Nvidia's Bold $2B Investment Boosts AI Ambitions
Weekly Nvidia Brief — January 22–28, 2026
A concise roundup of truly new developments this week. Each item notes how it fits Nvidia’s bigger vision and what it means for AI overall.
- Nvidia puts $2B more into CoreWeave to speed AI factory build‑out
- What’s new: On January 26, Nvidia invested $2B in CoreWeave and broadened the partnership so CoreWeave adopts Nvidia CPUs, storage, and multiple GPU generations while targeting over 5 GW of AI data centers by 2030. CoreWeave’s shares jumped on the news. (nvidianews.nvidia.com)
- Nvidia’s bigger vision: Directly advances the “AI factory” strategy—aligning chips, networking, and software with a cloud partner that can soak up demand and help global customers stand up capacity fast. (nvidianews.nvidia.com)
- AI as a whole: Adds near‑to‑midterm compute supply for model training and inference, easing capacity crunches and supporting broader deployment of agentic and domain‑specific AI. (ft.com)
- China reportedly clears purchases of Nvidia’s H200 chips
- What’s new: After a week of customs holdups, multiple reports today say Beijing has approved several hundred thousand H200 units for major tech firms—potentially around $10B of orders—under defined conditions. Markets viewed it as a positive catalyst for Nvidia. (wsj.com)
- Nvidia’s bigger vision: A conditional reopening of China sustains demand while the company scales Rubin and compliance tooling; it also reduces uncertainty around near‑term shipments. (barrons.com)
- AI as a whole: Geopolitics continues to gate where top‑tier compute lands. If approvals hold, Chinese labs regain access to state‑of‑the‑art training hardware, accelerating regional model development. (wsj.com)
- Davos follow‑through: “AI buildout” is accelerating—even older GPUs are tight
- What’s new: In Davos conversations this week, Jensen Huang said GPU rental spot prices are rising—even for two‑generation‑old parts—framing this as the largest infrastructure buildout in history. The WEF recap underscored Nvidia’s “five‑layer” view: energy, chips, cloud, models, apps must scale together. (pcgamer.com)
- Nvidia’s bigger vision: Validates the end‑to‑end platform push (from GPUs and networking to CPUs and software) to remove bottlenecks and keep utilization high across the stack. (weforum.org)
- AI as a whole: Persistent compute scarcity shapes model choices and deployment economics, pushing the industry toward efficiency (smarter routing, MoE, better inference stacks) and diversified suppliers. (pcgamer.com)
Sources
- Nvidia–CoreWeave investment and partnership details: NVIDIA Newsroom; Financial Times; Business Insider; TechCrunch. (nvidianews.nvidia.com)
- China’s reported H200 approvals and market reaction: Wall Street Journal; Barron’s; Times of India. (wsj.com)
- Davos remarks and framing of the AI buildout: PC Gamer recap; World Economic Forum summary. (pcgamer.com)
End of brief.