Nvidia’s grip on the world’s fastest machines is as tight as ever, yet a Chinese upstart has just snatched the outright performance crown. The company powered 80% of the 500 most powerful supercomputers on the latest TOP500 list, and for newly installed systems that share jumps to 90%. But the top spot now belongs to LineShine, a purely CPU-based system from China that became the first machine of its kind to break the two-exaflop barrier — without a single Nvidia chip inside.
That contrasting picture set the tone for today’s annual shareholder meeting. Investors are clamoring for details on the Blackwell production ramp and how the company plans to deploy its ballooning cash flows. Nvidia is also forging ahead with its biggest European capacity expansion yet: 35 new AI supercomputers are under construction across the continent, slated to go online by year-end and serve millions of researchers.
On the same day as the shareholder gathering, Nvidia unveiled its most ambitious pharmaceutical play. At the BIO International Convention in San Diego on June 23, the company launched the BioNeMo Agent Toolkit, a platform that bundles more than a decade of Nvidia libraries, tools and open models for biosciences. Kimberly Powell, Nvidia’s vice president for healthcare, framed the move as a turning point for the $300 billion R&D industry: the question is no longer whether AI can help science, but whether science has the right tools to use it.
The toolkit is designed to be vendor-agnostic and is already being deployed by a striking list of early adopters: Dassault Systèmes, Databricks, Lilly, Schrödinger, Snowflake and UW Medicine Institute for Protein Design. Anthropic and OpenAI are integrating it to bring agent-based workflows into drug discovery. More than 50 companies are using BioNeMo for tasks ranging from protein structure prediction to generative chemistry and biomarker identification. The platform is available immediately on Nvidia’s developer resources and GitHub, with no geographic restrictions — China included.
Should investors sell immediately? Or is it worth buying Nvidia?
On the financial front, the numbers remain enormous. Data center revenue hit $75.2 billion in the first quarter of fiscal 2027, up 92% year over year. Nvidia has guided for $91 billion in the current quarter. Full fiscal 2026 closed at $215.9 billion, a 65% jump from the prior year. Yet the stock has struggled to reflect that strength. Shares changed hands at around €176 recently, roughly 13% below the 52-week high of €202.50 set in May. The entire chip sector — including Nvidia, AMD and Micron — is under a sell-off cloud.
One worrying sign comes from the market for computing power. On the Kalshi prediction platform, traders are betting on falling prices for Nvidia’s B200 chips. Rental costs have slipped from over $6 per hour in late May to around $4 per hour now. A sustained decline could squeeze profit margins and amplify investor unease.
Two structural risks linger in the background. Supply constraints on HBM memory and advanced packaging continue to throttle production. And Nvidia’s next leg of growth depends on whether AI investments shift from pure training towards revenue-rich inferencing applications. Meanwhile, US export restrictions on China remain a wild card for revenue geography.
Despite the headwinds, Bernstein has held onto its buy rating. The real test will come with the next earnings report for the second quarter of fiscal 2027. If Nvidia delivers on that $91 billion target and addresses the China export question, today’s sell-off may well be remembered as a healthy correction. For now, the company is betting that drug molecules and supercomputers can both keep the growth engine humming.
Ad
Nvidia Stock: Buy or Sell?! New Nvidia Analysis from June 24 delivers the answer:
The latest Nvidia figures speak for themselves: Urgent action needed for Nvidia investors. Is it worth buying or should you sell? Find out what to do now in the current free analysis from June 24.
Nvidia: Buy or sell? Read more here...
