Sam Lessin: VCs are burning money on AI, Meta is best-positioned hyperscaler, and creators are the new seed founders

Summary

Sam Lessin, co-founder of Slow Ventures, argues that the current AI investment cycle is a supercycle of money burning — and Slow has put a billboard on Highway 101 to say so. The creative reads: a burning pile of money around a castle, tagline "AI is not your moat." It's a literal version of Lessin's two-year position: LLMs are powerful technology, but powerful technology and a good business are different things. LPs have too much capital to deploy, GPs get paid fees regardless of outcome, and the result is structurally irrational capital allocation at scale.

His preferred frame is what he calls "AI cherry on top" businesses — companies that are fundamentally sound and happen to get meaningful new leverage from AI, rather than companies whose entire thesis is the disruption itself.

Meta as the safest AI bet

Among the hyperscalers, Lessin singles out Meta as the best-positioned. The logic is straightforward: unlike Google, Meta has no core business that AI can disrupt. It owns attention, sells ads, and LLMs make both the targeting and the creative better. The AI capex, while large in absolute terms, is a small percentage of market cap — essentially cheap insurance that ensures enough of the innovation stays open-source and accessible that Meta never gets cornered. Apple, by contrast, is in a genuinely strange position. Its Apple Intelligence text summaries are, in Lessin's words, "comically bad" — sometimes returning the opposite of what a message says. He floats the idea, only half-jokingly, that Apple Intelligence could be exhibit A in an antitrust consumer-harm case. The saving grace is the hardware moat: making phones is hard, the camera is good, and no one likes green bubbles.

On Google, Lessin thinks Gemini is actually a strong product that almost no one uses, and NotebookLM is genuinely good. The distribution and marketing machine is the failure, not the technology. His read is that Google's path forward probably requires Larry Page and Sergey Brin to step in with founder-level conviction rather than relying on Sundar Pichai to manage the tension between protecting search revenue and cannibalizing it. He draws the Kodak parallel directly: Kodak invented the digital camera and destroyed itself anyway, because the institutional incentives made it almost impossible to act.

On SoftBank and OpenAI, Lessin is blunt: "Whenever SoftBank buys, you sell." He argues OpenAI is caught in the middle — unable to outspend the hyperscalers in a capital-intensive race, and equally exposed if the future is open-source and DeepSeek-style efficiency gains commoditize the model layer.

The rollup thesis

Slow has been doing tech-enabled rollups before it became fashionable. The two anchors are TeamShares, where Slow seeded the company from zero and bought 10% for $400,000, and Metropolis, where they were early seed investors. The underlying argument is that SaaS is frequently the wrong business model for transformative technology: if your software generates 50% profit for a customer, they will pay you a SaaS fee and capture the rest themselves. Buying the customers directly and deploying the technology into owned assets is a better structure — but only if the technology genuinely outperforms. Rollups without real tech differentiation, in Lessin's view, make no sense.

Creator fund

Slow has launched a $65 million vehicle focused on what Lessin calls "seed creators" — people with roughly one million followers, around $1 million in top-line revenue, growing quickly, but without investable capital to build actual companies. The structural problem Slow is trying to solve is that in creator businesses, the equity and community value sit outside any legal entity. If the creator pivots or the community drifts, investors in the C-corp have nothing. Slow's answer is to invest in the creator the same way it invests in founders — first check, no board seat, aligned if they win — and then help them build C-corps and businesses on top of the community they already own.

Lessin is deliberately uninterested in ad revenue. There's no enterprise value in it, and Slow isn't in the business of taking a creator's last dollar. The one carve-out: if Slow's capital helps a creator grow 10x and they then redirect that scale into pure ad plays, Slow will want its share of that upside.

On MrBeast, marked at roughly $5 billion privately, Lessin is skeptical of the business even while crediting Jimmy Donaldson for expanding creator ambition. The audience skews young and low-income, which constrains LTV. The media contracts — a reported $100 million Amazon deal, for instance — reportedly cost more to produce than they pay. His characterization: if it goes public, it trades as a meme stock, not a money-printing machine.

The broader creator investment thesis Lessin is running is that the most interesting opportunity is not the loudest, most over-capitalized creator deals, but the seed layer — where capital is genuinely scarce and the gap between community value and investable capital is largest.