The $2.5 Trillion Tech Selloff Put AI Valuations Under Pressure

The Nasdaq Composite lost an estimated $2.49 trillion in market value on Nov. 5, its steepest one-day decline in the chart shown on air. John Coogan described the session as a roughly 5% drop and the worst day since June 5, when the index shed $1.87 trillion. The scale of the dollar losses partly reflects how much the market has grown, he said, but the move was still unusually large. By the next morning, the Nasdaq was up 1.38% and holding its gains.

The hosts did not settle on a clear explanation for why the market sold off when it did. Coogan noted that stocks had rallied on Wednesday, despite the midterm results and other political developments, before selling sharply on Thursday. He suggested investors may have needed time to digest the news. AppLovin offered a striking counterpoint: Jordi Hays said the stock had risen 67% in five days and gained another 10% after earnings, a sharp swing for a company he described as worth $131 billion.

The possibility of a buying opportunity remained conditional. Hays paraphrased investor Mike Bird’s view: the decline could be an opportunity if stocks bounced back, but not if they fell further. Coogan made the same uncertainty concrete with the old warning against catching a falling knife. Their studio demonstration involved plastic cutlery, not a test of the proverb’s investment wisdom.

The selloff also raised a more specific question: whether investors were abandoning AI companies or reassessing parts of the infrastructure trade. Coogan said the pain appeared concentrated in semiconductors, neocloud developers and politically exposed companies such as SpaceX, while some large technology companies had held up better. He argued that hyperscalers may be in a different position because they already have substantial data-center businesses and other sources of revenue. Apple, he added, was relatively insulated from the political debate and had navigated the moment effectively.

That distinction matters to the argument about Anthropic’s possible IPO. Coogan said the rumored valuation had moved from $2 trillion to figures around $1.5 trillion, then toward $1.2 trillion or $1.3 trillion. The uncertainty was not just about the company’s prospects: it was also about whether this was a sensible moment to enter the public market. He cited a Polymarket estimate that the chance of an IPO by Nov. 24 had fallen to 49%, from roughly 70%–75% during the latest round of rumors.

Coogan described a tension in the available signals. He cited Anthropic’s strong financial position and substantial cash, and a reported $80 billion annual recurring revenue figure, while also saying that Q3 had been weak. The company was still advancing in science and capabilities, he said, but the wider market was turbulent. Hays offered the counterargument: if investors had strong conviction in Anthropic four months earlier, a broad market decline and falling infrastructure stocks should not automatically weaken that conviction. A lower IPO valuation could give people who wanted to invest a chance to buy in at a discount to the prices they might have faced if the AI infrastructure trade had continued rising.

The hosts treated the debate over Anthropic’s valuation as part of a wider sorting process. John Coogan said the selloff looked worse for chip stocks than for some of the larger technology companies that use their products. The distinction echoed James Chanos’s warning, read aloud by Coogan: selling high-beta technology stocks already down 40%–70% to buy low- or no-growth consumer companies at 30–100 times earnings might not be a better trade. Jordi Hays joked about Monster Energy trading at a 42 price-to-earnings ratio; the exchange made the point without resolving what investors should buy instead.

The political risk was also less categorical than the market reaction might suggest. Coogan read comments from Texas governor-elect Gina Hinojosa saying that the state would remain pro-business but needed rules to protect residents in communities where data centers were being built. Her position, as he presented it, was not opposition to data centers outright, but a demand for protections and a more beneficial partnership with Texans. Investor Tae Kim’s reaction, Coogan said, was simply: “Buy stocks.”

Hays recalled hearing an argument that markets may be entering an era of stock-picking: after a period in which buying broadly had worked, the recent declines could make selection more important. Coogan connected that possibility to the histories of successful venture firms that invested in companies such as Google and Amazon through the dot-com crash. Even in a major market reset, he argued, some companies could emerge as the next generation of hyperscalers.

A chart shown on air compared the dot-com and AI market cycles, aligned by trading session. Coogan said the paths looked similar for a time, before diverging sharply: the dot-com line eventually reached about 151 times its starting value, while the AI line stood at about 16 times its baseline at the latest close shown. He cautioned that the comparison depended on where the starting point was set. The chart was a reason to consider how a few eventual winners could outlast a correction, not evidence that today’s market would follow the same path.

Some investors were already positioned for a different version of the trade. Coogan cited a post claiming that Lead Edge Capital’s public fund was up 32% gross year to date and down just 1.2% on the selloff day, with the author attributing the resilience to owning technology companies at single-digit ARR and low double-digit cash-flow multiples. That sat alongside reports of hedge funds struggling with the swings. Coogan and Hays discussed October losses, staff cuts and capital returns, including a report that one fund had fallen 33%. Coogan noted that leverage could make the next few days difficult for firms exposed to semiconductors, though they did not establish which funds had been forced to liquidate.

The stress showed up outside portfolio returns. Coogan shared a report that hedge-fund trading coach Denise Shull had clients cancel appointments amid the market chaos. Hays suggested the cancellations made sense: in the middle of a crisis, a manager may need to solve the immediate problem rather than spend an hour talking it through. The contrast was between investors with different exposures, valuations and ability to withstand a fast move.

The October jobs report came in well above forecasts. John Coogan said employers added 165,000 jobs, compared with an economist consensus of 59,000, and unemployment fell to 4.1%. A post cited on air said the result exceeded all 80 forecasts in Bloomberg’s survey.

Jordi Hays was reluctant to celebrate a single report, saying his default assumption was that the numbers would later be revised down. Coogan replied that recent reports had not yet been revised negatively and said figures from payroll companies ADP and Gusto were consistent with the direction of the data. He treated the report as positive, while Hays’s caution remained part of the exchange rather than disappearing into the headline number.

For Coogan, the composition of employment mattered as much as the total. The report included 19,000 new manufacturing jobs. He contrasted that with gains in health care, which he described as important but less indicative of a new growth engine. Manufacturing, in his view, could support new technologies, vehicles, capabilities and defense.

He also pointed to rising labor-force participation as a useful counterpart to the lower unemployment rate. A falling jobless rate is less encouraging if people have stopped looking for work; in this case, he said, more people were entering the labor force and finding jobs. Coogan cited Apollo economist Torsten Slok’s observation that unemployment had remained below 5% for roughly five years, approaching the 64-month stretch from 1965 to 1970.

Coogan connected that unusually sustained strength to the possibility of much faster economic growth, including a scenario above 20%. He compared the potential economic role of artificial intelligence to oil: a productive input that could be deployed across many parts of the economy. But he did not present that outcome as automatic. Models would need to improve, adoption would have to expand, and there was still much to deploy.

Tesla’s Cybertruck was another case where ambition and market response diverged. John Coogan said production would wind down the following year, describing the move as part of Tesla’s shift from selling conventional vehicles toward autonomous vehicles and robotics. The original targets had been far higher: Jordi Hays recalled a 2019 plan for annual production of 250,000, with a longer-term target of 500,000. Hays put deliveries since late 2023 at about 73,000.

The hosts revisited the gap between the original pitch and the product that reached customers. Coogan recalled an entry-level price of $39,900 and claims about performance, a bullet-resistant body and armored glass. When sales began in 2023, he said, prices started around $61,000—about $21,000 above the original claim. At the launch, a metal ball shattered the truck’s windows during a demonstration. Hays called the incident an example of how even prominent technology leaders can have a badly botched product demo.

Hays owns a Cybertruck and gave a mixed assessment. He said the vehicle’s build quality was poor and described features he found dangerous, including sharp edges that made him warn people nearby to keep their fingers clear. He said visible problems could be spotted from a distance. At the same time, he liked the vehicle’s unusual appearance and said it drove very well when the computer was driving, even if it was not pleasant to drive himself. “It really should never have shipped,” he said, while adding that he was glad it did.

Coogan had hoped the truck might lead to a three-row SUV on the same platform. Instead, the production wind-down sharpened the question of what Tesla would continue to sell as it moves toward autonomous vehicles. The Cybertruck’s trajectory did not erase the fact that it was built and shipped, but it left Hays with a vehicle he said he would cherish—and with fewer choices for buyers who wanted one.

The Wallis Annenberg Wildlife Crossing, built over the 101 freeway in Los Angeles, had its first documented animal crossing before its ribbon cutting: a coyote, recorded by a trail camera. John Coogan described the project as a long-running megaproject costing well over $100 million. Jordi Hays said he had mistakenly assumed the cost was closer to $10 billion, adding that he would support more crossings at the lower figure.

The hosts discussed the bridge’s design as an attempt to make a large piece of infrastructure feel like ordinary terrain to animals. Hays wondered whether it needed to blend into the hillside rather than look too narrow or artificial. Coogan questioned whether a series of smaller bridges might be more sensible; Hays responded with a deliberately outsized alternative, proposing that California cover its freeways or turn the 101 into a continuous tunnel. Coogan pointed out that the state was not in a position to spend hundreds of billions of dollars on the idea.

The exchange was playful, but the cost and scale of the crossing remained the substance: a major public works project, now used by at least one coyote. Coogan said the first crossing was welcome news. Hays’s immediate reaction was that, if the bridge cost about $100 million, he would build many more.

Developers in Texas, Utah and Arizona are using artificial surf lagoons as amenities for high-end residential communities. The appeal, as John Coogan described it, is a reliable session without waiting for waves or dealing with sharks, jellyfish, rocks or currents. At the Cabo Real Surf Club in Mexico, one prospective buyer had put down a $600,000 deposit on a $1.8 million lot and planned to build a $3 million house across the street from the beach, beside a private wave lagoon.

Jordi Hays likes wave pools but questioned the logic of buying a home in Mexico near a chlorinated pool rather than the ocean. Surfing was only one part of what he valued about the coast; he pointed to walking on the beach and experiencing the ocean more broadly. For him, a wave pool’s convenience did not replace the setting that makes a place by the sea appealing.

The developments also carry operating costs. Coogan described the $130 million Zion Shores project in Utah, where half of 56 residential lots had sold within a year of development starting. Homeowners faced a $500 monthly homeowners-association fee, plus surf fees of $30 to $200 an hour. Hays explained that the range reflected peak and off-peak sessions, as well as wave settings: larger waves use more energy and allow fewer people in the pool.

For Hays, the bigger risk was technological obsolescence. Wave-pool technology was improving quickly, he said; pools he had surfed two years earlier no longer seemed as enjoyable compared with a newer one. Anyone buying into a surf community had to trust that the operator would keep updating the equipment. That makes the amenity an ongoing technology investment whose future quality depends on the operator’s willingness and ability to keep pace.

The dispute is not whether artificial waves can be fun—Hays said he loves them—but whether they can substitute for a natural coastline and justify the costs attached to a residential development. The hosts read opposing reactions from readers: one surfer and scuba diver said the pools had no appeal compared with ocean or lake swimming, while another argued that the joy of surfing was in witnessing nature, not just riding waves.