Earnings call transcript: Blend Labs Q4 2025 sees revenue beat, EPS miss

Blend Labs reported Q4 2025 revenue of $32.37M, up 7% YoY and 1.86% above consensus, while EPS was $0 versus $0.0038 expected (100% negative surprise). Non-GAAP operating income was $5.4M, free cash flow was $1.3M in Q4 ($2.8M for the year), and the company ended with $68.3M cash and zero debt; shares jumped 8.98% after-hours to $1.82, supported by the launch of Blend Autopilot and a new $50M buyback authorization. Management guided Q1 2026 revenue of $28.5–$30M (6–12% YoY) and non-GAAP operating income of $2–$3M but disclosed a material weakness in revenue controls and ongoing profitability challenges, creating execution and accounting risk.

Analysis

Blend’s Autopilot push is strategically sensible: embedding agentic workflows into an originations platform converts high-frequency transactional signals into a proprietary, labeled dataset that compounds with volume and raises switching costs. That dynamic disproportionately favors a success‑based pricing vendor over seat‑based incumbents because revenue scales with lender ROI, not user counts — creating a structural advantage if Blend can maintain data exclusivity and low-latency integrations.

Near-term financial noise is the real wildcard. Changes in capitalization rules and the disclosed revenue control weakness will compress headline margins and create model divergence between cash flows and reported non‑GAAP metrics for several quarters, amplifying volatility around each print. Regulatory and compliance scrutiny is a three- to twelve‑month tail risk for any agentic product in lending: misclassifications, auditability gaps, or an adverse QC finding would force costly product rework and slow enterprise rollouts.

For competitors and ecosystem players the second-order effects are already forming: cloud and AI infrastructure vendors win as customers standardize on agent stacks (benefiting large cloud providers), while legacy LOS and manual servicing vendors face accelerated rationalization or consolidation. The most actionable readthrough is that Blend’s optionality on margin expansion is real but binary — success will show up as step‑function uplift in EVPFL and per‑customer spend over the next 3–12 months; failure or regulatory pushback will reintroduce pronounced lumpiness and valuation compression.

Key Decisions for Investors

• BLND — tactical long on measured pullback: Buy BLND shares or a 6–12 month call spread after a >10% intraday pullback tied to an earnings or guidance reset. Rationale: asymmetric upside if Autopilot adoption converts N early enterprise wins into higher EVPFL; downside limited by capital buffer and buyback optionality. Risk management: stop-loss at 25% below entry; target 2.5x reward in 6–12 months.
• BLND — defensive hedge / event trade: Buy a 3–6 month put spread (protective) sized to cover 25–50% of position notional ahead of Q1 results and Autopilot adoption milestones. Rationale: insulates against near‑term downside from accounting restatements, revenue controls, or regulatory headlines while retaining upside exposure. Expect cost = small premium; reward = large tail protection.
• MSFT — tactical core long to play enterprise agent adoption: Buy MSFT or a 6–12 month call spread to capture sustained cloud and copilots spend as banks scale agentic deployments. Rationale: durable demand for secure, compliant LLM infrastructure and tooling; low single-stock execution risk. Timeframe: 6–12 months; target modest outperformance vs SPX with stop at 15% drawdown.
• Pair idea — high conviction, lower net beta: Long BLND (smaller size) financed by buying put protection or pairing with a large-cap AI infra long (e.g., MSFT) to hedge macro/tech beta. Rationale: isolates idiosyncratic bet on Blend’s product execution while retaining exposure to secular AI lift. Rebalance after 3–6 months based on adoption datapoints (customer activation rates, EVPFL changes, QC audit outcomes).