing is great for, discovering applications, right?
00:31:32.529 --> 00:31:34.480
Because some of these applications are not known.
00:31:35.329 --> 00:31:38.809
And it's not that you're going to use quantum computing to figure those out.
00:31:38.809 --> 00:31:45.779
You actually need to have, accelerated computing or AI to support that discovery process.
00:31:46.099 --> 00:31:48.980
much like what we see with, drug discovery, right?
00:31:48.980 --> 00:31:55.099
So there's a role there and, your question, Carl, how are the likes of Nvidia?
00:31:55.099 --> 00:31:58.670
Because I assume that they're gonna be more players than just Nvidia in this game.
00:31:59.210 --> 00:32:00.470
How will they monetize?
00:32:00.470 --> 00:32:06.380
Well, they would monetize by supporting these important activities in not only scaling.
00:32:07.025 --> 00:32:16.724
Quantum computing itself from in terms of progressing all these different modalities, but also the discovery of.
00:32:17.275 --> 00:32:22.605
the applications, but then also the design of the systems, right?
00:32:22.605 --> 00:32:23.505
And the circuits.
00:32:23.565 --> 00:32:26.025
That was the other thing that I took away as well.
00:32:26.025 --> 00:32:27.255
I think in general, you're right.
00:32:27.355 --> 00:32:29.694
it was really interesting article.
00:32:29.694 --> 00:32:31.464
I believe it's Wall Street Journal Suite.
00:32:31.974 --> 00:32:36.714
But it was talking about how Jenssen's investment strategy is not to pick winners.
00:32:36.785 --> 00:32:37.714
Don't pick winners.
00:32:37.755 --> 00:32:42.789
Invest in everybody who has a shot and eventually you'll hit one or two that really.
00:32:43.244 --> 00:32:44.714
Give you a good return on investment.
00:32:45.305 --> 00:32:47.345
and I think that's what he's doing right now with Quantum.
00:32:47.375 --> 00:32:50.315
He's saying, look, I don't know which one's gonna be best.
00:32:50.644 --> 00:32:55.444
In fact, I don't know which ones are gonna be best at which problems to Jim's point earlier.
00:32:55.934 --> 00:33:02.615
I'm gonna invest in all of them by giving them the tools they need to accelerate, fault tolerant computing era.
00:33:02.920 --> 00:33:08.230
Get there sooner, get there with mo with higher quality qubits and lower costs.
00:33:08.599 --> 00:33:10.849
right now cost is not in the equation, right?
00:33:10.849 --> 00:33:10.970
It is.
00:33:10.970 --> 00:33:12.019
Just can you get there?
00:33:12.019 --> 00:33:12.079
Yeah.
00:33:12.440 --> 00:33:17.890
But eventually, if you can get there, well then it'll, we'll determine the winner is well, how much it costs.
00:33:18.490 --> 00:33:22.359
And how manufacturable is it and how high quality is it?
00:33:22.359 --> 00:33:23.440
How dependable is it?
00:33:23.440 --> 00:33:24.609
All that stuff, right?
00:33:25.069 --> 00:33:29.819
all the Illities will come into play, but right now capability, is king.
00:33:29.880 --> 00:33:30.865
And so he's investing in everybody.
00:33:31.230 --> 00:33:31.289
Yeah.
00:33:31.779 --> 00:33:33.940
And expects to be able to make money along the way.
00:33:34.660 --> 00:33:35.815
Carl, You're getting soft.
00:33:36.039 --> 00:33:37.240
I'm out in the open.
00:33:37.240 --> 00:33:38.410
You can take a shot.
00:33:38.410 --> 00:33:38.769
Come on.
00:33:39.710 --> 00:33:44.640
one of the interesting things is that, with any new technology, there usually comes new business models.
00:33:44.670 --> 00:33:47.519
We don't know what those business models look like at this point in time.
00:33:47.910 --> 00:33:50.279
That's going to be, have to evolve over time.
00:33:50.309 --> 00:33:54.880
And, and like Nvidia, we saw them do an end-to-end solution for ai.
00:33:54.910 --> 00:33:55.295
We see.
00:33:55.730 --> 00:33:58.700
IBM doing the same thing for quantum computing.
00:33:59.029 --> 00:34:03.200
They're the only company I know of that has, well, D-Wave and others.
00:34:03.200 --> 00:34:09.500
But really, most of the modalities are relying on kind of a ecosystem partnerships to go through there.
00:34:09.500 --> 00:34:10.670
And I'm not knocking that.
00:34:10.670 --> 00:34:14.744
I think that's great too, but IBM puts so much in end to end from.
00:34:15.150 --> 00:34:25.360
The refrigeration solutions to the systems, to the wiring, to, they have invested billions and continue to in the entire thing, and now they've got a partnership with a MD.
00:34:25.644 --> 00:34:25.704
Yeah.
00:34:26.364 --> 00:34:29.425
Now, if you don't know, these two companies have worked together for decades.
00:34:29.844 --> 00:34:29.905
Yeah.
00:34:29.905 --> 00:34:33.204
On copper interconnects and silicon on insulator technology.
00:34:33.204 --> 00:34:42.324
They have helped pioneer, these two of these companies have helped pioneer some of the major breakthroughs we've had in semiconductor manufacturing and design over the past three or four decades.
00:34:42.625 --> 00:34:47.684
So having those two work together, I think is gonna be even, is gonna be a big opportunity.
00:34:48.135 --> 00:34:53.135
And trying to see what, because once again, now you've got another ai company that's.
00:34:54.320 --> 00:34:55.070
Fray there.
00:34:55.429 --> 00:34:59.929
And it'll be interesting to see how everything kind of meshes together as we get farther on down the road.
00:34:59.980 --> 00:35:05.460
and it's funny'cause some of the quantum company, quantum computer companies are promoting on-prem systems.
00:35:05.969 --> 00:35:09.719
I don't know if that's going to be the preferred business model.
00:35:09.719 --> 00:35:19.929
I still think that for the vast majority, cloud-based, kind of rental solutions are probably going to be the most likely scenario for quantum computing, especially for the next decade.
00:35:21.139 --> 00:35:23.965
I think the exception to that's probably in two areas.
00:35:23.965 --> 00:35:26.934
One is in financial services and the other is, government.
00:35:27.414 --> 00:35:30.985
you look at a lot of these companies right now, they're, they have two primary.
00:35:31.929 --> 00:35:33.550
Markets, customer segments.
00:35:33.695 --> 00:35:39.190
One, one is, is researchers and the other is national defense.
00:35:39.579 --> 00:35:39.670
Mm-hmm.
00:35:40.059 --> 00:35:49.695
and so national def both those areas, I think there is a role for on-prem quantum computing, even though it may be an internal self-hosted cloud.
00:35:50.175 --> 00:35:51.644
so kind of a hybrid model.
00:35:52.394 --> 00:35:52.635
Yeah.
00:35:52.635 --> 00:35:58.425
And you, it's important to note that, there's obviously a lot of investment from the research community too, matter of fact.
00:35:58.425 --> 00:35:58.514
Mm-hmm.
00:35:58.875 --> 00:36:07.184
IBM has been putting some of their systems out there in the research community, like, uh, believe the University of Cleveland or something like that, or mm-hmm.
00:36:07.425 --> 00:36:10.635
They've gotta be universities and research organizations around the world now.
00:36:10.905 --> 00:36:13.545
So it's gonna be interesting to see how this all plays out.
00:36:13.545 --> 00:36:17.054
And matter of fact, two of the companies are spin outs from Oxford University.
00:36:17.380 --> 00:36:29.760
So, it's, I I think we're probably, maybe not to the point where we saw transformer models and we saw LLMs with, ai, but we're getting there very quickly and by the end of the decade it's gonna get exciting.
00:36:30.210 --> 00:36:30.570
Yeah.
00:36:31.110 --> 00:36:32.639
Sounds a lot like super computing.
00:36:34.295 --> 00:36:34.585
It's
00:36:34.804 --> 00:36:35.034
how it's,
00:36:35.110 --> 00:36:36.730
super computing on steroids.
00:36:36.730 --> 00:36:37.630
Steroids, yeah.
00:36:37.630 --> 00:36:38.920
It's quantum computing.
00:36:38.920 --> 00:36:41.440
So, Carl, do you think we're gonna have quantum factories.
00:36:43.110 --> 00:36:44.039
Quantum factories.
00:36:44.099 --> 00:36:47.429
It sounds like Jim described, quantum infrastructure.
00:36:47.429 --> 00:36:47.670
Right.
00:36:47.670 --> 00:36:52.900
The whole supply chain and the build out large, data centers and or quantum centers.
00:36:52.949 --> 00:36:53.280
yeah.
00:36:53.324 --> 00:36:59.014
I think there will be quantum factories, because the cost of these systems is still quite high.
00:36:59.465 --> 00:36:59.554
Yeah.
00:36:59.554 --> 00:37:02.085
And so you'd amortize those costs over as many.
00:37:02.880 --> 00:37:05.159
Use cases and customers as as you can.
00:37:05.699 --> 00:37:19.824
And so building quantum factories where you have CPUs, GPUs, xus, and Q all basically providing services for applications we can barely imagine today.
00:37:20.525 --> 00:37:21.425
This has been great.
00:37:21.425 --> 00:37:23.485
This has really been, informative.
00:37:23.704 --> 00:37:30.480
I know I've learned a lot from both of you in this 30 minutes or so that we've had this chat.
00:37:30.480 --> 00:37:33.780
And, it's great to have an opportunity also to talk about Quantum Day.
00:37:34.199 --> 00:37:37.320
I, I thought, that it was really well done program.
00:37:38.144 --> 00:37:41.494
even some of the detail they went into Was incredible.
00:37:41.525 --> 00:37:41.795
I,
00:37:42.034 --> 00:37:42.335
yeah,
00:37:42.364 --> 00:37:47.764
I need to go back and watch it probably a dozen more times because some of the mathematics, especially in the last presentation,
00:37:48.695 --> 00:37:48.905
yeah.
00:37:48.905 --> 00:37:49.894
I was in transit.
00:37:49.894 --> 00:37:52.625
I was trying to listen to this stuff as I was on a flight.
00:37:52.625 --> 00:37:53.284
So,
00:37:53.344 --> 00:37:56.255
yeah, the professor from Israel was just, oh my god.
00:37:56.255 --> 00:37:58.715
He, his mathematics depth was incredible.
00:38:00.815 --> 00:38:02.434
So, gentlemen, Thank you so much.
00:38:02.434 --> 00:38:03.300
It was really our pleasure.
00:38:03.485 --> 00:38:05.224
Really, really great, great stuff.
00:38:05.224 --> 00:38:18.364
And, I think the, audience will undoubtedly find it incredibly grounding and, a good orientation as to how they should, view the aqua, the Cambrian explosion.
00:38:19.034 --> 00:38:20.445
That is quantum explosion.
00:38:20.954 --> 00:38:21.764
The what?
00:38:21.974 --> 00:38:23.144
The quantum explosion.
00:38:23.144 --> 00:38:24.855
The quantum explosion.
00:38:25.215 --> 00:38:27.315
Nah, I gotta make it Cambrian.
00:38:27.510 --> 00:38:28.000
Cambrian.
00:38:29.585 --> 00:38:30.144
Cambrian.
00:38:30.489 --> 00:38:32.170
Cambrian Quantum explosion.
00:38:32.559 --> 00:38:32.860
I've got
00:38:32.860 --> 00:38:33.219
all these qu
00:38:33.550 --> 00:38:34.329
the qc,
00:38:34.960 --> 00:38:35.289
I've got all
00:38:35.289 --> 00:38:37.570
these Cambrian fossils.
00:38:37.570 --> 00:38:39.519
I can't change'em all to quantum fossils.
00:38:43.514 --> 00:38:56.534
So, thank you gentlemen and everyone, if you're interested in tapping into, Jim's research calls research, reach out to both of these gentlemen and their respective firms.
00:38:56.534 --> 00:39:01.445
Of course, Jim McGregor is with tedious research.
00:39:02.409 --> 00:39:02.829
How's that?
00:39:02.829 --> 00:39:02.980
It's
00:39:03.144 --> 00:39:03.465
pretty good.
00:39:03.905 --> 00:39:03.925
Pretty good.
00:39:04.449 --> 00:39:04.659
Great.
00:39:04.659 --> 00:39:04.719
Yeah.
00:39:05.320 --> 00:39:05.739
I mean, sad.
00:39:06.280 --> 00:39:12.309
Hey, I'm an honorary Scotsman now because my nickname, or at least my Scottish name is, Liam, right?
00:39:12.400 --> 00:39:13.030
Yes, exactly.
00:39:14.650 --> 00:39:17.320
I gotta live up to that honorary designation.
00:39:17.320 --> 00:39:19.059
So yes, reach out to him at cheer.
00:39:19.280 --> 00:39:22.940
research@www.teusresearch.com.
00:39:22.940 --> 00:39:28.440
He's also on Forbes and he's quite active He Times and other publications.
00:39:28.519 --> 00:39:36.565
And then of course we have Carl Fre of Cambrian Quantum AI Research.
00:39:37.309 --> 00:39:39.530
Put Hy keep putting hyphens on the company name.
00:39:39.530 --> 00:39:42.949
Cambrian AI hyphen Quantum hyphen Research.
00:39:43.000 --> 00:39:49.264
So catch him at www.cambrianaium.com.
00:39:50.349 --> 00:39:51.159
Did I get that right?
00:39:51.545 --> 00:39:52.085
That's right.
00:39:52.420 --> 00:39:52.900
Yeah.
00:39:52.900 --> 00:39:54.579
And then, he has a substack.
00:39:54.610 --> 00:40:00.079
He has, a, he's on Forbes and he's on LinkedIn.
00:40:00.139 --> 00:40:04.860
He's also public, well published, across a number of publications and outlets.
00:40:05.309 --> 00:40:08.840
And of course, please subscribe to our podcast here.
00:40:08.889 --> 00:40:18.880
We do a, A, an episode every month to cover everything that's happening that you should know about, in the world of.
00:40:20.034 --> 00:40:23.335
AI, semiconductors and quantum.
00:40:23.335 --> 00:40:43.994
So, make sure to check us out on YouTube, subscribe, like, share, and then check out the audio version on Buzz Sprouts and listen to us on your favorite podcast platform and subscribe to the next curve research portal@wwwnextcurve.com for the tech and industry insights that matter.
00:40:44.235 --> 00:40:45.315
Gentlemen, thank you so much.
00:40:45.315 --> 00:40:45.639
This is awesome.
00:40:46.500 --> 00:40:46.949
Thank you.
00:40:47.309 --> 00:40:48.900
We will see you by the end of the month.
00:40:49.179 --> 00:40:49.900
talk to you soon.
00:40:49.900 --> 00:40:50.139
Yep.
00:40:50.235 --> 00:40:50.525
Yeah.
00:40:51.184 --> 00:40:51.605
Cheers.
00:40:52.105 --> 00:40:52.525
Cheers.
