Tenstorrent

Run superclusters for high margin AI use cases including agentic workflows, real-time systems, and long context reasoning.

Tenstorrent Galaxy superclusters run video models and generate videos faster.

Sparse expert models are a concept dating back thirty years that is re-emerging as an architecture in deep learning.

Tenstorrent, a hardware start-up developing next generation computers, announces the addition of Matthew Mattina as VP of Machine Learning.

Tenstorrent President and CTO Jim Keller gives the keynote at Stanford's System X conference

Tenstorrent builds computers for AI and is headquartered in the U.S. Its teams work in computer architecture, ASIC design, RISC-V technology, advanced systems, and neural network compilers.

Preferred Networks and Tenstorrent secure a five-year research grant from Japan’s New Energy and Industrial Technology Development Organization (NEDO) to develop chiplet-based xPU technology. The collaboration focuses on advancing specialized processing units for high-performance computing applications. NEDO selected the project as part of its broader efforts to support next-generation hardware innovation.

Oct 2026TNGlobalNews article

GitHubRepository owner

Tenstorrent raises $1.5B to reach a $6.2B valuation, focusing on AI inference chips. The startup, founded over a decade ago, demonstrates its Galaxy servers can run DeepSeek models. Funds will support scaling chips for open-weight AI and video generation, competing with Nvidia’s dominance. Backers include Trace Capital, Cadence Ventures, and Fidelity.

Oct 2026TechCrunchNews article

Tenstorrent raises **$1.5 billion** in a Series E round led by Trace Capital and Innovation X, reaching a **$6.2 billion valuation**. The AI hardware startup, led by semiconductor legend Jim Keller, offers **RISC-V-based accelerators (Tensix Neo) and CPUs (Ascalon)** across **Galaxy servers, TT-QuietBox workstations, and Blackhole cards** for laptops. Its **open-source software stack (TT-Forge, TT-Metalium, TT-LLK)** enables developers to deploy AI workloads on its hardware. Galaxy servers power on-premises AI trading at **Virtu Financial** and are available via **Equinix**, while TT-QuietBox provides liquid-cooled desktop AI compute. Funds will accelerate production, research, and global hiring to compete with Nvidia’s GPU dominance.

Oct 2026SiliconANGLENews article