Cerebras

AI chipmaker Cerebras Systems just filed to go public on Nasdaq under ticker “CBRS.” The headline numbers look incredible: $510 million in revenue, a swing from a $485 million net loss to $87.9 million in net income, and $24.6 billion in remaining performance obligations. But buried in the S-1 is a financial arrangement so tangled it deserves far more scrutiny than it’s getting. OpenAI isn’t just Cerebras’s biggest customer. It’s also an investor, a lender, and the single reason this IPO is happening at all. ...

The biggest bottleneck in AI isn’t training anymore. It’s inference — the moment a model actually does something useful. And AWS just partnered with Cerebras Systems to attack that bottleneck with an approach nobody has tried at this scale. The deal: Cerebras’ massive wafer-scale CS-3 chips will sit inside AWS data centers, accessible through Amazon Bedrock. The promise: 5x faster inference. The method: tearing the inference pipeline in half. Splitting the Brain Traditional AI inference runs both stages on the same GPU. You send a prompt, the chip processes it (prefill), then generates a response token by token (decode). One chip, both jobs. ...