Six-month-old chip startup Nuvacore is seeking hundreds of millions of dollars at a roughly $2.5bn valuation, despite not yet having a product. The reported fundraising is a striking bet on demand for data-centre CPUs as AI infrastructure expands, and on a design strategy that has yet to produce a chip.
Watch Desk analysis
What happened
Reuters reports, citing two people familiar with the fundraising, that the round has not closed and that its size and valuation could change. Nuvacore was previously reported to be raising $200m or more. The San Jose startup was founded by Gerard Williams, John Bruno and Ram Srinivasan; Williams previously led Nuvia, which Qualcomm acquired for $1.4bn in 2021.
Nuvacore says it plans to develop a data-centre processor by building its core functionality before committing to an instruction-set architecture such as x86 or Arm. The company argues that this approach could help engineers optimise for common data-centre and AI workloads. It has not yet developed a product, and declined to comment to Reuters. Read the Reuters report.
Why it matters
The fundraising points to investors backing a new CPU contender before it has silicon to show. CPUs already handle work around AI accelerators, including coordinating parts of data-centre operations and running software agents. Nuvacore is entering a market long dominated by Intel and AMD, while Arm-based processors from companies including Amazon and Microsoft are also making inroads.
Reuters reports that investors put about $10.7bn into semiconductor startups in the first five months of 2026, already approaching the $12.2bn raised across all of 2025. That is a lively backdrop for a company seeking a multibillion-dollar valuation, but funding appetite is not the same thing as a working processor or a customer order.
Our read
The interesting wager is not simply that AI needs more chips. It is that a new company can design a competitive data-centre CPU by delaying a choice that has long shaped processor development. That could give Nuvacore room to optimise for its target workloads, or leave it with the difficult job of turning an unusual design plan into manufacturable hardware. At this valuation, the engineering milestones matter rather more than the pitch deck.
What to watch
- Whether the fundraising closes, and at what valuation and size.
- When Nuvacore commits to an architecture and publishes technical details.
- Whether it produces a working processor and wins data-centre customers.
Discussion spark: Should investors back a chip startup at a multibillion-dollar valuation before it has a product, or is that a reasonable bet on the AI infrastructure market?
Sources and evidence
- Source update (11 October 2026, 12:49 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.