FILE PHOTO: A sign is pictured outside a Google office near the company's headquarters in Mountain View, California, U.S., May 8, 2019. REUTERS/Paresh Dave/File Photo
(Reuters) - The cloud computing arm of Alphabet Inc said on Thursday it had formed a partnership with startup Hugging Face to ease artificial intelligence (AI) software development in the company's Google Cloud.
Following the launch of generative AI tools focused on consumers from the likes of Microsoft and Google parent Alphabet, large and small businesses have become interested in deploying AI for their own purposes - such as boosting an existing product or serving an internal need. The need to create AI software that is optimized for specific tasks and needs prompted the partnership.
"It shows we are moving into a world where there's not just one model, but there as many models as companies, Hugging Face Chief Executive Clem Delangue told Reuters in an interview. "Every company is training, fine tuning and optimizing their own models, which is the direction that the field is taking."
Rampaging demand for cloud-based AI computing suggests that at its current trajectory it will overtake the traditional cloud software market in the future, Google Cloud CEO Thomas Kurian said.
"It always starts infrastructure up because you first got to put in the machines that reflect the demand - and grow from there," Kurian said.
The companies reached a revenue-sharing agreement, but did not disclose the terms.
The arrangement will allow developers using Google Cloud's technical infrastructure to access and deploy Hugging Face's repository of open source AI software, such as models and data sets. Hugging Face has grown into a central hub for open-source AI software, and signed a similar arrangement in February last year with Amazon.com's cloud unit.
The partnership coalesced because cloud customers have become more interested in modifying or building their own AI models versus using off-the-shelf options, Kurian said. One such example is ensuring against specific safety frameworks in place to secure against data leakage, or evaluate the quality of the model running.
Reflecting the amount of interest in building AI software, the New York-based Hugging Face's repository of AI-related software will roughly double in size in the next four months, when it previously took four years to reach its current state, Delangue said.
Developers will be able to use an advanced version of Google's in-house built AI chips, or tensor processing units. And Google will complete the technical integration required to use Nvidia's coveted advanced AI chips, the H100, within "weeks" not "months," Kurian said.
(Reporting by Max A. Cherney in San Francisco; Editing by Leslie Adler)
