AI & Enterprise

Cerebras compresses 163 seconds into 5 seconds, says GPU era is over

AI chip designer Cerebras put the 1 trillion-parameter open-weight model Kimi K2.6 into its enterprise inference service and achieved 981 tokens per second, a pace it says is the world’s fastest. It also cut the time to complete 500 output tokens from a 10,000-token input to 5.6 seconds, versus 163.7 seconds on the official Kimi endpoint. The company is pursuing an IPO and reported 2025 revenue of $510 million and net profit of $238 million.

AI & Enterprise

Low-cost AI spread could threaten OpenAI, Anthropic IPO plans

The spread of mass-market AI models and companies’ efforts to cut costs could become an obstacle for OpenAI and Anthropic as they pursue IPOs at high valuations. Both are valued at more than $800 billion, but cost efficiency is becoming as important as performance in the enterprise AI market. Companies including Meta, Shopify, Spotify and Pinterest say higher AI and inference costs are squeezing margins.