Brokers tend to talk about MLS fragmentation as a tax on growth – the cost of joining one more system every time a team pushes into new territory. For the brokerages large enough to pay their way around it, fragmentation also works as a competitive barrier against everyone who can’t. Close to 500 MLSs still operate across the United States, and more than 200 of them have fewer than 800 members, according to Nick McElmurry, chief strategy officer of My State MLS. Anyone who wants a complete national picture of listing data has to assemble it one system at a time.
Only the largest brokerages can afford to do that. McElmurry describes companies that belong to as many as 350 MLSs and spend millions of dollars, along with a dedicated development team, to build a single database of listings nationwide. “A lot of the large brokerages like that because then it’s an exclusive benefit that they can provide,” he says. Smaller competitors can’t match the investment, so the same fragmentation that frustrates them shields the companies at the top. What may finally change that, McElmurry argues, is AI – specifically, agents’ growing expectation that their AI tools can read the data MLSs control.
Not every large brokerage has settled for working around the barrier. Compass, which now includes the Anywhere brands, has gone after fragmentation directly, partnering with Chicago-based MRED to take that MLS nationwide; CEO Robert Reffkin has said the goal is a national MLS that replaces local ones. The two strategies point in opposite directions, but they rest on the same advantage. Only the largest brokerages have the resources either to buy their way around fragmentation or to reshape the system themselves.
The industry’s recent attention has gone elsewhere. At the start of 2026, the National Association of Realtors stopped requiring Realtor membership for MLS access and left the decision to each MLS, and several – MRED among them – have since dropped the requirement. Many agents welcomed the change after the commission lawsuits left them angry at NAR. McElmurry says the change hasn’t altered the conversations his company has with prospective members, because it leaves the underlying structure intact. “The whole regional component and having to belong to 10 MLSs is still alive and true regardless of the NAR requirements,” he says. What may finally change that structure, he argues, is AI – specifically, agents’ growing expectation that their AI tools can read the data MLSs control.
Before AI can pry the structure open, it has to overcome the inertia that keeps agents inside the system, and that inertia runs strongest in big cities. In a large metro, thousands of agents share the same MLS, the same showing tools, the same contracts and the same forms. No one can maintain relationships with 5,000 colleagues, so the shared system stands in for them. “These 4,998 people are in the system,” McElmurry says. “And if you want to do things mildly differently outside of the system, you are going to get so much pushback because they only play in the system.”
That pushback can reach as far as offers. McElmurry says members who leave a big-city MLS sometimes find that other agents won’t look at their offers because they arrive on something other than the locally mandated form – even though every agent has a fiduciary duty to present offers to the client. In his account, local rules can end up overriding legal obligations in practice, and the cost of stepping outside the system lands on the agent who left.
Outside the big metros, the calculation changes. In rural markets with 20 brokerages whose agents all know one another, business runs on direct conversation rather than automated MLS workflows, so leaving a local system costs far less. Agents near state lines feel fragmentation most directly: an Omaha agent licensed in both Nebraska and Iowa would otherwise need memberships in two state boards and two local boards to serve a market that spans a short drive. Those agents can work around fragmentation, but their workarounds leave the large metro systems – and the data they hold – untouched.
The pressure that rural and border agents can’t apply, AI may. Today most agents use ChatGPT or similar tools conversationally – drafting listing descriptions, brainstorming, consulting chatbots. McElmurry expects that to change as MCP servers – protocols that let AI applications connect directly to outside data sources and software – become common. Once agents hook their AI into every system they run their business on, they will expect it to see listing data too.
That expectation puts MLSs in a bind, because they act as the industry’s data gatekeepers. Agents already use AI tools to pull listing data in bulk, which in practice means bots scraping MLS sites, and the natural instinct of any company that holds data is to shut that down. An MLS could build its own AI tools around its data, but for larger, more technically sophisticated brokerages that wouldn’t go far enough – the MLS would have to replicate all the software those brokerages use outside it. So far, few have moved. UtahRealEstate.com, one of the country’s largest MLSs, has launched an MCP server, and MLS software vendor FBS has tested one, but they remain exceptions among close to 500 systems.
He doesn’t expect MLSs to hold that position indefinitely. “Once the norm is MCPs, allowing your AI to at least view all of the components of your business – to view it is eventually going to be a necessity,” he says. He compares the position MLSs will be in to having a gun to their head – dramatic, he concedes, but a fair description of pressure he sees as unavoidable.
The brokerages with the most to gain from open data are the ones fragmentation currently shuts out. Large brokerages already hold the national dataset they paid to build. AI access to MLS data would give smaller competitors a version of that advantage without the millions of dollars or the development team, and McElmurry expects pressure on MLSs to build as those smaller brokerages grow more technically capable.
That shift would reverse the current economics. Today fragmentation rewards scale, because only scale can pay to overcome it. If MLS data flows into agents’ AI tools as a matter of course, the exclusive benefit the largest brokerages built narrows to something any capable competitor can reach. The timeline depends on how quickly agents move past chatbots, and for now that remains early: McElmurry says agents who connect AI to all their business software in a secure way are still uncommon.
While MLSs and brokerages sort out who gets access to data, McElmurry worries the industry has its attention pointed in the wrong direction. Much of that attention has gone to the fight over private listings, with Reffkin as its most prominent voice. McElmurry agrees with parts of Reffkin’s critique, including his challenge to what both see as monopolistic MLS structures. His concern lies with what the debate crowds out.
Consumers have soured on agents since the commission lawsuits, and a prolonged internal fight does little to repair that. Meanwhile, McElmurry argues, a company with sophisticated enough AI could enter the market and undercut agents in a lot of ways – a possibility he says almost no one in the industry discusses. AI, in other words, cuts both ways. The same AI pressure that could level data access among brokerages could also hand that data to competitors outside the industry, and the brokerages and MLSs still arguing over listing rules would meet that competition unprepared.
About the Expert: Nick McElmurry is chief strategy officer of My State MLS, a nationwide multiple listing service. He is based in Grand Rapids, Michigan.
This article is intended for informational purposes only and does not constitute legal, financial, or investment advice. The views and opinions expressed herein reflect those of the individuals quoted and do not represent an endorsement of any company, product, or service mentioned. Readers should conduct their own due diligence and consult qualified professionals before making any investment decisions.
Explore similar articles from Our Team of Experts.
A sudden rise in mortgage rates has not landed evenly across Hampton Roads, Virginia. Over three weeks to a month, rates climbed almost a full percentage point, according to Bubba Hutchison, who leads Hutchison & Co. at RE/MAX Connect in Yorktown. Freddie Mac’s weekly survey tracked the same climb: the average 30-year fixed rate rose ...
For most of the past two decades, the conventional playbook for condominium buyers in San Francisco followed a predictable arc: buy a studio, sell in five to seven years, move up to a one-bedroom, repeat. Appreciation was reliable enough that buyers could treat each purchase as a stepping stone. That assumption no longer holds, according ...
Harford County, Maryland, never gave back its pandemic-era gains. Prices in the rural county 30 minutes north of Baltimore rose slowly and without a dip from 2020 until about May of this year, according to Lee Tessier, who leads the Lee Tessier Team of eXp Realty in Bel Air. Over the same stretch, he says, ...
Data center construction is pushing smaller buyers to the back of the line for sand, gravel, and stone. Wherever large campuses go up, they claim local supply before anyone else places an order. When a quarry commits its stock to one of those projects, the homeowner building a pool or the small developer pouring a ...
