How We Got Fred Wilson, Benchmark and Index to Invest $94M | Why Robinhood's Strategy is Wrong

Why does it matter? Because the standard playbook for building a consumer trading app is exactly backwards.

FOMO raised from 140 angels not to fund the company but to solve a distribution problem — then gave early non-founder engineers the equity percentage that normally only founders get, while the entire team worked for free for eight months. That's how Paul Erlang built a $550M trading app with 17 people, zero hierarchy, and sharp opinions about why Robinhood's entire strategy is wrong.

• Consumer angel rounds should be sized around distribution reach, not runway — your best users become owners first, advocates second
• Non-founders who receive founder-level equity (2-3%) become an extended founding team; the ownership culture that results is a more durable moat than headcount
• "Everything app means not intentional" — financial super apps without a unifying thesis lose to focused, conviction-led products
• AI company valuations depend entirely on one number: developer token spend as a percentage of salary (currently ~3.8% at Salesforce; at 20%, OpenAI and Anthropic are $5 trillion companies)

For consumer products, the angel round isn't a capital event — it's a distribution weapon

Consumer products die from lack of initial users willing to care, not from lack of capital. FOMO solved this structurally: 140 angels, chosen not for check size but for distribution reach. "When we raised the initial round, the goal was to create distribution and we think that our best users should have some ownership in the product." Not all of them were traders. "There are definitely builders in the industry and we've been able to leverage tons of them as we continue to build."

The lesson for consumer founders: select angels based on their value as users and distributors first, check size second. Structure round size around distribution reach, not runway math. Great consumer products can launch into silence and die there — the structural fix is to make your best prospective users owners before you ship.

B2B fundraising advice doesn't translate. A B2B company hires a sales team and runs outbound. Consumer has no equivalent lever — unless you engineer one by turning investors into advocates with skin in the game.

Giving non-founders founder-level equity — zero salary for eight months included — creates an extended founding team that no incentive structure can match

The first eight months at FOMO, nobody took any pay. What they received instead: "We gave non-founders a percentage of the company that usually founders get." Concretely, 2-3% to a handful of core engineers. "If those five to seven to 10 people build this business for the next 10 years, there's literally nothing stopping us."

The result is what Paul calls an extended founder team — engineers who don't need managing, who ship on weekends not because they're told to but because they own the outcome. FOMO is 17 people building at a $550M valuation. That isn't just lean. It's a different model of what a company is.

In an AI-native world, this scales. When 17 people can build what once required 100, the math on equity changes. Headcount isn't the moat anymore — depth of ownership is. "Which is why it's okay to give more equity early. That's how we saw things." Standard 0.25-0.5% option grants were designed for a world of large teams. That world is dissolving, and the equity structure should dissolve with it.

'Everything app means not intentional' — financial super apps fail without a unifying conviction

Bundling products without a reason for them to coexist isn't a strategy. Paul's critique lands hard: "everything app means not intentional. It means let's just throw everything in there for the user to access. What is the glue between these things?"

FOMO's answer is the social graph and thesis-driven markets. A user should be able to express a single market conviction — say, that the Strait of Hormuz will close — across asset types in one place: buy oil on Hyperliquid, short US equities exposed to oil, purchase the prediction market. "The reason these different market types exist is to express conviction on a belief, whereas all these other apps are just everything super apps and there's no intentionality behind why all those things have to exist in the same place."

Robinhood's horizontal expansion made sense for saturating the US market. But bundling retirement accounts with sports betting and prediction markets doesn't cohere around a story. The winner in financial services won't be the most comprehensive platform. It'll be the one with a single-sentence answer to why every product in the bundle belongs there. Without that sentence, it's a super-app trap.

The entire question of whether AI companies are worth trillions hinges on one number almost nobody is tracking

3.8%. That's the current figure — developer token spend as a percentage of developer salaries at Salesforce, extrapolated from Marc Benioff's disclosure that they spend $300 million on Anthropic. Paul's frame is unsparing: "If it stays there, paying a trillion dollars for OpenAI and Anthropic is grossly overvalued. If it goes to 20%, which is what many think it will, they're $5 trillion companies."

Can FOMO see itself reaching 20% of dev salary on tokens? "Definitely. Yeah... I think 20% is definitely within reason." The bet underneath that answer is commoditization — tokens get cheaper as energy and compute decline, models converge, spend percentage climbs.

But right now, 3.8% is the honest baseline. Investors and CFOs who want a real signal on AI fundamental value should track developer token spend as a percentage of engineering payroll — not revenue multiples, not GPU shipments, not benchmark scores. That single ratio, watched quarter over quarter against a 20% threshold, tells you more about AI company valuations than anything else on an earnings call.

Product-market fit is momentum — and the correct response to traction is urgency, not a breath

BeReal had product-market fit. Clubhouse had product-market fit. Both collapsed. The failure mode wasn't the product — it was treating traction as a plateau.

"Everything is about momentum. When you have momentum, instead of taking the gas off the pedal — it's like no, you need to double down 10 times harder. Once you lose momentum the boulder just starts rolling down the hill." At FOMO, the framing is almost anxious: every day, ship features or lose everything. That's not neurosis — it's clarity about how consumer gravity works.

BeReal required users to act every single day. "As soon as you lose that you lose momentum very quickly." Clubhouse amplified with celebrities, diluted the core user base that actually cared, and never recovered the original energy. Both failed the same test: when they had it, they stopped pushing.

PMF is a forcing function for acceleration. Founders who treat it as a signal to optimize org structure or stabilize headcount are making the same mistake — momentum lost is catastrophically expensive to rebuild.

When a content format converts, iterate it to perfection — chasing the next format is the most common growth mistake

The instinct the moment something works: go find the next thing. "That is completely wrong." The correct move is to freeze format exploration and build a systematic optimization loop around every variable of the format that's converting.

"Users are more likely to convert if this is the type of font, if this is the color of the font, if this is the placement of the font, if it's this person talking versus this person talking." FOMO manages 30-40 creators in house, constantly pruning underperformers and doubling on what converts. The metric is acquisition cost against lifetime value — conversion-driven, not impression-driven.

One nuance worth flagging: increasing CAC can still be rational. If a harder-to-reach user carries the same LTV, spending $3 to acquire them instead of $0.80 is correct — as long as CAC stays below LTV. Growth teams shouldn't optimize only for the lowest-hanging fruit. But they should never diversify away from a format that's working. Freeze format exploration the moment something converts. Replicate at scale before introducing anything new.

Delay your round announcement until you're ready to raise again — PR is a fundraising tool, not a milestone

Standard advice: close the round, announce it immediately, get the press. The problem: "as soon as you announce a round, you get tons of inbound from other investors and it takes up time to kind of tell them, 'No, we're not raising right now.'"

The fix is clean: decouple round closing from round announcement. Hold the announcement until you're ready for the next raise. Then the inbound you field is pre-warmed, timely, and convertible — not a time tax. "If you really want to raise capital in the near future, you can just wait to announce your round until you're ready."

The corollary matters equally: just because a VC wants to meet doesn't mean they want to do your round. VCs' entire job is to meet companies. Founders who mistake investor meetings for investment signals waste cycles that should go toward building. Timing the announcement yourself — rather than letting the close date determine it — hands that leverage back to the founder.

The 17-person company at a $550M valuation isn't a curiosity — it's the template

FOMO is 17 people, planning to stay under 25. No hierarchy, no one-on-ones, 2-3% equity grants to a core team that has never needed managing. The model only works because AI has changed the production function for software — and the implicit bet running through this entire episode is that the best companies of the next decade look less like organizations and more like extended founding teams.

Incumbents can respond by cutting headcount and reorganizing. But building ownership culture from scratch is harder than inheriting one that was never diluted by scale.

They won't move fast enough.

---

Topics: fintech, consumer startups, fundraising, equity compensation, AI productivity, social trading, product-market fit, creator growth, venture capital, crypto