AppLovin

We see a multi‑moat architecture led by marketplace network effects plus switching costs, reinforced by proprietary models and data scale.

Evidence of network effect strength is the step‑function in monetization quality: with AXON as the demand optimizer and MAX as the supply auction, net revenue per installation increased 72 percent in 2025 on just 3 percent install growth, a hallmark of a learning network.

Switching costs are material for developers who integrate MAX SDKs, tune waterfalls/bidding and automate monetization; moving mediation platforms is non‑trivial and can risk yield.

Intangibles include the AXON AI stack and brand relationships across advertisers and publishers; Adjust adds attribution and Wurl adds CTV reach, broadening data and utility. Cost advantages stem from at‑scale infrastructure and very low incremental delivery costs, reflected in cost of revenue of 12 percent.

Efficient scale is present in parts of the in‑app bidding niche but remains contestable given capable rivals.

Component scores: network effects 85/100, switching costs 80/100, intangibles 75/100, cost advantages 70/100, efficient scale 60/100; weighted (35/25/20/15/5) to 82. Key sources: 2025 10‑K descriptions of AXON, MAX and revenue model (agent, net of inventory costs), and 2025 performance disclosures.