Fifth Wall’s Brendan Wallace Sees AI Expanding Proptech Into Real Estate Operators

AI is changing what counts as a proptech company—and, in Brendan Wallace’s view, where venture capital can find durable businesses. The co-founder and CEO of Fifth Wall sees opportunity shifting from software built only for real estate toward horizontal AI tools and companies that own and operate physical assets.

That shift matters for founders and investors because familiar software products may be easier to recreate in-house, while businesses that combine technology with real estate operations can offer a different path to scale. Wallace also sees persistent openings in housing finance, construction efficiency and the use of data to assess property.

Proptech extends beyond property software

Founded in 2016, Los Angeles-based Fifth Wall has roughly $3 billion in commitments and capital under management. Its portfolio includes Procore, OpenDoor, Blend and Hippo. ServiceTitan went public in 2024, while CBRE took full ownership of flexible workspace provider Industrious in January 2025 at a valuation of more than $800 million. CBRE, Hines, Hilton and Marriott are among the firm’s investors, connecting portfolio companies with potential customers.

Wallace’s perspective spans real estate finance and startups. Before co-founding Fifth Wall, he worked at Goldman Sachs and Blackstone, co-founded workforce analytics startup Identified, which Workday acquired in 2014, and was involved with ridesharing company Cabify.

He thinks conventional proptech tallies can miss a growing share of relevant capital. Some investors are backing horizontal products that serve multiple sectors; others are financing businesses that directly build or operate real estate. Both may fall outside a narrow definition centered on software made specifically for property companies.

Fifth Wall’s investment in Higgsfield illustrates the first category. The visualization software can turn blueprints for an unbuilt home into a scene that lets a prospective buyer see the space and move through it. Wallace said real estate agents and commercial leasing brokers were already using the product to represent properties. In his view, it had surpassed industry-specific tools for visualizing space.

The second category includes businesses that use autonomy, AI or robotics to reshape physical operations. Fifth Wall invested $135 million in Atoms, which Wallace described as an example of a company that may not register as proptech under a conventional label, even though it connects to real estate.

That distinction helps explain why a funding count can tell an incomplete story. Capital may be directed toward products used by property businesses without those products being built for the sector alone. It may also go to companies that are themselves real estate operators, rather than software vendors. A tally focused on dedicated property software could miss both.

AI raises the bar for software startups

Wallace sees AI as a major force behind the move toward horizontal products. When companies can reproduce a basic application internally in a short time, they have less reason to buy a third-party tool. He cited tenant engagement software as an example of a product a real estate company might now recreate over a long weekend.

That pressure makes highly capable products harder to replace more attractive. Wallace does not think real estate companies will build software like Higgsfield themselves, but he expects more scrutiny of simpler applications. He also suspects property companies’ spending on foundation models may be approaching or surpassing what they spend on proptech applications, though he said he did not have a statistic to support that estimate. Anthropic, he observed, is not usually classified as a proptech company, yet foundation model spending may absorb money that once flowed through property software.

For venture-backed software startups, the test is whether a product can maintain a defensible advantage as major AI developers improve their tools. Wallace said simple point solutions are becoming harder to finance when a large model provider or the customer itself can build a comparable interface. Adding AI to an existing workflow may not be enough if the product’s underlying intellectual property can be copied readily.

He is less convinced that specialized knowledge will protect a company indefinitely. Domain expertise can matter now, but general models may adapt to specific uses with increasing speed. Wallace pointed to the changes in ChatGPT and Claude between 2023 and the present as evidence of how quickly capabilities can advance. A niche advantage that looks durable today could erode as models gain stronger memory, personalization and ability to follow natural language instructions.

At the same time, Wallace sees room for AI development firms that build custom systems inside companies. Fifth Wall backed Antares, which he compared to a Palantir for real estate. Its teams work with strategic limited partners, identify a problem suited to an AI tool, develop software tailored to that organization and leave the resulting intellectual property with the client. The business model is closer to paid development work than licensing a standard product.

Physical operations bring a different risk profile

Construction technology remains part of Fifth Wall’s focus, though Wallace says it receives less attention than before. He sees the firm’s stronger performance in pure software businesses—such as marketplaces and consumer or enterprise applications—and in full-stack companies. Businesses that physically do the work carry a substantially different risk profile.

Construction still attracts capital because the market is enormous and inefficiency remains a substantial problem. A company that makes building more efficient could address a large opportunity, which helps explain why sizable funds continue to pursue the sector. The scale of the prize, however, does not remove the operational challenges that have made construction technology a difficult area for venture-backed companies.

Wallace’s preferred opportunities increasingly involve building real estate businesses themselves. He believes it has become possible to create a venture-scale company that buys multifamily properties, reduces costs and uses AI to run more of the processes, operations and management. He said that model would have been less likely to fit a venture mandate five years ago.

Data is another opening. Wallace described real estate information as unusually opaque compared with other large financial markets, and said the past decade of proptech has not fundamentally changed that. Generative AI could help find useful signals across broader or different datasets, potentially improving decisions about where capital goes and what kinds of properties may perform better.

He argues that income-producing real estate can be considered alongside assets such as stocks and bonds, even though property is not always analyzed in the same way. Computational capacity that was previously unavailable could make quantitative approaches more practical. The opportunity depends on whether better data analysis can produce useful signals from a market where information remains difficult to compare.

Land, new asset classes and housing costs

Some of the opportunities Wallace follows have little to do with AI. Land is an input to most real estate, yet many building owners would prefer not to own the land or accept the associated risk. He is interested in financial structures that could separate ownership of land from ownership and operation of the building on it.

He also expects AI, robotics and autonomy to influence how people use space and what kinds of properties businesses need. Past technology shifts have helped create new real estate categories: e-commerce supported the growth of modern industrial property; cloud computing and the internet helped drive data center development; mobile phones created demand for cell towers; and genome sequencing contributed to lab space. Wallace sees the three current technologies as connected forces that could support further asset classes.

That view has implications for the venture market. Wallace expects fewer funds focused narrowly on consumer or enterprise software for a single industry, with less capital going to that category. He anticipates more venture investment in real estate companies themselves, including businesses that combine physical assets with technology and operational control.

Housing remains the problem he would most want to tackle. Wallace said housing costs weigh on the U.S. economy and that affordability remains unresolved. Potential approaches include building homes faster and more cheaply, reducing the cost and friction of home financing, and making city processes more efficient. He described getting a mortgage as painful, slow and expensive despite substantial past investment in the field.

Wallace also pointed to work at the intersection of technology, city processes and consulting as a possible way to improve access to housing, while acknowledging that entrepreneurs cannot solve the problem alone. For founders, his read is that the strongest openings may come from tackling the underlying economics of property—how it is financed, built and operated—rather than simply adding another software layer.

Related reading

• Global Venture Funding To Proptech Startups In 2026
• Global Physical AI Venture Funding In 2026
• Sector Snapshot: Proptech Funding Holds Up, But Investors Are Placing Different Bets

Illustration: Dom Guzman