SANTA CLARA, California, April 17, 2026, 11:42 PDT.
Nvidia took a bigger step into robotics this week, expanding its partnership with Cadence Design Systems. The AI chip giant is searching for new growth avenues outside of the data-center servers that fueled its climb. TSMC and ASML, meanwhile, released new forecasts showing demand for AI chips and their production gear remains strong.
Investors are turning up the heat on Microsoft, Meta, Amazon and Alphabet, demanding evidence that all their AI investments are paying off. As Reuters noted, those giants are on track to pour over $600 billion into data centers this year. Meanwhile, TSMC projected its 2026 revenue will jump more than 30% in dollar terms. ASML also hiked its 2026 sales target to between 36 billion and 40 billion euros, just ahead of Big Tech earnings set for April 29.
Nvidia was up roughly 1.2% to $200.79 late Friday morning in California, pushing its market cap to around $4.53 trillion. Shares had previously topped the $5 trillion mark on Oct. 29—the same day the Nasdaq set its all-time high, a level it revisited this week.
Speaking Wednesday at CadenceLIVE in Santa Clara, Nvidia CEO Jensen Huang said his company is teaming up with Cadence “across the board” on robotics. The collaboration matches Cadence’s physics engines—tools for simulating material behavior—with Nvidia’s AI models, letting robots practice inside computer simulations before they’re let loose outside. Reuters
Cadence CEO Anirudh Devgan says improved generated data translates into better models. The company added in a statement that its expanded partnership stretches into digital twins—virtual stand-ins for real-world systems—and engineering workflows, which, according to Cadence, might see certain tasks speed up by as much as 100 times.
Nvidia is eyeing physical AI—think robots and autonomous systems that interact with the real world—as its next big growth driver, right up there with training and inference chips. Inference, the process where a trained AI responds to live prompts, is seeing a surge: Reuters reported this week that the need for advanced processors tackling those tasks is rising.
The demand signal isn’t just a Nvidia story—AMD and Broadcom also depend on TSMC’s top-tier capacity. TSMC CEO C.C. Wei described AI demand as “so strong” that capacity remains tight. ASML’s Christophe Fouquet warned that chip demand could soon outpace supply. Still, Reflexivity’s Giuseppe Sette said ASML’s latest results give semiconductors “a favorable picture,” bubble concerns notwithstanding. Reuters
Nvidia’s growth hasn’t escaped notice in Washington. This week, Senator Elizabeth Warren questioned the Energy and Defense departments over Nvidia’s acquisition of SchedMD. SchedMD’s Slurm software runs about 60% of the world’s supercomputers, and Warren warned the deal might hand Nvidia “disproportionate control over a chokepoint.” Nvidia responded that Slurm remains open source and said it’s still investing in its development. Reuters
There’s a straightforward risk here: AI investments must start delivering real returns. Should cloud firms come up short on that front—or if export restrictions and production slowdowns deepen—then Nvidia’s bets on robotics and software might not be sufficient to offset a sluggish main chip business.
The core business remains massive. Nvidia posted a record $68.1 billion in quarterly revenue back in February, with $62.3 billion coming from data center alone. New moves in robotics, simulation, and software are building on top of an already sprawling AI infrastructure operation.
