David George just published a16z’s inaugural State of Markets analysis, and one pattern should catch every marketer’s attention:

The market is bifurcating. Top quartile companies are pulling away. Revenue growth for top performers is exploding while everyone else flatlines.

But here’s what the data doesn’t explain: why are the winners winning?

The answer lies in a 19th-century economic principle that most marketers have never heard of.

Jevons’ Paradox Is Real

The a16z deck addresses the “is this a bubble?” question with a striking observation: 7-8 year old TPUs are still running at 100% utilization. Cheaper compute hasn’t reduced demand—it’s increased it.

This is Jevons’ Paradox in action.

In 1865, economist William Stanley Jevons observed that as steam engines became more efficient, coal consumption increased rather than decreased. Better efficiency didn’t mean using less coal—it meant coal became viable for more applications.

The same dynamic is playing out with AI.

Cheaper tokens don’t mean companies use less AI. Cheaper tokens mean AI becomes viable for more use cases. Usage explodes upward.

The Marketing Implication

Here’s where most marketers get it wrong.

Conventional thinking:

“AI makes content cheaper → we can produce the same content for less money”

Jevons’ Paradox thinking:

“AI makes marketing cheaper → we can now afford to market across channels we couldn’t staff before”

The first approach optimizes cost. The second approach expands capability.

The companies in a16z’s top quartile? They’re thinking Jevons, not cost reduction.

”Built Different” Means Structurally Different

The State of Markets deck notes that today’s winners are “just built different.” This isn’t marketing speak—it’s a precise description of what Clayton Christensen called disruptive positioning.

In The Innovator’s Dilemma, Christensen explains that disruptive companies don’t just do the same thing better. They operate on a different curve entirely.

Sustaining Innovation	Disruptive Innovation
Better blog posts with AI	AI discovering new channels
Faster content production	Automated channel testing
Same channels, lower cost	More channels, more learning

The gap between top quartile and everyone else isn’t about execution quality. It’s about structural approach.

The winners aren’t doing the same marketing better. They’re doing different marketing entirely.

The Channel Discovery Gap

Gabriel Weinberg, founder of DuckDuckGo, identified 19 distinct traction channels that businesses can use for growth. Most companies use three or four—the ones they already know.

This creates what we call the AI Results Gap: marketers apply AI narrowly (content generation) while ignoring the other 15+ channels where AI could help.

The a16z market data shows this gap playing out at scale:

Company Type	AI Usage	Channel Coverage	Result
Top quartile	Capability expansion	Testing many channels	Pulling away
Bottom 75%	Cost reduction	Stuck in 3-4 channels	Falling behind

The concentration of market success mirrors the concentration of channel discovery.

The Compounding Effect

Eric Ries describes the Build-Measure-Learn loop in The Lean Startup: faster iteration leads to faster learning, which compounds over time.

Apply this to channel discovery:

Winners: Rapid Test-Learn cycles across 8+ channels. Each iteration compounds.
Everyone else: Slow cycles across 2-3 familiar channels. Learning stalls.

The winners aren’t just ahead—they’re accelerating ahead. Each cycle of validated learning across multiple channels compounds their advantage.

This is why the a16z data shows winners “keep winning.” It’s not momentum—it’s compounding learning.

What This Means for 2026

The State of Markets analysis covers public and private markets, but the implications for marketing are clear:

1. Cost reduction is a trap

Using AI to do the same marketing cheaper puts you in the bottom 75%. The winners use AI to expand what’s possible.

2. Channel exploration is the opportunity

The channels you’re not using—affiliate marketing, speaking engagements, trade shows, business development—are where the asymmetric opportunities hide. AI makes them accessible.

3. Speed of learning wins

The companies pulling away aren’t just doing more. They’re learning faster. Build-Measure-Learn across more channels creates exponential advantage.

4. “Built different” is achievable

You don’t need to be a unicorn to think like one. The structural shift—from cost reduction to capability expansion—is available to any company willing to make it.

The Opportunity

Jevons’ Paradox tells us that cheaper AI won’t reduce marketing activity. It will increase it—dramatically.

The question is whether you’ll be the one doing more marketing, or the one being outmarketed.

The companies in the top quartile of a16z’s analysis didn’t get there by doing the same thing for less money. They got there by doing more things—testing more channels, learning faster, and letting AI handle the execution across a scope that would have been impossible with human teams alone.

That’s what an AI CMO enables. Not cheaper marketing. More marketing. Across channels you’d never try on your own.

The market is bifurcating. Which side are you building toward?

The full State of Markets deck from a16z is worth reviewing. For more on channel discovery and the AI Results Gap, see our posts on 19 Traction Channels and The AI Results Gap.