How Kaz Nejatian Is Rebuilding Opendoor

Summary

• Opendoor was EBITDA-positive as of April 1, 2026, and Matt Reustle’s opening frames it as on track to be adjusted-net-income-positive by year-end. Kaz says it is important for Opendoor to be funded by cash flow rather than repeatedly reaching for capital markets. His diagnosis of the old company: it “had too little discipline… because it has been able to reach for capital markets over and over again,” was “basically being run by outside consultants” with almost no stake in outcomes, and its OpEx “was just honestly stupid… spent on the wrong things.”
• Nejatian’s core reframe of the Opendoor thesis: “Opendoor is a market maker, not a prop desk.” Calling it a real-estate asset manager with software “is like saying Amazon in its early days was a warehouser of books that just happened to have software” — the leverage comes from software-enabled market making and underwriting, not holding assets for profit. Its underwriting engine is “the best underwriting engine in the business… nothing close to it.”
• Velocity beats spread because information is the point. “Card counters at a blackjack table play every hand… the information is the point”: high transaction flow gives Opendoor a “90 to 120 day advantage” over anyone scraping MLS data, while wide spreads create a negative feedback loop — offer $300K on a $400K home and the seller may know things Opendoor does not. Tighter spreads mean more flow, more data, and still “significant room for margin.”
• The attach opportunity is what Nejatian says he most underestimated, and the fee stack is the sizing exercise: 6-7% transaction cost, 1-2% title & escrow, 300-400bps mortgage margin, 100-200bps insurance, plus warranty, home care, and solar — a fragmented stack of low-NPS providers who all pay CAC that Opendoor doesn’t. “There isn’t a $100 billion market cap real estate company in the public markets in the U.S.” That does feel like a flaw in the matrix.
• Product roadmap runs through a “thin waist” — checkout for real estate: title & escrow first (already owned and growing quickly), mortgage second (live in Colorado, more states soon), then home warranty and insurance. He also sketches a solar-panel leasing model and says the related financing risks can be evaluated together.
• When the host pushed that attach could add friction and slow velocity, Nejatian first said, “I think you’re just straight-up wrong,” then added, “Actually, maybe you’re not wrong.” He called point-solution complexity “FUD”; “for 80% of people, 100% of these services are relatively vanilla,” and the goal is a “one-click mortgage, title, escrow, home-buying, and home-selling experience” for “the teacher in Kansas City” — the $14 million-mansion buyer with no U.S. residency and global trusts can go elsewhere.
• The operating model is leanness plus AI: Opendoor has fewer than 70 engineers. “If you cannot, using Claude or ChatGPT or Codex or Grok, write a SQL query, you should not be working at a tech company” — engineers build leverage systems for non-engineers, and “we don’t have the luxury of waste.”

Deep dive

1. What surprised the would-be buyer-turned-CEO: the models were in better shape, consultants and spending were the problem

• Nejatian’s origin story, lightly corrected: in February 2025, still at Shopify, it was his wife who “encouraged me to think about” selling their possessions and levering up to take Opendoor private — he got the CEO seat the traditional way instead.
• Two upside surprises from inside: the company was “structurally in much better shape than I thought” — its underlying models, databases, and processes were in good shape despite a “doom loop,” like Raiders of the Lost Ark: “there were a bunch of people here who were still guarding the cup.” And he “significantly underestimated the attach opportunity” despite having “done this whole attach-services thing my whole career.”
• The downside, delivered without varnish: the company had effectively been run by outside consultants, and the people making decisions had “almost no stake in the outcome.” He called the resulting OpEx “stupid,” spent on the wrong things, and “offensive.”

2. Market maker, not prop desk — and why velocity and tight spreads matter

• The host’s East Coast-skeptic framing led to Kaz’s distinction: a prop desk holds assets for profit; a market maker’s “fundamental job is to not hold assets for profit.” Opendoor still needs to be very good at underwriting — “damn right we do” — but its leverage comes from software-enabled market making and flow, not carrying assets.
• The velocity logic chain in full: buying and selling lots of homes quickly yields live feedback on clearance prices, renovation, and demand — anyone scraping MLS data is “90 to 120 days behind us.” Wide spreads create a negative feedback loop: the only scenario in which a seller accepts a $300K offer on a $400K home is if the seller knows things Opendoor does not. Tight spreads that account for Realtor fees, taxes, holding costs, fall-through risk, and 90- to 120-day hold periods can offer sellers something faster, cheaper, and more certain while preserving margin.
• The customer mix is quickly shifting away from distress — “death, divorce, delinquencies… under stress” — toward families like the Watsons from the earnings call, who moved from California to Colorado without a deadline and bought a home, took a mortgage, and sold a home through Opendoor “because it was objectively a great deal.” The end state: customers should think of Opendoor the way they think about Uber, Amazon, or a market maker.

3. The attach stack: jamming a fragmented, low-NPS industry into one checkout

• The profit-pool math as told: 6-7% transaction cost, 1-2% title & escrow, mortgage margins of “300, 350, 400 basis points,” insurance at 100-200bps, plus home care, satellite dishes, and solar. Underwriting a home for acquisition, a mortgage, and insurance involves identical underlying work. Kaz calls real estate “the single largest market in the world,” significantly bigger than the stock market, and says the absence of a $100 billion market-cap real estate company in U.S. public markets feels like “a flaw in the matrix.”
• Prioritization follows the “thin waist” principle — e-commerce’s is checkout, while real estate has none. Opendoor owns a title-and-escrow business that Kaz says is growing quickly; title and escrow come first, “once we solve that,” then mortgage (live in Colorado), home warranty, insurance, and other services.
• His solar sketch: lots of homes Opendoor buys lease their panels from someone else; Opendoor could buy those panels and lease them to Next Solar. He then goes down the financing stack, asking how correlated the risks of solar leasing, mortgages, and insurance are. His analogy for today’s home purchase: cars were once bought with the engine separate from the carriage; homes still involve “a couple of dozen different parties in the system. And they all hate each other.”

4. Reustle’s friction pushback — met head-on with a Shopify precedent

• Reustle’s challenge: each attached product, such as mortgage or insurance, is a potential point of failure that could slow velocity. Kaz’s first reaction was, “I think you’re just straight-up wrong,” followed by, “Actually, maybe you’re not wrong.” He called the complexity argument “FUD” from point-solution providers, citing Shopify Payments and Shopify Tax, both of which had extremely high penetration and were used by basically everyone.
• The 80/20 resolution: “For 80% of people, 100% of these services are relatively vanilla.” Those 80% are the margin opportunity that pays for the 20% of complexity. The average American in South Dakota should not subsidize the $14 million-mansion buyer with no U.S. residency, trusts all over the world, and multiple cars — “all the dudes… with their Maseratis can go to someone else.”

5. Discipline as strategy: the weights-on-the-bedroom-door company

• Matt opens with the earnings framing that Opendoor was EBITDA-positive as of April 1, 2026 and on track to be adjusted-net-income-positive by year-end. Kaz’s operating answer is discipline: his wife put the workout bench against the bedroom door so he couldn’t ignore it — “Opendoor requires some of that energy.” He says repeated access to capital markets allowed too little discipline and that it is important for Opendoor to be funded by its cash flow.
• He concedes the cost openly: “Am I forgoing an area of growth that I would not forgo if I had lots of profits? Yes, I am. For sure I am.” But “this will hopefully be the last job I ever have,” and the goal is to make Opendoor “the most disciplined and most aggressive tech company in the market.”
• The leanness proof points: Opendoor has fewer than 70 engineers. Kaz cites an excellent team, including an ex-signals-intelligence head of data and multiple YC founders, and says he knows what every engineer is working on. He found an entire service maintained by an engineer whose job was to calculate employee RSU allocations: “I can’t tell you how nutty a waste of time that is.”
• The model is not simply fewer people: engineers build systems that let non-engineers create leverage. The PR consultants “all rage quit” when he joined; the head of internal comms now does the work with Claude and spends most of her time on it, “so does everyone else in the company.”