The Board Room
DeepSeek is rewriting its core code for Huawei's CANN framework
Jensen Huang is publicly alarmed. Simultaneously, insurance carriers are quietly exempting AI workloads from cyber and E&O coverage, meaning your organization is now self-insuring every AI-related liability — potentially without knowing it.
The CUDA Moat Cracks — DeepSeek Moves to Huawei Chips
DeepSeek is migrating core code to Huawei's CANN framework for the Ascend 950PR. Jensen Huang is publicly alarmed. If V4 runs competitively, US export controls lose their teeth and Nvidia's 95%+ GPU share era ends. Every AI infrastructure strategy needs a Plan B.
AI Risk Goes Uninsurable — Carriers Drop Coverage
Insurance carriers are categorically excluding AI workloads from cyber and E&O policies, citing unpredictable outputs. This isn't a pricing adjustment — it's a withdrawal from AI risk transfer. Enterprises running AI at scale are now self-insuring every AI liability without realizing it.
AI Coding: The 3-8x Productivity Illusion
Waydev data across 50 companies and 10,000+ engineers: AI-generated code shows 80-90% initial acceptance but only 10-30% after revision churn — a 3-8x gap between boardroom metrics and production reality. Organizations scaling AI coding tools are accumulating technical debt while reporting productivity gains.
Compute Surplus Becomes M&A Currency
xAI is converting its compute stockpile into acquisition leverage — selling capacity to Cursor while positioning for vertical integration. OpenAI acquired TBPN and eyes consumer hardware. Vox Media selling brands piecemeal. AI infrastructure surplus is becoming the new M&A currency; if you have enterprise distribution, you're a target.
Agent Identity & Commerce Rails Forming
Two competing agent payment protocols emerged: x402 (Coinbase, integrated by Google/Cloudflare/Vercel) and MPP (Stripe, $0.003/txn). Actual volume is $1.6M/month after filtering 93% wash trading. World ID signed Zoom, Tinder, DocuSign, Ticketmaster, Eventbrite — 'proof of human' is becoming infrastructure, not a feature.
DeepSeek on Huawei Chips: The Most Consequential AI Supply Chain Threat Since Export Controls
Jensen Huang's public alarm about DeepSeek is not corporate posturing — it's a CEO watching Nvidia's most durable competitive advantage face its first credible threat. DeepSeek is actively rewriting its core code for Huawei's CANN framework, preparing to run its V4 multimodal model on the Ascend 950PR chip. If it runs competitively, the cascade effects are severe.
If China's leading AI lab can build frontier models without American chips, US export controls lose their strategic teeth — and every company that assumed Nvidia infrastructure was the only game in town needs a Plan B.
Why This Is Different From Previous China Chip Narratives
Previous attempts to build non-CUDA AI stacks failed because the software ecosystem was too shallow. DeepSeek is the first frontier-class lab committing engineering resources to making a non-Nvidia stack work at the model level, not the research level. The CUDA flywheel — where developers build on CUDA because everything else is inferior, which makes CUDA more dominant — faces a scenario where a top-tier model proves you can cross the moat. That proof point changes the calculus for every other lab and every government evaluating chip sovereignty.
Second-Order Effects to Model
- US export controls as leverage: If Ascend 950PR proves sufficient for frontier training, the primary policy instrument the US uses to maintain AI leadership loses efficacy. The geopolitical implications extend beyond tech into trade negotiations and alliance structures.
- Nvidia's market share: The 95%+ GPU dominance era doesn't end overnight, but the perception of inevitability cracks — and perception drives infrastructure procurement decisions 12-18 months out.
- Multi-vendor AI infrastructure: The Cerebras IPO filing further confirms investors see room for multiple AI chip players. Your infrastructure team should be testing portability assumptions now, not after a market shift forces it.
- Talent and knowledge flow: DeepSeek's engineering effort creates institutional knowledge about non-CUDA optimization that will diffuse across China's AI ecosystem, compounding the threat over time.
What Makes This Actionable Now
DeepSeek V4 hasn't shipped yet — this is still a developing scenario. But the planning window is now, not after results are published. Companies that audit chip dependencies, test model portability across frameworks, and build vendor-diversification clauses into infrastructure contracts this quarter will be positioned. Those that wait for V4 benchmarks will be reacting from behind.
The convergence with AI startup valuations decoupling from fundamentals — Cursor at $50B, DeepSeek's first outside round at $10B, a 4-month-old company raising $500M at $4B — suggests that capital markets are pricing in a world with multiple viable AI hardware ecosystems. Whether that's prescient or premature, your infrastructure strategy should hedge for both outcomes.
The US AI supply chain moat is cracking — DeepSeek migrating to Huawei chips is the first credible proof that frontier AI can be built without American hardware — while at home, insurance carriers are quietly dropping AI coverage from cyber policies, your AI coding productivity metrics are 3-8x inflated versus production reality, and compute surplus is becoming the new M&A currency. The three audits you need this quarter: chip dependency, insurance exposure, and real (not reported) AI coding productivity.