COMESA, WhatsApp Business, and Antitrust in Search of a Theory

Key Takeaways

• COMESA uses SLC test instead of abuse‑of‑dominance standard
• Essential‑facilities doctrine unlikely for WhatsApp due to alternatives
• Self‑preferencing claim weak; WhatsApp not presented as neutral gateway
• Tying plausible if regional AI firms rely on WhatsApp API
• Importing contested EU doctrines risks over‑regulation of digital markets

Summary

Meta’s October 2025 amendment to WhatsApp Business API barred third‑party AI providers while favoring Meta AI, prompting antitrust probes by the EU, Italy and now COMESA. COMESA’s investigation invokes Regulation 36 but mistakenly applies the substantial‑lessening‑of‑competition (SLC) merger test rather than an abuse‑of‑dominance standard. The commission’s notice leans on refusal‑to‑deal and essential‑facilities arguments, yet establishing WhatsApp as an essential facility or self‑preferencing platform is legally tenuous. A tying theory may be the most viable, but it hinges on defining a relevant market and proving regional AI firms’ foreclosure.

Pulse Analysis

Meta’s decision to block third‑party AI access to the WhatsApp Business API has ignited a wave of antitrust scrutiny, with the COMESA Competition and Consumer Commission joining the EU and Italy in probing potential abuse of dominance. The COMESA notice, however, misapplies the substantial‑lessening‑of‑competition (SLC) test—a standard reserved for merger control—rather than the abuse‑of‑dominance framework prescribed by Regulation 36. This doctrinal slip risks distorting the analysis from the outset, as the commission must first define the relevant market and assess whether Meta’s conduct is exclusionary or exploitative before invoking any competition test.

Legal scholars have debated several theories of harm. An essential‑facilities claim appears weak because alternative messaging platforms and distribution channels dilute WhatsApp’s uniqueness, making it hard to prove indispensability for AI competitors. Likewise, a self‑preferencing argument falters; Meta has not positioned the WhatsApp Business API as a neutral gateway, and rivals have not historically depended on it for market access. The most promising avenue is a tying theory, which requires showing that WhatsApp (the tying product) and Meta AI (the tied product) are distinct, that users are coerced into the tie, and that rivals are foreclosed. While global AI giants can bypass the API, smaller regional AI firms may rely on WhatsApp for customer reach, offering a narrow but defensible tying case if market evidence supports substantial foreclosure.

The broader policy lesson for COMESA is caution against importing contested EU doctrines wholesale. Over‑reliance on frameworks like the Google Android Auto decision could chill platform openness, discouraging firms from experimenting with interoperable architectures for fear of perpetual regulatory obligations. A balanced approach—grounded in solid economic analysis, clear market definitions, and respect for a platform’s right to evolve—will better serve Africa’s digital economy, fostering both competition and innovation without imposing undue administrative burdens.