NIO's Massive Software Update, Li Auto Goes All-In on Robots & Memory Chip Crisis

NIO ships to 460K vehicles, Li Auto goes all-in on robots, memory crisis hits everyone. This is execution vs vision vs reality.

NIO NWM UPDATE (460K+ VEHICLES):

• Major "human-like" driving update using closed-loop reinforcement learning
• Learns from REAL human driving, not just experts
• Battery swap navigation: industry-first piloted driving to 2,000+ stations
• Shenji in-house chips (no NVIDIA delays)
• EXECUTION AT SCALE despite sales struggles

LI AUTO ROBOT PIVOT (LEAKED INTERNAL MEETING):Jan 26 all-hands: Li Xiang announces humanoid robot push

Key Points:

• 2026 = last year to become top AI company
• Only 3 global companies will master foundation models + chips + OS + embodied intelligence
• Li Auto will be one
• Restructuring: cars + robots = "hardware ontology team"
• Aggressive hiring: "bring back employees who left for robot startups"
• Multiple robot R&D roles posted

Context: Sales 500K (2024) → 400K (2025), -20%. Pure EV struggling. Is this genius or desperation?

MEMORY CHIP CRISIS (AFFECTS ALL):DDR4/DDR5 prices +40-70%, adding 1,000-2,000 yuan per vehicle

Stats:

• Li Auto: <50% fulfillment rate possible 2026
• Lei Jun: 70% Q1 increase projected
• 200K yuan EV: +800 yuan memory cost
• Automotive = <10% global DRAM (AI gets priority)

Why: AI eating HBM supply, DDR4→DDR5 transition, car memory needs +40% (ADAS)

Will prices rise? NO (short term). Competition too fierce. Companies absorbing... for now.

THE CONNECTIONS:

• NIO: Executing but bleeding costs (memory + swap + international)
• Li Auto: Pivoting because core struggling
• Memory: Great equalizer, 500K+ scale = survival edge

Not hype phase. GRIND phase. Who actually survives?

COURTSIDE FINANCIAL - Hosted by Obi