A thesis has been spreading recently. Now one founder with a laptop and a subscription to AI tools can do what a team of dozens of people used to do.
The share of startups with one founder has indeed increased. According to estimates, from about a quarter in 2019 to more than a third by the beginning of 2025. But here it is worth asking another question. Is the total cost of launching a product really decreasing, or is it simply postponed to a later point, when the prototype turns into a real business?
What exactly has changed and what has not
Illustrative examples of recent years are stories of projects where one developer brought an idea to a working product in a matter of months. One such case is an AI app development platform that grew to hundreds of thousands of users in six months and ended up with a multi-million dollar acquisition deal with less than a dozen people on the team.
Another example is an infrastructure service for developers that grew annual revenue to hundreds of millions of dollars on the wave of AI coding.
In both cases, the threshold for entry has really changed from idea to first working prototype. What used to require months of approvals, budgets, and hiring can now be done by one person in weeks.
Where costs don’t disappear
The problem arises in the next step. Mykhailo Zborovskyi, who has experience in developing iGaming products, draws attention to this very moment in his recent column, that putting together a prototype on your own is one level of complexity, but taking the product through licensing, compliance, payment security and 24/7 support is a fundamentally different one, and no AI tool has yet taken on this part of the work.
Simply put
AI does remove the barrier to entry, but it does not remove the barrier to scaling. The costs that previously went to the development team do not disappear, they simply move to the stage when the product already has users and requires legal, financial and operational infrastructure.
According to the expert, it is at this transition that most solo founders in gambling overestimate their own strengths. A prototype that looks good on a demo is not yet a product ready to work with real user money under the supervision of a regulator. Mykhailo Zborovskyi notes that AI greatly reduces the time from idea to first version, but does not reduce the time required to pass the inspection of a licensing authority or payment partner at all.
Another point that the columnist draws attention to is the illusion of a zero team.
Formally, a company may have one founder, but in fact a network of contractors, consultants, and service providers already appears around him, who are simply not shown on the organizational chart. The costs for them do not disappear anywhere, they just cease to look like full-time employees.
AI reduces the cost of a mistake at the idea stage, but does not reduce the cost of a mistake at the real business stage. Mykhailo Zborovskyi
According to him, it is precisely the lack of understanding of this difference, and not the lack of technical skills, that most often becomes the reason why one-person AI projects in gambling stop at the working prototype stage and do not grow into a stable operator.
What AI really replaces and what it doesn’t
AI is good at:
- Rapid prototyping of game mechanics or interfaces;
- Processing large data sets for odds analytics;
- Generating base code and visual elements for a specific request;
- Routine user support tasks at the entry level.
AI does not yet replace:
- Licensing and compliance with the requirements of specific jurisdictions;
- Payment system security on a real transaction scale;
- 24/7 support for complex cases and disputes;
- Legal responsibility for product decisions before the regulator.
Previously, to test new game mechanics, a small team was needed. A person to calculate the RTP, a backend developer, an artist for visual elements. Today, specialized AI platforms allow one founder to get working code and graphics for a specific request literally in a week, which really changes the startup budget at the start.
But when it comes to a full-fledged operator, and not a demo version of the product, another level of complexity appears. The operating system, compliance, legal requirements, and this is where, according to the columnist, a solo project hits a limit that one person physically cannot overcome on their own.
AI is truly changing the economics of the early stage of a startup. It is now much faster and cheaper to test a hypothesis and collect the first users than it was five years ago. But this does not mean that the cost of building a full-fledged business has disappeared, it has shifted to the stage of scaling, compliance and operational stability. Mykhailo Zborovskyi
Scaling checklist: where the deferred bill arrives
A prototype proves demand. An operator proves reliability. The gap between them is mostly security and process work.
- Security baseline. According to ISO/IEC 27001 and the NIST Cybersecurity Framework, risk management must be documented, repeatable and owned by named people — not by a single tool.
- Payment and app layer. OWASP recommendations on web and API security are a sensible starting point before a payment partner audit.
- Operational resilience. ENISA and CISA guidance stresses incident response and continuity planning — exactly the 24/7 part that AI does not cover yet.
Why you can trust this analysis
We cover AI, product economics and iGaming from a practitioner angle. Expert opinions are attributed by name. Figures are linked to their original source.
How we verify information
Every statistic is checked against the primary report. Security claims are cross-checked with public frameworks such as NIST, OWASP and ISO/IEC 27001, and with industry write-ups like the Cloudflare Blog and Microsoft Security Blog.
Editorial principles
No paid conclusions. Clear separation of quotes and analysis. Corrections are made openly. This material is not legal or financial advice — licensing rules differ by jurisdiction.
FAQ
Does AI make launching an iGaming startup cheaper?
At the prototype stage — yes, noticeably. Licensing, compliance, payment security and round-the-clock support still cost money; they just arrive later.
Can one founder run a licensed operator?
One founder can build a demo. A licensed operator relies on lawyers, compliance officers, payment partners and support teams — often as contractors rather than staff.