Tenstorrent raises $1.5 billion after takeover interest from Intel and Qualcomm

Tenstorrent has closed a $1.5 billion Series E round at a $6.2 billion post-money valuation, the company said Monday. It chose fresh capital over a sale. Earlier this year it drew takeover interest from Intel and Qualcomm at valuations of as much as $5 billion, according to Bloomberg reporting cited by TechCrunch.

Chief executive Jim Keller told Bloomberg the company has 1,150 employees, spread across offices in Santa Clara, Austin, Toronto, Belgrade, Tokyo and Bangalore, according to the company. Tenstorrent plans to spend the money on Galaxy server manufacturing and commercial deployments, new generations of its AI and RISC-V CPU designs, and global hiring, TechTarget reported.

Keller told Bloomberg the decision to stay independent came after Tenstorrent showed its Galaxy servers could run DeepSeek's models. Cailian Press, citing that interview, dated the demonstration to May and described it as the turning point.

The price has moved fast. Tenstorrent's December 2024 round raised more than $693 million at a $2 billion pre-money valuation, led by Samsung Securities and AFW Partners, according to Reuters. TechCrunch puts that round's valuation at $2.6 billion; TechTarget calculates an implied post-money figure of roughly $2.7 billion. On either number, the valuation has more than doubled in under two years. It remains modest beside Etched, another AI chip startup, which was fielding investment offers at valuations of $40 billion or more, TechCrunch reported this month.

Trace Capital, a firm founded this year by Quiet Capital co-founder Ben Mahdavi, led the round with Innovation X. Fidelity, T. Rowe Price, Morgan Stanley, ARK Invest, Sandisk, Cadence Ventures and Tether Investments also took part, according to the company's announcement. One name on the investor list is also a customer: Virtu Financial, which has deployed Galaxy servers on-premises to run AI agents supporting trading and operations, TechTarget reported.

Tenstorrent is aiming at inference, the work of running trained models, rather than trying to displace Nvidia across the whole AI market. Its systems are pitched at open-weight models and video generation. Beyond Virtu, named deployments include Equinix, which offers Galaxy systems through its Distributed AI Hub, and Smallest.ai, which runs a voice model on Galaxy servers for customers that want it on-premises. In Japan, ai& has deployed more than 120 Galaxy systems for sovereign AI infrastructure, according to Tenstorrent. The company also licenses its TT-Ascalon RISC-V CPU and Tensix AI cores to chip designers including Samsung and LG.

Keller's pitch to customers, as he described it to Bloomberg, is lower prices, a willingness to take orders too small for the largest suppliers, and hardware and software customized to the buyer.

That breadth is also where the strain shows. Selling complete systems demands manufacturing and customer support; licensing chip designs demands design enablement and software support for outside silicon teams, TechTarget noted. "The open software stack helps both sides, but execution risk rises if the company tries to be excellent at full systems, chiplet IP, and developer tooling at once," said Stephen Sopko, practice lead for semiconductor and deep tech at HyperFrame Research.

Tenstorrent says its Galaxy Blackhole systems are in production and shipping in volume. Sopko's test for what comes next is blunter: "Building a few hundred servers is one thing; predictable multi-thousand unit ramps are another."