Microsoft is making a $1 billion bet on OpenAI, the company formed by notable founders including Elon Musk and Sam Altman three years ago with the aim of doing research and development work to steer the growth of artificial intelligence toward the “friendlier” end of the spectrum, in order to help mitigate what Musk sees as a potential existential threat from AI if it isn’t developed responsibly while it grows in capability.
On Monday, Microsoft and OpenAI announced a multiyear “exclusive computing partnership” that will include the two companies building new AI supercomputing technologies for Microsoft’s Azure cloud platform, and OpenAI will also port its existing services to work on Azure. Microsoft will also now be “OpenAI’s preferred partner” when it comes to the commercialization of new AI technologies it develops in the future.
All the talk of “exclusivity” and “preference” in this announcement is particularly interesting because one of OpenAI’s founding principles was to “freely collaborate” among other AI researchers, and make both its work and patents available to others. But there are some caveats, including that there is OpenAI Inc. the nonprofit organization, and its for-profit corporate subsidiary OpenAI LP, and that its current charter includes a provision that out of “safety and security concerns” it may reduce its public publishing of its work as it moves forward.
The goal of this partnership for Microsoft seems to be to provide it an edge in building out a broad-scale Azure AI platform, and ensure its supercomputing technologies are involved in the development of artificial general intelligence. OpenAI benefits because Microsoft will be party to its principles around developing advances in AGI safely and with humanity’s interest in mind — and there’s the $1 billion, too.
This $1 billion isn’t a lump-sum up-front investment — instead, it’s a total amount that can be delivered anytime over the next decade, in installments that correspond with funding efforts on OpenAI’s side that feed back into Microsoft’s own AI ambitions, as detailed by The New York Times.
Update (11:30am PT): long after the press releases hit the wire at 6am this morning, OpenAI contacted us with a bit more detail about the announcement. Here is the statement we received from OpenAI co-founder and CTO Greg Brockman: “It’s a cash investment into OpenAI LP. It uses a standard capital commitment structure, to be called as we need it. We plan to spend it in less than five years, and possibly much sooner.”
At its launch, OpenAI noted that it had $1 billion committed from Musk, Altman and co-founder Brockman, as well as Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services (this makes the Azure angle here particularly interesting), Infosys and YC Research, though it did not anticipate spending that much in the ensuing few years.