OpenAIplans to add Alphabet’sGooglecloudservice to meet its growing needs for computing capacity, threesourcestold Reuters, marking a surprising collaboration between two prominent competitors in the artificial intelligence sector.
Thedeal, which has been under discussion for a few months, was finalized in May, one of thesourcesadded. It underscores how massive computing demands to train and deployAImodels are reshaping the competitive dynamics inAI, and marksOpenAI’s latest move to diversify its computesourcesbeyond its major supporter Microsoft, including its high-profile Stargate data center project.
It is a win forGoogle’scloudunit, which will supply additional computing capacity toOpenAI’s existing infrastructure for training and running itsAImodels,sourcessaid, who requested anonymity to discuss private matters. The move also comes asOpenAI’s ChatGPT poses thebiggest threattoGoogle’s dominant search business in years, withGoogleexecutives recentlysaying that theAIrace may not be winner-take-all.
OpenAI,Googleand Microsoft declined to comment. Alphabet’s stock was up 2.1% on Tuesday afternoon following the news, while Microsoft shares were down 0.6%.
Scotiabank analysts called the development “somewhat surprising” in a note on Tuesday, highlighting the growth opportunities forGoogle’sCloudunit, while expressing caution regarding competition from ChatGPT.
“Thedeal… underscores the fact that the two are willing to overlook heavy competition between them to meet the massive computing demands. Ultimately, we view this as a big win forGoogle’scloudunit, but … there are continued worries that ChatGPT is becoming an incrementally larger threat toGoogle’s search dominance,” the analysts wrote.
SinceChatGPT burst onto the scenein late 2022,OpenAIhasdealt with increasing demand for computing capacity – known in the industry as compute – for training large language models, as well as for running inference, which involves processing information so people can use these models.OpenAIsaid on Monday that its annualized revenue run ratesurged to $10 billion as of June, positioning the company to hit its full-year target amid booming adoption ofAI.
Earlier this year,OpenAIpartnered with SoftBankand Oracleon the $500 billionStargate infrastructure program, andsigneddealsworth billions with CoreWeavefor more compute. It is on track this year to finalize the design of its first in-house chip that could reduce its dependency on external hardware providers, Reutersreportedin February.
The partnership withGoogleis the latest of several maneuvers made byOpenAIto reduce its dependency on Microsoft, whose Azurecloudservice had served as the ChatGPT maker’sexclusivedata center infrastructure provideruntil January.GoogleandOpenAIdiscussed an arrangement for months but were previously blocked from signing adealdue toOpenAI’s lock-in with Microsoft, a source told Reuters. Microsoft andOpenAIare also in negotiations to revise the terms of their multibillion-dollar investment, including the future equity stake Microsoft will hold inOpenAI.
ForGoogle, thedealcomes as the tech giant isexpandingexternal availabilityof its in-house chip known as tensor processing units, or TPUs, which were historically reserved for internal use. That helpedGooglewin customers including Big Tech player Appleas well as startups like Anthropic and Safe Superintelligence, twoOpenAIcompetitors launched by formerOpenAIleaders.
Google’s addition ofOpenAIto its customer list shows how the tech giant has capitalized on its in-houseAItechnology from hardware to software to accelerate the growth of itscloudbusiness.
GoogleCloud, whose $43 billion in sales comprised 12% of Alphabet’s 2024 revenue, has positioned itself as a neutral arbiter of computing resourcesin an effort to outflank Amazonand Microsoft as thecloudprovider of choice for a rising legion ofAIstartups whose heavy infrastructure demands generate costly bills.
Alphabet faces market pressure to demonstrate financial returns on itsAI-related capital expenditures, which are expected to hit$75 billionthis year, while maintaining its bottom line against the threat of competingAIofferings, as well asantitrust enforcement.
Google’s DeepMindAIunit also competes directly withOpenAIand Anthropic in a race to develop the best models and integrate those advances into consumer applications.
Selling computing power reducesGoogle’s own supply of chips while bolstering capacity-constrained rivals. TheOpenAIdealwill further complicate how Alphabet CEO Sundar Pichaiallocates the capacity between the competing interests ofGoogle’s enterprise and consumer business segments.
Googlealready lacked sufficient capacity to meet itscloudcustomers’ demands as of thelast quarter, Chief Financial Officer Anat Ashkenazi told analysts in April.
Although ChatGPT holds a large lead overGoogle’s competing chatbot in terms of monthly users and analysts have predicted it couldreduceGoogle’s dominant search market share, Pichaihas brushed aside concerns thatOpenAIwill usurpGoogle’s business dominance.
