By Hamish Armstrong (Senior Communications Officer), Published

OpenAI’s latest wave of deals with chipmakers including Nvidia, AMD and now Broadcom have generated significant publicity due to their scale. Coupled with agreements to build and develop data centres, OpenAI’s deals are now reported to be worth up to $1.5 trillion.

Concentration of investment around a handful of major players in the industry, however, is both unsustainable and economically irresponsible, according to Professor Feng Li, Associate Dean for Research & Innovation at Bayes Business School.

Commenting on the eye-watering sums of money involved, Professor Li offers a stark warning to industry regulators that as many as three bubbles are set to burst if action isn’t forthcoming on scrutinising such deals – putting taxpayers and savers at perilous risk.

“As the cliff edge approaches, AI regulators are asleep at the wheel,” he said.

“OpenAI’s latest partnerships with Nvidia, AMD and now Broadcom have pushed valuations and commitments into a realm where money circulates among a handful of dominant firms that are increasingly detached from economic fundamentals.

“These are not normal business transactions. OpenAI lacks the capital to fund the deals while its partners rarely have the facilities to deliver what is pledged. Each depends on the others’ inflated valuations to justify their own. Nvidia, for instance, profits handsomely from chip sales, yet most buyers cannot turn those chips into profits.

"This circular flow of promises and paper wealth amplifies systemic risk without creating commensurate real value.

“Ant Group’s IPO was halted because regulators feared its extreme leverage could destabilise China’s financial system. Again, circular and opaque financing behind OpenAI’s trillion-dollar infrastructure commitments now exposes a comparable global risk on a far larger scale.”

Commenting on whether the AI bubble was set for a crash, Professor Li, a member of the Bank of England's Artificial Intelligence Consortium, said there could be dire consequences unless regulators woke up to reality.

“Three bubbles are inflating at once – financial, infrastructural, and utility. When they burst, taxpayers and savers will ultimately pick up the tab.

“When trillion-dollar AI infrastructure deals no longer raise eyebrows, something is seriously wrong. Regulators who fail to act now will be guilty of a dereliction of their most basic duty: to safeguard financial stability before it is too late.”


All quotes can be attributed to Professor Feng Li, Associate Dean for Research & Innovation at Bayes Business School. Professor Li is a member of the Bank of England’s Artificial Intelligence Consortium and is available to provide further comment on AI development and how is redefining work.

For inquiries, please contact the City St George’s, University of London press office.