Alexis Ohanian on women's sports investing, the 'barnacle economy' around AI foundation models, and why social media peaked in 2022

Summary

Alexis Ohanian is making a clear directional bet: sports, live content, and the application layer of AI are where durable value gets built next, while foundation model competition is effectively over for new entrants.

Women's Sports and Live Content

Ohanian's Athlos women's track and field event drew over 3.5 million viewers this year, distributed across YouTube, X, ESPN, and The Zone. He traces his conviction on women's sports back to a March 2019 tweet arguing the sector was undervalued, a call the market has since validated. His broader thesis is that live sports is the last category of content immune to AI displacement. Celebrity athletes derive fame from performance, not from distribution, making them structurally resilient in an era where entertainment and music content will be largely AI-generated or assisted.

The 'Barnacle Economy' Around Foundation Models

Ohanian endorses a framework for early-stage AI investing that centres on companies selling picks and shovels to the large foundation model labs. Several portfolio companies from his 776 fund, seeded in the 2021–2022 vintage, pivoted hard toward this model after the generative AI disruption hit, and are now generating revenue strong enough that he has been asked not to publicise them. The adjacent opportunity is the application layer, where scrappy consumer product teams can build differentiated experiences on top of commoditising model infrastructure. He cites a virtual try-on startup as an example, arguing that large platforms like Google lack the product culture to execute well here.

Social Media Peaked in 2022

Ohanian validated recent Financial Times data showing social media time-on-platform peaked in 2022, with younger cohorts cutting back first. His read is that the combination of COVID normalisation, the proliferation of bot-driven and AI-generated content, and what he calls the Dead Internet dynamic has eroded the authenticity premium that originally drove social engagement. The response he sees in consumer behaviour is a drift toward verifiably human, live, or in-person experiences, citing the rise of run clubs as a concrete offline proxy. The next generation social platform, in his view, either leans into radical intimacy and decentralisation or doubles down on live formats, where Western apps like Whatnot have only scratched the surface relative to Chinese equivalents.

Subreddit Manipulation as a Signal of Platform Decay

Ohanian described a case from roughly eight to nine months ago in which a founder told him he had effectively acquired control of a niche subreddit by purchasing the accounts of its seven moderator accounts. The motive was generative engine optimisation, seeding the subreddit with product-positive posts to influence AI training data and LLM-driven search results. Ohanian treated the anecdote as a leading indicator of how degraded online signal has become, reinforcing his view that live and verifiably human content commands a growing attention premium.

AI Investing Framework: Capital Discipline Over Headcount

On early-stage AI broadly, Ohanian flagged concern that founders raising large rounds at elevated valuations face a discipline test. His preferred model is a small, Navy SEAL-style team that deploys capital primarily into compute rather than headcount, treating the current environment as structurally different from the blitzscaling era. He did not specify fund size or check amounts but indicated 776 continues to seed at early stage.

TikTok and the Frozen Product Problem

Ohanian, who had publicly expressed interest in the TikTok deal, said his DMs remain open but acknowledged the deal was the kind where insiders likely knew the outcome well in advance. He has been publicly critical of TikTok as a CCP vehicle for years and said he welcomes American ownership. The more pointed observation is that TikTok has appeared frozen in product development for roughly two years amid regulatory and ownership uncertainty, which creates a window for competitors. He noted the platform's core ML and recommendation infrastructure was best-in-class but that this advantage has not yet been leveraged into generative AI features.

Retro Gaming and Mod Retro

Ohanian disclosed a portfolio investment in Mod Retro, founded by Torin and Palmer, describing it as a 776 company. He is developing an indie game of his own, budgeted at tens of thousands of dollars total, covering developer, designer, and soundtrack costs. He frames the constrained pixel-art format as analogous to Twitter's character limit, forcing design creativity, and is bullish on a broader Kickstarter-style crowdfunding revival around low-cost indie games.

AI Companions and the Voice-First Generation

Ohanian expressed little enthusiasm for funding AI companion or adult-oriented AI products, noting the market will find its level without his capital. The use case he finds compelling is AI as an always-on executive coach and educational tutor. He uses AI tools with his eight-year-old daughter for trivia and creative storytelling, always supervised. His longer-term hardware thesis is that the AI-native generation will default to voice input over QWERTY interfaces, which he views as an artefact of typewriter-era mechanical constraints. He expects consumer hardware designers, specifically citing Palmer, to build toward this voice-first interaction paradigm.