The latest player in the race to acquire AI unicorns among cloud providers is World Labs, the business founded by Fei-Fei Li. In a recent decision that might cost hundreds of millions of dollars, the corporation selected Google Cloud as its main computing source for training its artificial intelligence models. Google has made it clear that Li’s prior position as Chief Scientist of AI at Google Cloud had no bearing on the choice.
It was revealed at the Google Cloud Startup Summit that World Labs plans to lease GPU servers on Google Cloud using a substantial amount of its recent funding in order to create AI models that are “spatially intelligent.” The business, valued at more than $1 billion, has secured $230 million to fund the creation of AI world models that can generate, interpret, and interact with geographical and video data.
Well-funded AI firms like World Labs, OpenAI, and Anthropic are excellent targets for cloud providers.
These businesses frequently spend millions on computing services, and if their AI models become more complicated, they might require even more resources. With World Labs as a partner, Google Cloud is better positioned in this cutthroat market.
According to James Lee, Google Cloud’s General Manager of Startups and AI, World Labs chose to work with Google Cloud because of the platform’s scalability and infrastructure. Lee stressed that the startup looked at a number of possibilities before deciding to use Google’s services, despite Li’s prior affiliations with the company.
Despite the fact that Google Cloud sells exclusive AI hardware like TPUs, World Labs chose to train its AI models using Nvidia GPUs for the time being. Lee pointed out that as the company’s product pipeline develops, this choice may change.
Although World Labs may look into other cloud providers, this agreement is not exclusive. Google Cloud presently manages the majority of the startup’s workloads, and the company wants to keep that partnership going ahead.