AppLovin's Adam Foroughi: Axon self-serve ad platform launches in two days — from 2,000 to millions of advertisers

Summary

AppLovin is preparing to launch its self-serve Axon advertising platform within two days of this recording (September 29, 2025), a move that could expand its advertiser base from roughly 2,000 managed accounts to potentially millions. The company currently processes more than $11 billion in annualized ad spend — a figure that has grown across recent quarters — while operating with fewer than 1,000 employees total and a core advertising business team of just 300 to 350 people.

The Axon launch is the defining near-term catalyst. CEO Adam Foroughi frames the opportunity explicitly against Meta's trajectory: when Facebook opened Ads Manager it was generating $55 billion in annual revenue; AppLovin already exceeds double that run-rate in ad spend volume, yet has never offered self-service access. The Q4 timing is deliberate and carries execution risk. The company has spent the past year building fraud controls, automated compliance tooling, and end-to-end payment infrastructure to handle the transition without the managed-service guardrails it has relied on since founding.

The audience and inventory profile are frequently misunderstood. AppLovin's supply is built on casual mobile games — solitaire, match-three, word puzzles — with more than one billion daily active players. The demographic skews toward adult women who are household heads, on premium devices, with average ad viewing times of 35 seconds, and up to 60 seconds for opt-in rewarded placements. Foroughi argues this is among the highest-attention ad inventory available anywhere, and is directly comparable to what streaming pre-roll can offer.

The ecommerce pilot is the proof of concept Axon is built on. A few hundred direct-to-consumer brands, including Ridge Wallet, have been running on the platform over the past year. Foroughi says those advertisers are seeing spend levels comparable to Meta. The revenue-based pricing model, where advertisers set a target return and scale spending only when that return is met, means AppLovin does not operate a traditional sales force and carries no budget-request dynamic.

Infrastructure is deliberately constrained. AppLovin runs on Google Cloud, expensing compute costs rather than capitalizing them. Foroughi explicitly rejected hypothetical large-scale on-premise GPU investment, arguing that compute scarcity forces engineers to optimize model efficiency. He describes a tradeoff where a one-point accuracy loss might reduce compute costs by 90%, a discipline baked into company culture from its bootstrapped origins.

Headcount discipline is central to the business model. Most coding is now done by LLMs. The ecommerce go-to-market team is 30 people. Foroughi says the company churns employees who cannot sustain high individual contributor output and has no hiring quotas. He frames this as both a cultural and strategic asset, one that makes acquisitions structurally difficult.

AppLovin did bid on TikTok and the industrial logic, deploying its ad model against TikTok's inventory, was considered sound internally. The bid was ultimately set aside in favor of organic growth, with Foroughi citing the distraction cost to a sub-1,000-person team and the scale of the Axon opportunity as the deciding factors.

On LLM-driven ad monetization, Foroughi expects search ad dollars to migrate toward LLM output placements as users bypass the discovery phase entirely. He views this as deflationary for bottom-funnel advertising and additive to top-funnel platforms like AppLovin, since cheaper conversion costs free advertiser budgets for discovery spend. He is skeptical that LLM interfaces themselves are viable ad surfaces given the inability to interrupt long-form query experiences meaningfully.

The venture community passed on AppLovin in 2012, consistently citing Google and Facebook as existential threats. The company grew without institutional capital, and Foroughi attributes the competitive durability partly to the revenue-based pricing model that made AppLovin additive to both Google's bottom-funnel search business and Meta's top-funnel social inventory, rather than directly substitutive.