Sequoia Capital-backed chip startup Nuvacore is raising funds at a valuation of about $2.5 billion, sources told reporters on October 10, in the latest sign that investor appetite for AI hardware has fully returned after a decade of venture capital treating chips as too slow and too capital-intensive to bother with.
Nuvacore's founder is a former Apple executive who sold his prior startup, Nuvia, to Qualcomm for $1.4 billion in 2021. Earlier this year, the company announced a seed funding round led by Sequoia Capital. The new raise would value a company that has not yet released a product at more than double the price Qualcomm paid for its founder's last startup, a number that would have been unthinkable in the chip-startup winter of the late 2010s.
The timing is no accident. Investors poured roughly $10.7 billion into semiconductor startups in the first five months of 2026 alone, already outpacing the $12.2 billion raised globally in all of 2025, according to Crunchbase data. Demand pressure is visible at the incumbents too. Intel said earlier this year that it could not manufacture enough chips to meet demand and was selling parts it had previously judged too faulty to ship, while AMD described swelling CPU sales on a recent earnings call. Shares of both companies have risen more than 180 percent this year.
Nvidia's own trajectory is part of the backdrop. The company launched a CPU called Vera this year, and CEO Jensen Huang has said Nvidia expects to sell $20 billion worth of Vera CPUs in the fiscal year ending in January. When the category leader is booking tens of billions from a new processor line, investors start looking for the next founding team with a credible chip roadmap.
The $2.5 billion figure should still be read with care. A valuation on a pre-product chip startup is a bet on the team and the roadmap, not on shipped silicon, and hardware timelines remain brutally long compared with software. But the broader signal is hard to miss: capital has decided that compute is the scarce commodity of the AI era, and it is willing to pay software-scale multiples to own a piece of the silicon that delivers it.