SAN FRANCISCO — Digital mortgage technology provider Blend Labs demonstrated resilience and operational discipline on Thursday as it reported its second-quarter financial results. Surpassing expectations on profitability amidst a stubbornly high-interest-rate environment, the San Francisco-based firm pointed to expanding customer relationships, rigorous cost management, and the commercial rollout of its cutting-edge artificial intelligence product, Autopilot, as primary drivers of its ongoing turnaround.
The company’s latest financial disclosure paints a picture of a tech provider successfully streamlining its internal operations while delivering tangible efficiency gains to mortgage lenders struggling with low industry volumes. By leaning heavily into agentic AI frameworks—dubbed internally as "Blend 3.0"—Blend is attempting to rewrite the playbook on software development efficiency and mortgage fulfillment speed.
Main Facts
Blend Labs posted a total second-quarter revenue of $33.8 million, marking a 7% increase compared to the same period last year and an improvement over its first-quarter figure of $30.8 million. The revenue beat came in near the high end of the company’s internal guidance range.
The growth was anchored by its core software platform business, which generated $31.4 million—also up 7% year-over-year. Professional services contributed an additional $2.4 million, climbing slightly from $2.2 million in Q2 of the previous year.
On the bottom line, Blend significantly narrowed its GAAP operating loss to $1.6 million, an improvement from the $4.8 million operating loss reported in the second quarter of 2025. Non-GAAP operating income saw even stronger expansion, jumping to $7 million from $4.6 million a year prior. Furthermore, the company’s non-GAAP gross margin expanded to 78%, up two percentage points from 76% in Q2 2025.
Blend concluded the quarter in a healthy liquidity position, reporting $44.9 million in cash, cash equivalents, and marketable securities, with zero debt on its balance sheet. This financial cushion allowed the company to aggressively return capital to shareholders, repurchasing 11 million shares for $18.2 million during the quarter and leaving $13.2 million remaining under its existing share buyback authorization.
Chronology of Events
The trajectory leading up to Blend’s second-quarter earnings report involved several pivotal strategic milestones executed over the past year and culminating in the summer of 2026:
- Late 2025 – Early 2026: Blend begins an internal restructuring of its engineering workflows, integrating generative AI agents to handle foundational coding tasks. This initiative, later branded as Blend 3.0, systematically increases internal throughput without expanding the company’s headcount.
- Q1 2026: Blend reports foundational stability with $30.8 million in revenue, setting the stage for a stronger spring push into automation products.
- June 30, 2026: The second quarter closes with solid financial metrics, outperforming management’s expectations despite macroeconomic headwinds in the housing sector.
- July 1, 2026: Blend’s heavily anticipated AI-powered product, Autopilot, officially becomes commercially available. Simultaneously, six early-adopter lenders—including Onity Mortgage Corp.—are integrated into the system, having signed on prior to or at launch.
- Thursday (Earnings Announcement): Blend publicly releases its Q2 2026 financial results, details the initial success of Autopilot, and issues forward-looking guidance for the remainder of the year.
Supporting Data and Financial Metrics
A deep dive into Blend’s financial disclosures reveals steady optimization across multiple operational metrics:
- Revenue Breakdown: Total revenue reached $33.8 million, led by software platform revenue at $31.4 million and professional services at $2.4 million.
- Losses and Margins: GAAP diluted net loss from continuing operations attributable to common stockholders remained flat at 3 cents per share year-over-year. On a non-GAAP basis, diluted net income was break-even, identical to Q2 2025. Non-GAAP gross margins ticked upward to 78%.
- Balance Sheet: The company ended the quarter with $44.9 million in liquid assets and marketable securities, entirely debt-free.
- Share Repurchases: Blend bought back 11 million shares for $18.2 million, with $13.2 million still authorized for future buybacks.
- Customer Growth: The firm added or expanded 14 customer contracts during the quarter, with six specifically tied to the newly launched Autopilot platform.
- Autopilot Impact Metrics: Preliminary internal data from early-adopter loans processed via Autopilot indicates a 10% to 15% improvement in pull-through rates, a 2- to 4-day reduction in loan cycle times, and an average savings of 4.5 hours of fulfillment tasks per loan.
- Forward Guidance: Blend forecasts Q3 revenue between $31.5 million and $33.5 million, with non-GAAP operating income projected between $3.5 million and $4.5 million. For Q4, management anticipates a market-wide volume drop of about 11% year-over-year at the midpoint due to a "higher-for-longer" interest rate environment.
Official Responses and Executive Commentary
Company executives emphasized that the quarter’s success was engineered through strict fiscal discipline and a revolutionary pivot toward AI-native corporate operations.
"We delivered the quarter we aimed for—revenue near the high end, profitability above the high end, and a strategic return of capital through our share buyback," said Jason Ream, Blend’s head of finance and administration.
Nima Ghamsari, co-founder and head of Blend, highlighted the dual application of artificial intelligence—both externally for customers and internally for the company’s own workforce. He characterized the quarter as "disciplined" and "profitable," noting that Blend 3.0 represents a fundamental shift in corporate identity.
"Blend 3.0 is an agentic-first company, and that doesn’t mean just for our products. That also means how we work internally," Ghamsari told investors during the earnings call. He revealed that the engineering team’s throughput has surged by a factor of 3.6 since January with roughly the same headcount, as internal AI agents take a first pass on routine code changes and technical tasks.
"If we continue at this pace, we could be doing 10 times as much throughput on the engineering team by the end of this year as what we did at the end of last year," Ghamsari projected.
Addressing the macro environment, Ghamsari noted that the housing market has provided little relief to mortgage tech providers. "We delivered Q2 with revenue near the high end and non-GAAP operating income above the high end of our guidance, and we did it in a market that isn’t giving us much help," he remarked.
Regarding Autopilot’s early reception, Ghamsari pointed to Onity Mortgage Corp. and five other lenders who jumped at the chance to use the software on day one. "We’re starting to see evidence [that] Autopilot is driving faster clearance times, higher conversion rates, and potentially reducing fulfillment costs," he said.
Industry Implications and Future Outlook
Blend’s strategic pivot carries significant implications for the broader mortgage technology sector. As interest rates remain stubbornly elevated—fostering a "higher for longer" monetary policy cycle—mortgage originators are under immense pressure to slash operational costs per loan. Traditional scaling methods, which rely heavily on hiring and firing processing staff to match cyclical volume swings, are increasingly giving way to automation.
By automating an average of 4.5 hours of manual fulfillment tasks per loan, Blend’s Autopilot directly targets one of the most expensive bottlenecks in mortgage origination. If early metrics holding a 10% to 15% improvement in pull-through rates and multi-day cycle time reductions are validated across a wider customer base, demand for agentic AI tools in lending could accelerate rapidly.
However, management remains realistic about the immediate macroeconomic hurdles. Looking toward the final stretch of the year, Jason Ream noted that Blend is adopting a conservative stance on total mortgage volumes for the fourth quarter. Citing skepticism over refinance activity amid persistent interest rates, the company expects Q4 market volumes to drop roughly 11% year-over-year at the midpoint. Blend anticipates its own funded loan volumes to decline by 10% to 15%, though these projections remain fluid depending on broader economic shifts.
Despite these near-term headwinds, Blend Labs has positioned itself to weather the cyclical downturn through disciplined cost controls, robust liquidity, and a high-margin software platform that leverages the latest advancements in artificial intelligence. As the lending industry searches for ways to operate profitably in a low-volume era, Blend’s bet on agentic automation may well define the next generation of digital mortgage infrastructure.