In Part 1 of this two-part editorial series, I argued that Real’s $880 million REMAX deal is a bet on the agency at a time when consumers are leaving. And Tamir Poleg argued that the deal had to be made anyway because the alternative agency was stuck as a mature eXp in a winner-take-all market.
What happens to fees when consumers finally know what they are paying for?
In the REMAX deal, Real and in the Anywhere deal, Compass didn’t buy the broker, but the equipment surrounding it, such as franchise fees, divisions, agent caps, mortgages and title attachments. All of that is at the top of the committee.
Participate in the INMAN Intel Index Survey
But AI is doing the same thing for fees as it does for all other bundled services: breaking them down into parts that consumers can price.
Selling a $400,000 home is about the same as a $1.5 million home. Fees vary widely. 3% of $400,000 is $12,000. 3% of $1.5 million equals $45,000 with the same listing entry, same MLS adjustments, and same showings.
This discrepancy has been true for 40 years and has been maintained because there was no way for consumers to question it the moment they signed the listing agreement.
Things changed with the launch of ChatGPT in November 2022.
Sellers can now consult with AI before consulting with an agent. They ask, what can a 6% fee actually buy? They receive a list. AI does the work you already do for free, including listing entries, comps, marketing copy, MLS coordination, scheduling, and paperwork. They ask which parts of the bundle still require humans.
They get a short list. Negotiations, on-the-spot decisions, deal closing, and legal responsibilities. They ask how much those parts themselves are worth.
They get a number, but the number is not 6 percent (or 5.44 percent).
Then ask the agent to list the items. They negotiate fees for the listing side. They ask why we also pay a buyer agent.
The sellers in that conversation are not acting like the sellers when REMAX wrote the franchise agreement, when Compass set up the split, or when Real set the fee cap.
None of these models are designed for sellers who know math.
Realtor.com’s October 2025 survey found that 82% of buyers and sellers are using AI in their housing decisions. You’re not making housing decisions, you’re making actual decisions about what to pay, what to offer, and what to negotiate. ChatGPT (67%) and Gemini (54%) lead the way. The study’s most important findings are buried beneath the headlines. For the question “Become smarter about the market,” respondents already rated AI a 61, while agents rated it a 62.
Three years after the launch of ChatGPT, consumer trust in AI as a market intelligence source has effectively captured agents.
The national average fee of 5.44% won’t disappear tomorrow. But within five to seven years, perhaps sooner, that will no longer be the default expectation.
Flat-rate and à la carte plans account for a large proportion of transactions, primarily because consumers stop paying for services they no longer need. Discount brokerages undergo second review. And because full-service agents are no longer the only path forward, revenue per trade for all brokerages will decline.
There is no version that doesn’t.
Companies in such positions do well. Companies where the entire stack is structured to attach to existing fees are not.
Real-REMAX falls into the second category. Compass-Anywhere is similar. Both deals make perfect sense if the existing fee structure is maintained. Both trades present a much more difficult math problem than would otherwise be the case.
Rocket is the only one of the three deals that doesn’t require a retention fee because it has acquired a customer relationship that will survive whatever happens to the traditional agent-brokered model.
A homeowner with a loan serviced by Mr. Cooper will continue to be a homeowner with a loan serviced by Mr. Cooper, whether the posting fee is 6 percent, 4 percent, or a flat fee of $3,000. Buyers who use Rocket Mortgage for financing will be using Rocket Mortgage regardless of whether the agent gets the listing at 3% or 1.5%. This relationship is upstream of the committee. This is why rocket bets work on any futures and real bets only work on one of them.
This is a difficult discussion that the industry does not want to have. Consolidation in residential real estate is currently being driven by two forces simultaneously.
The first is that scale is starting to matter again. The number of agents, brand awareness, investment in technology, etc. are all attractive when you get bigger. The second is that per-transaction economics are about to be compressed, and the only way to survive with low profit margins is to have more agents. Both forces push all operators toward the same playbook. Buy more agents. Further reduce costs. No readjustment required.
The problem is that this playbook only works if we accept that mediation, as we know it, is the unit of the future.
That’s what both Compass-Anywhere and Real-REMAX assume
Rocket made another bet, about $16 billion, that customer relationships are the unit of the future and that brokerage is part of a long arc.
Tamir Poleg is one of the best operators in this business and he had to make this deal. Wall Street’s growth calculations left Real with no choice but to achieve the same wall experience points. If done correctly, the synergy will help his bottom line for years.
A more difficult question is whether his $880 million is going toward customer relationships that will last for the next 10 years.
Rocket bought out homeowners – people who lived in their homes, paid their mortgages, refinanced them, took out HELOCs, eventually sold them, and eventually bought them again. Real bought out agents, people who show up for deals, get paid from commissions and go home.
As the consumer-trade relationship continues to change, only one of these bets will become stronger. And consumer relationships are rapidly changing.
Amit Kulkarni is the co-founder of Alloy Advisors and currently serves as Interim CEO of Homes for Heroes. Connect with him on LinkedIn.
