What OpenAI must do to stay on top

OpenAI is the buzziest company in the field of AI.

Within two months of launching ChatGPT in 2022, the chatbot had accrued some 100 million monthly users — setting the record for the fastest-growing user base of any consumer application in history.

The company also tripled its valuation to over $80 billion in the last year, thanks in part to investments from tech giants like Microsoft, which has funneled some $13 billion into it. Apple is integrating ChatGPT into its leading products. And the company is now gearing up to launch its most advanced model yet, GPT-5.

Armed with a roster of top talent and over $11 billion in support from investors outside Microsoft, OpenAI has comfortably positioned itself ahead of its competitors.

But the company is not infallible. Altman’s recent projections show that the company will hit $3.4 billion in annual revenue this year but incur at least $5 billion in losses. Those losses and a lack of any clear path to profitability in the near future are beginning to make investors nervous.

Investor anxiety exists around other Big Tech companies, too, like Meta, Google, Apple, Amazon, and Microsoft. Huge spending on AI coupled with disappointing earnings contributed to a sell-off in the tech industry earlier this month.

So, how much runway will investors be willing to give OpenAI before it can start making money? Is OpenAI too dominant for investors to give up on? We asked the experts.

OpenAI needs to retain its talent

Investors are pouring money into OpenAI in part because it’s an incubator of some of the brightest minds in the AI industry — so retaining them is key to keeping its backers happy.

OpenAI first launched in 2015 as a nonprofit dedicated to developing artificial general intelligence — a theoretical future AI that can reason like humans — that is “most likely to benefit humanity as a whole.” But the company also added a for-profit arm in 2019 to fund its capital-intensive mission.

The move led to some disagreements between its aggressive profit-driven faction and the more cautious, safety-conscious side. It came to a head last year when the OpenAI board briefly, yet unsuccessfully, ousted CEO Sam Altman.

Altman was saved, in part, by public support from OpenAI’s employees.

The company has lured some of the AI industry’s most talented minds over the years, sometimes aggressively poaching employees from its competitors, including Google. An analysis by BI found that at least 44 ex-Googlers jumped ship to OpenAI since the beginning of the year alone.

The Information, a tech publication, reported that OpenAI is set to spend $8 billion in 2024, with at least $1.5 billion going into hiring talent.

Ajay Agrawal, a professor at the University of Toronto who co-authored a book on the economics of AI, told BI that hiring the world’s best engineers and researchers is part of OpenAI’s edge and the cash burn is strategically sound in a still nascent market.

“I would say that their strategy, which is making very large investments than their current revenues can support, seems to make a lot of sense because they’re competing in a market that is new,” Agrawal said, adding that the potential size of the AI market is still unknown.

Investors and partners will likely be empathetic toward OpenAI’s capital needs to maintain talent, Mahdi Raza, cofounder and a managing partner of Exponent Founders Capital, said.

“Investors will continue to be excited to fund OpenAI as long as they continue to accrue a compounding customer, compute, and talent advantage,” Raza said.

But to retain that talent, the company must maintain the culture and mission that first attracted it.

“As they become more commercial, partner more with companies like Microsoft and Apple, and as their valuation goes up, there’s real risk that the culture changes,” Brad Porter, the CEO of robotics company Collaborative Robots, said.

Several high-profile researchers have already left the company. OpenAI’s former chief scientist, Ilya Sutskever, left in May because he didn’t agree with Altman’s aggressive approach to rolling out new products. He’s since launched his own venture, Safe Superintelligence Inc., which bills itself as taking a more responsible approach to AI.

Jan Leike, who helped lead OpeanAI’s superalignment team, which worked to ensure that the company’s AI technology was safe for human society, also jumped to its rival Anthropic. Its cofounder, John Schulman, announced recently he, too, was leaving for Anthropic. Researchers Daniel Kokotajlo and William Saunders also departed earlier this year.

“Anyone at OpenAI can start their own company now or join another, so their position is more precarious than ever,” Porter said.

So, whether top researchers stay — or go — might be the clearest test of the company’s strength.

“If, internally, they continue to know they have the best AI and truly feel that they are the closest to realizing the vision of AGI, then most will want to stay to see that vision come to life,” Porter said.

OpenAI’s first-mover advantage

Because of OpenAI’s aggressive investments, the company’s technology is often seen as the gold standard for large language models.

But it has competition. And that could cause investors to start looking elsewhere, experts told BI.

“Given the velocity of new development, it would not be unforeseen for more nimble competitors to spring up and some of the Big Tech AI incumbents to encroach and threaten their business,” Nicos Vekiarides, the founder and CEO of digital media validation company Attestiv, told BI in a written note.

Mark Zuckerberg’s Meta released its latest iteration of the Llama model in July. Google launched Gemini last year, and Microsoft, an OpenAI partner, recently identified the ChatGPT-maker as a competitor.

OpenAI’s models are industry-leading, but companies know they have other options, Agrawal said.

“I think there’s an increasing dependence on OpenAI models. I think their competitors are trying to reduce the dependence,” he said, pointing to Microsoft as an example.

In a recent SEC filing, Microsoft acknowledged the “long-term partnership” with OpenAI but also called the company its competitor.

“Our AI offerings compete with AI products from hyperscalers such as Amazon and Google, as well as products from other emerging competitors, including Anthropic, OpenAI, Meta, and other open source offerings, many of which are also current or potential partners,” the company wrote in the filing.

Agrawal added that few corporations, if any, have “extremely high dependencies” on OpenAI that would prevent the use of substitute models.

Raza agreed, noting that it’s also unclear whether OpenAI’s models exhibit a real technical defensibility against competitors. “Any model advantage so far has proved temporary,” he said.

AGI could make or break OpenAI

OpenAI’s ultimate mission is to achieve artificial general intelligence, a hypothetical form of AI that mirrors human capabilities.

It’s the company’s wild card. Its potential is part of what excites its investors.

“I think AGI will be the most powerful technology humanity has yet invented,” Altman said in an interview with Time. He added that he expects to see major progress in the next decade.

However, some experts believe OpenAI might be better off narrowing its focus — especially since its competitors are also barreling toward the same aspirational goal.

“The wide-net approach to AGI and product might come back and bite them, but Sam Altman is a fantastic CEO,” Krithik Puthalath, the founder of full-stack AI startup Zyphra, told BI. “They know how to play the game, so I don’t think time is up for them to demonstrate profitability. They will be able to raise a few more high-priced rounds.”