There’s a cohort of companies that have convinced everyone – including themselves – that they are incumbents. They are not.

They have the revenue ramps, the massive funding rounds, the enterprise contracts. They look like market leaders. But they’re cosplaying as incumbents while operating with fundamentally early-stage risk profiles. 

These are false-positive incumbents – companies that mirror the revenue patterns of historical market dominators, but operate in a fundamentally different context where those patterns don’t guarantee the same outcomes. The underlying game has shifted, and there’s no assurance these companies will maintain their “dominance” over the coming years, or even months.

Some of this is real: we are seeing companies build sustainable and rapid growth. But some of this is theater. You need to realize how much of this is theater. The critical insight is recognizing how much of the current market positioning is performative. 

Companies are architecting entire strategies around belief network effects – where perception of success becomes self-reinforcing – and AI has become the ultimate amplifier for this dynamic. Once you see it, you can’t unsee it. 

The good news is that, once you see it, it opens up much more opportunity. Basically, the companies optimizing for today’s fundraise may be vulnerable to competition tomorrow, or two years down the line. The second mover advantages in AI are coming, if you have the poise to see this market for what it is, play the game just enough, and optimize for the long-run.

The Power of “Playing Incumbent”

Early success does not always equate to sustainable advantage. AI has turbocharged a classic error: mistaking ‘first to market’ for ‘best to market.’ 

AI makes it fuzzier to spot the difference because today’s growth metrics look amazing from the outside. 

We’ve seen ARR metrics that look like they’re on steroids right now. And, if you see a company scaling to $100M ARR in two months, it makes competition seem impossible. They attract the press, the fundraising – it’s almost like hacking preferential attachment: the idea that the richest nodes in the network get richer. It feels like you’re playing in a sea of incumbents. 

Again, some of this growth is real. But a lot of this is not. 

Why do we suspect many companies are false-positive incumbents? 

First, there is no guarantee that these revenue metrics – which are supposed to be a measure of repeatable growth – are actually repeatable. Enterprise buyers are under immense pressure to adopt AI and digital transformation tools. They’re making purchasing decisions based on FOMO rather than proven ROI, creating artificial demand that looks like genuine market pull. 

This can work for a while – but hype driven customers are not likely to stick around long term. If your customers aren’t expanding usage, or looking to align your service with future roadmaps…that should concern you. It may suggest they’re getting similar functionality from an all-purpose tool, like ChatGPT, and may churn away in a few months. 

That FOMO leads to “stripper pole” revenue ramps. Growth that would have taken incumbents decades now happens in <12 months, in some cases. But this speed often comes from heavy discounting, subsidized customer acquisition, or land-and-expand strategies where the “land” is basically free. Some of this revenue is likely coming from AI pilot programs, which may evaporate as users churn away. 

TLDR; not all revenue is created equal. Without medium-to-long-term retention data, we don’t know if these revenues represent sticky, expanding relationships or just expensive experiments that customers will abandon. The revenue might be real, but the value might be ephemeral.

We suspect many VCs and founders know this. Which is why we have seen the rise of “kingmaker capital.” Multi-stage funds are deploying massive checks to anoint winners before they’ve actually won (you know who you are). Without network effects or long-term defensibilities in place, VCs are trying to manufacture dominance in an industry that’s still rapidly changing. 

The result of all these forces? A feeling of artificial scarcity that forces companies to prioritize the next fundraise over fundamentals. This creates the perception that capital and growth alone can be a moat, when history has shown us it likely can’t be.

The Historical Reckoning of False-Positive Incumbents 

Real incumbents get stronger under pressure. When capital becomes scarce, their network effects become more valuable, their data moats widen, and switching costs increase as customers consolidate with “proven” providers. They benefit from flight-to-quality dynamics.

False-positive incumbents collapse under that same pressure. 

Strip away the subsidies and you realize their “network effects” are just expensive user bribes. 

Their “data advantages” are cash-burning data collection programs. 

Their switching costs are actually switching incentives – customers only stayed for the discounts.

We see examples of this all over startup history.

First we have our graveyard of companies that were outlier growth machines with huge funding rounds, and virality that ended up being expensive user bribes. 

In 2010 Groupon was the fastest growing startup ever. At its peak, it seemed to have everything: explosive growth, endless capital, a $1.2B valuation, and what even appeared to be a network of users and businesses. But as competition intensified and capital became more scarce, Groupon became weaker, not stronger. 

Their “network effects” were revealed to be expensive customer acquisition via unsustainable discounts. Merchants weren’t loyal to the platform, because users weren’t loyal to businesses they found via Groupon. Without constant cash flow to subsidize artificially low prices, the marketplace evaporated. 

On the flipside, we see plenty of category-defining companies that weren’t first, but still won. They killed the false-positive incumbents of their time. 

Meta (Facebook) wasn’t the first social network. But they recognized that real identity and college-centric networks created stronger engagement loops than pseudonymous or open platforms

Google wasn’t the first search engine. It was the best search engine – by a lot. In this case, quality trumped mediocrity, and translated to distribution advantages over time. 

Ramp wasn’t the first corporate card. Brex had raised ~$300m, had a freshly minted ~$2.6b valuation (which would peak at $12b in 2021), and dominated startup spending. But when venture funding dried up in 2022-2023, Brex’s model began to crack. Their growth was fueled by high-burn startups with loose expense controls – customers who vanished when funding became scarce. In contrast, Ramp focused on expense management and cost reduction rather than rewards. As companies tightened belts, Ramp’s value proposition strengthened while Brex pivoted away from startups entirely. Brex is still an incredible company, but Ramp proved that first mover ≠ best mover.

The pattern is consistent: these companies moved fast and raised money when it made strategic sense, but they also built increasingly defensible positions over time. They aimed to control more of their destiny – not less. Each funding round, each product launch, each acquisition served to deepen their moats rather than just extend their runway to the next milestone.

Everyone is chasing cheap, early growth but growth alone doesn’t equal lasting success. 

First-movers validate markets, but winners build sustainable advantages. AI does not invalidate this, it just exacerbates the “success” of the first-movers. It’s harder to imagine that a second-mover advantage exists today, but it does. And it will accrue to people who built something enduring.

The Bullshit Test

Let’s be fair. Some of these revenue ramps and kingmade companies are going to win…So how do you spot the difference between a false-positive and an upstart incumbent? 

The siege test: What happens if this company couldn’t raise money for 24 months? Real incumbents adapt and survive. False-positives implode – regardless of the size of their war chest. 

Customer behavior: Would customers pay more for this product over time? Real incumbents build increasing willingness to pay. False-positives depend on maintaining artificially low pricing.

Technology depreciation curves: Real incumbents build tech advantages that compound over time. False-positives rely mainly on execution speed that competitors can copy (likely faster, too) once the market is proven.

Competitive response: When real incumbents enter a space, existing players change strategies significantly (e.g., pricing, who they’re selling to, etc.). False-positives only “win” in greenfield markets or where everyone is playing with house money; once a capital-efficient competitor enters, their “network effects” reveal themselves as expensive user rental programs.

Margins still matter: Real incumbents have a path to expanding margins (fixed costs amortizing, pricing power increasing, etc.). False-positives see margins stay flat or compress – the more customers you onboard to a GPT wrapper, the higher your costs get. You’re essentially selling $1.50 for $1 – you’re paying the provider of an underlying model more than you can sustainably charge for your product. 

False-positive incumbents optimize for the next funding round rather than profitability, so they leave all these problems unsolved.

What Early Stage Companies Should Be Doing

So far, we have outlined the problem. So what should companies be doing today? 

Don’t fall for false-positive incumbent dynamics. Second- (or third-) mover advantages exist and will emerge. You may have to play the short-term virality game too, but don’t believe that this is your ticket to success.

If you do the inverse of the bullshit test above, you’re likely on the right track.  

Retention curves > AI tourism – focus on engagement and retention metrics, not just ARR. Are customers integrating your product into daily workflows or just experimenting with AI after a top-down directive? Real usage depth matters more than vanity metrics.

Create differentiable value props – Most startups don’t begin defensible, but the best build it in over time (see this Elad Gil piece). Start solving real workflow problems, not just showcasing AI capabilities. 

Focus on value creation over value capture – Don’t make short-term optimizations for revenue ramps instead of baking in immense value creation for your customers over the long haul. Heavy discounting doesn’t just train customers to expect commodity pricing – it trains you to build a commodity product. 

It’s better to have fewer customers who genuinely value what you’re building than many who see you as interchangeable. 

Map out a path toward margin expansion – Healthy unit economics don’t have to exist today, but the path should be visible. The light at the end of the tunnel can be small, but it should be there.

Avoid optimizing for just the next fundraise – Look, we get it, fundraising is important for many reasons (thankfully – we wouldn’t have a job otherwise). But it is not the be all and end all. Kingmaker capital creates perverse incentives to optimize for fundraising theater over fundamentals. The fundraise should be the byproduct of achieving your goals, not the goal itself.

Big Caveat

We can already hear people arguing in the comments. Distribution is the only thing that matters, they say. And if this is all true, what about OpenAI or the other major AI labs? They don’t pass this test with flying colors, and yet, they’re the biggest existential threat to many startups. 

These companies burn capital at rates that would kill traditional startups in months. They don’t have clear paths to margin expansion. Their moats are talent hoarding and compute scale — both of which bleed money. (And talent hoarding is getting more expensive by the day, thank you Meta). In a 24-month funding drought, they’d likely face existential pressure despite their valuations & war chests. 

And yet, no one believes there will be a world where OpenAI folds because of funding shortages or technology depreciation curves. Why? 

They’ve transcended the normal rules by becoming too big to fail. OpenAI has achieved something unprecedented – belief network effects so powerful that funding them has become a geopolitical imperative, not just a business decision. When your product becomes infrastructure that entire industries depend on, different economics apply. 

If you are building a company with potential to become a geopolitical imperative, definitely come talk to us. 

(Stablecoins and longevity are good places to start).

Why We Might Be Wrong…

Nobody’s perfect. Not the darlings of early stage venture, nor the pontificating venture capitalist (that’s us, thanks for listening). So we’ll be honest: this framework has holes. 

“Fake it till you make it” sometimes works! Amazon subsidized growth for years while building real infrastructure. Netflix burned cash to create content moats. Some players might be using their subsidized phase to build genuine network effects and data advantages.

Market creation vs. market taking. These companies might not be subsidizing their way into existing markets – they might be creating entirely new categories where traditional incumbent playbooks don’t apply.

Capital efficiency evolution. Maybe the old rules don’t apply when software creation & distribution costs approach zero. These companies might have genuinely different unit economics that look unsustainable by historical standards but actually work.

The music might not stop. The whole thesis assumes a correction is coming. But what if cheap capital persists and these growth patterns become the new normal?

So What?

Look around your market. Half the “incumbents” you’re worried about are just well-funded startups with unsustainable unit economics. The other half are building real moats while everyone else chases fundraising metrics. 

The false-positive incumbent framework isn’t perfect, but the alternative is assuming every company with a $100M (or $1B) valuation has actually earned it when that’s likely far from the case. 

Some will eventually earn their status, but many won’t. Build like you’re in the minority that will.