Why The AI Sector Is Crashing? (No, It's Not Jensen's Fault)
The real reason AI stocks are crashing isn't a spy scandal — it's circular financing, unsustainable capital expenditure, and unit economics that just don't add up.
If you have been scrolling through the financial news today, you would be forgiven for thinking the AI market is collapsing because Nvidia's CEO, Jensen Huang, allegedly fell into a geopolitical honeytrap. The media is having an absolute field day with the spy scandal that supposedly tanked the market in a matter of hours.
Let us look at the reality of the tape: Nvidia suffered its worst single-day fall since early 2025, plunging 9.26% in a single session and wiping out roughly $447 billion in market value. The stock is now down nearly 25% from its recent peak. The wider semiconductor index dropped by 11%, and competitors like AMD have shed over 17%.
The tabloid headlines are fantastic for generating clicks, but they are a complete distraction. Jensen's romantic entanglements are not the reason AI stocks are taking a beating. The real culprit is much less theatrical and far more dangerous: a colossal, structural unravelling of circular financing and unsustainable capital expenditure. Investors are panicking, withdrawing billions from private credit funds, and the flow of money into AI has suddenly hit a brick wall.
The Why?
To understand what is actually breaking underneath the surface, we need to look at the economics of AI — and the illusion of recent tech profits.
Instead of genuine organic growth, what we are witnessing is an echo chamber of capital. A mega-cap tech company raises debt or equity and invests a massive sum into a private AI startup. That startup then takes that exact same capital and routes it straight back to the mega-cap company to pay for cloud computing and servers. The tech giant logs this as accelerating revenue growth, while simultaneously writing up the paper value of their initial investment in the startup.
More than half of Google's recent earnings, and nearly 46% of Amazon's profit, have come from revaluing their investments in companies like Anthropic and SpaceX. These are paper gains — one-offs that mask a deteriorating core business.
Meanwhile, the actual cost of running these AI models is accelerating way faster than the revenue they generate. Anthropic recently reported a 47% quarter-on-quarter revenue increase, but their compute costs surged by 77%.
Then there is the physical bottleneck. The real world is rejecting the digital dream. AI data centres are consuming so much power that local grids are buckling. Electricity prices in the US have spiked 20%, and citizens are pushing back. In the recent midterm elections, nearly every ballot measure designed to restrict or ban data centre construction was passed. You cannot build a multi-billion pound AI empire if the locals will not even let you plug it into the mains.
The Evidence
The mathematical evidence of a bubble is stark. Over the last nine months, Google and Amazon alone have claimed roughly $195 billion in combined paper profits through the revaluation of their own investments.
When you remove these "fake" profits from the equation to calculate the true operating earnings, the S&P 500's Price-to-Earnings (PE) ratio jumps from a benign 25.1 to a highly concerning 30. We have only seen sustained PE ratios above 30 immediately preceding, or immediately following, major market crashes like the 2000 dot-com bubble. Furthermore, Anthropic is structurally committing to hundreds of billions in hardware spend over the next decade for a product whose profit margins are currently deeply negative. The unit economics simply do not scale linearly.
How We Navigate This Sector
Volatility is just the toll we pay for market participation, but blind faith in inflated metrics is how you destroy your capital. Here is a pragmatic, unemotional checklist for navigating this AI unravelling:
1. Strip Out the Paper Gains
When you evaluate mega-cap tech earnings, completely ignore investment revaluations. Focus strictly on operating cash flow and core revenue growth. If the numbers only work because of accounting gymnastics, categorise the stock as highly speculative.
2. Focus on the Infrastructure Bottleneck
The ultimate ceiling on AI is not code; it is electricity and copper. Keep an eye on energy constraints and regulatory pushback on data centres. The companies that solve the power grid problem will be the actual long-term winners, not necessarily the ones training the largest language models.
3. Pound-Cost Average Defensively
Do not try to catch falling knives in the semiconductor sector. Maintain your discipline. Continue to pound-cost average into your ISA, but ensure your allocations are heavily diversified into broad, non-tech weighted global trackers that are insulated from this Silicon Valley echo chamber.
4. Ignore the Headlines
Whether a CEO is caught in a tabloid scandal or a company promises a shiny new 100-gigawatt server farm, tune out the noise. Let the underlying unit economics dictate where you deploy your capital, not the panic of the herd.
Related Posts
Midterms, Musk, and Market Panic: Staring Down a Sea of Red Without Fear
Between the sharp correction and the Democrats' midterm sweep, the tech sector has split violently. Here's why abandoning ship right now is a terrible idea.
How AMD and Broadcom Are Exploiting Nvidia's Chaos?
With Jensen Huang out and Nvidia distracted, AMD and Broadcom are making their move. Here's exactly how each rival is positioning to capitalise on the leadership vacuum.
Can We Be Bullish on Nvidia?
Jensen Huang's abrupt exit sent Nvidia shares tumbling, but the fundamental investment case remains entirely intact. Here's why the panic selling is creating a generational buying opportunity.
Disclaimer: This article is for educational purposes only and does not constitute financial advice. Past performance is not indicative of future results. Always consult a qualified financial adviser before making investment decisions.