For years, the founders sitting across the table from us had a fairly predictable set of worries. Could they actually build the thing? Could they find a developer they could trust? Could they afford the months of design and engineering standing between an idea and something real?
That conversation has quietly moved on. Most weeks now, founders come to us with a working prototype they stitched together over a weekend using an AI coding tool. The question isn’t “How do I build this?” anymore. It’s “Why isn’t anyone using it?”
That shift tells you something important. The bottleneck has moved. Building the product used to be the hard part. AI has made that dramatically easier.
Now the bigger challenge is everything that comes after: finding the right customers, positioning the product in a way that makes people care and building a business that investors actually want to back.
Building has never been this accessible
A custom mobile app used to run a founder anywhere from fifty thousand to half a million dollars, depending on scope and who you hired to build it. That’s still roughly what a traditional development shop will quote you, though what AI app development actually costs in 2026 looks very different now. A non-technical founder can describe a product in plain English and have a simple prototype working within a day or two, for the cost of a monthly subscription rather than a developer’s salary.
We see the upside of that every week. Founders who once needed a technical co-founder just to test an idea can now build a rough version themselves and see if it has legs before committing serious time or money. That’s a genuinely positive shift for founders.
What it also means is that building is no longer where a startup earns its edge. Everyone has access to roughly the same tools now. A founder in Melbourne and a founder in Manila are both one prompt away from a passable version of almost anything. We’ve watched the Indie Hackers community wrestle openly with this over the past year, with more than a few founders admitting their AI tooling bill has crept past a hundred dollars a month while their product still has zero paying customers. It’s become common enough on that forum to be something of a running joke, and also a fairly honest signal of where the real cost has moved to.
The data backs up what founders are already noticing
There are numbers behind it too. The Financial Times’ chief data reporter, John Burn-Murdoch, published an analysis this year drawing on research into productivity across generations of AI coding tools, and the pattern is stark. Since agentic coding tools became mainstream, monthly iOS app releases have climbed to somewhere around eighty per cent above their 2024 baseline. Over that same stretch, app store reviews have fallen, and the share of apps that go on to achieve any meaningful ongoing usage has slipped as well, with some estimates putting it at roughly three-quarters of where it started.
Supply has exploded. Demand hasn’t followed it anywhere close. The analysis has been picked up widely by economists and technology commentators since it was published, largely because it puts a number on something a lot of people in this industry had already sensed: that shipping more software is not the same thing as creating more value, and a flood of new apps landing in the App Store each month says very little about whether anyone downstream actually wants them.
When cheap building outruns everything else: the Medvi story
The clearest illustration of both sides of this landed in the news earlier this year. Matthew Gallagher launched Medvi, a telehealth business selling GLP-1 weight loss prescriptions, out of his Los Angeles home in September 2024 with twenty thousand dollars and no employees beyond his brother, Elliot. The pair used a stack of AI tools, including ChatGPT, Claude and Grok, to handle the company’s website, marketing copy, customer service and analytics, while outsourcing the licensed clinical and pharmacy work to partners, CareValidate and OpenLoop Health, who carry the regulatory responsibility that Gallagher’s two-person team never could.
The New York Times profiled the company in April, verifying financials that showed Medvi generated 401 million dollars in sales in its first full year, on roughly 250,000 customers and a 16.2 per cent net margin, a figure that made national headlines because it beat the margins of an established, well-staffed competitor several times its size. Forbes, PYMNTS and several other outlets followed with their own coverage of the same numbers. By most measures, it’s the most concrete real-world example yet of the “one-person billion-dollar company” that AI leaders have been predicting for a couple of years.
It’s also a genuinely useful cautionary tale. Within months, the company had received a formal warning letter from the FDA over misleading product claims, and OpenLoop Health, one of its clinical partners, disclosed a data breach affecting patient records. None of that erases what Gallagher built. It does show what happens when growth this fast outpaces the compliance, governance and risk foundations a business needs to hold up under scrutiny, whether that scrutiny comes from a regulator, an investor doing due diligence, or a customer who was given the wrong answer about their own medication. Cheap building and clever structuring of who carries the risk can produce extraordinary revenue very quickly. Whether it survives the following eighteen months tends to depend on the parts of the business nobody was optimising for tokens.
When positioning is done properly: the other kind of story
The clearest counter-example we can point to is one of our own founders, rather than someone else’s headline. Thomas McCarthy came to us with deep domain expertise in Hollywood VFX and crypto, and a genuinely ambitious technical vision: a canonical ownership layer for media, with automatic payment routing back to rights holders. The idea was sound. What it lacked was a way to explain itself to anyone who wasn’t already a specialist.
By his own account, the work that mattered most wasn’t the build. It was turning a complex technical vision into something a non-technical investor or creator could understand in one conversation, and structuring the thesis so it could be validated and tested rather than taken on faith. That’s the part we spent the most time on with him: sharpening the product narrative, and working out how to position something genuinely complex in simple, compelling terms.
The startup, JubJub, went on to secure venture funding from Antler, deploy smart contracts on Base, and complete the first end-to-end streaming payment between a viewer and a rights holder, picking up guidance from Disney executives along the way. The narrative and positioning work sitting underneath the technology is what turned it into something investors were willing to fund.
Distribution is becoming the real differentiator, and it isn’t free either
Across the venture and go-to-market commentary we’ve been reading this year, a similar view keeps surfacing: as features become trivial to copy and AI compresses build time down to almost nothing, brand, audience and distribution have become the more durable form of advantage, because they’re the parts a competitor genuinely can’t clone over a weekend. We agree with the substance of that, with one important caveat. Distribution is not staying cheap either. One industry analysis published this year put the rise in customer acquisition costs across B2B SaaS at more than 220 per cent over the past eight years, and estimated the median SaaS company now spends close to two dollars in acquisition cost for every new dollar of recurring revenue it brings in. That’s part of why we spend so much time with founders on how to build traction without relying on a bigger ad budget as the fix.
So the honest picture for a founder in 2026 looks like this. Building got cheap. Getting attention got more expensive, and it’s getting more crowded by the week as every other founder reaches for the same AI tools and, often, the same three marketing channels to promote what they’ve built. Securing a slice of someone’s attention and turning it into paying, retained customers has become the scarce skill of 2026.
Where this leaves the value we provide founders
This is the part of our own thinking that’s shifted the most this year, and it’s worth being direct about it. The value in working with a studio like ours has always included helping someone work out what to build in the first place, and stress-testing whether that’s a product or an actual business, well before real money gets spent on it or a single line of code gets written. What’s changed is how much earlier and how much more seriously the positioning and distribution conversation now needs to happen, because a product with no credible path to customers isn’t fundable, no matter how well it was engineered.
That’s why our Accelerate Process treats go-to-market planning, early adopter targeting and traction building as part of the core product and business work, running in parallel with defining what gets built rather than being bolted on once development is finished. It’s also why Launch Ready, our AI-assisted development process, is built around delivering a landing page designed specifically to support early traction alongside the live product itself, with experienced developers overseeing the AI-generated work rather than shipping it unreviewed. AI App Development is one of the areas we spend the most time in with founders precisely because the speed on offer is real, and because speed without a plan for who’s going to use the thing is how you end up as one more data point in the pattern the Financial Times documented, another release that added to the supply without adding to the demand.
What this actually looks like in practice
A few things we’d genuinely encourage any founder building in this environment to do differently this year:
- Test your distribution assumptions before or alongside building, not after. If you can’t describe, in one sentence, where your first hundred customers are actually going to come from, that’s worth solving before you spend another dollar on development.
- Start building an audience, or at minimum a warm list of early adopters, well before launch day. A pre-launch audience of a few hundred genuinely interested people will outperform a cold launch to nobody, every time.
- Treat positioning as a real deliverable with its own budget of time and attention, not marketing copy you write the week before you go live.
- Remember that AI tools give you execution speed. They don’t give you judgement about what’s worth building or who actually wants it. That part still needs to come from customer conversations, market research and, ideally, people who’ve been through this build cycle before.
- Keep your compliance and risk foundations proportionate to what you’re actually promising customers, especially if you’re anywhere near health, finance or anything else regulated. Speed is only an advantage when the foundations underneath it can actually hold.
If you want a practical starting point for turning that list into a working system rather than a set of good intentions, we’ve laid out a simple marketing engine that works for early-stage founders.
Where founders go from here
Building fast remains a smart instinct, and AI-assisted development deserves its place in every founder’s toolkit. It’s a genuine gift to founders who previously needed two hundred thousand dollars and a technical co-founder before they could test whether an idea had legs. What’s changed is which part of the job actually determines whether a startup makes it through its first eighteen months. For most of the founders we work with, that part was always going to need support from people who’ve done it before. It has simply become a lot more obvious this year that it was the part that mattered most all along.
If you’re sitting on a working prototype and wondering why the market’s gone quiet, that’s usually a positioning and distribution conversation, not a technical one, and it’s worth having before you spend more on ads or more on tokens. Book a free startup strategy session with our team, and we’ll tell you plainly whether the gap is the product, the positioning, or the plan to reach the people who’d actually pay for it.
Sasha Reid is the founder and CEO of Hyper, an Australian startup studio helping founders validate, launch and grow startups. Since 2015, he has worked with over 1,500 founders across AI, SaaS, fintech and consumer tech.
