Blend Labs: A $4 Price Target Chasing a $2 Stock With Negative Equity

# Blend Labs: A $4 Price Target Chasing a $2 Stock With Negative Equity

Here's a fun exercise: find me another company where the sell-side consensus says "buy" unanimously, the price target implies 65% upside, and the balance sheet shows negative $55.9 million in equity. That's Blend Labs right now — a stock that wants you to believe in the AI-lending-software future while its debt-to-equity ratio (1.25) reminds you it hasn't finished paying for the past.

Blend reported Q2 2026 earnings on August 6 — two trading days before this column — and by the numbers it was a genuinely fine quarter. Revenue came in at $33.8 million, up 7% year-over-year and near the top of guidance. Non-GAAP operating income beat expectations. Free cash flow turned positive, at $6.2 million. And the AI Autopilot product — the company's push into agentic AI for mortgage pre-underwriting — got real airtime on the call, with management pointing to it as the next leg of growth after processing 25,500 loans in a recent stretch. Blend even rang the NYSE Closing Bell the day before earnings, which, sure, doesn't move fundamentals, but it does tell you the company is in full "look at us" mode.

So why is the stock still parked around $2, roughly where it's traded for a while, while Citizens JMP reiterates a $4.00 target and the broader analyst consensus sits near $3.33 — a 66% gap that's persisted through an earnings beat?

Because the market isn't stupid, and the fundamentals underneath the "beat" are thinner than the headline suggests. Net margins are still deeply negative, at -19%. Gross margins are excellent — 74.6%, which tells you the software itself has real pricing power — but the company is burning through operating expenses fast enough that "beat guidance" and "profitable" remain two very different sentences. Forward EPS estimates sit essentially at breakeven, and one estimate puts forward revenue growth at just 2.1%, which is a far cry from the "AI is eating the mortgage industry" story the bulls want to tell. Zacks put BLND on its Strong Sell list on July 31 — a week before the earnings report that was supposed to change the narrative. That's not nothing.

And then there's the balance sheet, which is the part of this story nobody wants to lead with. Negative equity of nearly $56 million and a debt-to-equity ratio above 1.2 means Blend is levered against a business that still hasn't proven it can be consistently profitable on a GAAP basis. This is a company whose fate is tethered to mortgage origination volumes — a cyclical, rate-sensitive business it doesn't control — trying to grow its way out of a structurally fragile capital position. That's a tough two-front war.

The technical picture is genuinely interesting, I'll give it that. The stock's been in an uptrend since May, RSI is pushing into overbought territory near 65, and short-term forecasts point toward $2.23. But the model behind those forecasts has a directional accuracy of just 64% and a MAPE of 19% — translation: don't confuse "trending" with "predictable." This is a stock where a decent chart is doing a lot of the emotional heavy lifting that the income statement can't.

Here's my read: the bull case isn't fantasy — Autopilot is real, gross margins are real, and if MeridianLink integration and AI adoption actually convert into durable revenue growth and margin expansion, $3.50-plus is not a crazy destination. But "not crazy" and "likely" are different words, and right now the fundamentals score (call it a 3.8 out of 10 on our internal read) is fighting the growth story to a draw. The bear target of $1.65 and the bull target of $3.01 aren't close to each other for a reason — this is a binary-outcome stock dressed up as a steady compounder.

My take: this is a trade for people who like volatility and have position-sizing discipline, not a conviction holding for people who like sleeping at night. Watch the debt-to-equity line, not the price target. The stock will tell you it's healed when that number starts moving toward 1.0 — not before.