Big Event Media/Getty Images for HumanX Conference
- Vercel's CEO, Guillermo Rauch, said the days of partnering with one AI lab are over.
- Now, companies are turning to different labs for different parts of their AI stack, he said.
- Rauch's comments come as companies are thinking of how to spend on AI more effectively.
Vercel's CEO said companies are no longer relying on a single AI lab for all their needs.
"Last year, there were a lot of people picking one lab partner — saying they would build everything on OpenAI or Anthropic," chief executive Guillermo Rauch said in a Monday interview with TechCrunch.
But he said that companies now understand how each part of the AI stack works — from model, harness, data platform, sandbox, and gateway — and "every piece is plug and play."
"You can use OpenAI, you can use Anthropic, or you can use Gemini," he added. He said he's seeing a lot of growth in Gemini in particular, because Gemini models have "awesome price/performance characteristics" when scaling up.
Chinese models like DeepSeek and Z.ai's GLM-5.2 are also seeing rapid adoption, he said.
Rauch added that last year was "all about prototyping" for AI, where everyone built AI agents. Now, companies are "getting into the realities of agents in production, and some of the challenges."
Vercel, which is based in San Francisco, is a cloud platform that helps developers host and launch websites and apps.
The executive's comments come as companies are coming around to the realization that money spent on AI is not directly translating to more value shipped for customers. The days of urging employees to burn as many AI tokens as possible are over, and companies are now thinking about how to cut back on spending or use AI more efficiently.
Coinbase CEO Brian Armstrong said in an X post in June that he was experimenting with using Chinese LLMs like GLM-5.2 and Kimi AI's K2.7 as defaults, which are cheaper than those by American AI labs like Anthropic and OpenAI.
He also talked about model routing — routing his engineers' prompts to the most appropriate models for the task at hand, so as not to use expensive frontier models for simple tasks.
Rauch's comment on companies partnering with different AI labs for different tasks is similar to how companies used to pick one major cloud provider, such as Amazon Web Services or Microsoft Azure, but have since adopted multi-cloud strategies to avoid reliance on a single vendor and optimize costs.
Read the original article on Business Insider