EliseAI, which builds AI for apartment operators and medical practices, said that it raised $350 million at a $4 billion valuation. Andreessen Horowitz (a16z) and Bessemer Venture Partners co-led the round. Ontario Teachers' Pension Plan joined as a new investor, and Sapphire Ventures and Navitas Capital returned. The round came 13 months after a
$250 million Series E that valued EliseAI at $2.2 billion.
When co-founder and CEO Minna Song
sat down with EO last October, the numbers were $2.2 billion, a little over 10% of US apartments, and $100 million in ARR after four straight years of doubling. Today they are $4 billion,
one in six US apartments, and more than $200 million in ARR after a fifth year.Few startups get here the way EliseAI did: nine years in one industry nobody else wanted, building integrations one operator at a time. But a company at this size faces the same problem every company at this size faces. The next doubling has to come from somewhere, and it is not obvious yet where.
What Actually Changed
The company crossed $200 million in ARR in June 2026 and now says it handles roughly
5 million phone calls a month across housing and healthcare. On September 3 it launched
Apollo, a single agent that can perform any task inside the Elise platform.
The technology under the product has also shifted. In March 2025,
an OpenAI customer story said EliseAI powered its platform with OpenAI's APIs. EliseAI still works with OpenAI, and Song told Fortune that the company tests OpenAI's new voice features at scale.
But it has also started training its own models.In a later case study, the AI infrastructure company
Baseten wrote that EliseAI's closed-source APIs had "hit a ceiling on cost, speed, and controllability," and that EliseAI set out to build custom models of its own. It now fine-tunes small open-source models for tasks such as pulling a renter's budget and move-in date out of a conversation, and its voice agent runs on one.
"From the outside, it looks like a chatbot that texts renters about apartments. Underneath, it's some of the hardest applied AI work there is."
How a Chatbot Got Here
None of that explains how EliseAI got here. Execution does: staying in an industry investors ignored and selling to operators who wanted more than the product could do. Song told
EO that
the company's first fundraising round drew more than 100 rejections."A lot of the rejections were the same: 'We're not interested in this market. We only look at horizontal SaaS, or we only look at enterprise SaaS.'"
Housing was not a market venture investors wanted, which left EliseAI in its first years with little competition from the largest operators in the country. It won those operators by building whatever they asked for.
Song recalled a three-hour meeting with 12 executives at one of the largest apartment owners in the US, then a two-person company pitching to run their entire portfolio. "Honestly, I think I said for 3 hours straight, 'No, it doesn't do that,'" she said. The next day the COO called back: "We're going to do this together. We're going to help you build the product." Her lesson from that period: "Quite honestly, I believe your first customers, you will have to service forever."
Two of its earliest large customers were AvalonBay Communities and Equity Residential. In August 2026, they merged into
Vivmark Residential, a landlord with more than 184,000 apartments. In their merger presentation, the two companies said AI handles
about 90% of their prospect workflows. Nine years of building to order produced the company's largest customer and its largest concentration.
Where the Next Doubling Comes From
At $4 billion on $200 million of ARR, the valuation assumes at least two more years of doubling. The company says more than 75% of the 50 largest US apartment managers already use it. Its largest customer shows what saturation looks like. Vivmark already routes about 90% of its prospect workflows through AI, which leaves little room to sell it more of the same product. The rest of the top 50 are on a similar path.
That leaves selling more to the customers it already has, which means taking over work that other software does today. EliseAI's AI depends on data that sits in those other systems, such as the CRM that holds renter records and the property management system that holds leases and rent. When EliseAI starts competing with them, their owners can restrict its access.
That has already happened once. Once EliseAI started pushing its own CRM, Funnel Leasing, a CRM company it had long integrated with, pressed it in late 2025 to limit what that CRM could do in exchange for continued access. EliseAI refused, and
the dispute became public in January 2026.
The other path is
healthcare. For about three years, EliseAI has sold to specialty physician groups, handling work from the first patient call through referrals, scheduling, and insurance checks. But it does not disclose how much revenue that business brings in. Until it does, there is no way to tell whether healthcare can carry the next doubling.
Asked by Fortune about a public offering, Song said, "I don't think we have any specific timelines or definitive outcomes." The line can be read two ways. A founder who sees something larger ahead does not need a timeline. Neither does one whose next path is not yet settled.
Which reading is right will show up in where the $350 million goes, whether the company starts reporting healthcare separately, and how far it pushes into the systems it once only plugged into.