Tech stock wipeout

[WendyBG]

https://www.wsj.com/finance/stocks/investors-zero-in-on-runaway-tech-spending-putting-dent-in-ai-trade-1da74e98?mod=hp_lead_pos11

$890 Billion Tech Wipeout Puts Focus on Runaway AI Spending

Concerns about artificial-intelligence spending hit the market value of the Magnificent Seven hard on Thursday

By Hannah Erin Lang , Tina Li and Caitlin McCabe, The Wall Street Journal, July 23, 2026

1. Both Alphabet and Tesla reported soaring revenue, but investors worried about other metrics.
2. Alphabet and Tesla reported negative free cash flow as they increased spending on artificial intelligence, autonomous vehicles and robotics.
3. Analysts expect Meta and Amazon to also post negative free cash flow, while Microsoft is the only major AI spender with positive cash flow.

Wall Street is reckoning with a new reality: The biggest tech companies are no longer cash-printing machines.

Concerns about Alphabet’s(https://www.wsj.com/market-data/quotes/GOOGL) and Tesla’s (https://www.wsj.com/market-data/quotes/TSLA) earnings results spread to other major tech stocks on Thursday, with investors dialed in to the implications of ramped-up capital spending, putting a dent in the artificial-intelligence trade.

Alphabet’s shares slumped some 7%, costing the company more than $293 billion in market value, its largest one-day market-cap loss on record. Shares of Tesla tumbled 15%, the stock’s worst post-earnings performance ever…

Both Alphabet and Tesla reported soaring revenue, but investors instead zeroed in on their AI spending. The phrase of the day: free cash flow—which turned negative at both…

Those companies are racking up unprecedented debt to fund their AI infrastructure plans. At the same time, cheaper open-source AI systems threaten the business models of leading AI labs whose compute and chip needs are driving hyperscale spending… [end quote]

I wrote about this new, thriftier paradigm a few days ago.

A Thriftier Paradigm for AI

We all know that the AI hyperscalers are planning to spend hundreds of billions of dollars on massive data centers for generative AI. But Gemini told me that a new paradigm is developing for business-oriented AI. AI is already used in dedicated applications. This is orders of magnitude less expensive than generative AI which tries to absorb and interconnect all of human knowledge. Dedicated AI applications are vastly less expensive to run and maintain than generalized frontier Generative AI m…

I ended with, “What will the stock market do when the numbers show up in the financial reports of the hyperscalers? Only time will tell.”

Time has told.

The stock market doesn’t like negative free cash flow at all. The stock market doesn’t like incredibly high spending on capital-intensive utility-like data centers which will create competitive utilities when stock multiples reflect the hyperscalers’ SaaS-like low fixed asset business model.

Most of all, the stock market knows how quickly tech can become obsolete and understands how the new paradigm can quickly overwhelm the old. We have all seen this many times in tech. The Return on Invested Capital (ROIC) will plummet if the investments are huge while the demand for the service (frontier AI) plummets as customers swing toward cheap distilled AI. Especially since the huge investments in complex chips will depreciate fast since those chips physically deteriorate in 3 - 5 years.

This is the pin that will pop the tech bubble.

Wendy

[pauleckler]

I still say the real concern is what return they get on that investment. Anything else is anticipation of possible trouble. As always the top two or three are likely to be profitable. The trick is to identify that group.

No one wants to be left behind. All are investing for position but most will end up with crumbs. Much potential for losses or mediocre returns—unless they specialize. But much potential for specialized ai in select fields such as legal.

The data centers w outdated or defunct chips are still an asset. They will likely be refurbed if they are profitable. How much ai capacity will ultimately be required. In the digital world mountains of new data are generated every minute. Capacity required to store and process it is likely to be infinite.

It’s too early to shut down the patent office because we have all we need.

[intercst]

The S&P 500 hit its all time high of 760 on Jun 02. It’s currently 740, that’s 2.6% off it’s high.

The much anticipated 'Tech Stock Wipeout" is showering the rest of the economy with cheaper AI. And you’re seeing that in the broad market indices.

They’re still hiring AI prompt engineers who know how to get the best from AI.

intercst

[McLovin1981]

Alphabet’s earnings were a little misleading at face value. Because of the stupid GAAP rules that require companies to report unrealized gains as earnings, Alphabet reported a huge unrealized gain on their SpaceX investment. So the topline earnings report was kind of screwy.

But it was still a blowout. Revenues were up 24%. This is remarkable for a company approaching half a trillion dollars in annual revenue. Margins increased. Cloud revenue was up 82% and is growing much faster than AWS or Azure. Cloud backlog is over half a trillion dollars. Search, Youtube, etc. all had double digit revenue growth. Huge enterprise uptake with Gemini, 90% of Fortune 100 companies are subscribers. Alphabet shot the lights out last quarter.

The dark cloud was that capex spending drove free cash flow negative for the first time in the company’s history. But how bad is that? It seems to me that that if capex investment results in 82% growth you should be spending every last dime on it. Alphabet has a P/E of 16, but if you exclude the unrealized gains it is more like 26. But that’s a fair price for a company with 24% revenue growth.

Tesla did not shoot the lights out, but it wasn’t bad. Tesla sold 480,000 vehicles, which was a great quarter. Accordingly revenues were up BUT margins were down. They are making less money on every car, which has been the trend for a while.

Free cash flow was down due to increased CAPEX spending, but they said last quarter to expect that. Energy business growth was very good, but that’s still a small part of the company. On the face of it the earnings report was, in the worlds of Supervisor Dyatlov, not great, not terrible.

Tesla has some major operational headwinds, however. For the first time on the earnings call, Tesla admitted they can’t roll out autonomy in a general way. Their AV ride hailing program remains tiny. It is still limited to small, geomapped areas, daylight hours, good weather, and no freeway driving.

The Cybercab appears to be a flop. They can’t be used as robotaxis and they can’t be sold to the public. A few hundred have been built and are just rusting in parking lots. The spent billions in capex for a car they can’t make money with. The Cybercab might be a bigger flop than the Cybertruck.

In previous calls, Musk made huge predictions for Optimus robots, which was supposed to be in full production this summer. Instead, he talked about the many engineering and production challenges they were facing. He mentioned the first production run would be for internal testing and development purposes. So Optimus as a revenue stream seems very far off.

The Semi, which was also supposed to be in full production by now was barely mentioned. Which leads me to believe they didn’t have any good news to share. No talk of customer orders or planned production numbers.

Tesla has a P/E of 287. It is priced like a very high growth company. But that is based on the hope of revenue streams which have not arrived and may never arrive.