The AI boom has turned global markets into a kind of high-voltage arena where money, hype, and fear increasingly share the same stage. That was the mood in New York this week as two veteran finance executives described a financial system hurtling into an AI-driven future without fully accounting for the risks forming beneath its feet.
Speaking at the Reuters Momentum AI 2025 conference, Matthew Danzig, managing director at Lazard, and Joanna Welsh, chief risk officer at Citadel, painted a picture of markets where enthusiasm is running far ahead of fundamentals. AI, Danzig said, has become “the number one topic of conversation” among both investors and corporate executives. And with that obsession comes a predictable scramble: companies racing to define an AI strategy, hunting for proprietary data, and acquiring technology they cannot realistically build on their own.
“Every company that’s a potential target is figuring out their AI angle,” Danzig said, describing an acquisition landscape where nearly any firm can pitch itself as an AI play.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Valuations have drifted into historical extremes as investors pay for future potential, not present income.
“It’s markets willing to pay for the future,” he said.
Those expectations are already visible in the numbers. McKinsey & Co., Danzig noted, estimates that the industry will need roughly $7 trillion in capital by 2030 just to finance data-center expansion. It’s an astonishing figure that dwarfs the capital requirements of previous tech cycles. Yet despite the sheer volume of leverage flowing in, investors have shown limited interest in the absence of revenue needed to support that debt.
The tension surfaced again this week in the trading pattern of Nvidia, the silicon powerhouse whose chips have become the backbone of AI infrastructure. After the $4.5 trillion company posted record revenue and a 65% year-over-year jump in net income for its fiscal third quarter, the stock initially surged. Then the sentiment jolted. By Thursday afternoon, Nvidia shares were down 2.2% at $182.46, pulling down other tech names with it as worries about a potential AI bubble re-entered the conversation.
Under the sparkle of record earnings, Welsh warned, lies a market architecture that is becoming increasingly fragile.
Citadel, which manages $71 billion in assets, has been modeling scenarios in which shocks propagate through markets faster than most investors expect.
“Markets are just faster,” Welsh said. “These volatility spikes and pulses, they hit harder, they fade faster, they repeat more often.”
It’s a dynamic that becomes even more concerning when paired with credit-market behavior that Welsh and others have been tracking for months. She said risks are now “starting to converge and stack” with the AI boom, especially as companies issue 30- and 40-year bonds on assets that depreciate in roughly four years. The mismatch is stark because firms are locking themselves into decades of debt for technology that may be obsolete halfway through the current business cycle. Such distortions, she said, strain cash flow and deepen systemic vulnerabilities.
At the lower end of the credit spectrum, the picture is no more reassuring. Welsh pointed to a surge in zero-coupon convertible bonds issued by less creditworthy tech firms — instruments that give investors equity upside in boom times and priority in bankruptcy, but offer no coupon payments. The renewed appetite for these bonds is a signal she has seen before.
“Zero-coupon converts are having a big issuance year, same as they did in 2001, the same as they did in 2021,” Welsh said, invoking two eras that preceded sharp downturns. And when combined with capital flowing into illiquid private credit, she added, the mix becomes combustible. “You can see how there’s some portfolios where… a brush fire could be pretty healthy.”
What she meant was that a small shock, a wobble in tech valuations, a batch of weak earnings, a sudden liquidity squeeze, could ignite far larger disruptions. The market, she implied, has created its own dry tinder.
The broader message from both executives was not that AI’s economic potential is exaggerated. It’s that the infrastructure supporting the boom — from data-center financing to corporate capital structures — may not withstand the speed of the cycle it has created. Investors are pouring money into the future at a pace that leaves little margin if that future arrives more slowly than expected.