Extracts: headline, IR url · metadata or bounded excerpt
Retention: D1 event history with canonical source link and deduplication metadata.
Access basis: Public endpoint; no project credential required.
Last ingested: 2026-11-17 · run status: success with data
Every customer served. Every product sold. Every commit merged. Every agent task completed. That’s inference. Training creates the possibility. Inference creates the value. The more we ask of AI, the more inference it takes. And inference is becoming the bottleneck. Groq was buil
Cards Add to, or config your own rig. Passive, active and liquid cooling options. Starting at $999 Multi-server configurations for AI workloads – video, code, LLM – run fast. Near-linear scaling, preserving quality and interactivity. Whisper-quiet, liquid cooled - run models up
Chip performance that pays for itself. The most profitable way to run models. Inference is going to be the center of the AI economy SambaNova CEO Rodrigo Liang joins CNBC's Squawk on the Street to explain why the economics of inference, and not training benchmarks, now decide wha
High-throughput chips for LLMs Our goal is to make the best chips physically possible for the large model needs of frontier labs. The [MatX One](/research/series_b) chip delivers higher throughput than any announced product while also matching the best latencies of any products.
Gavin Uberti Co-Founder & CEO Harvard Thiel Fellow. World math champion, Math 55 alumnus, expert in AI compilers, developed the Cortex-M backend for TVM. We co-design chips, racks, software, and manufacturing methods so frontier models can run with best-in-class throughput, laten