Five podcasts, three explanations for the AI selloff — and none of them is just the election

The trigger

The week's business podcasts agree on one thing: the Democratic sweep in the midterm elections repriced the AI trade. Colin Graham of Robeco put it bluntly on his firm's Market Flash: "The democratic sweep was the spark that caused the sell-off." The Nasdaq lost roughly $2.5 trillion in market capitalisation in a single day after the elections, and the Magnificent Seven stocks lost $1 trillion between them in one day, according to the hosts of CNBC's Squawk Box podcast. m:…9312 m:…3348

But Graham also said the spark landed on dry ground. Credit markets had flagged rising debt costs for hyperscalers and their suppliers months before the equity sell-off, he said, and the broader economy remains fundamentally sound, with strong earnings outside technology. "We haven't seen wholesale selling yet", Graham said. Investors, in his telling, are rotating into value and defensive stocks, not leaving the market. m:…9312

The numbers

Lou Miller of Goldman Sachs gave the correction its scale. AI data centre stocks have fallen 36% from their highs earlier this year, he told The Markets, with memory stocks down 42%, optical networking down 46% and neocloud stocks down 57%. Multiples on the data centre trade have dropped to roughly half their June levels. Miller's read is that the fall has removed valuation as the main obstacle, and that the next leg of the AI trade is healthcare, a defensive sector he described as historically uncorrelated to AI. m:…2711

Miller's prescription follows from that read: the trade for the next six months is the S&P 500 excluding AI, built on high-quality balance sheets, and he expects a relaxed Santa Rally into year-end. Traditional defensive sectors, by contrast, look expensive to him: consumer staples earn and pay out less than high-yield bonds and Treasuries yield. m:…2711

The disagreement

Andrew Sheets of Morgan Stanley disagreed with the alarm, not the facts. Bond markets have grown comfortable with the election result, he told listeners, because the result itself removed uncertainty. "Uncertainty is now lower. After all, we know the result", he said. He argued investors may read the result as a signal that inflationary policies such as tariffs could be moderated, and that more regulation of AI could stretch out the build-out and ease the pressure on bond supply: AI spending has generated record amounts of bond issuance this year. His exhibit: "Following the last 11 midterm elections from 1982 to 2022, the S&P 500 was higher one year later after all of them, regardless of the party that won." m:…9727

Odd Lots took the longest view and the darkest one. Skanda Amarnath of Employ America argued the Democratic victory in both chambers is structurally significant, ending a long Republican advantage in the Senate, and that filling Federal Reserve board vacancies, including the replacement for Jay Powell, will now be significantly harder. He also argued AI capital expenditure is cooling and markets are rewarding capital efficiency over raw spending, and that data centre construction has become politically toxic in many states, forcing governors to change their message. AI-sensitive spending now accounts for nearly 8% of US GDP, he said: a large share to cool. m:…9092

The side argument

On Squawk Box, the argument turned to who saw it coming. Tarek Mansour, chief executive of Kalshi, said prediction markets called the midterms with 98% accuracy, outperforming polls and pundits. "The markets are not creating information. They're weighing all the information that's out there", he said. Ron Baron, meanwhile, kept his eye on the long term, predicting data centres in space will be significant by 2028. m:…3348

So the podcasts agree on the spark and split on the fire. One camp reads a correction that has made valuations healthier and a bond market that has already moved on. The other reads a structural shift: dearer debt, cooler capital spending, a Senate that will fight over the Fed, and data centres that have become a political problem. The disagreement is the story, and the next earnings season, and the next Fed appointment fight, will say which camp was right. m:…2711 m:…9727 m:…9092