OpenAI’s ChatGPT is quickly moving from exhilarating novelty to enterprise cornerstone—and spotlighting a fast-emerging power dynamic in the process.
Microsoft officials announced Monday that the tech titan’s Azure cloud customers are getting broad access to the generative A.I. tech developed by OpenAI, including the systems behind the ChatGPT chatbot and DALL-E 2 text-to-image generator. Microsoft initially made much of OpenAI’s tech available via Azure in late 2021, but a limited number of customers were granted access.
The development marks a rapid escalation of Microsoft’s foray into generative A.I., illustrating its ambitions for integrating OpenAI’s tech throughout the company. Microsoft is OpenAI’s “preferred partner” for commercializing the startup’s tech after investing $1 billion in the outfit in 2019. Fortune’s Jessica Mathews and Jeremy Kahn reported this month that Microsoft has since committed $2 billion more to the company and is deliberating an additional investment worth up to $10 billion.
The offering should boost the attractiveness of Microsoft’s Azure offering, the world’s second-largest cloud computing platform, as it competes with market leader Amazon Web Services and third-place Google. Microsoft’s Azure clients will soon be able to incorporate OpenAI offerings into their own applications, helping them automate repetitive tasks, write basic text, and analyze large databases, among other chores.
“Large language models are quickly becoming an essential platform for people to innovate, apply AI to solve big problems, and imagine what’s possible,” Eric Boyd, Microsoft’s corporate vice president for A.I. platforms, wrote Monday.
Microsoft’s move raises several questions for the future: Will generative A.I. actually prove profitable? When will Google start commercializing its own advanced A.I. products? Is Microsoft acting responsibly by releasing generative A.I. technology that its own creators call “incredibly limited”?
For now, though, I want to look back at how we reached this point and what it says about Big Tech’s influence over machine learning.
While OpenAI is hardly a rags-to-riches story—it began as a nonprofit with $1 billion in committed funding from tech luminaries like Elon Musk and Peter Thiel—its operations bear the hallmarks of a scrappy startup. The company generates minimal revenue, employs only a few hundred people, and preaches about benefiting “all of humanity.” (Insert Silicon Valley meme here.)
But whereas the Jeff Bezoses and Mark Zuckerbergs of yore could gradually build their products over time, relying on incrementally rising amounts of venture capital funding before going public, high-level machine learning technology requires massive amounts of upfront investment in computing power. It also can’t be easily commercialized without major enterprise partners.
In turn, organizations with deep pockets, sprawling business interests, and vast computing infrastructure—namely Microsoft, Amazon, and Google—are well-positioned to swallow up innovators in need of support. To wit, OpenAI opted to take Microsoft’s money, an undisclosed chunk of which comes in the form of Azure credits, in exchange for first dibs on its tech. Blockchain startup ConsenSys struck a similar deal in early 2022, albeit at a lower price tag.
“Microsoft partnership has been one of OpenAI's secrets to success — we work very closely with Azure to produce AI training & serving infrastructure that can scale to our cutting-edge (& entirely unprecedented!) needs,” OpenAI president and co-founder Greg Brockman tweeted last month.
Big Tech behemoths using their money and might to crowd out the competition is hardly a novel concept. Nearly a decade ago, Google reportedly spent more than $500 million to acquire A.I. pioneer DeepMind, which just started turning meager profits. At the same time, many aspiring A.I. companies are going the old-fashioned route, staying independent and using their venture capital fund to cover infrastructure vendor costs.
But if Microsoft can spin gold out of its OpenAI tie-up, the success story could pave the way for deeper Big Tech investments in A.I., further concentrating power in the hands of the few. As DeepMind founding investor Humayun Sheikh said in 2020: “Commercialization for any A.I.-based company is very difficult, unless you get absorbed by a big corporate.”
Want to send thoughts or suggestions to Data Sheet? Drop me a line here.
Jacob Carpenter
Editor's note: A previous version of this article incorrectly listed the reported price of Google's acquisition of DeepMind.
