AI digest: Enterprise race heats up
Big money flows into AI enterprise services whilst technical advances promise better efficiency.
The enterprise AI gold rush is hitting full stride. Meanwhile, some proper technical advances are cutting through the noise.
Sierra bags $950M to own enterprise AI
Sierra just raised nearly a billion dollars to become the “global standard” for AI customer experiences. That’s serious money for what’s essentially AI-powered customer service. The race to own enterprise AI deployments is getting expensive fast.
OpenAI and Anthropic both go enterprise shopping
Both OpenAI and Anthropic launched joint ventures with financial partners this week to push enterprise services harder. OpenAI raised over $4 billion for “The Deployment Company” whilst Anthropic partnered with Blackstone and others. Turns out selling AI to businesses needs more than just good models.
Cerebras targets $40B valuation on OpenAI partnership
AI chip maker Cerebras is heading for a blockbuster IPO that could value it at $26.6 billion or more, leaning heavily on its deep relationship with OpenAI. The chip shortage premium is real, but that valuation feels frothy even by AI standards.
Zyphra’s parallelism breakthrough delivers 2.6x speedup
Zyphra introduced Tensor and Sequence Parallelism (TSP), a hardware-aware training strategy that cuts both parameter and activation memory whilst delivering 2.6x throughput over standard approaches. This kind of efficiency gain matters more than flashy demos when you’re actually running models at scale.