Algorithmic Pricing Antitrust Risk 2026: When AI Pricing Becomes Illegal Collusion
Two competitors who never speak to each other can still be sued for price-fixing if they both use the same AI pricing algorithm. Here's how the RealPage litigation reshaped antitrust exposure for anyone licensing a shared or vendor-supplied pricing tool.
Why Antitrust Law Applies to Software That Never Talks to a Human
Section 1 of the Sherman Act prohibits agreements between competitors that restrain trade — the classic example is a phone call where two rivals agree to fix prices. Antitrust enforcers have long recognized that competitors don't need an explicit agreement to run afoul of the law; a "conscious parallelism" of independently reached decisions is generally legal, but coordination facilitated by a shared mechanism is not. AI pricing software is now that mechanism.
The RealPage litigation crystallized this theory: landlords who separately chose to license the same revenue-management software, which recommended rents using pooled data from other landlord clients, were alleged to have effectively agreed to a common pricing strategy without ever negotiating directly. Courts have allowed these claims to proceed past motions to dismiss, and the theory has since been picked up by plaintiffs' firms and state AGs targeting pricing software in other sectors.
The Fact Pattern That Creates the Most Exposure
The Algorithmic Pricing Antitrust Compliance Checklist
- ☐Identify whether your pricing tool is built in-house or licensed from a third-party vendor
- ☐Ask the vendor directly whether the algorithm uses pooled data from other clients, including competitors
- ☐Confirm whether any competitor-derived data is real-time, non-public, or client-specific versus aggregated and delayed
- ☐Retain and document the ability to override or reject the algorithm's recommended price
- ☐Avoid vendor contract terms or incentive structures that discourage deviation from suggested pricing
- ☐Keep records showing independent business judgment was exercised, not just accepted output
- ☐Ask how many of your direct competitors also license the same tool
- ☐Ask whether the vendor markets the tool's ability to 'increase win rates' by predicting competitor behavior
- ☐Review any vendor materials suggesting the tool helps avoid 'leaving money on the table' relative to competitors — a phrase that has appeared as evidence in prior litigation
- ☐Route any new shared or vendor pricing tool through antitrust counsel before rollout, not after a subpoena
- ☐Reassess existing vendor relationships against the RealPage fact pattern, not just new purchases
- ☐Build a documented, periodic review process rather than a one-time sign-off
Why This Is Different From Discrimination-Based Pricing Risk
Algorithmic pricing discrimination claims (surveillance pricing, disparate impact on protected classes) focus on how an algorithm treats individual consumers. Antitrust claims focus on a completely different harm: whether the algorithm functions as a coordination device between competing businesses, harming the market as a whole through artificially elevated prices. A pricing tool can be low-risk on the discrimination axis — charging every customer the same fair price — while still creating serious antitrust exposure if it coordinates that price across an entire industry using shared competitor data.
Businesses often run privacy and consumer-protection review on a new AI pricing tool without ever routing it through antitrust counsel, because the discrimination framing is more familiar. The RealPage line of cases makes clear that's an incomplete review — the two risks require separate diligence.
Frequently Asked Questions
Is it illegal to use AI pricing software at all?
No. Using AI or software to set prices based on your own data is not itself illegal — dynamic and algorithmic pricing built on proprietary data is a normal, legal business practice. The risk arises specifically when the tool pools non-public data from competitors, or when competitors' shared reliance on the same recommendation functions as a substitute for an illegal price-fixing agreement.
Can we be liable even if we never override the algorithm and just happened to follow it like everyone else?
Potentially, yes — and following the recommendation without deviation is itself a fact plaintiffs point to as evidence of an effective agreement. The less independent judgment a business exercises over the algorithm's output, and the more it relies on data the algorithm draws from competitors, the stronger a plaintiff's coordination argument becomes.
Does it matter if the pricing algorithm is 'AI' versus a simpler statistical model?
Not for antitrust purposes. The legal theory doesn't turn on whether the tool uses machine learning versus a simpler rules-based formula — it turns on whether competitors coordinate pricing through a shared mechanism using non-public competitor data. Marketing a tool as 'AI-powered' doesn't create risk by itself, and calling it something else doesn't remove it.
Related Guides
Audit Your Pricing Vendor Before a Subpoena Does It For You
The fastest way to reduce algorithmic pricing antitrust exposure is to ask your vendor the questions a regulator would ask: what data feeds the recommendation, who else licenses the same tool, and how much independent judgment your business actually exercises over the final price.
Get that diligence done now, while enforcement in this area is still expanding beyond real estate — retrofitting a defense after a state AG opens an investigation is far more expensive than reviewing the contract today.