Nvidia’s Two-Front Strategy: Nuclear Energy and a $30 Billion Data Bet

Nvidia is widening its footprint beyond the chip business, striking a nuclear energy partnership while pouring capital into a data infrastructure startup valued at $30 billion. The moves come as the company’s stock hovers near a 52-week high, with investors weighing the broader tech selloff against a string of strategic announcements.

The chipmaker has teamed up with nuclear startup Oklo and the Los Alamos National Laboratory to power next-generation data centers with small modular reactors. The collaboration will use Nvidia’s AI models to accelerate materials research and reactor construction, addressing a pressing need for stable, high-capacity energy sources as hyperscalers expand their AI clusters. Oklo shares jumped 5% on the news, though Nvidia’s own stock barely budged.

On the data front, Nvidia participated in Vast Data’s Series F funding round, which closed on April 22, 2026, at a $30 billion valuation — more than triple the startup’s worth at the end of 2023. The $1 billion-plus round was led by Drive Capital and Access Industries, with Fidelity and NEA also joining. Vast Data, founded in 2016, builds software infrastructure for managing massive datasets tailored to AI workloads. Its customers include CoreWeave, Mistral, the U.S. Department of the Air Force, and Cursor. The company’s cumulative order backlog exceeds $4 billion, with annual recurring revenue north of $500 million.

The investment fits a broader pattern for Nvidia, which has systematically built stakes across the AI technology stack this year, participating in rounds for OpenAI, Anthropic, xAI, Nscale, and Wayve. Vast Data has positioned itself for an IPO, hiring former Shopify CFO Amy Shapero about 18 months ago. The CEO expects the company to be ready for a public listing by the end of 2026, though the final decision remains pending. For Nvidia, a successful Vast Data IPO would validate its strategy of financing the entire ecosystem that makes its chips indispensable.

Should investors sell immediately? Or is it worth buying Nvidia?

Meanwhile, Nvidia wrapped up its presence at Hannover Messe 2026, the industrial trade fair that ran from April 20 to 24. Alongside partners including Siemens, SAP, ABB, and Microsoft, the company demonstrated how AI is transforming factory operations, robotics, and engineering processes. Hexagon Robotics, for instance, uses Nvidia’s physical AI platform to train robots faster, with one system slated to take over assembly work at BMW’s Leipzig plant.

Internally, Nvidia is running a notable experiment: more than 10,000 employees are already using OpenAI’s GPT-5.5 model through the Codex application for automated programming. The infrastructure runs on Nvidia’s own GB200 NVL72 systems from the Blackwell generation. Debugging processes that once took days now complete in hours, with the GB200 systems delivering 50 times more token output per second per megawatt than previous models.

Despite these developments, Nvidia shares slipped about 2% on Thursday, closing in Frankfurt at €172.48. That’s roughly 13% higher than a month ago but still nearly 4% below the 52-week high. The relative strength index sits at 68, suggesting a mildly overbought condition. Rising oil prices and a sharp pullback in the software sector weighed on sentiment, while competitive pressure mounts: AI chip startups have raised about $8.3 billion in 2026, nearly matching the full-year record of $8.5 billion, with many developing specialized chips for inference applications — the very area where Nvidia must defend its dominance long-term.

The company reports quarterly earnings on May 20, having guided for revenue between $76.4 billion and $79.6 billion in the current quarter — well above initial analyst expectations of roughly $66 billion. Globally, AI companies have raised around $280 billion in capital this year, underscoring the scale of the market Nvidia is betting on.