An MVP that cost 200,000 euros in 2022 costs 30,000 to 60,000 euros in 2026 if the team uses the modern AI tooling stack properly. This is not a marketing claim. It is the operating reality of every product company I run, and it is the most under-priced shift in the software economy right now. Company owners who internalize the new cost curve can fund three product bets where they used to fund one. This is the math, the mechanism, and the implication.
What the cost curve actually looks like
Take a representative project. A multi-tenant B2B SaaS MVP with three workflows, basic integrations, and a real production deploy. Six years ago this was a six- to nine-month build with a team of four engineers, a product manager, and a designer. Total cost ranged from 250,000 to 500,000 euros depending on geography.
The same MVP today, built by a strategist-builder using Claude Code, Cursor, n8n, and a modern AI stack, ships in eight to fourteen weeks with a team of one to three people. Total cost lands between 30,000 and 90,000 euros. The gross compression is roughly 4x to 8x on cost and 2x to 3x on time.
The compression is not uniform. Some kinds of work compressed dramatically. Some kinds of work barely moved. Founders who understand the difference allocate budget where it actually buys speed.
What got dramatically cheaper
Three categories of work collapsed in price.
Boilerplate and scaffolding. Authentication, CRUD, deployment configs, basic admin tooling, standard integrations. Work that used to consume the first month of a project now takes a day or two with modern AI coding tools. The effect is to push the start of real product work from week 4 to week 1.
First-pass content and copy. Marketing copy, in-app text, user-facing language, basic visual layouts. Work that used to need a content writer and a designer at the start can now be drafted by a senior product person with AI assistance and refined by specialists later. The effect is to allow the early stages to run with fewer specialist roles.
Repetitive engineering. Form handling, validation, simple state management, minor refactors, common bug patterns. The 30 to 50 percent of an engineer's time that used to disappear into routine work now compresses to 5 to 15 percent. The engineer is doing the harder thinking that they were always meant to do.
What barely moved
Three categories of work cost roughly what they did before.
Discovery and customer understanding. Talking to customers, mapping workflows, sitting in on real work, defining the product. AI does not compress this and probably never will. It is the human work of figuring out what to build.
Architecture decisions on hard systems. Multi-tenant data architecture, security models, integration design with legacy systems, performance under scale. AI helps at the implementation layer. The decision layer is still human, and skipping it produces fragile systems.
Customer adoption and change management. Getting real users to actually use the product. Training. Documentation. Trust. The product is built faster, but the customer's calendar has not gotten less crowded. Change management is paid in patience, and AI does not move that line.
The mechanism behind the compression
The naive explanation is that AI writes code faster. That is part of it. The bigger part is that AI removes coordination overhead. The traditional five-person team spent 30 to 50 percent of its time coordinating among themselves. Meetings, handoffs, code reviews, design reviews, scope debates. A team of one or two builders does not have most of that overhead.
The risk in stating this is making it sound like teams are bad. They are not bad. They are necessary at scale. But for the first 12 to 18 months of a product, the throughput of an integrated mind beats the coordinated throughput of a small team. The math is not subtle. It is the difference between a project that ships in three months and a project that ships in nine.
The implication for company owners
The most important consequence is portfolio. The cost of a single product bet has fallen enough that a company that previously could afford one bet at a time can now afford three. Three bets in parallel produce three real signals. One bet at a time produces a single signal that is heavily contaminated by noise. The strategic value of running three bets and killing two is enormous.
The companies that win the next decade will run more bets, kill more bets, and find their breakthroughs faster. The companies that staff one bet at a time the way they always have will fall behind not because their bets are worse but because they do not get to run enough of them.
Why most companies have not adjusted
Three reasons.
The hiring template lags. Most product organizations are still organized around a 2022 staffing template. PM, designer, four engineers per product line. When a senior person says they can run three bets with three strategist-builders, the org cannot find the boxes on the chart for that.
The procurement and budget templates lag. Budget approval cycles assume the costs of three years ago. Founders and CFOs hear 30,000 euros for an MVP and reflexively assume the team is cutting corners. They are not cutting corners. The cost curve moved.
The risk perception lags. Decision-makers who have lived through software projects associate small budgets with high risk. The risk model needs updating. A small budget executed by a strategist-builder using modern tooling is lower risk than a large budget executed by a 2022 team, because the iteration speed catches problems faster.
What to do about it this quarter
Three concrete moves.
Run one product bet at the new cost curve. Pick a small, contained product idea. Staff it with one strategist-builder, full autonomy, eight to twelve week budget. See what happens. The signal will tell you more than any spreadsheet.
Map your portfolio against the new cost curve. Look at the products you would build if each one cost a third of what you assume. Which projects cross the threshold from too expensive to fundable. That list is your strategy for the next 18 months.
Adjust your hiring template. Add the strategist-builder role. Not as a replacement for traditional PMs and engineers, but as a category that runs the early stages of new product bets and hands off to specialist teams when scale arrives.
The window
The companies that adopt the new cost curve in 2026 will outpace the companies that adopt it in 2027. The advantage compounds because each bet that ships earlier produces signal that informs the next bet. Founders and operators who delay are not preserving cost discipline. They are foregoing optionality.
If you are a company owner thinking about how to use the new cost curve, write to me. I respond within 48 hours.