CoStar Launches Homes AI as Real Estate Bets on Domain-Specific Intelligence
CEO Andy Florance positions voice-enabled assistant as data-moat play against generic AI tools—and a wedge against Zillow, Redfin in proptech arms race.
CoStar Group launched Homes AI on February 17, a conversational property-search assistant powered by Microsoft Azure OpenAI and integrated directly into Homes.com, marking the company’s most aggressive move yet to leverage its proprietary data advantage against larger residential rivals.
The tool allows users to search for properties via voice or text in real-time, accessing Business Wire reports indicate CoStar’s property database, Matterport 3D digital twins, proprietary school and neighborhood data, and market intelligence. CEO Andy Florance described the launch as Bloomberg covered, a “fully integrated native AI solution designed specifically for the sector.” Unlike chatbot plugins built atop third-party platforms, Homes AI runs entirely within CoStar’s ecosystem—positioning it as a differentiated alternative to general-purpose AI assistants that lack Real Estate domain expertise.
The strategic timing is deliberate. CoStar recorded HousingWire reported 100 million average monthly unique visitors to Homes.com in 2025, with organic traffic up 134% year-over-year in January 2026. The platform also achieved 2.1 billion views and secured its position as the second-largest U.S. residential portal by traffic. Early engagement metrics are stark: in the first week post-release, users engaging with Homes AI spent 16 minutes 50 seconds on the site versus 4 minutes 24 seconds for non-AI users. AI users conducted nearly four times as many searches, favorited seven times as many properties, and submitted seven times as many email leads.
The Data Moat Thesis
CoStar’s pitch centers on proprietary data accumulated over decades in commercial real estate, recently augmented by the Wikipedia documented $1.6 billion acquisition of Matterport in February 2025 and the $156 million purchase of Homes.com in 2021. The company operates CoStar (commercial data and analytics), LoopNet (commercial marketplace), Apartments.com (multifamily), and a suite of hospitality, auction, and international platforms—generating over 143 million average monthly unique visitors across its portfolio in Q3 2025.
Florance emphasized that Business Wire reported the system is “purpose-built for real estate,” contrasting it with implementations that rely on external large language models trained on generic datasets. CoStar’s data remains within its proprietary ecosystem and is never used to train external AI models, addressing a common concern among real estate professionals about data leakage. The company plans to extend the capability across Apartments.com, LoopNet, Land.com, and BizBuySell, creating a unified conversational layer across commercial and residential assets.
Competitive Landscape Hardens
The launch arrives as Proptech incumbents scramble to integrate AI without ceding control to external platforms. Zillow, which commands MarketMinute noted 221 million monthly users, saw its stock decline roughly 15% earlier this month following a cautious 2026 outlook and concerns over CoStar’s AI push. Zillow has deployed AI-assisted CRM tools for agents and launched a ChatGPT plugin, but its approach remains largely algorithm-driven rather than conversational. Redfin, acquired by Rocket Companies in mid-2025 and now operating as “Redfin Powered by Rocket,” has introduced conversational search but lags in traffic scale compared to CoStar and Zillow.
CoStar’s competitive posture is explicit. The company filed copyright infringement lawsuits against Zillow, is embroiled in antitrust litigation alongside Zillow and Redfin over rental syndication, and faces activist investor pressure over Homes.com’s profitability. Despite recording HousingWire stated $900 million in Q4 2025 revenue (up 27% year-over-year) and the 59th consecutive quarter of double-digit growth, net income fell to $7 million for 2025 from $139 million in 2024, attributed to acquisition costs for Matterport and Domain. The company authorized a $700 million share buyback and signaled 2026 as the year of “earnings acceleration,” banking on reduced marketing spend and AI-driven engagement to drive margin expansion.
“This innovation signals a shift as significant as the emergence of online search itself.”
— Andy Florance, CEO, CoStar Group
Industry Adoption Challenges Remain
Real estate remains structurally resistant to technology adoption, despite billions in proptech investment. PwC analysis indicates AI is moving from experimentation to measured implementation in 2026, with organizations beginning to adopt solutions that streamline operations and reduce costs. Yet the transition from pilot to production remains uneven. Proptech venture funding reached $16.7 billion globally, representing 68% growth as investors prioritize platforms delivering operational efficiencies, but ICSC survey data shows over half of firms deploying AI have yet to realize revenue or cost benefits.
The proptech AI market itself is nascent. Cogito Tech projects the global AI in PropTech market will rise from $20.5 billion in 2023 to approximately $159.9 billion by 2033, expanding at a 22.8% CAGR. JLL Research found that among 7,000 global PropTech companies, about 10% (700 companies) currently provide AI-powered solutions, with 62% venture-backed and 83% generating revenue. Top use cases include document sorting, portfolio analytics, and client communication—domains where CoStar’s integrated data infrastructure provides structural advantage.
CoStar’s residential push has drawn comparisons to Bloomberg’s dominance in financial data. Founded in 1987 by Andy Florance, the company went public in 1998 and has delivered approximately 20% CAGR over 25 years. The firm spent $35 million on Super Bowl LVIII advertising to challenge Zillow and Realtor.com, and maintains an aggressive acquisition strategy that has consolidated commercial data while expanding into residential at scale.
Agent Relations and Revenue Model
CoStar emphasized that Homes AI is designed to enhance—not replace—real estate professionals. Under the company’s “Your Listing, Your Lead” commitment, inquiries from buyers go directly to listing agents, contrasting with Zillow’s lead-generation model that charges agents for buyer access. Homes.com reached HousingWire reported over 26,000 agent members, up nearly 150% since Q3 2024. Florance argued the AI creates “meaningful conversations” between buyers and agents by surfacing properties aligned with natural-language preferences, giving agents clearer records of consumer intent during follow-up.
The revenue model hinges on subscription growth and lead quality, not advertising volume. CoStar’s commercial segment (CoStar and LoopNet) continues growing at 20% clips, providing cash flow to fund residential expansion. The firm targets MarketMinute indicated $5 billion in revenue by 2030, requiring Homes.com to maintain its #2 traffic position and convert engagement into subscriptions at scale. Q1 2026 adjusted EPS guidance fell short of analyst expectations due to front-loaded AI development spending, but management framed the investment as critical infrastructure for long-term margin expansion.
What to Watch
CoStar’s ability to convert AI-driven engagement into subscription revenue will determine whether the data moat translates to sustainable competitive advantage. Key metrics include agent membership growth, subscription ARPU (average revenue per user), and retention rates as the company scales from 26,000 to a target base that can support $5 billion annual revenue. The planned rollout across Apartments.com and commercial platforms will test whether conversational search drives similar engagement lifts in non-residential segments.
Competitive response from Zillow and Redfin will clarify whether domain-specific AI creates defensibility or whether large language models commoditize search experiences. Zillow’s 221 million monthly users and incumbent brand remain formidable, while Redfin’s integration with Rocket Mortgage provides a vertical moat CoStar lacks. The proptech funding environment—where AI-native companies attract premium valuations while non-AI solutions face constrained capital—will shape M&A dynamics and platform consolidation through 2026.
Regulatory scrutiny around data usage, Fair Housing compliance, and antitrust concerns will intensify as AI systems influence property discovery and valuation. CoStar’s commitment to keeping data within its proprietary ecosystem may limit model sophistication relative to competitors training on broader datasets, creating a classic trade-off between control and capability. The company’s litigation with Zillow over photo copyrights and ongoing FTC scrutiny of rental syndication deals signal that competitive dynamics will play out in courtrooms as much as product releases.