Trebellar, a software company building artificial-intelligence agents for corporate real estate teams, has raised an $18 million Series A led by Blossom Capital, the company said in a post on its blog dated Sept. 24. Haystack, Alt Capital, 1Flourish and Bynd Venture Capital also participated.
The buyer here is not a broker or a landlord. It is the in-house team at a large company that decides how much office space the employer holds, where it holds it, and what happens when leases come due β the discipline the industry calls corporate real estate, or CRE occupier work. Trebellar named Meta, Uber, Merck and Cohesity among its customers.
Co-founder and chief executive Diego Ferreiro Val framed the opportunity as a gap in enterprise software rather than a gap in real estate. “Legal, HR, and finance all have AI-native platforms built for how they work,” he wrote. “Corporate real estate, the second-largest expense for most enterprises, hasn’t had one until now.” The claim about relative expense is the company’s own; Trebellar did not cite a source for it.
The problem it describes
Ferreiro Val’s description of what occupier teams actually do all day will be familiar to anyone who has worked a portfolio review. “Leases, space, headcount, and utilization data sit across systems that don’t talk to each other, so every major decision starts with weeks of reassembling the picture by hand,” he wrote. “Dashboards show what happened. Consultants deliver a report that’s outdated on arrival.”
Trebellar’s answer, per the announcement, is to pull that scattered information into a single live view of a portfolio β how much space a company has, what it costs, and whether it is working for the people using it β and then layer on commute times, transit access and employee sentiment. The company says the system also records why each decision was made and whether it held up.
On its product pages, Trebellar lists automated lease abstraction, renewal and expiration tracking, scenario modeling for supply and demand, and location scoring using demographic and drive-time data. The company describes its approach as agents that “plan, execute, and deliver a finished result” rather than chat interfaces that answer questions.
Where the money goes, and who is advising
Ferreiro Val said the round will fund engineering and go-to-market hiring, deeper AI capabilities and expansion to more enterprise customers. He co-founded the company after leading engineering at Salesforce.
The announcement also disclosed an advisory hire with more direct real estate weight than most proptech board additions: Dave Radcliffe, who the company said “led the expansion of Google’s real estate footprint under Eric Schmidt, Larry Page, and Sundar Pichai.” Trebellar said further detail on the appointment would follow.
A crowded year for real estate AI money
The round lands in a stretch of steady venture funding for software that automates document-heavy real estate work. Cushman & Wakefield’s experience is a useful marker of what buyers are after: a vendor said this year that the brokerage had cut lease abstraction from days to minutes using AI, the same task sitting at the base of Trebellar’s product.
Capital has flowed to adjacent problems as well. Buildots raised $130 million for construction AI aimed at the data center buildout, and homebuilders D.R. Horton and PulteGroup backed Scaffold’s $15 million round to untangle builder data.
What distinguishes Trebellar’s segment is who pays. The customer is the tenant, not the landlord and not the brokerage. On our reading that is the harder sale, because the budget sits inside corporate facilities and workplace teams rather than in a real estate P&L, and those teams have to justify the spend against the same portfolio costs the software is meant to cut. An $18 million round with Meta and Uber already on the customer list is a reasonable test of whether enterprises are ready to buy it.
