New York, February 4, 2026, 07:25 EST — Premarket
Nvidia (NVDA.O) shares edged up roughly 0.4% to $181.05 in U.S. premarket trading Wednesday, rebounding slightly after a 2.8% drop the day before. The decline came amid a broader selloff hitting software and data stocks, dragging down major tech names.
The wobble matters because Nvidia has become a bellwether for spending on artificial intelligence hardware, and investors are now parsing a shift in focus from training AI models to running them. OpenAI has been exploring alternatives for some “inference” work — the phase where a trained model generates answers — even though Nvidia still dominates chips used to train large models. Nvidia insists customers pick its chips for inference due to “best performance and total cost of ownership at scale,” highlighting the overall cost of running systems, not just the chip price. Reuters
Another key driver is Nvidia’s significant stake in OpenAI. According to a source, Nvidia is close to sealing a deal to pump around $20 billion into OpenAI’s newest funding round. The ChatGPT creator aims to raise up to $100 billion, targeting a valuation near $830 billion.
Nvidia CEO Jensen Huang, addressing a Cisco AI conference in San Francisco, dismissed concerns that AI will replace software tools. He labeled those fears “the most illogical thing in the world,” saying new AI systems will build on existing tools instead of starting from scratch. Reuters
The backdrop remains shaky. U.S. stock index futures held mostly steady as investors remained cautious following a steep drop in software and cloud stocks, while markets awaited more earnings from mega-cap tech firms. “The move was more about traders stepping back from an ostensibly crowded AI trade,” said Jake Behan, head of capital markets at Direxion. Reuters
Chip stocks saw mixed signals from earnings across the sector. Advanced Micro Devices (AMD.O) dropped almost 7% in premarket after its quarterly sales forecast missed expectations, raising doubts about how fast it can close the AI accelerator gap with Nvidia. Bernstein analyst Stacy Rasgon pointed out that near-term AI figures “are not really inflecting.” Reuters
Intel (INTC.O) CEO Lip-Bu Tan confirmed the company’s move into GPUs — the chip type Nvidia made famous — targeting data centers. “It’s tied in with the data center,” Tan told Reuters, noting Intel is collaborating with clients to pinpoint their requirements. Reuters
The downside risk for Nvidia remains clear: should major clients accelerate efforts to diversify their inference hardware, or if the shift from richly valued AI stocks intensifies, the shares could drop sharply without any fresh company news. On top of that, a significant OpenAI investment—if it goes through—might raise fresh doubts about how Nvidia is allocating capital, especially as investors are increasingly focused on the returns from AI investments.
Nvidia investors are eyeing February 25, the date when the company will release its fiscal 2026 fourth-quarter and full-year results.
