Why Are Tech Stocks Suddenly Crashing?

The best assistants don't multitask their attention across a hundred tools. Neither does Catch. It's an AI agent that focuses on one thing — the admin work you'd rather not touch — and does it exceptionally well.

Scheduling, flights, restaurants, follow-ups, vendors, clients. You hand it over; Catch handles the back-and-forth and comes back with it done.

No context-switching. No dropped balls. Just your admin, quietly cleared — so your focus stays on the work only you can do.

Meet the agent built for admin, and it'll be ready to work before your next meeting.

Get started at catchagent.ai — and give your attention back to what matters.

Start now for free

On 23 July 2026, US markets experienced one of the most confusing sell-offs in recent history. Unlike traditional crashes—where weak earnings, slowing growth, or corporate losses trigger panic—this one happened despite record financial performance.

So why did investors panic?

The answer isn't weak earnings. It's the realization that AI spending has reached a scale where investors are beginning to question whether future profits can justify today's extraordinary investments.

To understand the market's reaction, we first need to distinguish profit from free cash flow.

Imagine a coffee chain called Brev.

Although the business remains profitable, its cash position falls:

The company isn't losing money—it is investing heavily—but cash is leaving the business faster than it is being generated.

This is exactly what happened at Alphabet.

Although Google generated $39 billion in operating cash (up 41% YoY) and reported $12 billion in quarterly profit, it spent $44.9 billion on AI infrastructure.

Result:

This marked Alphabet's first negative free cash flow since 2004.

Even more importantly, Alphabet disclosed approximately $811 billion in future purchase commitments and contractual obligations.

For years, Wall Street rewarded rising AI spending as a sign of confidence. On July 23, investors suddenly interpreted the same spending as evidence that AI investment may be becoming too expensive.

According to calculations cited from JP Morgan, investments of this scale require roughly a 10% return, implying AI needs to generate around $650 billion annually.

Current projections fall well short.

The transcript estimates:

Combined annual revenue:

Even then, these companies are estimated to collectively lose $20–30 billion.

This creates a troubling gap between today's infrastructure spending and tomorrow's expected cash flows.

If AI companies aren't yet profitable, how are they signing trillion-dollar infrastructure deals?

The transcript explains this using a simple analogy.

Suppose Dave manufactures commercial espresso machines costing $12,000 each.

Instead of asking cafés to pay upfront, Dave charges only $1,000 initially, allowing them to pay the remaining $11,000 only after becoming profitable.

Suddenly, Dave appears to have sold 100 machines, generating $1.2 million in revenue.

But he hasn't actually secured revenue.

He's made 100 bets.

Each café now needs to sell roughly 200 cups of coffee every day to make its payments.

If all the cafés succeed, Dave becomes a visionary.

If they fail, his revenue disappears because it depended on future profits that never materialized.

This financing model is known as vendor financing.

The transcript argues that Nvidia is effectively operating inside a similar financing ecosystem.

Current reported negotiations include:

For perspective:

Nvidia's annual revenue is approximately $216 billion.

The capital cycle works like this:

Money effectively circulates within the same ecosystem.

The concern isn't that Nvidia's technology is weak—it's that suppliers, financiers, lenders, and customers are becoming increasingly interconnected.

Wispr Flow turns your voice into clean, professional text inside any app. Emails, Slack, client updates — speak once, send without editing. 4x faster than typing.

Try Wispr Flow free

OpenAI has reportedly committed around $1.4 trillion in infrastructure while generating approximately $25 billion ARR.

CEO Sam Altman has acknowledged that even $200/month Pro subscriptions remain unprofitable.

If future custom chips like Jalapeño, alongside newer AI models, dramatically reduce inference costs, these investments could eventually become extraordinarily profitable.

If not, today's commitments become far harder to justify.

Legendary investor Michael Burry highlighted another concern.

Imagine buying an ₹80,000 phone.

Nothing about the phone changed.

Only the accounting assumption changed.

Similarly, a $30 billion AI server fleet:

Reported profit instantly increases by $5 billion annually without generating any additional revenue.

According to the transcript, companies have lengthened estimated server lives:

Whether these assumptions prove reasonable won't become clear until 2028.

Building AI infrastructure may cost over $3 trillion, with JP Morgan estimating over $5 trillion once power infrastructure is included.

Few companies can finance that directly.

The transcript cites Meta's approach.

Instead of borrowing $29 billion itself, a newly created entity—Binet Investor LLC—borrowed approximately $27.3 billion from bond investors (maturing in 2049).

The shell company owns the data center.

Meta leases it while owning only about 20% of the venture.

This structure keeps debt off Meta's balance sheet while still giving investors confidence through Meta's involvement.

Collectively, Meta, Alphabet, Amazon, Microsoft, and Oracle reportedly account for roughly $1.65 trillion in AI-related debt commitments.

The transcript ends with a signal many equity investors overlook.

A Credit Default Swap (CDS) is essentially insurance against loan default.

Example:

Higher CDS prices indicate higher perceived risk.

After Nvidia's OpenAI financing news on 27 July 2026, insurance costs on Nvidia's debt recorded their largest intraday jump, suggesting credit markets had become significantly more cautious.

As the transcript puts it:

The transcript doesn't conclude that AI is a bubble.

Instead, it argues that the market is wrestling with a fundamental uncertainty.

If AI token costs continue falling over the next three years—as they have over the previous three—AI could become embedded across virtually every enterprise.

In that scenario:

However, if profitability never arrives, today's enormous spending, circular financing structures, debt commitments, accounting assumptions, and financing guarantees could all become significant vulnerabilities.

The market's July sell-off wasn't driven by collapsing businesses.

It was driven by a simple question:

Can hundreds of billions of dollars in AI investment eventually produce trillions of dollars in economic value?

For now, investors remain divided—and that uncertainty explains why some of the world's strongest companies experienced one of the most confusing market declines in recent history.

Read more here

Trade on real-world events you already follow, from elections and inflation to sports, tech, and more. Choose “Yes” or “No” based on what you think will happen and see how your prediction plays out.

Pick your market and start trading what’s next.

Trade $25, Get Up to $500

_{Bonus credit varies from $15 to $500. Terms apply.}

Disclaimer: The information provided on this website is for educational and informational purposes only and does not constitute financial, investment, or trading advice. Investing in securities involves risk, including the potential loss of principal; always conduct your own research and consult a qualified financial professional before making investment decisions.