Centicorns Reshape Markets Through Scale-Driven Feedback Loops

"The winners are compounding faster than ever, which means the costs of not being a winner are higher than ever."

-- Thomas Laffont

We’re not just in a new tech cycle--we’re in a regime change. Thomas Laffont’s appearance on All-In reveals that the AI-driven economy isn’t merely accelerating growth; it’s restructuring the rules of value creation, competition, and capital allocation. The most non-obvious implication? The system is now self-reinforcing: the biggest players grow not just by outperforming, but by reshaping the markets they operate in, turning scale into a structural moat. This isn’t about better products--it’s about control over entire profit pools. Anyone making investment, strategic, or entrepreneurial decisions in tech, finance, or infrastructure over the next five years needs to understand this shift. The advantage? Seeing where capital is forced to flow, not where it wants to go--and positioning accordingly.

Why the Obvious Winners Are Also the Hidden Market Makers

Most analysis stops at “AI is growing fast.” Laffont goes further: he shows how the fastest-growing companies are not just beneficiaries of a trend--they’re becoming the trend. When OpenAI or Anthropic grow at a pace that outstrips Adobe, Salesforce, and now even Google Cloud, they don’t just capture revenue. They redefine what’s possible in enterprise software, consumer behavior, and infrastructure demand.

This creates a feedback loop: growth attracts capital, which funds more growth, which increases market confidence, which lowers their cost of capital, which allows them to reinvest more aggressively. The result? A compounding velocity that smaller players can’t match--not because they’re less innovative, but because the system now rewards scale above all else.

"If you're a centicorn--$100 billion or more--the odds of you having a 10x are 31%. For unicorns, it's 8%."

-- Thomas Laffont

That’s not just a statistic. It’s a map of the new power law: once you cross the $100B threshold, the odds of explosive growth increase dramatically. This flips conventional VC wisdom on its head. In the past, early-stage investors bet on outlier potential; now, the outlier potential is concentrated in the already-outlier. The implication? Capital is rationally shifting toward de-risked, late-stage giants. Why bet on a 1-in-12 shot when you can bet on a 1-in-3?

But here’s the hidden consequence: this creates a capital gravity well. As more LPs, hedge funds, and even retail investors pile into the proven compounders, early-stage innovation gets starved. Not because it’s unimportant--but because the return asymmetry now favors patience over risk. The system routes around uncertainty by crowding into certainty, even if that certainty is priced at 50--100x revenue.

The 18-Month Payoff Nobody Wants to Wait For

Laffont highlights Cerebras as a case study in delayed payoff. The company endured “multiple years of no new capital,” grinding through technical development before landing a transformative contract with OpenAI. That moment didn’t just validate the tech--it multiplied the company’s value overnight.

This is where conventional wisdom fails. Most investors, especially in the ZIRP era, optimized for speed to scale. But Cerebras proves that endurance is becoming a competitive advantage. The companies that survive long enough to plug into the AI ecosystem’s core--whether through compute, data, or integration--get pulled into the gravity well of the giants.

The system responds: as foundational models demand more specialized hardware, memory, and data infrastructure, a new tier of beneficiaries emerges. These aren’t the flash-in-the-pan apps built on top of GPT--they’re the enablers who built the rails. And because their value is unlocked only after the core platforms scale, they require a holding period most investors aren’t structured to endure.

This is where others won’t go. The pain of waiting--years with no liquidity, no headlines, no momentum--is precisely what creates the advantage. And when the breakout comes, it’s not incremental. It’s step-change.

How the System Routes Around Your Solution

One of the most underappreciated points Laffont makes is that disruption isn’t sector-specific--it’s profit-pool-specific. Starlink isn’t just a satellite company. It’s a telco disruptor. By enabling phone calls anywhere on Earth, it targets the entire global wireless and broadband profit pool, estimated at $200--400B.

That’s not competition. That’s replacement.

And the same dynamic applies to AI. When 25% of Google and Meta’s ad revenue is already AI-enabled--and that’s expected to hit 100%--the ad tech stack isn’t being optimized. It’s being rewritten. The AI model becomes the interface, the decision engine, the personalization layer. The old ad tech middlemen? They’re being routed around.

This is systems thinking in action: Laffont doesn’t ask “What does this company do?” He asks, “What profit pool does this displace?” And the answer determines not just valuation, but strategic inevitability.

The feedback loop here is brutal: the more AI integrates into core workflows, the more data it generates, the better it gets, the more indispensable it becomes. The system doesn’t just adopt AI--it becomes AI. And once that happens, competition shifts from features to access and integration depth.

Where Immediate Pain Creates Lasting Moats

SpaceX is the ultimate example of a company that turned operational difficulty into a structural moat. Early on, launch cadence was erratic. But as they launched more, the business model improved in quality--not just in scale.

Laffont breaks it down:
- Phase 1 (Pre-constellation): One-off government contracts. Unpredictable. Low margin.
- Phase 2 (Initial ramp): Starlink begins. Recurring revenue. Subscriber growth.
- Phase 3 (Scale): Multiple constellations. Platform effects. New businesses (space data centers, lunar logistics).

The key insight? The more they launch, the more valuable each launch becomes. It’s not linear growth--it’s exponential optionality. And the market knows it: SpaceX’s valuation per launch has increased as launch frequency increased.

This is counterintuitive. Normally, scaling drives commoditization. Here, scaling drives differentiation. The pain of mastering reuse, reliability, and cadence--the years of explosions, delays, regulatory battles--created a moat that’s nearly impossible to replicate. Competitors aren’t just behind. They’re playing a different game.

And now, with Starlink poised to go public, that moat is about to be stress-tested in the public markets. But the real test isn’t valuation--it’s whether the market understands the platform shift. Because if it does, the IPO won’t be a liquidity event. It’ll be a capital magnet.

Key Action Items

• Over the next 6--12 months: Monitor the IPO filings of SpaceX, OpenAI, and Anthropic not just for valuation, but for revenue composition. Where is the money really coming from? That reveals the next profit pools at risk.

• Within the next quarter: Reassess early-stage AI bets through the lens of enabler vs. app. Apps are vulnerable to platform shifts. Enablers--hardware, memory, data infrastructure--have longer, more predictable payoff curves.

• Start now: Build relationships with companies operating in adjacent infrastructure to AI (semiconductors, power, cooling, data logistics). These are the hidden beneficiaries of the AI scale-up.

• Over 12--18 months: Expect pricing pressure in core AI services. With massive capital reserves, OpenAI or Anthropic may trigger a price war to dominate adoption--just as AWS did in cloud. The one with the lowest cost of capital wins.

• Begin immediately: Shift LP allocation strategy to account for the centicorn premium. Waiting for $100B+ companies may be the most rational (if unglamorous) way to capture AI upside with lower risk.

• Long-term (2+ years): Watch for profit pool collisions. When AI meets autonomous vehicles, biotech, or energy grids, the disruption won’t be incremental--it’ll be systemic. Position for the intersection, not the trend.

• Right now: Accept that the old VC playbook is broken. The power law isn’t just alive--it’s accelerating. The fastest way to fail is to apply 2010s-era diversification to a 2025s-era concentration economy.