Following OpenAI and SoftBank’s surprise announcement of the $500 billion Stargate project designed to facilitate the construction of data centers across the United States, Microsoft lost its status as the ChatGPT maker’s exclusive cloud provider. In an interview with CNBC, Salesforce CEO Marc Benioff (well-known for criticizing Microsoft’s AI efforts) indicated:
“I think it’s extremely important that OpenAI gets to other platforms because Microsoft is building their own AI and I don’t think Microsoft will use OpenAI in the future.”
Over the past few months, Microsoft’s multi-billion dollar partnership with OpenAI has seemingly been strained due to several factors. As you may know, most of Microsoft’s AI-powered models are backed by OpenAI’s technology. However, a report by Reuters suggested that the Redmond giant could be moving away from OpenAI’s AI products, like its GPT-4 model, because it’s too expensive and isn’t fast enough to meet its enterprise customers’ requirements.
The move is reportedly part of the tech giant’s broader plans to reduce the cost of its enterprise services, including GitHub Copilot, and pass on the savings to consumers. On the other hand, OpenAI has complained about Microsoft’s failure to meet cloud computing needs, delaying the launch of some of its next-gen AI models. Interestingly, the AI firm claimed that its rivals could first achieve the coveted AGI benchmark because of Microsoft’s computing shortcomings.
Perhaps more interestingly, Benioff claimed, “Mustafa Suleyman and Sam Altman aren’t best friends.” He attributed his deductions to last year’s Davos conference, where both executives were on the panel. He added that they both seemed uneasy around each other, potentially highlighting friction between Microsoft and OpenAI.
The best tech bromance strays as OpenAI races to build AGI
To this end, Microsoft’s partnership status with OpenAI remains unclear. However, both parties signed a new agreement that gives Microsoft “the right of first refusal.” For context, Microsoft will be the first option to host OpenAI workloads in its cloud infrastructure and services. However, if it can’t meet the requirements, OpenAI can source the services from competitors.
Additionally, the new agreement retains 4 key elements of the original agreement, including rights to OpenAI’s IP for its AI products, OpenAI’s API exclusivity to Azure, revenue sharing agreements that flow both ways, and Microsoft remains OpenAI’s major investor.
SoftBank CEO Masayoshi Son indicated that AGI is coming very soon with the launch of Stargate. Interestingly, the executive’s remarks reiterate Sam Altman’s remarks that AGi could be achieved sooner than anticipated with current hardware. But we might have to wait a bit longer after the executive dismissed reports alluding AGi would be deployed next month. “We are not gonna deploy AGI next month, nor have we built it,” indicated Altman. “We have some very cool stuff for you but pls chill and cut your expectations 100x!”
However, it seems OpenAI might be close to achieving AGI. Sam Altman claimed that OpenAI knows how to build AGI and could shift its focus to superintelligence. Additionally, the ChatGPT maker is reportedly planning to scrap a stringent clause that would sever its ties with Microsoft after achieving AGI to secure additional funding and access to its cloud computing services.
A leaked document revealed OpenAI would only hit the AGI benchmark after developing an AI system capable of generating up to $100 billion in profit. This seems like a reach considering OpenAI’s recent predicament, placing it at the cusp of bankruptcy with projections of $5 billion in losses within a year. Market analysts claim OpenAI could suffer an additional $44 billion loss before turning a profit in 2029, partly due to its Microsoft tie-up.
