Just before the turn of the year, Japan-based holding company SoftBank Group completed a $41 billion investment in generative AI pioneer OpenAI, boosting the ChatGPAT maker’s market valuation to $500 billion.
The deal, which closed in late December, was the multinational conglomerate’s fourth billion-dollar investment in OpenAI. Just two weeks later, the contract with OpenAI was signed AI chip maker Cerebras $10 billion for compute power to boost the inference and performance of OpenAI’s generic AI models.
All this activity follows a difficult year in which OpenAI featured prominently in five of the 10 largest deals in the fast-growing AI industry: $500 billion stargate project with SoftBank and software-hardware giant Oracle; a separate $300 billion long-term deal with Oracle; $100 billion investment in OpenAI by AI chip market leader Nvidia; And the partnership between OpenAI and AWS is valued at approximately $38 billion. Meanwhile, tech vendors and investors spent more than $61 billion last year on construction Hundreds of AI data centers Around the world to provide computation to run AI models now and in the future.
OpenAI seen as weak
And yet, despite the confidence shown in its future by some of the tech sector’s biggest players, countless critics say OpenAI is destined to fail.
Despite its grand ambitions, the former nonprofit research lab’s expenses dwarf its revenues – an estimated $20 Arab By OpenAI’s CFO in 2025. Meanwhile, as OpenAI was struggling to make serious inroads into the enterprise AI market, its rival, Anthropic, was forming. Preferred generic AI vendor for business users and codersHowever, it has also recorded very low revenues. OpenAI continues to subsidize most of its operations with private investment subsidies.
“The amount of money that is needed to justify the fixed costs that occur downstream is because different AI companies today don’t have a business model and they don’t have a cost structure today that works economically,” he said. andy wooAssociate Professor of Business Administration at Harvard Business School. “So overall, most AI companies lose huge amounts of money.”
“The problem is that the commitments to infrastructure by all companies combined are around $1 trillion,” said Johna Till Johnson, CEO of research and advisory firm Nemertes. “So, you have to pay about $800 billion a year to cover the interest on your debt to get your infrastructure. It won’t work.”
Visions of Fall
In the eyes of AI skeptics, the collapse of OpenAI will trigger or contribute to the massive burst of what they see as a fragile AI bubble, a term that means Become part of the everyday AI vernacularEven amid outward signs of an AI boom.
According to AI skeptics, this bubble has arisen due to highly anticipated demand for AI services. Add to that circular financing in which vendors and customers exchange money, unaccounted for excess revenues, highly valued tech stocks, depreciation of AI chips, and unsustainable debt incurred by large AI infrastructure spenders, particularly Oracle and neocloud GPU-as-a-service vendors. corewave.
In that scenario, the bursting of the AI bubble would throw financial markets into chaos and destroy market values, with ramifications for technology and society as a whole. The bubble has not burst yet. Indeed, AI vendors keep spending and borrowing to build infrastructure and create new, bigger, and smarter generative AI models and agentic AI systems that are proliferating in the enterprise world.
AI boom outlook
Those who see AI as a transformative technology similar to the electric grid and the Internet reject most arguments about an AI bubble.
He argues that, unlike the dot-com bubble of the late 1990s, most large AI vendors, especially hyperscalers, are in good financial shape with multiple revenue streams beyond AI.
As far as Nvidia is concerned – that’s with OpenAI At the center of the financing cycle – He said that as the world’s most valuable company, it has cash to spend and is smartly growing its ecosystem by investing in customers. Some say Amazon suffered similar losses in its early years, but weathered shareholder demands for profits while building its customer base and demand.
“We’re saying it’s not a bubble. That’s not to say we don’t think there’s an area of hype,” said Alison Porter, portfolio manager at Janus Henderson Investors, which runs several funds heavily invested in AI. “We’re moving from this generative AI-driven demand situation with larger language models and asking questions of ChatGPT to broader industry use of agentic AI. And as we move toward agentic AI, we see this exponential growth in the number of tokens and the amount of compute supply required.”
Other drivers of demand for AI beyond consumer use are making corporate processes automated and more efficient, including industrial and physical AI in manufacturing, and scientific and drug discovery and medical research, which see an AI boom rather than a bubble.
As far as Oracle and its commitment to building an AI data center to provide OpenAI $300 billion price calculation Spanning about five years, Porter said there is risk involved in the deal – Oracle is expected to borrow $100 billion to bring the deal to fruition – but that the entire AI field will not be brought down if it is not successful. “Is there incremental risk at Oracle? Yes. Is that systemic risk? No,” she said. While Oracle has a strong software and database business, it is not as diverse as other tech giants, as many have noted.
Porter said she sees similar dynamics at work with neocloud vendors, some of which are heavily burdened with debt As they build data centers to meet anticipated demand. “Is there a risk to the stock prices of many of those new clouds? Yes. Does that mean there’s a systemic risk to overall AI development and capital spending? No,” she said.
Porter also sees a runway for OpenAI and Anthropic to meet revenue targets, noting the progress both vendors have made over the past three years. Although still far from the benefits, OpenAI went from Revenue will grow from $2 billion in 2023 to $20 billion in 2025An increase of 1,000%. Anthropic’s revenue is projected to grow from less than $1 billion in 2023 to $5 to $7 billion in 2025, which is also a steep growth curve.
“But they are rapidly burning cash and dependent on large capital investments. We see a tension between real technology breakthroughs and skyrocketing valuations,” said Kashyap Kompela, CEO of RPA2AI Research. “AI valuation is in speculative territory that only leading companies may be able to sustain over the long term.”
Much of the bubble-boom debate hinges on how much demand exists for enterprise AI and whether businesses are still seeing ROI in the early implementation phase of generative and agentic AI.
A recent survey of 350 medium-sized to large enterprises by Omdia, a division of Informa TechTarget, found that organizations are starting to see that ROI. Omdia analyst Mark Becque said the majority of companies surveyed in October reported “very good” to “exceptional” returns on their AI investments.
“Everyone talks about the bubble,” Becque said, referring to an influential MIT study in August that said essentially the same. AI boom was pure hype And most organizations were seeing zero ROI from generative AI technology. “Our data must differ.”
Amid the debate, some on both sides see AI at a stage where elements of both approaches contribute to a kind of middle ground as the Herculean rate of investment and R&D in generative and agentic AI technology continues unabated, with the fear of crash appearing more acute and widespread than ever.
“There is duality in the current AI wave,” Kompela said. “This demonstrates both boom and bubble dynamics. Bubble elements – such as sky-high startup valuations, the fact that some AI investment demand is driven by investment money itself in a circular loop and signs of adoption friction – suggest that a correction may be on the way.”
“Not all players will survive the promotion and some capital will be misallocated,” he said.
