Whatnot founder Grant LaFontaine on 80 min/day engagement, food's next frontier, and why live commerce isn't gambling

Summary

Whatnot, the live-stream shopping marketplace co-founded by Grant LaFontaine in 2020, has never grown revenue below 90% year-over-year across its entire history, reaching a valuation of $11 billion on just under $1 billion raised. The company operates in nine countries, expanded internationally roughly two and a half years ago, and generates average daily engagement of 80 minutes per user, a figure that compares favorably to the most addictive consumer apps in market.

Origin and Product Evolution

Whatnot launched as a static social marketplace for Funko Pop collectibles, a niche that deliberately avoided the broader China live-commerce playbook. The pivot to live video came from observing a seller running informal auctions in Instagram Live comments, where 30-plus seconds of latency made the experience obviously broken. LaFontaine and co-founder Logan shipped a low-latency live auction product in four weeks, building their own streaming infrastructure from scratch. LaFontaine's first personal livestream sold $5,000 worth of Funko Pops in two hours, and the team redirected all resources to live commerce the following day.

For the first three to four years, Whatnot grew entirely through referrals and word-of-mouth within collectibles and enthusiast communities. Revenue accelerated through COVID, briefly decelerated post-pandemic as organic tailwinds faded, and then re-accelerated as the team became more deliberate about growth. Fashion, particularly multi-brand overstock and prior-season inventory sold at roughly 60% below retail, became a major subsequent unlock.

Business Model and Seller Ecosystem

Whatnot takes a marketplace cut from sellers who operate what LaFontaine describes as "digital brick-and-mortar stores." Top-tier sellers now run operations with 100-plus employees. The platform's trust and safety and customer experience function is the company's largest team, a direct response to scrutiny around whether randomized product reveals constitute gambling. LaFontaine is explicit that gambling is prohibited, and the team enforces platform policies accordingly.

Labubu, the collectible toy line from Pop Mart, is generating hundreds of millions of dollars in annual sales on Whatnot despite the company having no direct commercial relationship with the brand's parent. Whatnot does maintain partnerships with Marvel and works with brands in the toys and comics category.

Logan Paul came in as an investor via an introduction through Andreessen Horowitz, which led Whatnot's Series A. Paul's early and active involvement with Pokémon cards made him a natural strategic fit.

Competitive Positioning

LaFontaine argues that purpose-built platforms structurally outperform entertainment platforms that bolt on commerce. TikTok's massive spend on live shopping is cited as an example of the mismatch between a video-discovery intent and a transactional one. Whatnot's response to competition has been consistent execution rather than counter-marketing, a posture reflected in LaFontaine's near-total absence from public media until this appearance.

A meaningful share of daily viewers never transact. On any given day, the vast majority of Whatnot's audience watches without purchasing, yet still averages 80 minutes of daily engagement, suggesting the platform functions partly as entertainment and parasocial content consumption independent of commerce intent.

Priorities and Emerging Categories

Food and drink is flagged as the next major category bet. A farm near San Diego already sells organic citrus on the platform with next-day delivery. LaFontaine envisions direct-from-source seafood and fresh meat as longer-term opportunities, contingent on resolving logistics infrastructure. Cars remain an aspiration but are currently constrained by payment and shipping requirements built into the platform's existing policies.

On AI, Whatnot is deploying LLMs in customer experience, trust and safety, and content understanding, feeding live-stream metadata into ML ranking systems. Core recommendation and ranking models remain traditional machine learning. LaFontaine is cautious about AI features that surface visibly to users, noting that Whatnot's value proposition is explicitly human interaction and that any erosion of that dynamic creates a negative community response.