OpenAI CEO Sam Altman has committed to spending more than $1 trillion to build out AI infrastructure, totaling 26 gigawatts of compute capacity from tech companies including Nvidia, AMD, and Oracle, according to the Financial Times‘ calculations. And that’s just over the last month.
Recouping these astronomical costs isn’t just OpenAI’s problem, either — vast portions of the wider economy are relying on OpenAI’s success, highlighting growing concerns over an AI bubble that could have sweeping implications for the rest of the economy.
For now, the company is desperately trying to build out new revenue lines and taking on even more debt to keep things running. It’s looking at advertising online, monetizing its new text-to-video AI generator app Sora, and leaning into online shopping. The company is also working on a new personal device with the help of former Apple designer Jony Ive.
But whether any of that will sufficiently grow its income to make a dent in the enormous amount of money it’s planning to spend in the next five years remains anything but certain.
The numbers are certainly far from confidence-inducing. OpenAI boasts that it has 800 million ChatGPT users. That’s impressive, but only a tiny sliver — roughly five percent — of them are actually paying for a subscription, according to the FT‘s sources at the company.
Despite running the most popular AI chatbot service by far, in other words, it’s struggling to convince the vast majority of users that it’s worth $20 per month, let alone $200 per month for a pro subscription.
In an apparent effort to boost engagement, Altman recently reversed course on a prior promise, announcing that the company would soon allow “mature apps,” despite boasting two months earlier that ChatGPT didn’t host any “sexbots.”
This all comes as OpenAI is still bleeding cash at an alarming rate. In the first half of the year, OpenAI lost $8 billion, despite doubling revenue compared to the year before, per the FT. And insiders are worried that there isn’t a cohesive plan to turn OpenAI into a viable business that isn’t bleeding billions of dollars long-term.
“[Investors] expect you to have a five-year model,” a senior OpenAI executive told the FT, but “right now I’d say there’s lots of fuzz on the horizon, and as it gets closer and it’s going to start to take real shape.”
Altman has also indicated that it’s still far from a concrete plan to monetize Sora, a resource-hogging app that could quickly turn into a major expense.
And a personal AI chatbot device, even with the help of Ive, may also prove to be an uphill battle, considering the fate of previous attempts, including the disastrous Humane AI pin.
In the meantime, the company is looking to lean on its partners, including Oracle, to take the hit on spending on AI infrastructure, giving it “time to build the business,” the executive added.
OpenAI leadership is hoping that with more compute, revenue will continue to grow, a belief that has quickly turned into a major point of contention.
The stakes are incredibly high. Capital expenditures for AI have contributed more to the growth of the US economy as of August than all of consumer spending combined, as the Wall Street Journal reported earlier this year, raising the terrifying possibility that a crash could take down the US economy with it.
Analysts have also voiced concerns after OpenAI announced it would be using investment money from AI chipmaker Nvidia to buy hardware from the company, one of several circular deals that are heightening concerns of a crash.
Perhaps the most eyebrow-raising reality is that currently, stopping the bleeding isn’t a main priority at OpenAI. Altman doesn’t seem all too concerned about breaking even, telling reporters earlier this month that it’s “not in my top ten concerns, but we obviously someday have to be very profitable.”
More on OpenAI: Two Months Ago, Sam Altman Was Boasting That OpenAI Didn’t Have to Do Sexbots. Now It’s Doing Sexbots
The post If You Were Bankrolling OpenAI, the Percent of ChatGPT Users Willing to Pay for It Might Make You Break Out in a Cold Sweat appeared first on Futurism.
