Dark cloud over ChatGPT revolution: The cost
WASHINGTON: The explosion of generative AI has taken the world by storm, but one question is rarely asked: Who can buy it? The Information reports that OpenAI lost about $540 million last year while making ChatGPT, and the company says it needs $100 billion to reach its goals.
Sam Altman, the founder of OpenAI, recently told a group, “We’re going to be the most capital-intensive start-up in the history of Silicon Valley.”
When people ask Microsoft, which has put billions of dollars into OpenAI, how much its AI adventure will cost, the company says that it is keeping an eye on its bottom line.
To make something even close to the size of what OpenAI, Microsoft, or Google give, you would need to spend a lot of money on cutting-edge chips and hire prize-winning researchers.
“People don’t realize that it takes a lot of computer power to do a lot of AI tasks like ChatGPT. “It can cost tens of millions of dollars to train these models,” said Jack Gold, a free-lance researcher.
“How many businesses can actually buy 10,000 Nvidia H100 systems, which cost tens of thousands of dollars each?” asked Gold.
The answer is pretty much no one, and in tech, if you can’t build the infrastructure, you rent it. Many companies already do this by outsourcing their computer needs to Microsoft, Google, and Amazon’s AWS.
With the arrival of generative AI, this reliance on cloud computing and tech giants will grow, keeping the same people in charge, experts say.
Stefan Sigg, Chief Product Officer at Software AG, which makes business software, said that the uncertain costs of cloud computing are a big problem that many companies don’t take into account enough.
Sigg compares cloud costs to energy bills and says that companies that don’t know better will be in for “a big surprise” if they let their engineers run up bills in their rush to build tech, including AI.
Azure is Microsoft’s most popular cloud service, and some people think the company’s big bet on AI is really a way to protect Azure’s success and make sure the cash cow will be around for a long time.
Azure has been the giant’s boring moneymaker for years, making a lot of money but not getting as much attention as the iPhone or social media, which go straight to the customer.
Dan Ives of Wedbush Securities said, “The golden goose for Microsoft is monetizing the cloud with Azure, because it could be worth $20, $30, or $40 billion a year in the future if the AI bet works.”
Satya Nadella is the CEO of Microsoft. He says that generative AI is “moving fast in the right direction.” Ives said that Nadella, who is well-liked on Wall Street, will have six to nine months to show that his bet was a good one.