Antbuilder is a venture studio. We don't back companies — we create them, in-house. Ten ventures in motion and five already live in 85+ countries — eight of them launched on $23,000 of total capital, by one operator. That efficiency is the whole thesis.
Flagship Pioneering created Moderna. Sutter Hill created Snowflake. Studios don't pick winners — they manufacture them. Here is ours, and what it costs to run.
A real problem with existing demand, reachable without permission.
Live product, payments, analytics. No pilots, no decks — real users, real money.
Every step measured. We kill what doesn't convert instead of arguing about it.
Each venture gets a dedicated owner. The studio keeps 40–60% and moves on.
The old constraint on this model was engineering capacity, and it made studios expensive. That constraint is gone. What money buys us now is operators, not engineers — and every operator we add is another venture running in parallel.
This isn't a forecast. It's the only reason one person in Buenos Aires has five products live in 85+ countries.
Intelligence became a commodity and the bottleneck moved. It is no longer engineering capacity — it is taste, speed, and distribution. Every incumbent still carries the old cost structure. We never had one.
Every venture above is one more grain on the pile.
Every generational portfolio was assembled in a narrow window when the cost of building fell and the incumbents hadn't noticed yet — microcomputers in 1977, the web in 1995, mobile in 2008. We are inside the next one right now.
In each case, the returns went to whoever was already compounding when the window opened — not to whoever wrote the best memo about it afterwards. We started in January and never stopped shipping.
The last letter of the alphabet. The end of the line. There's no generation after us to hand the problem to.
And the world you've handed us is, frankly, insane — stalled, financialized, and rotting at the institutions. We didn't break it. But we're done waiting for someone else to fix it. So we will — the only way anything ever gets fixed. By building.
And for a while, it was arriving. Then, somewhere in the 1970s, we stopped building it.
In the 1960s the world moved forward. We put people on the Moon. We built rockets, reactors, supersonic flight — real things, at the edge of the possible, and the progress was shared.


And then something broke. You can argue about the cause — the end of Bretton Woods in 1971, the oil shocks, globalisation, the end of cheap energy — but you cannot argue about the shape. For twenty-five years, what workers produced and what workers were paid rose together, almost perfectly. In the 1970s the two lines separated, and they never came back together.
Indexed to 100 in 1948. Source: Economic Policy Institute productivity–pay tracker. The shaded area is the gap that never closed.
From 1948 to 1973, output per hour and the pay of a typical worker both roughly doubled. Build more, earn more. That was the deal, and it held for a generation.
Since then productivity has more than tripled. Pay has barely moved. The shaded area is the part of what we built that stopped reaching the people who built it.
The same decade shows the same fracture almost everywhere you look: nuclear plants stopped getting finished, aircraft stopped getting faster, infrastructure cost per kilometre began climbing and never stopped. Whatever happened in the 1970s, it did not happen to one industry.
41 reactors started in 1973. The last U.S. order was placed in 1978. Then, for forty years, essentially none.
Concorde flew at 1,350 mph in 1976. It was retired in 2003, and we went back to the speed of a 1958 Boeing 707. Travel got slower.
Building the same mile of road got roughly five times more expensive — without getting five times better.
Real data, 70 years of it. Productivity grew 1.09% a year from 1954 to 1973, then 0.48% for the fifty years since — less than half. Had the old pace held, we would be 36% richer in productive capacity than we are. The shaded wedge is 28 years of progress that never happened.
Sources: U.S. Energy Information Administration (reactor construction starts); manufacturer cruise-speed specifications; Brooks & Liscow, Infrastructure Costs (2023); and Penn World Table via FRED series RTFPNAUSA632NRUG (total factor productivity, 1954–2023) — downloaded and computed directly, not quoted. Break dates are the conventional ones and are debated — the shapes are not.
Then, quietly, it financialized. Ambition migrated from the launchpad to the spreadsheet, and a growing share of the entire economy went to simply moving money around. You can watch the moment it took over:
Source: Philippon (2008). The economy tilted from building things to pricing them.
Money replaces honour and adventure as the objective of the best young men… The object of the young and the ambitious is no longer fame, honour or service, but cash. No longer do schools aim at producing brave patriots ready to serve their country — parents and students alike seek the qualifications which command the highest salaries. — Sir John Glubb, The Fate of Empires
And so, one percent at a time, we stopped building things.
We refuse this. Antbuilder is a revolt — against a world that stopped building, and against the corruption hollowing out our institutions. The answer isn't nostalgia. It's to build again.
Each tier rests on the one below it. We start at the wide base and build toward the summit.
The summit. Acquire the land, charter the institutions, build a land of knowledge.
A fund and institutions that point real talent and real money at real problems — and see them through.
Physical products. Slower, more expensive, harder to fake — and far harder for others to copy.
Apps and AI tools, shipped to real users across 85+ countries. The training ground — where we learn to finish and to sell.
Venture returns follow a power law: one outcome pays for everything else. So we don't place one bet. We run a portfolio, build it in-house, and hold every venture to a billion-dollar ceiling.
We need one. And we have ten shots, one operator who ships daily, a burn most startups spend on catering, and a structure where every venture makes the next one cheaper to build. Nothing here is a slide — five of them already take money from real people in 85+ countries.
No single investor has to believe in all ten. The studio raises operating capital; each venture raises its own round from its own specialists.
You're underwriting the creation engine and its hit rate — not one product. Capital converts directly into operators, and each operator is another venture running in parallel.
At $4,600 per launch, $500k funds six new ventures and four operators for two years — while the five live ones keep running.
Fit: family offices, permanent capital, operators who've built studios.
Personalised neoantigen cancer vaccines, manufactured in Argentina. Merck/Moderna's Phase 3 read out positive in melanoma on 19 August 2026 — the science is now de-risked. The open problem is cost: ~$45k per patient, and manufacturing is the stated bottleneck.
Swiss-standard science on an Argentine cost base. If GMP lands at a third of Basel's cost, the same trial budget treats 3× the patients.
Fit: biotech specialists, cancer foundations, mission capital, families touched by melanoma.
Organic distribution is solved. Monetisation is not — yet.
Where it goes: most of it buys reach — creator-led UGC in the US and LatAm, the channel that already sends us traffic for free. A first slice funds offer testing against the audience we have, and the ad budget stays locked until checkout clears our conversion threshold. We scale a working funnel or we don't scale at all. Full funnel data shared under NDA.
Live on the App Store. Subscription mechanics already built.
The math: at $12/mo and 6-month average retention, LTV is $72. That makes any channel with CAC under $24 profitable at 3×. $80k funds ASO, creator seeding and retention work to find one such channel.
Self-funding. This is the studio's operating floor.
The math: 3 clients × 8h/week × $200/hr = $6,400/mo. Tromen Peak alone covers the studio’s running costs — it needs no capital, only calendar.
A marketplace where young people get paid to help older people with technology.
Where it goes: marketplaces are won by solving one side first. The budget buys supply — recruiting and vetting young helpers in one city — then demand through the channel older families actually trust: their own children. One operator, one city, one repeatable playbook before we open a second.
One kitchen per city. Every meal you eat, built to your macros, for a flat monthly price.
Where it goes: the sector's graveyard died of one thing — cooking to guess demand, and throwing away 16% of everything made. A subscription cuts that to ~2%, which in this business is the entire margin. The budget buys a founding cohort in one city through contracted kitchen capacity, not capex. Own facility only after retention proves out. Autonomous delivery at ~$1 a trip is what eventually makes daily fresh food cheaper than cooking it yourself.
A hedge fund built on one thesis the West keeps mispricing: China's growth is not over.
Where it goes: anchor capital to open the book and build a track record — data, execution, compliance and the first audited years. Mispricings this large only exist while the story is unpopular, which is exactly the window. Strategy memo on request.
Personalised neoantigen cancer vaccines, designed and manufactured in Argentina.
Where it goes: $220k to a co-founder scientist and two researchers; $80k to contracted preclinical validation — we rent the wet lab rather than build one; $40k to patent the selection method before publishing; $30k to compute. No facility, no capex, no permits until the data exists. Swiss-standard science on an Argentine cost base — if GMP lands at a third of Basel’s cost, the same trial budget treats three times the patients.
Every country is badly run. We would like to build one that isn't.
Where it goes: nowhere, yet. This is the summit of the pyramid and it is funded by everything below it. We name it here because it is the honest reason the rest exists — and because the people we want beside us are the ones who read this and lean in rather than laugh.
The hard part is done. Five products are live in 85+ countries, the funnels are built end to end, and thousands of people arrive every week through channels that cost us nothing. What we have never had is the budget to finish the job.
Monetisation is a tuning problem: offer testing at volume, pricing ladders, and paid reach large enough to read a clean signal. Those are the three things money buys and effort cannot. Every ad budget on this page is staged behind a conversion threshold — we scale what proves out, and we cut what doesn't. Give us the reach and we will finish the funnel.
Five live products in 85+ countries, built for $4,600 apiece by one person. That was the hard part and it is already done. What capital buys now is operators — and every operator is another venture running in parallel. Back the engine, or back the single venture you care about most.
We spent this year proving we can build in the world of bits — five products, 85+ countries, eight launches on twenty-three thousand dollars. We are done with bits. Software was the training ground; it taught us to finish, to sell, and to move fast on almost nothing. The capital we are raising goes into the world of atoms — kitchens, robots, laboratories, land. That is where the problems that actually matter still sit unsolved, and it is the only part of the world that stopped getting better.
Deck & numbers on request — hello@antbuilder.com. We reply the same day.