MMOOZZGG.com

AI investment

How to raise money for an AI startup: who pays and what they check

5 min read2 sources

How to raise money for an AI startup in Russia depends on whom you approach: private funds gave 67% of venture money in the first half of 2026. Below are who gives the money, what investors check before a deal, which metrics they ask for and how to prepare for a round.

On this page · 4 sections

Who gives money

The Russian venture market in the first half of 2026 was $55 million, about 48% less than a year earlier, according to the Moscow Venture Fund. There were 54 deals with 33 investors.

AI is still in first place: $32 million and 33 deals in 9 months of 2026, by preliminary Venture Guide data. There is less money on the market, but AI gets more of it than other sectors.

Who gives money, by Moscow Venture Fund data for the half-year:

1. Private funds: 67%, or $37 million. They want growth and a clear exit from their stake in a few years.

2. Business angels: 16%. They come in at an early stage with smaller sums, and the decision often rests with one person.

3. Corporations: 9%. They look for technology for their own business and can become your first large customer.

4. State funds: 8%. They expect reports under the program terms and limit what the money can be spent on.

How to choose. If the product already works at customers and you need money for growth, go to funds. If the product solves a task of a specific industry, a corporation from that industry can give both money and a first large order. At the idea and first pilot stage, your investor is more often a business angel.

The choice of investor sets everything else: what will be checked, which metrics they will ask for and how long the deal will take. The largest deal of the first half-year, almost $13 million into ArchiTech AI, came from a corporate investor.

What investors check

Before a deal the investor runs due diligence, a comprehensive check of the company. For an AI startup it is wider than usual: besides finance and legal questions, investors look at whether the AI works on real data.

What they check in an AI company:

1. Product: whether the solution works on customer data, not only in a demo.

2. Data: where it comes from, whether there are rights to use it, and whether 152-FZ, Russia’s personal data law, is followed.

3. Dependence on other companies’ models: what happens to the product if the model vendor changes prices or closes access.

4. Team: which founder holds the key technology and whether that person will stay after the deal.

5. Revenue: whether customers pay and for what exactly, and whether they renew contracts.

The larger the sum, the deeper the check. A C+ round of $13 million, like ArchiTech AI’s, goes through a very different amount of checking than an early-stage angel round.

A common reason for a delayed deal: the investor finds something in the data or contracts that was not in the pitch deck. Check these five points yourself before the first meeting.

AI startup metrics

An investor looks at metrics that show the AI brings in money, not only that it works:

1. Revenue and its growth by month, the share of repeat payments.

2. The cost of serving one customer, including payment for models and computing. For an AI product it grows with the number of requests, and the investor will check whether the margin holds.

3. Accuracy on customer data: the share of answers or decisions that need no correction by a person.

4. Time to result for the customer: how many weeks after connection the customer sees an effect in money.

5. Retention: how many customers renewed their contract after a year.

Competition for money is high: AI had 33 deals in 9 months of 2026, and every investor compares you with other AI teams on the same metrics.

For an AI product, prepare a separate answer about data: what data the solution is tuned on and who owns it. A corporate investor asks this question among the first.

An investor treats metrics without numbers as a promise. Show them for recent months even if the sums are small: honest small growth convinces more than a forecast.

How to prepare for a round

Preparation takes months, so start before you need the money urgently. The order:

If you need a round in six months, start collecting documents now: the check in a large deal takes weeks, and fixing the problems found takes even longer.

1. Choose the type of investor from the list above: a fund, an angel, a corporation or a state fund. The document package depends on it.

2. Collect data for the check: financial statements, contracts with customers, rights to data and code, a description of dependencies on models.

3. Calculate the metrics from the section above on real numbers, and prepare an explanation of each.

4. Run the check yourself before the investor does: find the weak points before the investor finds them.

5. Prepare a term sheet with the main deal terms: the amount, the stake, the pre-money valuation and the investor’s rights.

6. Decide in advance what stake you are ready to give and which investor rights you will not accept.

The market is narrow: in the first half of 2026 it had 33 investors. Every conversation with a fund matters, and you may not get a second try with the same investor.

See who has already raised money and on what terms in the list of Russian AI startups.

Sources and fact-checking

Checked by the editors on 14 October 2026.

  1. Kommersant, citing the Moscow Venture Fund, 13 August 2026 (kommersant.ru)
  2. ComNews, citing Venture Guide (Moscow Venture Fund), 5 October 2026 (comnews.ru)

Found a mistake? Tell us

Publications

Mike Mozg

Agentic AI use cases at AT&T: nearly 500 processes now run on agents

AT&T no longer uses AI only to help staff work faster: it is rebuilding whole business processes around agents. SiliconANGLE reported on 14 November 2026 that nearly 500 agentic AI use cases at AT&T are already in production. Here is how the company rebuilds them, where a human still checks the agent, and what other large companies can take from it.

News4 min
Contact us