The ghost of 2000 is making a comeback. A Treasury report is sounding the alarm that the current AI investment mania bears a striking resemblance to the dot-com bubble, and that a sharp downturn in the sector could trigger widespread economic fallout across markets, including crypto.

Here’s the thing: the numbers backing up that comparison are getting harder to ignore. Expected long-term earnings growth for the S&P 500 hit 20.2% in mid-2026, surpassing the 18.6% record set during the actual dot-com peak. That’s not a fun record to break.

The investment surge that’s raising eyebrows

Global corporate AI investment reached $252.3 billion in 2024, according to the Stanford AI Index. Tech giants then pledged roughly $320 billion in capital expenditure for 2025, with a massive share earmarked for AI infrastructure.

The IMF has weighed in with its own warning, highlighting the risk of abrupt repricing in tech stocks if AI earnings disappoint. The fund flagged potential knock-on effects for broader macrofinancial stability, the kind of cascading damage that doesn’t stay neatly contained within one sector.

Gita Gopinath, the IMF’s First Deputy Managing Director, put a staggering number on the downside scenario. She estimated that an AI bubble bursting could erase over $20 trillion in US household wealth and $15 trillion for foreign investors. Ray Dalio has echoed similar caution, pointing to the dangerous gap between current valuations and proven returns.

What this means for crypto investors

Bitcoin miners saw their stock prices surge in 2025, driven in part by speculation around AI-related energy demand. The logic was that AI datacenters would need enormous amounts of power, and miners with existing energy infrastructure could pivot or benefit from the boom.

That thesis works beautifully in an up market. In a correction, those same mining stocks become doubly exposed, vulnerable to both a crypto downturn and an AI sentiment reversal simultaneously. Investors holding mining equities as an AI-crypto convergence play should understand they’re not diversified. They’re concentrated.

Crypto treasury firms managing Bitcoin and Ethereum holdings have shown more resilience to broader market swings, operating somewhat independently from the AI narrative. But “somewhat independently” has its limits when a $20 trillion wealth destruction event is on the table.

For investors navigating this landscape, understanding the difference between “AI will transform the economy” (probably true) and “AI will transform the economy on the exact timeline priced into current valuations” (much less certain) is the entire ballgame. With earnings expectations now officially exceeding the 2000 peak and multiple heavyweight economists flagging systemic risk, the smart money isn’t necessarily selling everything. It’s stress-testing portfolios against a scenario where the AI revolution takes twice as long as consensus expects.