Podcasts Agree: The AI Trade Has Split Into Winners and Losers

Dispersion, not a rising tide

The most consistent message across this week's business podcasts is that the AI trade has stopped moving as one thing. Lee Coppersmith, speaking on Goldman's The Markets, said AI demand remains strong but increased competition is causing significant dispersion in market reactions. Retail investor volumes in the AI trade are down approximately 20% compared to late June, options market sentiment has shifted toward a more two-sided debate with increased put buying, and credit market spreads are widening due to a supply-demand imbalance in AI-related financing. m:…0699

Jack Janasiewicz of Natixis Investment Managers put numbers on the split. Approximately 60% of S&P 500 market capitalization is related to the AI theme or its derivatives, yet the median constituent of the index is up 5.6% year-to-date while the index itself is up over 13%. Capital markets are becoming more discerning in funding AI-related companies, he said, leading to wider spreads for so-called neocloud providers. Natixis cut positions in the AIRR and PICK ETFs at the beginning of October to move toward direct AI exposure, and Janasiewicz argued that a slowdown in the AI capital expenditure cycle could lead to lower interest rates and a rotation into broader market leadership. m:…6399

The bottleneck is execution

If money is not the constraint, what is? Connor Teskey, chief executive of Brookfield, told Goldman Exchanges that the bottleneck for AI infrastructure is the operational capability to execute projects, not the availability of capital. Energy demand is the critical input to growth for every major company and country, he said, and there is a scarcity of it. Sovereign states are increasingly backstopping AI infrastructure build-outs in Europe and other regions to ensure productivity gains, and Brookfield's energy business is shifting from selling power to utilities to selling directly to corporate counterparties, with energy storage its fastest-growing segment. m:…2227

Franklin Templeton chief executive Jenny Johnson added a financing warning on the Sourcery podcast: long-term financing for data center infrastructure is risky because it bets on today's technology remaining static for 20 to 30 years. She also said AI is a disruptor that will eventually provide meaningful upside to company earnings, and pointed to Franklin Templeton's own use of AI agents in sales, which produced a 23% increase in client visits and an 11% increase in sales. The two points sit together awkwardly, and that is the point: even believers in the payoff are uneasy about how the buildout is being funded. m:…7466

Earnings draw the line

This week's earnings gave the dispersion thesis its exhibits. Alphabet shares dropped nearly 13% due to rising capital expenditure and the cannibalization of search advertising revenue by AI, according to the NAB Morning Call discussion, while Microsoft reported a surge in its cloud business while keeping capital investment in check. Caterpillar's third-quarter earnings only just beat street estimates, which the same discussion viewed as insufficient given its role as a data center equipment supplier. m:…7280

Intel fared worse. Its shares fell more than 20% following a revenue forecast miss, and its gross margin declined by more than 600 basis points quarter-over-quarter, according to Bloomberg's earnings discussion. Intel expects the PC market to decline 12% in 2026 and 10% in 2027, and semiconductor peers providing robust guidance more than two years out are creating peer pressure for Intel. Amazon, by contrast, merely met the bar: AWS net sales growth reached 39%, meeting analyst estimates, and Amazon projects AWS revenue growth above 40% for 2027 and 2028, even as its trailing 12-month free cash flow showed an outflow of $42.4 billion. Fast Money's discussion added that Alphabet's stock formed a death cross, with its 50-day moving average falling below its 200-day. m:…2150 m:…6151

Put the pieces together and the podcast consensus is less a verdict on AI than a warning about indiscriminate exposure. Demand is real, capital is available, and yet returns are dispersing, financing spreads are widening, and the physical work of building, powering and executing is now the binding constraint. The open question is whether the next leg of the trade rewards the companies doing that work, or the broader market waiting for the capital spending cycle to cool. m:…0699 m:…2227 m:…6399