Pinterest (PINS) Q2 FY26 earnings call: guidance and Q&A

Pinterest has transformed into an AI-driven shopping assistant, with every Pin personalized by AI, contributing to 12 consecutive quarters of record users and 11 straight quarters of double-digit user growth. The company's proprietary 'Taste Graph,' fueled by 80 billion monthly searches and 16 billion boards, provides unique signals for AI training. The new AI conversational layer, Pinterest Assistant, now available to most U.S. users, helps with later-stage research questions, enabling users to move from inspiration to purchase entirely on the platform. This assistant is built on Pinterest's visual-first approach and leverages open-source models for cost efficiency.

AI is improving advertiser performance across the funnel. Pinterest Performance+ campaigns, which automate setup, bidding, targeting, and creative optimization, are driving better ROAS and faster spend growth for advertisers. The new Smart Assembly feature, part of Performance+, delivered a 6% improvement in click-through rate in early alpha testing for advertisers without product catalogs. The Business Assistant, a conversational AI collaborator, is in beta to provide actionable recommendations and simplify Pinterest for advertisers, especially those without large dedicated teams. These initiatives aim to make AI-powered best practices the default for lower-funnel campaigns.

The company's go-to-market transformation, focusing on clear seller accountability, better packaging of commercial moments, and tighter product-performance-advertiser connections, has driven strong Q2 results, particularly in UCAN. UCAN revenue accelerated 5 points to 18% year-over-year, benefiting from increased retailer spend, growth in mid-market/SMBs, and emerging verticals. This success is now being applied internationally, with new leadership and testing of third-party demand expansion into Europe, though this is expected to be a multi-quarter journey.

Pinterest employs a blended model approach for AI, utilizing in-house compact models, post-trained open-source models, and limited closed third-party models. This strategy allows for significant cost savings, with open models achieving cost per transaction at less than 8% of comparable closed proprietary models. The company completed its full transition to GPU serving over two years ago, materially improving relevance and personalization. This disciplined approach enables investment in AI capabilities while expanding profitability, as evidenced by the raised full-year adjusted EBITDA margin outlook.

Pinterest generated $270 million in free cash flow in Q2, with a trailing 12-month FCF conversion of 94%. The company allocated $58 million towards share repurchases in Q2 and entered a capped call transaction for $99 million to protect against dilution. Year-to-date, Pinterest has repurchased over $2 billion of stock, retiring nearly 111 million shares at an average price of $18 per share. This proactive dilution management resulted in a 12% year-over-year reduction in net dilution for Q2, contributing to 30% non-GAAP EPS growth.