Volo Ventures

Raise venture only if you can name how you reach a billion dollars. Andrew Maguire's test is to ask how the company sells for a billion dollars before taking venture money. If you can't answer, and that outcome isn't your sole focus, venture will limit your options.

Funds are built on the power law: limited partners expect outlier returns, and a few wins must cover the losses. One founder at a few thousand dollars in MRR wanted to raise but 'could only grasp at straws' when asked for the billion-dollar path. Jacob Eiting: "Unless that is your goal and you're going to do nothing but focus on that outcome, do not raise venture capital."

Sub Club▶ 37:25

Match capital to the business's real ceiling, measured in years of your life. Some businesses can become massive with great execution and others can't, however well they are run. If you plan to spend five to ten years on something, know the best case before choosing how to fund it.

"Just align the capital strategy to reflect what type of company you're trying to build."

Sub Club▶ 39:55

Build a cash-flowing $10M business first, then raise if a bigger market appears. Running a profitable business of around $10M can show where the multi-billion-dollar company actually is, and then you can raise at a higher valuation with less dilution and better terms. If the big market never appears, you still own the whole business.

Curtis from Slups bootstrapped to about 12 people, reinvests every year and is happy with the result. The middle ground is thin: small-check funds like Tiny, NDVC and ComFund stopped, though cohort-based financing like General Catalyst's exists for large apps.

Sub Club▶ 38:58

Niche utility apps, not rivals to network-effect giants, are the bootstrapping lane. Disrupting chess.com or Strava is nearly impossible and a poor venture bet, but a niche chess app solving one problem could earn $10M a year. Near-zero build cost, billing infrastructure like RevenueCat, and paid acquisition funded by working capital let a small team scale without a VC.

Strava is about 15 years old and chess.com subscribers can stay for decades. Previously VCs wouldn't fund a chess.com rival, so these products never got built; now nobody needs $10M to build one. "There has never been a better time to indie developer a $10 million app."

Sub Club▶ 22:17

Consumer venture has faded because consumer apps rarely produce giant outcomes. Venture into early-stage apps is much less common than ten years ago, and outside fewer than ten social products no $100B consumer company exists. Duolingo, worth roughly $4B to $5B, is still venture-backable but harder to repeat in 2026.

Maguire calls venture a transient form of capital that suits fields where both costs and returns are high, like the Dutch East India Company, and fades once costs fall. Nearly all venture money now goes to the AI build-out.

Sub Club

Refuse any liquidation preference above 1X unless desperate. Preferred stock pays investors first in a downside sale, so a higher valuation doesn't guarantee a better founder outcome. On a $20M Series A, a $20M sale gives the investors all $20M.

A 2X or 3X preference can leave founders with little in a modest sale. "If you get a term sheet with anything but a 1X preference, do yourself a favor and don't take it unless you're desperate."

Sub Club▶ 39:09

Back low-churn products that improve for each user through usage alone. Consumer is investable when retention compounds, through network effects or deep personalization. AI makes personal data far more useful, so a product can be much better in two or three years because of usage rather than shipped features.

Health, nutrition and finance fit best, like a nutritionist who knows everything you eat over time. A plant identifier app turned a skeptic into a daily user with years of his plant data stored in it. Network effects are hard to create, and a utility often has to jumpstart the network.

Sub Club▶ 25:39

Venture fits viral products whose inference costs are high today. If a product has network effects, fast growth and addicted users but expensive inference, venture money can bridge the gap on the bet that costs fall while growth continues.

Timing matters: a cheaper model in two years versus five gives very different outcomes. RevenueCat estimated one LLM call per app per day would cost tens of thousands of dollars a month, and Opus pricing had not come down in six months. Unlike Facebook, whose costs rose with scale, an AI product pays inference costs even before it has users.

Sub Club▶ 33:30

Taste and judgment, not engineering, now cap how many great apps exist. Five years ago $20M could hire the best engineers, but without product intuition it was "$20 million on fire." Before product-market fit, more engineers don't speed you up because the work is hundreds of product and design decisions and talking to users.

Vibe coding multiplies leverage for people who already have good intuition; ATC, an air traffic control listening app, was built by a very small AI-assisted team. Quality also lowers acquisition cost, so a better product reaches scale more efficiently. Jacob Eiting first saw the taste argument as elitism, then changed his mind as the noise grew.

Sub Club▶ 4:30

AI can suggest ideas, but a human must pick one and hold a view. Asked for app ideas, ChatGPT 5.5 gave five, two with potential; choosing one and committing is the human's job. Labs deliberately trained models toward hyperneutrality for safety, and neutral answers won't win users at scale.

Models lack app taste because their feedback isn't curated for it: training on the whole app store is garbage in, garbage out, whereas Anthropic's model writes well because writers and psychologists guided its reinforcement learning. Maguire doubts one-shot AI app builders and prefers tools like portfolio company X1, which splits creation into onboarding, major features, brand and design so founders work through each decision.

Sub Club

Brand and opinion keep apps alive after features converge. Fitness apps like Ladder and Gravl differ in structure and opinion, and users like that opinion. Craigslist and Coca-Cola stay on top while many similar products exist.

Missing from the argument: the claim that apps won't be displaced by agents or self-built software rests on opinion and analogy, with no data.

Sub Club

Compare written one-star reviews to five-star ratings, not average stars. Star ratings are gamed by onboarding prompts that collect taps from people who don't care, so even weak apps sit near 4.9. Written reviews are harder to fake, making the ratio a better quality signal.

Maguire's own decade-old app sits at 4.5 stars, below many apps that look weak. Apple has recently cracked down on review farming.

Sub Club▶ 53:30