Tech Is Leading Rates, Rates Isn't Leading Tech

The rates trade is a CapEx trade

Keith Lerner, Truist's chief market strategist, told Bloomberg Surveillance that the usual direction of causality has flipped: "I'd make the argument that tech's leading rates, rates isn't leading tech. Tech's been pushing rates higher because of the CapEx build-out that's fueled much, much hotter nominal GDP in this country." Ten of eleven S&P 500 sectors are up, and the one that is down is tech — "10 out of 11 sectors are up. And one sector is down. That one sector is tech." The sector has more than doubled the index since late 2022, surviving five double-digit pullbacks along the way, and forward earnings estimates are up 18 percent over the last three months. The selloff, in Lerner's framing, is a reset inside a bull market, not the end of one. m:…9711

The macro desks see the same mechanism

Nomura's Jeremy Schwartz gave the economist's version on The Week Ahead: AI-driven capital expenditure is contributing to economic overheating, and a fundamental reassessment of that capex could reverse the pressure. His core PCE tracking estimate jumped from around 20 basis points to 31 or 32 — "Our core PC tracking estimate before the data was around 20 basis points, currently stands at around 31 or 32" — and he expects the Fed to hold in October even as the data keeps terminal-rate expectations for 2027 under upward pressure. The "animal spirits" behind AI investment show no sign of reversing: "people are still buying all the equipment they can get their hands on. They're still building every data center they can get a permit for." m:…2863

CNBC's Fast Money panel added the policy transmission: the AI build-out is now shaping monetary policy itself, with markets pricing a 17 percent chance of an October hike even as pipeline inflation shows up in PPI — energy accounted for 80 percent of the goods-side move, per the panel's read. High energy prices, they argued, will keep flowing through into CPI and PCE data. Nomura, for its part, flagged contagion from the AI sector into the credit market as a key risk the Fed is watching, even as credit conditions stay broadly accommodative. The inflation story and the AI story have merged into one trade. m:…1735 m:…2863

The financing brake

Morgan Stanley's Carolyn Campbell described where the brake actually engages. Highest-rated data center bonds yield close to 7 percent, up from below 5 percent at the start of the year, and "Our analysis suggests that when borrowing costs approach 8%, issuing these bonds becomes significantly more challenging under these current financing terms." Five-year Treasury rates are up more than a point since January. Operators have still managed about $15 billion of ABS and $7 billion of CMBS issuance this year — and Morgan Stanley expects $120 to $150 billion of data center supply over the next two years — but power availability could leave capacity 30 to 40 percent below projections, even as U.S. capacity is supposed to grow from 60 to 150 gigawatts by 2028. m:…9553

Dan Ives supplied the bullish counter on Bloomberg: the cycle is in the "third inning" — "demand's telling you we are third inning relative to this AI game" — with a $67 multiplier for every dollar of AI capex across the rest of tech, and 800 to 1,000 data centers expected to survive the political pushback even if 5 to 10 percent of planned projects get voted down. The desks are not arguing about whether the buildout continues. They are arguing about who can finance it — and at what price the bond market says no. m:…3854