Bank, bank said bank said

Business is one of the hubs of the 'AI bubble'

Although the shadow of OpenAI and its still unprofitable business remains in private hands, markets have been rocked throughout the second half of 2025.Speculation about an artificial intelligence (AI) bubble hasn't subsided, despite Nvidia reporting another successful quarter in November.

The question remains as it is open until it swings, on a hand that seems like an infinite love "to the data center that works through on red in black. This is the same question that the real one here COSO Samman answers with one word of Greatness in the form of a new podcast: "Enough."

The investment bank HSBC believes that "mega" means "mega" and believes it is "Megafafs" and is focused on its prediction, and if it is "Megafafs" it shows the deal from its prediction.This rigorous assessment requires more technology trends than infrastructure costs, competition, competition, demand for AI markets, financial investments and financial investments.

HSBC's AnistiSal analysis team, led by nicholas-Colisson, produced the numbers by saving the annual forkless for Micobon and A1 Octam and Amazon dates.It is important, according to HSUBC, that this problem is achieved without injecting new capital and in a series of expansions.The latest capacity sees the opening of 36 gigawats of compression power by the end of the decade.Assuming one pull could power about 750,000 homes, electricity on that scale would represent the needs of a state smaller than Texas and slightly larger than Florida.

However, Openii's long-term growth cash flow will be negative through 2030, with $207 billion of funding to be replenished through equity, debt capital, equity, or more aggressive revenue generation.According to the calculations of HSBC analysts, the costs of Openii Cloud and AI infrastructure will reach 2025 million.For 203 billion dollars, the total estimated liabilities are 2033 billion dollars, and in the next eight years, 1.4 trillion amount of credit plan has been determined.Data rental alone costs $620 billion.

Even so, the projected revenue – growing rapidly, to $213 billion by 2030 – may not be enough to fill the gap.

The bank points to several options for closing the gap, including dramatically increasing the proportion of paying customers (from 10% to 20% that could add $194 billion in revenue), capturing a larger share of digital ad spend, or extracting extraordinary efficiencies from computing operations.But even with an optimistic transition and monetization scenario, the company will still need fresh capital after 2030.

OpenAI's survival is closely tied to its backers and the AI ​​ecosystem.Microsoft and Amazon are not only cloud providers, but also major investors, and cloud companies like Oracle, NVIDIA, and Advanced Micro Devices have everything to gain—or lose—depending on OpenAI's fortunes.However, the risks are significant: unproven revenue models, potential saturation of the AI ​​subscription market, the threat of regulatory scrutiny, and the large volume of capital injection required.

HSBC suggests that OpenAI could raise debt to fund its computing needs, but that would be "potentially the most difficult option in current market conditions" as Oracle and Meta have recently raised a "significant amount" of debt to fund AI-related capital spending, "raising concerns in the market about overall AI funding".The bank notes that this is an exception because most so-called "hyperscalers" are funded with free cash flow, as JPMorgan's Michael Cembalest recently noted.

Like so many other banks that have written about the AI ​​revolution, "You're everywhere except for the pre-computer-age statistics of low productivity," which is an unfortunate feature of today's bloody economy."Some have made a significant return from the 30-year-old Internet revolution, which will convince the governor of a significant return"

Savita Subramanian, director of quantitative strategy and U.S. equities at Bank of America Research, told Fortune in August that she sees a "step change" in productivity coming from the 2020s economy in ways not fundamentally related to artificial intelligence.She says a combination of factors, including post-pandemic wage inflation, are driving companies to "do more with less" by connecting people to processes in a scalable and meaningful way.But one thing that got her thinking was the shift from a focus on small assets to a focus on heavier assets, as many of the most innovative tech companies discovered an almost insatiable desire for high-risk hardware: data centers.

A few months later, Harvard University economist Jason Furman did a quick calculation and found that without data centers, GDP growth would be just 0.1% in the first half of 2025. OpenAI seems to be asking the market a question: How long can AI growth—and a productivity revolution—come, based on the question of future returns, of which there are no guarantees?

This content is interpreted with the help of artificial intelligence tools and reviewed by our editorial team.Learn more in our AI Policy.

C.2024 Fortune way of Asto IP A limit

Distributed by the New York Times Licensing Group