AI Pricing Algorithm

State of play.

• DOJ has established the first federal enforcement template for algorithmic pricing. The proposed RealPage settlement prohibits use of competitors' nonpublic data in pricing models, mandates a court monitor, and is pending Tunney Act review — no financial penalties, no liability admission, but a compliance architecture that defines the floor .
• The FTC and House Oversight Committee are running parallel surveillance pricing investigations, with the FTC's Section 6(b) study ongoing and the House demanding documentation from travel and platform companies on revenue management algorithms and consumer data practices .
• State legislative activity has reached critical mass. More than 40 bills across at least 24 states have been introduced to regulate personalized algorithmic pricing; California's AB 325 is already in effect prohibiting competitor-data-reliant algorithms, and AB 2564 would go further with an outright ban carrying $12,500-per-violation civil penalties .
• AI vendor pricing models are shifting from fixed to variable, with Salesforce, Workday, and OpenAI moving to consumption-based structures — "agentic work units," "units of work," and token-based billing — creating new contract-drafting exposure for enterprise buyers where measurement methodologies remain undefined across the sector .
• For counsel advising companies that deploy algorithmic pricing or procure AI software, the practical baseline is a simultaneous exposure on two fronts: antitrust and consumer-protection enforcement targeting how pricing algorithms use competitor and consumer data, and contract risk from AI vendor pricing structures where cost caps and audit rights are not yet standard.

Where things stand.

• DOJ's RealPage settlement is the governing enforcement template. The settlement requires RealPage to stop using competitors' nonpublic data, limit model training to historical data at least 12 months old, and restrict geographic granularity to the state level — with a court monitor and Tunney Act review still pending .
• California AB 325 is already operative. Effective January 2026, it prohibits pricing algorithms that rely on competitor data — making California the first state with a hard statutory prohibition rather than an investigation or pending bill .
• The FTC's Section 6(b) surveillance pricing study, initiated in 2024, remains active. FTC leadership has told Congress that staff work continues and the agency is considering disclosure requirements for highly personalized, data-driven pricing .
• Regulators are drawing a line between traditional dynamic pricing and surveillance pricing. Market-condition-based dynamic pricing remains lawful; pricing tied to individual consumer data is the enforcement target — but the line is not yet defined by statute or final rule at the federal level .
• Federal legislative proposals are in play. The One Fair Price Act, introduced by Senators Gillibrand, Gallego, and Booker, would ban surveillance pricing nationally — adding a potential federal floor above which states are already moving .
• AI software contract architecture is in transition. The shift from per-seat to consumption-based pricing — confirmed across roughly 40 companies in a Goldman Sachs analysis — means enterprise AI contracts now carry variable cost exposure with measurement methodologies that vendors have not yet standardized .
• Antitrust scrutiny is extending beyond real estate. Healthcare, hospitality, retail, and grocery sectors are identified enforcement targets; the RealPage settlement is explicitly framed as a signal to other industries .

Latest developments.

• AI vendors shifting to consumption-based pricing — Salesforce "agentic work units," Workday "units of work," OpenAI token-based billing — with contract terms still undefined across the sector; a Goldman Sachs analysis of roughly 40 software and internet companies confirms the shift is sector-wide .

Active questions and open splits.

• Where does dynamic pricing end and surveillance pricing begin? Regulators have articulated a distinction between market-condition-based pricing and consumer-data-driven individualized pricing, but no federal statute or final rule defines the line — leaving companies operating in the gap .
• What compliance framework satisfies the RealPage template? The settlement's data-age, geographic-granularity, and competitor-data restrictions are the current floor, but the Tunney Act review may impose additional conditions — and the court monitor's scope is not yet detailed in public filings .
• Vendor liability vs. customer liability for algorithmic collusion. RealPage targeted the software vendor; enforcement posture toward landlords who used the software remains less defined. Whether customers of algorithmic pricing tools face independent Sherman Act exposure is an open question across sectors .
• How will conflicting state regimes interact? California AB 325 is operative; AB 2564 would go further; Maryland, New York, Tennessee, and Arizona have introduced similar measures. Companies operating nationally face the prospect of irreconcilable compliance obligations before any federal preemption framework exists .
• What contract protections are adequate for consumption-based AI pricing? Measurement methodologies for "agentic work units" and token-based billing are undefined by vendors; cost caps, audit rights, and dispute mechanisms are not yet standard — and the enterprise market has not produced a settled template .
• Will the One Fair Price Act or state bans survive constitutional challenge? A federal ban on surveillance pricing would face commerce clause and First Amendment scrutiny; state bans raise preemption questions if federal enforcement expands .

What to watch.

• Tunney Act review outcome for the RealPage settlement — whether the court imposes additional conditions beyond the proposed consent decree, and when the court monitor's scope is detailed in public filings .
• FTC action following its Section 6(b) study — whether the agency issues disclosure requirements, enforcement actions, or a rulemaking targeting surveillance pricing .
• California AB 2564 and parallel state bills — whether any enact into law, triggering a multi-state compliance obligation and potential preemption litigation .
• Whether DOJ brings a second algorithmic pricing enforcement action in healthcare, hospitality, or retail — the sectors explicitly flagged as next in line .
• Emergence of standard contract terms for consumption-based AI pricing — whether any major vendor publishes defined measurement methodologies or whether enterprise buyers begin demanding audit rights and cost caps as a negotiating baseline .