Nvidia挑战者Tenstorrent完成1.5亿美元融资，估值升至62亿美元，以保持独立并扩大开源AI芯片规模。
Tenstorrent, previously drawing acquisition interest, chose independence by raising $1.5 billion — pushing its valuation to $6.2 billion. The company is scaling production of open-weight AI chips optimized for Llama, Phi, and other mainstream open models, with early deployments in inference-heavy infrastructure stacks. Its roadmap includes full-stack tooling support and cloud provider integrations by late 2024. For early-stage AI application teams, this signals a tangible shift: affordable, auditable, and non-Nvidia inference hardware is moving from lab prototypes to deployable infrastructure within 12–18 months. Teams building on open models can now plan cost structures around real alternatives — reducing vendor lock-in risk and enabling earlier unit economics validation. Prioritize SDK stability and cloud availability over peak specs when evaluating adoption timing.