What does NVIDIA's $26B AI investment cover?

NVIDIA plans to invest $26B over 5 years in open-weight AI models, including AI Factory infrastructure (Blackwell GPUs, NVLink, liquid cooling), CUDA software ecosystem (400+ libraries, Dynamo AI OS, NIMs), and inference technologies (FP4, KV Cache orchestration).

Why did NVIDIA choose open-weight models?

Open models drive more GPU usage and deepen CUDA ecosystem lock-in, essentially using free model-layer offerings to bind users to NVIDIA's hardware and software stack.

How does this affect OpenAI?

When NVIDIA offers competitive open-source models, enterprise customers may have less incentive to pay premiums for closed-source APIs, creating indirect competitive pressure on OpenAI and Anthropic.

— AI DAILY

Sources

AI Digest
Investing.com