Nvidia announced on Monday that 35 of its AI supercomputers are currently being built across 23 European countries, marking what the company calls Europe's largest one-year increase in supercomputers. The infrastructure will grant over 3 million researchers access to advanced computing tools for fields including basic science, healthcare, clean energy, quantum computing, and climate science. The systems run on Nvidia's Blackwell and Hopper platforms, with over 800 AI exaflops of computing capacity deployed or disclosed since last year.
CEO Jensen Huang emphasized, "AI is the new instrument of science," as Nvidia also extends its reach into robotics safety with the Nvidia Halos platform, now adopted by Agility Robotics for humanoid robots used by Amazon and others. Additionally, Huang addressed the China International Supply Chain Expo via video, underscoring China's role as a global technology and industry hub.
Despite these expansions, prediction market traders have grown bearish on Nvidia's stock (NVDA). The price of B200 compute leasing—a barometer of demand—fell to $4.22 per hour down from $6.11 in May. Polymarket traders assign only a 3% probability of NVDA reaching $236, a 20% chance of hitting $220, and a 62% chance of falling to $204 by June 26. Kalshi contracts reflect similar caution. NVDA closed Monday at $208.78, down 1.20%, under its 20-day moving average but still above the 50-day and 100-day averages. Analysts maintain a Buy consensus with an average target of $323.83, while broader chip stocks rallied, with Micron gaining nearly 7%.
The next major catalyst is Nvidia's earnings report expected around August 26, 2026, with Wall Street forecasting EPS of $2.06 on $91.70 billion revenue.
