The $4M Seed Round You May Not Need Anymore
The $3 million to $4 million seed round was largely a budget to hire about eight engineers to build an MVP. If founders with AI coding tools can build that MVP themselves, the round should be rebuilt from milestones instead: a few key hires, inference costs, go-to-market spend and a buffer, which often totals far less.
The claim behind the $4M figure
A 22-second clip posted on November 15, 2026 by the Blunt Truths YouTube channel, cut from Press On Ventures' First Press show, carries the line. Its description attributes it to Willie Litvack, a founder of the Los Angeles early-stage firm: founders used to raise $3 million to $4 million in a seed round to hire eight engineers to build their MVP, and you do not need eight engineers to build your MVP anymore. So ask why you are raising at all, and do not raise unless you are certain the business is venture scale.
Whether to raise is its own question. This page takes the narrower one the clip implies: if the classic $4 million seed was a headcount budget, and the headcount is no longer needed, what should the round look like instead?
What a $4M seed used to buy
Take an illustrative budget, not a market benchmark. Suppose an engineer costs about $200,000 a year fully loaded, with salary, benefits, payroll taxes, equipment and recruiting. Eight of them cost about $1.6 million a year. Add two founders on modest salaries, an office, software, legal and accounting, and a small go-to-market spend, and the burn lands near $2 million a year. A $4 million round then buys roughly two years of runway, which was the usual target for reaching Series A metrics.
The structure matters more than the numbers. In that model, about three quarters of the round goes to people who build the first product. The round size follows directly from the hiring plan, and the hiring plan follows from the assumption that software takes a team to write.
What changed, and what did not
The assumption is what changed. Coding agents and AI development tools let a technical founder, or two, build and ship a working MVP that once needed several engineers. That removes the largest line in the old budget before the product exists.
It does not remove every cost, and it adds a new one. Products built on large language models pay for inference on every request, so a meaningful part of what used to be fixed engineering salary becomes a variable cost that scales with usage. Distribution has not become cheaper either: someone still has to find, sell to and support customers. And once the product works, you still hire; the question is when, not whether.
The market data points the same way. Carta's Q3 2026 data, as summarized in November 2026, shows 426 seed rounds on its platform, down 27% year over year, while dollars invested at seed rose 21% to $4.1 billion. Fewer companies are raising seed, and money is concentrating in larger rounds, many of them AI companies with heavy compute needs. A typical software company is not obliged to follow either trend.
Rebuilding the round from the budget
Start from milestones, not from a round size. Write down what has to be true at the next raise, or at profitability: revenue level, growth rate, number of paying customers, product scope. Then cost the shortest credible path there.
People. Founders, plus only the hires that agents and contractors cannot cover in the next 12 to 18 months. For many software companies that is one to three people, often in sales or customer success rather than engineering.
Inference and infrastructure. Model usage at your projected volume, with a buffer for price changes when you switch or upgrade models.
Go-to-market. The spend required to reach the customer milestone, tested in small amounts before it is scaled.
Fixed costs and buffer. Legal, accounting, insurance, tools, and six months of contingency.
The result is often well under $4 million. That smaller number means less dilution, a lower valuation bar at the next round, and more freedom to stay profitable instead of raising again.
When the $4M round is still the right size
Some companies still need it, or more. Hardware and robotics companies need prototypes, certification and manufacturing runs. Regulated businesses in health, finance or insurance need compliance staff and licences before revenue. Companies training their own models need compute. Marketplaces and sales-led enterprise products may need to buy distribution quickly in markets where the first strong player takes most of the value.
The risk on the other side is undercapitalization. A round sized for 12 months that takes 18 leaves you raising from weakness. Size for your real timeline plus a buffer, and if the budget genuinely needs $4 million, raise it; the point is that headcount for an MVP is no longer, by itself, a reason to.
Frequently asked questions
- How big should a seed round be in 2026?
- Big enough to reach the milestones for your next raise or for profitability, plus a buffer of about six months. Build it from a line-item budget rather than a market median. For a software company whose MVP is built by founders using AI tools, that number is often well below the $3 million to $4 million once considered standard.
- Can AI coding agents really replace engineers for an MVP?
- For many software MVPs, they reduce the team needed, which is the clip's point. They do not replace product judgment, security review or the work of operating a product in production. Most founders still hire engineers once the product has users; agents change when you hire and how many, not whether.
- What happens if you raise too small a seed round?
- You may run out of runway before reaching the metrics the next investors want, and then raise on weaker terms or bridge at a discount. Avoid it by sizing for a realistic timeline, tracking monthly burn against plan, and starting the next raise with at least six months of cash still in the bank.
- Does a smaller seed round hurt your Series A?
- Not by itself. Series A investors judge traction, growth and efficiency. A company that reaches the same revenue on less capital usually looks stronger, with a cleaner cap table and less dilution. The risk is only if the smaller round left you without the team or spend needed to show growth.