If OpenAI is going to float this year, it must sink in its business model. The wow factor surrounding the American company – the poster child of the boom in the AI industry that has sparked fears of the stock market – has been established for a long time, but when will the profits come? The party cannot go on forever.
Developer ChatGPT is one of the largest in the world, and is now valued at $850bn (£645bn). Meanwhile, it is reported to spend $600bn on infrastructure (the money it spends on datacentres and chips to power its AI models) by 2030. At least this is a decrease from the original estimate of $1.4tn.
Even though spending plans have been reduced, the startup is nowhere near profitable. In fact, if things keep going like this, they will burn half a billion dollars by the end of the decade. Boosters may point out that Uber, for example, spent billions before turning a profit – but it was $30bn, not $600bn.
OpenAI, led by Sam Altman, its chief executive, seems to be making decisions quickly, as the number of market types is approaching with a large float towards the end of this year. Three quarters of its business were liquidated last month; yet another has proven to offer the least promise.
In early March, OpenAI backed out of Instant Checkout, a project where customers would purchase goods directly within ChatGPT. This was after a five-month trial in which the company seems to have found that building a successful business platform is harder than it looks. “Like many of OpenAI’s startups, it felt more like a public demonstration of what the technology could do than a serious attempt to create a commercial enterprise,” said Niamh Burns, an analyst at Enders.
Then, last week, it dropped Sora, its video production platform, and with it a $1bn deal in which Disney would license OpenAI that was proposed to “open up new opportunities for storytelling”. This was OpenAI’s strategy, because Sora was a money pit. It was difficult for Disney, who is said to have learned that the stage would be closed an hour before the public.
Finally, last week, it also pulled the plug on provocative chatbots, the often delayed project announced last year to “treat adult users like adults” and allow them to have provocative conversations with ChatGPT. “This would have been an incredibly dangerous operation,” Burns said, especially with increased scrutiny on internet security. “It would have been an absolute nightmare from a product safety and PR perspective.”
Hopefully, all this represents a company that reduces fat before the initial public offering (IPO), in a competitive market where Anthropic, the creator of Claude chatbot, seems to be winning more loyal people among business customers. OpenAI is “under a lot of pressure to show strategic discipline”, Burns said. It casts a very wide net.
Adrian Cox, chief executive at the Deutsche Bank Research Institute, said OpenAI was making the right moves if, as reported, it was preparing to fly to take the business for $1tn. This compares with its annual revenue – an estimate based on short-term performance – of $25bn, which the company reported to have earned in early March.
“If OpenAI moves to an IPO and seeks a broader pool of investors, those investors will want to see real evidence of strong, sustainable revenue growth in the coming years,” Cox said. “By focusing its business model in this way, OpenAI is perhaps poised to grow in the best way possible.”
He also said that OpenAI seemed to have stopped fighting with opponents about “everything” in the business and was now reducing its focus.
“There was a concern about the lack of clear ways to monetize the consumer AI model,” Cox said. “Right now it seems to be making tough decisions that allow it to do its business better in the future. Many investors would say this is the best news they’ve heard from OpenAI in months.”
And the signature product of OpenAI, indeed of the entire AI boom, is gaining popularity. ChatGPT now has more than 900 million weekly active users and more than 50 million paying subscribers. OpenAI makes its money from these subscriptions – which account for 75% of its revenue – and offers businesses its ChatGPT business models, while allowing companies and startups to build their own products with its AI models.
But there is a feeling among critics that it should have gotten stronger earlier, especially as it burns billions of dollars every month on tests that end up being much more. A Forbes writer called OpenAI “the most messed up company in tech” after Instant Checkout collapsed.
Burns said: “We’ve seen a lot of consumer products being introduced, promising to disrupt the browser, online commerce, content creation, search…
Last week, OpenAI announced what seemed to be a victory in the midst of chaos: an advertising experiment in ChatGPT made $100m in annual revenue, which means it made about $12m in six weeks. Maybe this is a way to get a profit; ChatGPT, in fact, knows a lot about its users and can target ads in a unique way.
However, like everything else the company has tried, it will likely take more work to get right, Burns said. “It can start to feel awkward very quickly and pose a risk to the user and privacy issues.”
On the other hand, ads on ChatGPT won’t drive much business if they remain a “glorious ad below the answers” without focus, he said.
Nikhil Lai, an analyst at Forrester, said that the ad test was “better than expected”, but this did not mean that OpenAI was close to being able to make money from advertising.
Lai said it might be “two years before OpenAI gets there, if they get there”, adding: “They’re going to have to do a lot and they’re going to have to change a lot.”
The manufacturer of the world’s most advanced technology must find a way to make a profit from it and reduce the unsustainable waste of money. Investors are waiting for an answer.
An OpenAI spokesperson said infrastructure to run AI, or “compute”, was lacking and therefore prioritized investment.
“With user demand outstripping supply, computing is a critical tool when it comes to AI,” the spokesperson said. “In addition to locking in our long-term computing needs through our infrastructure strategy, we are also prioritizing the distribution of that computing ruthlessly wherever it drives long-term economic value: advancing frontier research, growing our 900m-plus global users, and strengthening enterprise use cases.
“As we continue to house larger and larger computers, focusing on where we use that computer allows us to scale, innovate faster and deliver more efficiently to businesses and developers.”
#OpenAI #float #stock #market #year #start #turning #profit