Cerebras just announced 6 new AI datacenters that process 40M tokens per second — and it could be bad news for Nvidia
1 min read
Summary
AI hardware start-up Cerebras Systems is expanding its data centre footprint with six additional AI data centres in the US and Europe to increase AI inference capacity to more than 40m tokens per second, as the company positions itself to become the leading provider of high-speed AI inference services.
The company has also announced partnerships with AI developer platform Hugging Face and financial intelligence platform AlphaSense to provide high-speed AI to developers and financial analysts.
Cerebras is challenging Nvidia’s dominance in the AI hardware market by providing ultra-high-speed AI inference, with its Wafer-Scale Engine (WSE-3) processor claiming to run AI models 10 to 70 times faster than GPU-based solutions.
The start-up is targeting real-time voice and video processing, reasoning models and coding applications as it aims to lead in the high-speed AI inference segment, even when up against the largest cloud providers.