Company🇺🇸 San Jose
Etched5 people
AI semiconductor company building frontier inference clusters through co-designed chips, racks, software, and manufacturing. Etched says its A0 silicon has returned from TSMC N4P, it has raised $800M across four financings, and it is preparing first rack shipments to fulfill more than $1B in customer demand. Investors include @Geoffrey Hinton, @Andrej Karpathy, @Fei-Fei Li, @Noam Brown, @Jerry Tworek, @Arthur Mensch, and more.
Company🇺🇸 Sunnyvale
Cerebras1 person
Cerebras builds wafer-scale processors, systems, and cloud infrastructure for high-speed AI training and inference. Its Wafer Scale Engine replaces clusters of conventional accelerator chips with a processor spanning an entire silicon wafer.
Company🇺🇸 Redwood City
Fireworks AI1 person
Fireworks AI operates an inference cloud for deploying and optimizing open and custom generative models. Its platform focuses on production serving, fine-tuning and model routing designed to improve application speed, quality and cost.
Company
Fixie.ai1 person
Fixie.ai built realtime voice AI products, including the open-source Ultravox speech model, before Justin Uberti joined OpenAI.
Company🇺🇸 Cupertino
Lepton AI1 person
Lepton AI built a cloud platform for deploying and managing AI workloads across GPU providers. NVIDIA acquired the company in 2025 and incorporated its technology into DGX Cloud Lepton, a platform and compute marketplace for training and inference.
Company🇳🇱 Amsterdam
Nebius1 person
Nebius builds full-stack cloud infrastructure for training, developing, and deploying AI models and applications, including large-scale GPU clusters and managed inference services.
Company🇫🇷 Paris
ZML1 person
AI infrastructure company building a production inference stack that runs models across multiple hardware targets from a single codebase.
Company🇺🇸 Reno, Nevada
Positron0 people
AI hardware company building purpose-designed accelerator systems for efficient, high-performance generative AI inference.