Before Revenue, Build Evidence

Early-stage founders may not have much revenue yet. But that does not mean they should arrive in front of investors without evidence.

One of the most interesting funding stories this week was Mantic.

The AI forecasting company raised a $25 million seed round after demonstrating its technology in the 2026 Metaculus Cup.

What makes the case interesting is not simply the size of the round.

It is what happened before it.

The Claim Came Before the Evidence

Every startup asks investors to believe something.

Maybe your product is faster.

Maybe your technology produces better results.

Maybe customers will change an established behaviour.

Maybe your software can reduce a cost that companies have accepted for years.

At an early stage, you may not yet have enough revenue, customers or historical data to prove the entire business.

But there is an important distinction:

You may not be able to prove the business yet.
You may still be able to test the claim underneath it.

That is where Mantic becomes interesting.

Its core proposition concerns AI-powered forecasting.

Rather than leaving that capability as a statement in a pitch deck, the company participated in an external forecasting competition where performance could be measured.

According to Reuters, Mantic's system outperformed every human participant in the Summer 2026 Metaculus Cup and all but one competing AI system.

The competition involved 51 published forecasting questions.

That does not prove product-market fit.

It does not prove that customers will pay.

And it does not prove that Mantic will build a defensible company.

But it does provide evidence for an important underlying claim:

the system can forecast unusually well.

That changes the fundraising conversation.

Pre-Revenue Does Not Have to Mean Pre-Evidence

Founders sometimes treat traction as binary.

Either:

We have revenue.

Or:

We are too early to prove anything.

There is a large space between those two positions.

Evidence can take different forms depending on what an investor needs to believe.

A founder might use:

• an independent benchmark;

• a controlled pilot;

• a blind comparison;

• measurable customer behaviour;

• technical validation;

• a signed commercial commitment;

• repeated usage;

• quantified cost savings; or

• performance against an existing alternative.

The objective is not to manufacture traction.

It is to identify the most important assumption in the investment case and find the strongest credible way to test it.

Start With the Investor Question

Instead of asking:

“How do I make my pitch more convincing?”

Ask:

“What does the investor need to believe for this company to become investable?”

Then go one step further:

“What evidence could make that belief less dependent on my words?”

That distinction matters.

A pitch deck makes claims.

Evidence reduces the amount of faith required to believe them.

Imagine a founder saying:

“Our technology produces significantly better predictions.”

The natural investor response is:

How do you know?

A benchmark begins to answer that question.

Or consider:

“Our software can materially reduce processing time.”

Again:

How do you know?

A customer pilot measuring processing time before and after implementation begins to answer it.

The stronger fundraising story is therefore not necessarily the one with the most claims.

It is often the one where the important claims have the shortest distance to evidence.

Build an Evidence Architecture

Before fundraising, write down the three or four things that must be true for your startup to work.

For example:

Customers experience the problem.

Our solution materially improves the current alternative.

Customers are willing to adopt it.

The economics can become attractive at scale.

Now look at each statement separately.

What evidence do you currently have?

What remains an assumption?

And what experiment, benchmark, pilot or customer behaviour could move that assumption closer to evidence?

You do not need to prove everything simultaneously.

But you should understand exactly where belief ends and evidence begins.

That is an important part of investor readiness.

Evidence Has Limits

There is another lesson in the Mantic case.

Good evidence should not be stretched beyond what it actually proves.

Performance in a forecasting competition provides evidence about forecasting capability.

It does not, by itself, demonstrate commercial demand, customer retention, attractive unit economics or defensibility.

Founders weaken otherwise strong evidence when they make it carry more weight than it can support.

A useful discipline is therefore:

Claim → Evidence → What it proves → What it does not prove

That makes the investment story more credible, not less.

The Founder Question

Before your next investor conversation, take the biggest claim in your pitch and remove the slide around it.

Then ask:

If I could not explain this claim to the investor, what could I show them instead?

That might be a customer.

A benchmark.

A pilot.

Usage data.

A contract.

A technical result.

Or a measurable change in behaviour.

Because when the company is early, investors will inevitably have to make assumptions.

Your job is not to eliminate every assumption.

Your job is to reduce the number of important things they simply have to take your word for.

Founder Capital Brief

This analysis is part of Founder Capital Brief, where I translate developments in venture capital and startup financing into practical implications for founders.

Sources: Reuters, AI startup Mantic raises $25 million for superhuman forecasting, 18 September 2026; Metaculus Summer 2026 forecasting competition.