Hello, 👋
Picture this: you’re two years into building something real. Revenue is growing. The team is strong. The product works. You walk into a Series A with genuine momentum, tell a compelling story, and then... the emails go quiet. The follow-ups are warm but vague. Nobody says no. Nobody says yes. And you’re left wondering what actually went wrong.
If this sounds familiar, or if you’re trying to make sure it never does, then this piece is for you.
Today we’re digging into one of the most misunderstood dynamics in startup fundraising. The gap between what founders think investors are evaluating, and what investors are actually doing on the other side of the table.
Here’s the short version before we get into it: investors aren’t valuing your company. They’re discounting it.
Let’s get into it.
In partnership with
Are your finances healthy? This (AI) CFO will tell you
Most companies don’t have a CFO - or theirs is stretched too thin for daily monitoring. Nume changes that.
The world’s first AI CFO onboards in 5 minutes and delivers proactive insights via Slack or email:
Monitors finances 24/7, identifies risks before they impact your business
Answers questions instantly: cash flow, runway, projections
Integrates with Xero, QuickBooks, and your bank accounts
If you’re interested in sponsoring this newsletter, please reach out here.
The Knight Distinction: Why “risky” and “uncertain” aren’t the same thing
Most people in startup land use the words risk and uncertainty interchangeably. They really shouldn’t.
Frank Knight was a University of Chicago economist who, in 1921, drew a line between the two that the industry has never fully absorbed. Risk is quantifiable. You can assign probabilities to outcomes, model the scenarios, price it into a spreadsheet. Actuaries do this all day. Insurance companies are built on it.
Uncertainty is different. It’s the category of things you can’t even enumerate outcomes for. You don’t know what you don’t know. No model helps you because you don’t have the inputs. The future isn’t probabilistically distributed, it’s genuinely open.
And here’s the thing. Early stage startups live in uncertainty, not risk. A seed-stage company doesn’t have a 34% chance of failure and a 66% chance of success. It has a set of fundamental questions that haven’t been answered yet, about the team, the product, the market, the economics, and the answers to those questions will determine what happens.
This is why a spreadsheet that produces a $2B valuation for a company with eight months of revenue isn’t doing analysis. It’s decoration. You can’t solve for uncertainty with better modeling. The only thing you can do is retire it. Answer the open questions. Make the unknowable answerable. Turn the fog into map.
That’s what venture capital is actually for.
So the number on your term sheet is what an investor thinks your company is worth, minus a penalty for everything that hasn’t been answered yet. Get that backwards, walk in thinking you’re being evaluated on your potential, and you’ll spend the next year confused about why your metrics look good and your raise isn’t moving.
The Option Stack: What’s actually happening when someone writes you a check
Here’s a frame that makes fundraising make a lot more sense.
Every funding round is an option purchase, not an investment.
Seed buys the right to participate in Series A, contingent on enough uncertainty being resolved to justify exercising that option. Series A buys the option on Series B. Series B buys the option on C. All the way through.
The value of each option depends on one thing: how much fundamental uncertainty has been retired since the last exercise, relative to how much remains. Not revenue projections. Not market size. The uncertainty delta.
This reframes a lot of investor behavior that founders find baffling.
When a VC says “we love the team, we love the market, we want to see a bit more traction” and then doesn’t follow up when you hit the traction number... what they were actually saying is: we can’t price the remaining uncertainty at a level that makes this option worth buying right now. If you hit the traction milestone but other layers are still open, the option still isn’t worth buying. They’ll find a different thing to wait on. This isn’t bad faith. Most investors don’t even have full visibility into their own reasoning. It’s option pricing doing its job.
And it reframes the question you should walk into every room prepared to answer. The Series A question isn’t “is this a big market?” That’s a seed question. By Series A the question is: has enough been resolved since seed that the next option is worth buying at this price?
Those are completely different questions. They call for completely different evidence.
The Uncertainty Layers: The five questions every investor is silently asking, in order
So what does “retiring uncertainty” actually look like in practice? It clusters into five layers. And the order matters as much as the layers themselves. That part is important.
Layer 1: Team execution. Not credentials, not pedigree. Evidence that this specific group of people can deliver under real pressure. Can they ship? Can they make decisions when the information is incomplete?
Layer 2: Product proof. Does it work at scale, not in a controlled demo? A product that performs beautifully with the founder hovering over it is a different product than one running in the wild.
Layer 3: Market demand. This means revealed preference, not stated preference. Someone paying and coming back. Not a waitlist. Not an LOI. Not an enthusiastic pilot.
Layer 4: Unit economics. Steady-state unit economics, not early cohort data with warm intros and founder discounts baked in. More on this below.
Layer 5: External risk. Regulatory exposure, platform dependency, geopolitical fragility. The things nobody thinks about until they’re the whole problem.
The reason the order matters: strong evidence at layer three means nothing if layer two is still open. A product that can’t scale doesn’t have real demand, it has a lab result. Strong economics mean nothing if the team can’t execute against them at speed. You can’t make the Series B argument until you’ve closed the Series A argument.
The mistake most founders make is trying to address multiple layers at once. A team still proving it can ship probably shouldn’t be running three simultaneous market experiments. A company without repeatable demand shouldn’t be hiring to scale unit economics that haven’t been proven yet. Every dollar spent on layer four before layer three is resolved is, in the technical sense, waste. You’re building complexity on top of open questions rather than closing them.
Staged financing exists precisely to enforce this sequencing. When investors want to wait for the next milestone, it usually isn’t hedging. It’s the option stack doing exactly what it’s designed to do.
The Demand Problem: Why 2026 has raised the bar
This layer deserves its own section because it’s where most companies quietly stall. And because the definition of what counts as proof has shifted in ways that catch founders off guard.
LOIs tell you a problem exists. Good NPS tells you people don’t hate you. A waitlist tells you the messaging works. A pilot tells you someone was willing to spend three weeks finding out if you solved something.
None of that is demand. Not really.
The distinction that actually matters here is between stated preference and revealed preference. Stated preference costs your customer nothing, it’s just telling you they want a thing. Revealed preference is paying, staying, expanding. And the gap between those two is where most early-stage companies are quietly living without realising it.
Now here’s what’s changed. Building costs have collapsed. A small team can ship something functional in weeks. Which means... “we built a thing” has basically inflated away as a signal. Everyone built a thing. The bar for what counts as real demand evidence moved right along with the build costs.
What investors are actually looking for in 2026 is cohort data. People paying repeatedly. Retention that holds. Usage that expands. And then, beyond the data, an explanation for why the cohort looks the way it does. Because the explanation is the thing that tells you whether it generalises to the next 10,000 customers, or whether it was just the particular magic of the first 50.
There’s also a timing shift worth understanding. Seed rounds are partly bets on founding team opportunity cost. What they walked away from to do this. Whether they’re the kind of people who figure things out in real time. That variable does its work at seed, then decays. By Series A the demand evidence carries it.
Founders who haven’t made that shift in their own mental model of what they’re being evaluated on tend to get a lot of enthusiastic second meetings that go... quiet.
The Cohort Trap: Your early numbers are not the business
Your first ten customers came through warm intros from people who wanted you to succeed. Support was handled by a founder who answered Slack messages at midnight. Infrastructure ran on cloud credits. Some customers got special pricing because you needed the logo.
None of that is the business. That’s the prototype of the business.
What investors are trying to understand by Series B is whether the economics work at arm’s length. When customers come in through a real sales motion, get onboarded by someone who hasn’t memorised their entire situation, run on infrastructure you actually pay for, at a price that reflects real positioning rather than desperation.
But here’s the thing that separates the companies that raise cleanly from the ones that get stuck. Investors aren’t really looking for a high cohort level. They’re looking for cohort improvement over time. Customers from Q3 outperforming customers from Q1, in ways you can explain, is evidence of a business that compounds and learns. It means the machine is getting better at itself.
A high average concealing deteriorating marginal cohorts is a completely different story. And experienced investors have more time than founders to stare at numbers and ask follow-up questions.
In a market now explicitly rewarding capital efficiency and earlier breakeven, this layer carries more weight than it did two years ago. That’s not going back.
The Bus Factor: The layer nobody thinks about until it’s the whole problem
Regulatory exposure. Platform dependency. Data residency requirements. Key-person concentration. Geopolitical supply chain fragility.
These live at the bottom of every founder’s list because they’re not urgent, right up until they’re catastrophic.
AI right now is not a hypothetical case study on this. The EU AI Act is in force. US federal governance has been inconsistent but is directionally toward more scrutiny regardless of who’s running things. This week, Anthropic drew two lines in a DOD contract. No domestic surveillance of Americans. No autonomous weapons. The Pentagon wanted all lawful uses. They couldn’t find a compromise. The response was a supply chain risk designation that has, to be clear, never been applied to an American company before. One that, if it held legally, could have effectively banned every major cloud provider from doing business with them.
A company valued north of $60 billion was days away from a potentially existential outcome because of a contract negotiation going sideways. That’s... wow. Politics is business strategy now, whether founders want to engage with that or not.
The founders who handle this layer well don’t treat it as compliance. They treat it as architecture. Build the flexibility in before you need it. Think through what operating under a different regulatory regime would actually require. Know which platforms you depend on and what the contingency is.
Name the buses. Investors know they’re out there. They’re just checking whether you’ve seen them.
What To Actually Do
Here’s the practical thing to carry from all of this.
Keep a live document. Not a deck, not a board update. A working document for yourself. The open questions that, if answered wrong, would materially change the trajectory of the company. What are they right now. How confident are you on each one. What’s the cheapest experiment that gives a real answer. Who owns it. What’s the timeline.
Review it before you spend money. If a dollar is going toward answering one of those questions, good. If it’s building on top of a question that hasn’t been answered yet, stop and ask why.
Capital is an option-buying mechanism. Every raise gives you the runway to retire uncertainty and create the conditions for the next raise at better terms. The founders who look materially cleaner at Series B than they did at Series A, not just bigger but actually cleaner, are almost always the ones who spent the in-between period doing that deliberately, as a practice, not as a reaction to investor pressure.
Valuation follows from it.
Thanks for reading.
If you enjoyed this issue, send it to a friend—it helps more than you think.
Back in your inbox Thursday,
Thanks for bringing up the difference between the risk and uncertainty, I haven’t thought about it before. Do you reckon, then, if VCs are truly funding the process of making the unknown knowable, or simply rewarding companies once uncertainty has already been reduced enough for others to get comfortable?