ChatGPT maker OpenAI shifts its focus to business users amid Anthropic pressure

The same ChatGPT chatbot that gave OpenAI CFO Sarah Friar a tilapia recipe for a recent Sunday night dinner at home also now performs its most mundane tasks at work, like summarizing her emails and Slack messages.

Friar and other company executives are banking OpenAI’s future on the latter, as it focuses on enterprise-focused products while shedding some of its consumer offerings to move toward profitability.

OpenAI says it will introduce a new artificial intelligence model for “high-value professional work”, as the company faces increased competition with rival Anthropic to get enterprise customers to adopt AI assistants in the workplace.

“You’ll see a new model coming from us soon. We’re very excited about it,” Friar said in an interview with The Associated Press.

OpenAI has more than 900 million weekly users of its core product ChatGPT, and Friar said about 95% of them “pay nothing” for the popular chatbot. But while all these interactions create habits and dependency, they also strain the expensive computing resources needed to power the company’s AI systems and highlight the need for large companies to help pay their bills.

OpenAI, valued at $852 billion, and Anthropic, valued at $380 billion, are both losing more money than they make, putting San Francisco-based private AI research labs in a fierce competition to generate more revenue as they race to become publicly traded on Wall Street.

A desire to improve the performance and sales of OpenAI’s commercially-oriented products – already Anthropic’s bread and butter – has pushed OpenAI to abandon some consumer initiatives, such as the AI ​​video generation application Sora.

“I think it was a little heartbreaking, but we’re like, OK, it’s not the main event right now,” Friar said. “We need to make sure our upcoming new model has enough compute.”

Codenamed Spud, OpenAI claims its “smartest model yet” delivers “stronger reasoning, better understanding of intent and dependencies, better tracking, and more reliable results in production.” This is part of OpenAI’s response to Anthropic’s new Claude Mythos, which Anthropic says is so “remarkably capable” that it limits its use to certain customers due to its apparent ability to outperform human cybersecurity experts in finding or exploiting computer vulnerabilities.

Friar, former CEO of neighborhood social platform Nextdoor, said enterprise customers made up about 20% of OpenAI’s revenue when she was hired in 2024 as CFO. She said it now represents 40% and is expected to account for half of OpenAI’s sales by the end of the year.

This is a drastic turnaround from late last year, when OpenAI co-founder and CEO Sam Altman was promoting a now-shuttered Sora partnership with Disney by launching a plan to sell ads on ChatGPT and floating the idea of ​​letting ChatGPT engage in erotica with paying adult users.

Altman said on the “Mostly Human” podcast earlier this month that a sharper focus is needed — and Friar agrees.

“Tech companies, when they grow, it’s a natural thing that happens. There are so many cool things you can do,” she said, adding that companies can end up doing “very poorly” if they do too many things, whereas “large companies are very good at, in a reasonable amount of time, doing that sorting out and refocusing and it’s extremely painful.”

The hiring three months ago of Slack CEO Denise Dresser as OpenAI’s first chief revenue officer signaled this shift.

Dresser said in a recent interview with AP that she has been focused on meeting with business leaders and positioning OpenAI as the go-to platform for workplaces employing AI agents to automate a variety of computing tasks.

“It’s very clear to me that companies have moved beyond the experimentation phase and are using AI to do real work,” Dresser said. “Business leaders recognize that AI is probably the most important change of their lifetime. »

But these leaders also have a choice, namely Claude from Anthropic, widely used by software professionals. Founded in 2021 by a group of former OpenAI executives who said they wanted to prioritize AI safety, Anthropic has positioned itself as the most responsible AI provider. The distinction gained attention when President Donald Trump’s administration punished the startup after a contract dispute over the use of AI in the military, and Altman used the opportunity to shore up OpenAI’s own deal with the Pentagon.

Consumer interest in Anthropic has increased, and the company said its annualized revenue reached $30 billion, a higher figure than OpenAI reported, although they measure it differently. Friar and Dresser declined to reveal OpenAI’s latest sales, but both suggested Anthropic’s figure was inflated because it doesn’t take into account the revenue it must share with cloud computing providers Amazon and Google.

The fact remains that it is a close competition that is also linked to the health of the stock market and the future of the economy.

“They’re probably pretty close,” said Luke Emberson, a researcher at the nonprofit Epoch AI Institute. “Certainly, the trends show that Anthropic is growing much faster than OpenAI. If this continues, they may cross paths soon.”

The urgency led Dresser to send a memo to OpenAI employees on Sunday, first reported by The Verge, that said Anthropic’s focus on coding “gave them an early advantage” but expressed confidence that OpenAI has the “real structural advantage” as AI use expands beyond software developers and OpenAI builds enough computing capacity to run its AI systems.

“Their story is built on fear, restrictions, and the idea that a small group of elites should control AI,” Dresser’s memo about Anthropic says. “Our positive message will prevail over time: build powerful systems, put the right safeguards in place, expand access and help people do more. »

But for skeptics about the financial viability of AI products like ChatGPT and Claude, the trajectory of the two losing companies is alarming, as smaller startups become increasingly dependent on their AI tools. Anthropic has already imposed pricing caps on heavy users, forcing some to wait hours to use Claude, and both companies have implemented service tiers that reward premium payers, said author and AI critic Ed Zitron.

“This is what I call the subprime AI crisis,” Zitron said. “People have built their lives and businesses on these companies that, in trying to save money, will start to turn the screws.”

One thing that AI executives and critics agree on is that it is an expensive technology, although it remains to be seen whether it is worth the cost in power-hungry AI computers.

“People will say, ‘Once it goes public, it’s safe.’ That’s not true,” Zitron said. “Public companies can and will die, especially those that depend on $100 billion to $200 billion every year, just to keep breathing. »