AI Bubble Could Burst in 2028, Triggering 40% Tech Stock Plunge and $20 Trillion Wipeout

AI Bubble Could Burst in 2028, Triggering 40% Tech Stock Plunge and $20 Trillion Wipeout

TREE NEWS reports: TS Lombard’s chief macro strategist Dario Perkins has outlined a detailed scenario in which the AI investment boom unravels in early 2028, leading to a sharp market reset. The report warns that the S&P 500 could fall 30%, tech stocks could plummet 40%, and roughly $20 trillion in market value could be erased, dragging the technology sector’s weight in the US market back to pre-ChatGPT levels. While the tech giants would survive thanks to strong balance sheets, weaker players in the AI ecosystem—such as emerging cloud service providers and independent model developers—face a wave of defaults, with estimated credit losses of $125 billion to $190 billion.

The core thesis is that the AI boom is built on fragile foundations. Current revenue growth largely stems from “capex recycling” within the tech ecosystem—one company’s capital expenditure becomes another’s revenue—and increasingly from debt financing, rather than from genuine end-user demand. Hyperscalers’ capital spending as a share of US GDP is near historic highs, but the proportion of revenue from outside the ecosystem has barely improved. For instance, about 80% of OpenAI’s and Anthropic’s revenue comes from just 1% of customers, most of which are loss-making AI startups dependent on venture capital.

Six Triggers That Could Pop the Bubble

The report identifies six potential catalysts, any of which could set off a chain reaction:

• Revenue shortfall: To justify current capex, AI industry revenue would need to reach $2.5–3 trillion by 2030—about 14 times current levels—a pace deemed unrealistic, especially with Chinese open-source models gaining share and consumer resistance to paid AI services.
• Rising credit risk: AI companies are expected to issue about $700 billion in new bonds in 2027 alone, with hyperscalers accounting for $400 billion. Off-balance-sheet liabilities across the AI ecosystem are estimated at $1.5–2 trillion, and some emerging cloud providers are already paying 15% interest.
• Teaser-rate trap and depreciation cliff: Data centers incur no depreciation during construction, but once operational, depreciation hits hard. Multi-year leases signed by AI labs are not expensed until the data center is live, potentially causing financial conditions to deteriorate within 12 months.
• IPO wave reveals truth: Planned IPOs from Anthropic and OpenAI (the latter delayed to 2027) could mark the cycle’s peak and expose opaque metrics like annual recurring revenue (ARR).
• Semiconductor cycle and supply chain risk: Nvidia’s vendor financing—essentially “balance sheet as a service”—and the bullwhip effect in chip demand could turn shortages into gluts once demand slows.
• Regulatory and safety risks: With 55% of Americans favoring slower AI development and AI-assisted cyberattacks on the rise, a major security incident could trigger political backlash and tighter regulation.

The Second Derivative Problem: Slowing Growth Is Enough

The report emphasizes that AI capex does not need to decline outright—a slowdown in growth is sufficient to create financial stress. This mirrors the subprime crisis, where collapsing home price appreciation, not falling prices, triggered the meltdown. Hyperscalers’ capex acts like a loan book: once it slows, the whole boom unravels.

Market Impact: $20 Trillion Wipeout, Mild Recession

In the crash scenario, the S&P 500 drops 30%, tech stocks fall 40%, and other sectors decline about 20%, erasing roughly $20 trillion in market value. The AI-related debt pile totals about $2.9 trillion, with 25% at risk. Assuming a 35% cumulative default rate from 2028 to 2030 and 25–50% recovery rates, credit losses would be $125–190 billion—significant but not systemic. Equity holders bear the brunt.

The economic fallout would be a mild recession similar to 1990 or 2001: AI data center spending declines shave about 2 percentage points off US GDP and eliminate 800,000 jobs, but healthy private-sector balance sheets and ample Fed easing space prevent a 2008-style crisis. Lower rates would revive the “K-shaped” economy’s lower tier, potentially fueling a new housing boom, just as after the dot-com bust.

Key Takeaways for Investors

• Don’t short tech or pile into duration yet: If the AI boom has years left, early positioning could underperform badly.
• Follow the “easy bubble” framework: Hold a diversified global equity portfolio that benefits if the boom continues and outperforms concentrated AI bets if it bursts.
• Watch for stress signals: Only gradually add government bonds when clear signs of stress emerge in the AI ecosystem.
• Geographic diversification: Europe and other markets less exposed to AI could outperform after initial selloffs; the dollar may weaken more than in past crises as capital flows to the euro, yen, and pound.
• Tail risk: An AI crash could accelerate automation, leading to a jobless recovery and deeper labor market scars.

The message is clear: this is not 2008, but it could be 2001. A 40% tech drawdown and $20 trillion in lost wealth would devastate any portfolio heavily concentrated in AI.