OpenAI Plans Grand Expansion While Losing Billions. Sound Familiar?

From its genesis as a small nonprofit startup in Silicon Valley in 2015, OpenAI has transformed itself into the top artificial intelligence (AI) company today, a status minted by ChatGPT, the most widely used AI chatbot in the world.

Its ambitions are nothing short of grand: Last month, OpenAI CEO Sam Altman stood with President Trump, SoftBank CEO Masayoshi Son and Oracle Chairman Larry Ellison to announce a $500 billion project called Stargate that would build AI data centers and other infrastructure.

San Francisco-based OpenAI has opened more offices globally — in New York, Seattle, London, Paris, Brussels, Singapore, Tokyo and Dublin — and it has diversified its operations across several industries. These include government contracts, the defense industry, and business automation across sectors. Earlier this month, it entered into a joint venture with SoftBank to integrate AI agents into the business workflows of enterprise clients.

If that’s not enough, OpenAI is also diving into hardware. It is designing its own custom AI chip to reduce its dependence on Nvidia, which has an 80% share of the AI chip market. Altman is also collaborating with former Apple design chief Jony Ive to create an AI-powered device, which is not a phone.

Altman has hobnobbed with heads of state around the world, which is rare for a startup CEO.

But cracks are showing: OpenAI has lost several co-founders as well as brilliant, key members of the original staff — such as Chief Scientist Ilya Sutskever, who was part of Nobel laureate Geoffrey Hinton’s select AI team at the University of Toronto. They were replaced by corporate types.

OpenAI is also bleeding money. It has raised $21.9 billion thus far, according to Crunchbase, but it is expected to lose $44 billion in total before becoming profitable in 2029, when revenue is projected to reach $100 billion, according to financial documents seen by The Information last October.

OpenAI expects to spend more than $200 billion by the end of the decade — 60% to 80% for training or running AI models and excluding stock compensation costs. Employee compensation topped $700 million in 2024 and is set to increase since OpenAI plans to increase hiring.

Now the question is, will OpenAI go the way of Amazon, which blazed a trail in eCommerce starting with books but didn’t make a profit in nearly a decade even after it went public? Or will it follow the trajectory of WeWork, another pioneer that once was valued at $47 billion but filed for bankruptcy?

AI experts contacted by PYMNTS have mixed opinions.

“OpenAI’s trajectory is reminiscent of Google’s early years — operating at a loss while aggressively scaling infrastructure and research to dominate a transformative technology,” Akash Nigam, CEO of Genies, told PYMNTS.

“Google invested heavily in data centers, search algorithms, and global expansion. OpenAI is doing the same in AI, a field poised to be as foundational as search was in the early 2000s, if not already,” Nigam added.

“The question isn’t just about profitability today, but about positioning for long-term market dominance,” Nigam said. “With growing adoption, enterprise partnerships, and diversified revenue streams, OpenAI is well on its way.”

Mithilesh Ramaswamy, an AI and security engineer at Microsoft, is likewise bullish.

“I look at OpenAI’s trajectory and I can’t help but think of SpaceX,” which was founded 23 years ago, Ramaswamy told PYMNTS. “They, too, burned through cash at an alarming rate, pushing the boundaries of what was thought possible, facing constant skepticism.

“OpenAI seminally is building the infrastructure for a new kind of intelligence, and that’s not cheap. Sure, the losses are eye-watering, but so were SpaceX’s early rocket failures, but long term they will be successful,” Ramaswamy added.

But others were not so sure.

Andy Cwik, co-founder and CEO of Hubub, told PYMNTS that OpenAI will have to compete with the similarly powerful open-source AI models like Meta’s Llama that are free.

“The real value appears to be in whatever [applications are] built on top of the model, such as a digital assistant,” not the underlying foundation model, which is what OpenAI develops, Cwik said.

“I do not think Sam Altman has a record of building great” apps, Cwik said. “OpenAI is therefore likely to fail, especially given the enormous amount of money they have raised, and the requisite returns those raises require.”

Nathan Brunner, CEO of Boterview, said he is “very pessimistic” about OpenAI’s future.

“There is no ‘secret sauce’ to their AI models,” Brunner told PYMNTS. “Open-source competitors like DeepSeek and Mistral are closing the gap. As AI models become commoditized, the value will ultimately shift to the application layer.”

In apps, Brunner doesn’t think OpenAI is “well-positioned to win the product war. They don’t have a powerful distribution system like Alphabet, Meta or Microsoft. Alphabet can distribute AI with Google, Chrome, Android and YouTube. Microsoft via Windows, Bing, Edge and LinkedIn. Meta via Facebook, Instagram, WhatsApp and Quest.”

“This simply looks like a losing battle for OpenAI,” Brunner concluded.

The post OpenAI Plans Grand Expansion While Losing Billions. Sound Familiar? appeared first on PYMNTS.com.