The next hundred billion dollars of software value
will not be built where the last hundred was.
Qanot is a first-cheque fund and venture studio for Central Asia. We write $150K into founders in Tashkent, Almaty and Bishkek, then build alongside them until the metrics are legible to a Series-A lead in Dubai or New York.
No warm intro required · Written rubric · Decision in 10 days
Eighty million people, and almost no venture capital.
Region scanCentral Asia holds roughly 80 million people with a median age under 30 — younger than India, younger than Brazil, and considerably younger than any market in Europe. Smartphone penetration crossed the majority years ago. Uzbekistan alone produces tens of thousands of engineering graduates a year, and pays them a fraction of Warsaw rates.
What it does not have is capital at the formation stage. A founder in Tashkent with a working prototype and early revenue has, in practice, nowhere domestic to raise a first institutional round. The gap is not talent and it is not demand. It is the absence of anyone whose job is to write the first cheque.
That is the entire thesis. Not that Central Asia will produce a hundred unicorns — that it has already produced two, with a formation-stage capital base close to zero. Fix the capital base and the base rate moves.
| Market | Population | Median age | Internet | $1B+ cos. | Qanot presence |
|---|---|---|---|---|---|
| Uzbekistan | 37.5M | 28.4 | 77% | 1 | Studio floor |
| Kazakhstan | 20.3M | 32.6 | 92% | 1 | Partner |
| Tajikistan | 10.4M | 23.1 | 39% | 0 | Scouting |
| Kyrgyzstan | 7.2M | 27.0 | 78% | 0 | Partner |
| Turkmenistan | 7.4M | 28.9 | 35% | 0 | — |
Figures are approximate 2024 estimates compiled for illustration. Source and date-stamp each row before launch.
Where billion-dollar companies actually come from.
Global datasetRoughly 1,400 companies have crossed a $1B private valuation. Two thirds of them sit in two countries. Understanding the shape of that distribution — how fast it forms, in which sectors, and how violently it responds to capital cycles — is the whole job of an early-stage investor.
Venture returns are not an average. They are a maximum.
ModelMost people assume a venture fund works like a diversified equity portfolio: spread risk, earn the mean. It does not. The returns of a venture fund are driven almost entirely by its single best outcome, and everything else is noise around that.
The consequence is counter-intuitive and it governs how we behave. We are not trying to avoid losses. A fund optimised to avoid losses will pass on exactly the companies that make it work, because those companies look reckless at the formation stage. We are trying to make sure we are on the register of the one company that becomes enormous.
That is why the cheque is small, the position count is high, the decision window is short, and the rubric rewards evidence of unusual slope rather than absence of risk.
What has to happen for this fund to work
Model assumes 60% of committed capital deployed as initial cheques, the remainder reserved for follow-on, and dilution already reflected in the ownership-at-exit input. Illustrative only — not a projection of returns.
What four partners score, independently, before they talk.
Published weightsMost funds will not tell you how they decide. We publish the rubric, the weights and the thresholds, because a founder who knows what is being measured can go and produce the evidence — and because a scoring system you can read is one you can hold us to.
| Dimension | Weight | Scored 0–10 on | What a 9 looks like |
|---|---|---|---|
| Evidence of demand | 30% | Paying customers, retention curve, unprompted inbound | Strangers pay, twice, without a discount |
| Founder slope | 25% | What you shipped in the last 90 days versus the 90 before | Visible acceleration, not a flat line of effort |
| Market shape | 15% | Domestic TAM, regulatory path, export optionality | Wins at home, ports to MENA without a rewrite |
| Unfair insight | 15% | Something true you know that the market does not | A non-obvious claim you can defend with data |
| Capital efficiency | 10% | Burn per unit of validated learning | Got to revenue on under $25K |
| Coachability | 5% | How you handled the hardest question in the interview | Updated your position live, with reasons |
Four stages. Ten days to a decision.
PipelineApply
A written rubric you finish in an evening. No deck, no warm intro, no NDA.
72 hoursScore
Four partners score blind against the published weights, then meet. You get the numbers either way.
10 daysBuild
Twelve weeks on the studio floor in Tashkent. Shared engineers, shared playbook, weekly revenue review.
12 weeksFly
$150K on the table, then introductions to the funds that lead your Series A in Dubai, Almaty and New York.
OngoingThe part that is not money.
OperatingEngineering bench
Four engineers and a designer on the studio payroll, assigned to portfolio companies for the twelve weeks where shipping speed decides whether there is a Series A at all.
Distribution desk
Warm routes into Central Asian enterprise and government buyers, and into the Series-A leads who price your next round. Introductions, not a directory.
Applied AI unit
Shared infrastructure for portfolio companies building on models — evaluation harnesses, inference cost tuning, and a first-party Uzbek speech and language stack.
Revenue review
One hour a week, same format, same metrics, every company. It is the single highest-signal thing we do and the one founders resist first.
Recruiting
A standing pipeline of regional engineering and sales candidates, pre-screened and scored, available to portfolio companies at no fee.
Corporate and legal
Delaware or ADGM incorporation, cap table hygiene, and the cross-border structure that lets a Tashkent company take money from a US fund without a six-month detour.
The deal, stated before you apply.
Standard- First cheque
- $150,000, on an uncapped SAFE with an MFN, or priced if the round is already priced by a lead we recognise.
- Target ownership
- 7–9% at the first institutional round, achieved through the initial cheque plus reserves — never by pressing for more at formation.
- Follow-on reserve
- Roughly 40% of the fund is reserved. We expect to follow into about a third of the portfolio.
- Board
- No board seat at first cheque. An observer seat only where the round's lead requests it.
- Pro-rata
- We take pro-rata rights and we use them. We do not take super pro-rata at the first cheque.
- Studio participation
- Optional and unpriced. Taking the cheque does not oblige you to sit on the studio floor, and sitting on the floor does not cost additional equity.
- What we will not ask for
- No exclusivity during diligence, no right of first refusal on your next round, no advisory shares, no fee of any kind charged to a portfolio company.
Before launch: have counsel confirm every line above against the fund's actual LPA and standard documents. Publishing terms is a commitment, and a mismatch between this page and the paper you send is worse than publishing nothing.
Companies on the register.
—Who scores your application.
Four voters[One line: the operating credential that makes this person's score worth something.]
[One line.]
[One line.]
[One line.]
Placeholder. Replace with the four real partners, each with a verifiable one-line credential and a link. On a fund site this section carries more diligence weight than the portfolio does — an anonymous partner page reads as a red flag.
Rolling. No warm intro required.
OpenSubmissions are read in the order they arrive. You will get a scored response inside ten business days whether we invest or not — and if we decline, you get the four scores and the dimension that held you back.
If you are pre-prototype, apply anyway and say so. The rubric handles it: Evidence of demand can be satisfied by a waiting list and a letter of intent as well as by revenue.