Tenstorrent, a startup designing AI inference hardware, has secured $1.5 billion in funding, pushing its valuation to approximately $6.2 billion. This latest round more than doubles the company’s worth from about $2.6 billion in late 2024. Investors leading the round include Trace Capital—co-founded by Quiet Capital’s Ben Mahdavi—and Innovation X, with participation from Cadence Ventures, Eclipse, Fidelity, SanDisk, and T. Rowe Price.
The capital will be used to scale up production of chips intended for open-weight AI models—those whose model weights are publicly accessible—and for hardware tailored to AI-powered video generation. As part of its latest advancement, Tenstorrent demonstrated that its Galaxy servers are capable of running DeepSeek model workloads, a notable milestone for the company in its bid to handle large-scale, open-weight architectures.
This funding round comes amid external interest from major semiconductor players. Five months prior, companies including Intel and Qualcomm were reportedly considering acquiring Tenstorrent at valuations up to $5 billion. However, Tenstorrent chose to remain independent, opting to raise new capital instead of selling.
Positioning & Competitive Landscape
Founded a decade ago, Tenstorrent has emerged as part of a broader movement of AI hardware startups challenging NVIDIA’s stronghold in inference workloads. Their focus lies in designing chips and systems that specialize in delivering high performance for open-weight model hosting, while also enabling AI-based video generation tools.
Its CEO and CTO, Jim Keller, has been instrumental in situating Tenstorrent within this competitive hardware space. Keller said the strategy hinges on staying independent and refining the company’s capabilities to support the increasingly demanding requirements of open-weight models. The Galaxy platform’s recent performance benchmarks are helping make that case.
Meanwhile, other competitors are also making waves. One fellow AI chip maker has been fielding investment offers pushing its valuation toward or above $40 billion, highlighting how investor appetite continues to grow in this sub-sector of AI hardware.
This round not only reflects investor confidence but also Tenstorrent’s belief in the future of open-weight AI models—and that there’s a commercial opportunity beyond the towering dominance of companies entrenched in closed-weight ecosystems.
What this means: Tenstorrent’s rise could signal shifting expectations in AI infrastructure, where transparency and flexibility—open weights—are becoming strategic advantages. The startup’s choice to scale via funding rather than exit suggests it believes the long-term gains of building its own platform outweigh immediate returns from acquisition. The next indicators to watch: benchmarks of its upcoming chips in real-world open-weight model workloads, the pace at which it can commercialize its Galaxy servers, and whether its pricing or performance can erode NVIDIA’s market leverage.
