Blend Labs heads into its third-quarter report with an awkward split personality. Its newest product, the Autopilot AI agent, moved from free beta to paid contracts faster than many expected, and profitability has been climbing. Yet management also cut its near-term outlook, flagged a fresh rise in customer churn notices and warned that higher mortgage rates would weigh on volume. The stock has since lost about a third of its value, so the November 5 release, due after the close, is really a test of whether the long-term AI story can hold up while the core business shrinks.
The Street expects a loss of $0.02 per share, an improvement from the $0.03 loss a year ago and in line with last quarter, on revenue of about $32.4 million. That would be a 1.4% decline from the prior-year period, a step backward after second-quarter sales grew 7% to $33.8 million. Management guided to $31.5 million to $33.5 million, and consensus has drifted down from roughly $33.9 million to sit just under the midpoint. That leaves the bar modest rather than demanding. A print near the top of the range would suggest the macro caution was overly conservative; a result near the bottom would confirm that rising rates and churn are biting harder than hoped.
Profit is where the comparison looks least flattering. Non-GAAP operating income came in at $7.0 million last quarter, a 20.6% margin that beat guidance, but management pointed to only $3.5 million to $4.5 million for the third quarter, partly because of the seasonal cost of its Blend Forum event. Beating that range would reinforce the idea that cost discipline and internal AI-driven efficiency are durable. Gross margin, which held at 78.3% despite Autopilot model costs, is another quiet gauge of whether the new product can scale without diluting economics.
The operating details matter more than the headline numbers. Economic value per funded loan has slid from $86 a year ago to $79, and another decline would signal pricing pressure as customers drift toward cheaper or free point solutions. Management estimated the churn notices at a low-single-digit percentage hit to annual revenue; investors need to know whether that figure is stable or growing. Consumer banking, guided to anywhere from a 5% decline to 1% growth after rising 6% last quarter, will show whether softness is spreading beyond mortgage. And any update to the fourth-quarter funded loan forecast of roughly 180,000 to 190,000, already down 10% to 15% as the final Mr. Cooper volume rolls off, will shape views on how deep the trough runs.
On the bullish side, Autopilot is the key. Six lenders signed paid, flat-fee contracts by the last call, including servicer Onity, and a large credit union became the first new logo to adopt it from day one. Management cited a 10% to 15% lift in pull-through and two to four days shaved off loan cycles. Revenue from the product is deliberately excluded from guidance, so the evidence to look for is contract count, conversions from the 65 activated lenders and progress in a late-stage pipeline that grew about 40% in a single quarter.
Sentiment has swung sharply toward optimism compared with a slightly bearish reading before the last report, even as the stock fell. At $1.16, shares trade well below the 200-day average of $1.66 but have bounced from a post-earnings low of $0.88. That suggests some investors see value, yet confidence remains fragile. Ultimately, the report will be judged on whether Autopilot momentum and margin resilience look strong enough to offset eroding unit economics and churn before 2027's promised payoff arrives.