Back in 2000, a founder told the Journal that everyone knew the Fed would raise rates in February. The venture money kept pouring in anyway. The Fed didn’t stop, and internet stocks crashed. Greg Ip says the same argument is back, and this time it’s about AI.
Ip’s Capital Account column ran in Wednesday’s Journal (A2). The pushback he hears is that AI demand doesn’t care about rates. Hedge-fund manager Bill Ackman asked on X whether a race with near-infinite returns can be slowed by borrowing costs. John Arnold of Arnold Ventures argued higher rates might even feed inflation, since companies pass costs on to customers. The Fed, Ip writes, is “entirely capable of deflating the AI boom and inflation.”
His case runs through liquidity, the easy money that lets investors take risks. A hike doesn’t just raise what AI companies pay to borrow. Credit gets tighter step by step, from government bonds out to the riskiest bets. Then new stock offerings, junk bonds (high-yield debt from shakier companies) and private credit get harder to sell. Ip lists busts with the Fed’s hand in them, from the 1929 crash to the housing bust. He adds that the Fed isn’t aiming at AI. Its target is 2% inflation. But if that road runs through AI’s money, he thinks the Fed would take it.
What the crowd prices for next week’s meeting
The Fed raised rates in mid-September to a range of 3.75% to 4.00%. It meets again Oct. 27-28. By our own pull on Friday, Oct. 23, at 9:22 a.m. PT, the Polymarket crowd put no change at 91.5% and a quarter-point hike at 7.5%. For December it put a quarter-point hike at 73%. Wednesday’s paper printed a 67% chance the Fed’s ceiling ends the year at 4.25%, which means one more hike. Friday’s paper printed a 97% chance of zero cuts in 2026. Those are crowd prices, not forecasts.
The same page ran the Journal’s survey of 71 economists, which we covered Oct. 14. This round adds two election questions. Asked which midterm result brings the strongest growth in 2027-28, the biggest group, 42.3%, said it makes no meaningful difference. The panel sees oil ending 2027 near $75 a barrel, up from $67 before the strikes on Iran began in late February. About three-quarters expected the 10-year to end this year at 5% or higher. Then Thursday happened. The 10-year fell to 4.952% from 5.145%, its biggest one-day drop since March 2023. Friday’s paper added that slower tech spending could cool inflation and ease the pressure on the Fed to raise rates.
Our read
Investments/Risk · IN04, with a Fixed Income · IN02 edge. Ip’s channel runs straight through the most borrowed-against corners of the AI build-out. IPOs, junk bonds and private credit are where that money lives. That’s one reason the desk keeps its safe money in Treasury bills and floating-rate paper (SGOV in 49 books, USFR in 14). It avoids long bonds. TLT sits in 4 older books, with nothing added. The money didn’t leave the market. It moved. We don’t know what the Fed does on Oct. 28, and neither does the crowd. We watch.
Here’s the check for your own plan this week. Open your 401(k) app and look at the bond side. Thursday showed how differently bond funds move. When the 10-year fell almost a fifth of a point, the long-Treasury fund TLT rose 1.87% and the total-bond fund BND rose 0.86%. The T-bill fund SGOV didn’t move. Know which kind you own, and whether that’s the job you gave it. Then put a review on the calendar before the Fed meets Oct. 27-28.