The state: strategy and programmes
| Indicator | Value | Source |
|---|---|---|
| AI products and services by 2030 | USD 1.5 billion | AI Development Strategy to 2030, Resolution PP-358 (in Russian) |
| Government e-services portal services based on AI by 2030 | up to 10% | PP-358 |
| AI research laboratories by 2030 | up to 20 | PP-358 |
| Government AI Readiness Index | target: top 50 | PP-358 |
| AI projects | more than 100; four data centres, two supercomputers, AI labs at 15 universities | Address to the Oliy Majlis, 26 Dec 2025 (in Russian) |
| Programme “Artificial intelligence — a partner for 10,000 enterprises” | 50% of AI implementation costs compensated; at least USD 100 million for the first stage | President’s press release, 18 Aug 2026 |
| Procedure for the 50% compensation | no regulation published as of 23 Sep 2026 | our check of lex.uz — programme explainer |
Law and fines
| Indicator | Value | Source |
|---|---|---|
| AI law ZRU-1115 | signed 21 Jan 2026, in force from publication | lex.uz |
| Legally significant decisions affecting a person’s rights and freedoms | may not rest solely on AI output | Art. 7¹, Law on Informatization (in Russian) |
| Fine for unlawful processing of personal data with AI | 50–100 base amounts: 22–44 million soum from 1 Sep 2026 | ZRU-1115, Art. 46²(2) of the Administrative Code |
| Base calculation amount from 1 September 2026 | 440,000 soum | Decree UP-115 (in Russian) |
| AI Ethics Rules No. 3787 | in force since 17 Jun 2026; the final decision is made by a person (para. 11) | lex.uz |
| Mandatory labelling of AI-generated content | not in the signed law; it was only in the draft | Gazeta.uz, 5 Nov 2025 (in Russian) |
| E-commerce contracts | may be concluded by a message (Art. 19); terms in the state language, other languages alongside it (Art. 21) | Law ZRU-792 of 29 Sep 2022 (in Russian) |
Market
| Indicator | Value | Source |
|---|---|---|
| Internet users | 33.1 million — 89.0% of the population; an estimate, mostly as of October 2025 | DataReportal, Digital 2026 |
| Mobile connections | 33.9 million | DataReportal, Digital 2026 |
| Share of cashless payments in retail and services by 2030 | target: 75% | Decree UP-246 of 10 Dec 2025 (in Russian) |
Lead response speed: what has been measured
| Indicator | Value | Source |
|---|---|---|
| Did not respond to a test inquiry (9,538 companies, 2013) | 47%; median time to first call 3 h 8 min | InsideSales, 2014 |
| Called within the first 5 minutes (same study) | 121 companies, about 1.3% | InsideSales, 2014 |
| Responded within the first 5 minutes (433 B2B companies, 2017) | 7%; 55% did not respond within five business days | Drift, 2017 |
| US car dealers with an ideal response to a web lead (3,290 dealers, 2026) | 51%; 14% never replied personally | Pied Piper via Business Wire |
| Unanswered calls to Russian developers’ sales teams (219,226 calls, Feb–May 2026) | 14%; missed callers were called back in 11% of cases | Yandex Real Estate as reported by Rambler (in Russian) |
| “Responding within 5 minutes makes you 100 times more likely to succeed” | from a 2007 InsideSales report, not MIT or Harvard | our analysis |
We found no open measurement of response speed for companies in Uzbekistan. The figures above are American and Russian.
AI pilots: what the research showed
| Indicator | Value | Source |
|---|---|---|
| “95% of organisations are getting zero return” from generative AI | preliminary MIT NANDA 2025 report: 52 interviews and 153 surveys | MIT NANDA, version 0.1 |
| Call centre with an AI assistant (5,179 agents) | +14% issues resolved per hour, +34% for novices | NBER working paper w31161 |
| Consultants on a task outside the model’s capabilities | 19 percentage points less likely to reach the right answer with AI | HBS Working Paper 24-013 |
| Experienced developers with AI (16 people, 246 tasks) | tasks took 19% longer, although they expected a 24% speed-up | METR, July 2025 |
| AI impact on EBIT (McKinsey survey 2026, 1,719 respondents) | 37% see at least some effect | The Register, 25 Aug 2026 |
What this means for your own pilot is covered in “95% of AI pilots fail”.
Infographic: what AI may not decide alone under ZRU-1115
How to cite
You can use these numbers without asking. Link to the primary source in the table or to this page: “AI in business in Uzbekistan: 2026 numbers and rules”, AI Capital Factory — https://aicapitalfactory.com/en/blog/ii-v-biznese-uzbekistana-cifry-2026/. Found an error or an outdated figure? Message us on Telegram — we will fix it and note the date.
This is not legal advice. We describe the rules as we read them for our own projects; details and caveats are in our explainers on ZRU-1115 and the 50% compensation programme.
How these rules play out in a project — what goes to the model, where the data lives, where a person keeps the decision — is on AI implementation for business.
If you want to find out which process in your company to hand to AI first, describe it in the questionnaire on the home page. We review one process for free within 48 hours.
