OpenAI is betting on a suite of new AI products, building its own data centres and a crucial partnership with Apple to supercharge its next phase of growth, as it targets reaching 1 billion users over the coming year.
The San Francisco-based group, whose popular ChatGPT chatbot has rocketed to 250mn weekly active users since its launch two years ago, plans to expand further through launching so-called AI “agents”, its own AI-powered search engine and ChatGPT’s integration with Apple devices.
“[In 2025] we will be coming into our own, as a research lab serving millions . . . hoping it can be billions of consumers around the world,” Sarah Friar, the company’s chief financial officer, told the Financial Times.
The goal comes as the nine-year-old start-up recasts itself as global technology giant and prepares for what founder and chief executive Sam Altman describes as the “Intelligence Age”.
Having raised more than $6bn of investment at a $150bn valuation in October — the highest for a start up in Silicon Valley’s history — Friar said OpenAI would continue to raise “more money”, including both equity and debt.
“In 74 days [since joining the company in June], we put ten billion of liquidity on the balance sheet. So that was my way of saying, hey, I’m going to get stuff done too,” she said.
She added: “We’re in a massive growth phase, it behoves us to keep investing. We need to be on the frontier on the model front. That is expensive.”
To achieve its goals, OpenAI plans to invest in building clusters of data centres in parts of the US midwest and south-west, according to Chris Lehane, OpenAI’s new policy chief.
This push to build its own AI infrastructure follows a similar strategy by Big Tech rivals such as Google and Amazon. Lehane said “chips, data and energy” are the critical resources required to succeed in the AI race.
OpenAI has transformed rapidly in the 12 months since Altman was ousted by the company’s board, and then subsequently reinstated as chief executive last November.
It has brought on its first financial and product leaders, increased headcount by five times to more than 2,000 people, and triggered a complicated transition from a non-profit to a for-profit business model.
While OpenAI has lost key executives across its research and safety teams, including three of its original co-founders this year and high-profile technical leaders, such as Ilya Sutskever and Mira Murati, it has made way for a wave of new engineers and leaders.
Many of the new team have an expertise in building and monetising consumer products. This has led to a dual focus: a long-term research vision and short-term product goals as it focuses on ramping up revenue-generating products to outpace its ballooning costs.
It is spending well over $5bn a year and “not close to breaking even” due to the costs related to building AI models, according to people with knowledge of the group’s finances.
The newer recruits say they are still guided by OpenAI’s “mission” of building and distributing artificial general intelligence — software with cognitive capabilities superior to humans — but are tasked with deploying real-world utility in the near-term.
“The last couple of years, we have had a really big inflection point in the quality of intelligence that can now be made into products that are actually useful for people,” said Srinivas Narayanan, vice-president of engineering at OpenAI, who joined last year from Meta. “That’s . . . why I’m here.”
The launch of AI agents — chatbot-like assistants that help execute tasks on the web, ranging from information gathering to booking or purchasing items — will be a key focus for 2025, according to Friar.
“Agentic has got to be the word of the year . . . It could be a researcher, a helpful assistant for everyday people, working moms like me. In 2025 we will see the first very successful agents deployed that help people in their day to day,” she said.
Rivals including Google, Anthropic and OpenAI’s biggest backer, Microsoft, have all signalled intentions to launch their own AI agents over the coming year.
Meanwhile, ChatGPT’s launch across Apple’s billions of devices, which began its rollout in the US last month, is core to driving a big jump in the number of users.
One of OpenAI’s major venture capital investors noted that the goal of 1bn users could be quickly reached because of this partnership.
“[OpenAI] are already at a few hundred [million] active users today without spending on marketing,” the investor said. “Apple has 2bn iPhones globally and want to push a new AI phone. The path to getting 1bn users with ChatGPT in their pocket is not that farfetched. If you get to that threshold, you’re competing with Google and Facebook.”
Meanwhile, OpenAI will also have to navigate an increasingly complex political landscape.
Lehane, a veteran political strategist who cut his teeth in the Clinton White House, will need to contend with the incoming President Donald Trump’s close adviser and OpenAI’s former co-founder Elon Musk, who runs his own AI company xAI, and is expected to help shape federal AI policy.
Musk recently filed a lawsuit against OpenAI and its backer, Microsoft, accusing Altman of “deceit of Shakespearean proportions” and seeking to void its commercial partnership with Microsoft.
“[Musk] is obviously a unique personality at this moment in time. I think the way we think about it is we control what we can control,” Lehane told the FT.
Despite the public conflict with Musk, he said OpenAI and the Trump team were aligned on AI’s role in national security and economic competitiveness.
Lehane wants OpenAI to take the lead in building US-led “democratic” AI at scale, compared to a Chinese-led version of the technology.
“We’ve had conversations with the transition team . . . both during the campaign and after,” he said.
“This administration has talked . . . about the imperative of . . . US-led AI prevailing over Chinese-led AI. And if you want that to happen as the US government . . . then OpenAI is going to have to be in the middle of that conversation.”
Lehane believes the next few years will usher in a global, historical transition — a period when technology evolves at a pace that societies will struggle to adapt to.
Governments will need to develop new public-private partnerships in AI, similar to an electric utilities model, to fairly distribute the technology and its benefits, he added.
“Part of this company’s responsibility and role, is to . . . potentially shape those conversations and form those conversations, and hopefully be able to find some of the answers as we move forward,” said Lehane.
