nickgallick SKILL 48: Antitrust & Competition Law
Purpose
Understand when platform growth triggers antitrust risk, what practices are per se illegal, and how to defend against competitor antitrust complaints.
When Antitrust Applies
- 30%+ market share: starts raising questions
- 50%+ market share: creates presumption of market power
- Even without dominance: certain PRACTICES trigger antitrust regardless of market share
Sherman Act §1 — Agreements in Restraint of Trade
Per Se Illegal (No Business Justification Defense)
- Price fixing: coordinating with competitors on fees
- Market allocation: agreeing with another platform "you take sports, we take politics"
- Group boycott: coordinating with other platforms to exclude a competitor
- Tying: "you can only use Agent Sparta if you also use our prediction market" — illegal if you have market power in one product
Rule of Reason (Justification Defense Available)
- Exclusive dealing arrangements
- Vertical restraints (between platform and participants)
- Most partnership agreements with non-compete provisions
- Defense: demonstrate pro-competitive justification + limited anticompetitive effect
Sherman Act §2 — Monopolization
What IS Legal
- Having a monopoly through superior product, business acumen, or historical accident
What is NOT Legal
- Maintaining a monopoly through exclusionary conduct
- Examples of exclusionary conduct:
- Exclusive contracts with all major AI labs (preventing competitors from accessing models)
- Below-cost pricing to drive out competitors (predatory pricing — must be below marginal cost AND with reasonable prospect of recoupment)
- Refusing to interoperate with competing platforms
- Self-preferencing: your own agents receive favorable scoring/visibility on your platform
- Denial of access to essential facilities
Platform-Specific Antitrust (Post-Epic v. Apple)
If Agent Sparta Controls a Market
- If Agent Sparta becomes the dominant/only AI competition platform: platform antitrust applies
- Self-preferencing risk: if Perlantir agents compete on Agent Sparta with better scoring algorithms → antitrust concern
- Fee structure: if excessive fees create competitive foreclosure → potential issue at scale
- Data advantages: if your platform gives you data competitors can't access → potentially exclusionary (see FTC v. Facebook)
Connector/API Standard
- Keep the connector/API open and non-discriminatory
- Document: any AI agent framework that meets the public spec can participate
- Do NOT make exclusive deals with AI providers to prevent competitors from using those models
- Open standards = antitrust protection (you're NOT restricting access)
Relevant Case Law
- United States v. Google (2024): Google liable for monopolization of search through exclusive distribution agreements with Apple, Mozilla, etc.
- Epic Games v. Apple (9th Cir. 2023): Apple's App Store restrictions analyzed under rule of reason; mostly upheld but anti-steering injunction issued
- FTC v. Meta (ongoing): attempted monopolization through acquisition of Instagram/WhatsApp
- Aspen Skiing Co. v. Aspen Highlands Skiing Corp. (1985): essential facilities / refusal to deal doctrine
- Verizon Communications v. Law Offices of Curtis V. Trinko (2004): limits on essential facilities doctrine — generally, no duty to deal with competitors
Antitrust Compliance Practices
- Keep Agent Sparta connector/API open and documented — any compliant framework participates
- No exclusive deals with AI providers that prevent competitor platforms from using those providers
- No predatory pricing (don't price below cost to kill competitors)
- No self-preferencing: Perlantir agents compete under the same rules as all others
- Document all pricing decisions with legitimate business justifications
- No coordination with competitors on fees, market allocation, or participant targeting
- If competitors request API access / interoperability → don't refuse without legitimate technical justification
Defense Framework (Competitor Antitrust Complaint)
- Challenge market definition: "AI agent competition" is too narrow — the relevant market includes all AI benchmarks, developer competitions, hackathons, and competitive programming platforms
- Deny market power: even if the defined market is accepted, market share + barriers to entry analysis
- Demonstrate open access: connector spec is public, any agent can participate, no exclusionary agreements
- Pro-competitive justification: all design decisions serve product quality, not exclusion
- No harm: competitor hasn't suffered actual antitrust injury (they can still operate)
This is legal research and intelligence, not legal advice. Consult qualified legal counsel before taking action.