Inside v0: How Vercel Is Reimagining Software, Org Charts, and AI Monetization

Vercel built its name by making it effortless for developers to build and deploy applications. Its second act, v0, extends that mission to a new generation of non-developers — product managers, designers, and consultants who build without writing code.

In this episode of Second Acts, Chargebee CEO Krish Subramanian speaks with Zeb Hermann, GM at v0, about how Vercel’s AI-native “startup within a startup” is redefining experimentation, organizational design, and pricing as AI reshapes software.

Listen to the full episode.

Building an Experimentation System

Zeb traces his approach to scaling back to lessons from Segment and Sequoia Capital.
At Segment, he saw how strong teams and deep understanding go hand in hand.

“There’s a mythology for leaders that you can delegate and just hire good people, and that they will get things done. What I saw instead was: ‘I’m going to go 15 levels deep into how our customer success ticketing system works.’”

At Sequoia, a single question reshaped how he thought about growth: What is the scale of your ambition? The answer, Zeb learned, is to think in portfolios of bets. Some will fail outright, a few will outperform, and the rest will sit in the middle.

That principle guides how v0 runs its AI initiatives. The company encourages a high volume of short-term experiments to allow the right ideas to surface quickly.

He points to a small example: a customer-support engineer built a refund bot using a large language model. What began as a hackathon project has evolved into a permanent tool, now maintained by Vercel’s automation team. It’s a system built to learn through doing.

Serving a Different Audience as Vercel’s First Customer

That audience includes product managers, designers, and consultants — non-developers who want to prototype and ship ideas quickly. To meet their needs, Zeb structured v0 to operate like an independent startup, building on top of Vercel’s own platform as one of its first and most demanding customers.

“We wanted to build on top of the same APIs that other applications built on top of Vercel use.”

The team develops directly on Vercel’s infrastructure, its AI Gateway, Sandboxes, billing, and account systems, and provides immediate feedback when tools don’t meet real-world needs.

“We think if we operate this more as its own startup and have that built on top of Vercel and give each other bi-directional feedback, we can move faster and better serve this different audience.”

Zeb leads an independent organization of approximately sixty people, spanning go-to-market, design, engineering, and DevRel. It is small enough to stay nimble yet close enough to the core platform to shape how Vercel itself evolves.

The Barbell GTM and Pricing Feedback Loop

v0’s growth follows a barbell pattern: millions of individual builders on one side and large enterprises on the other.

“About 60 percent of usage is nights and weekends… then they come to us and say, ‘I need this for every product manager at my organization.’”

That organic pull from self-serve users has become a reliable path to enterprise adoption.

Zeb’s early mandate at Vercel included pricing and packaging, where he found that innovation depends on proximity between product and pricing teams.

A notable example is the active CPU pricing model for Fluid Compute. Traditional infrastructure pricing charges for total runtime, with a charge for every second a server instance is up. Zeb’s team found that for AI workloads, much of that time is “dead”: servers wait idly while large language models process requests.

Instead of passing that inefficiency to customers, Vercel engineered a system to track and bill only for the time compute is actively working. It’s a more precise and customer-aligned approach, one that required close coordination between engineering and pricing to measure usage at that level of granularity.

“If you charge people for the time that the server is waiting, it’s wildly expensive for them. On our side, we can reuse that server time for other requests.”

Pricing, in other words, is treated as part of product design, not an afterthought.

Hybrid AI Monetization and Short-Cycle Vision

Zeb sees AI pricing evolving beyond simple token markups. A year ago, many expected model token costs to drop to near zero, but that hasn’t happened. Instead, he argues for a more sustainable approach:

“You sell tokens more or less at cost and capture value elsewhere. I think it’s the most customer-aligned.”

v0’s model earns margins on seat-based plans while passing through token costs, reflecting how different user types consume the product.

The result is a structure where customers pay in proportion to the value they create, not the raw resources they consume. Heavy users contribute more through seats and advanced features, while casual users benefit from lower token costs. Zeb describes this as a more aligned and transparent way to scale AI products — one that rewards efficiency on both sides.

That same pragmatism shapes how he thinks about planning. The speed of change in AI, he says, makes long-range roadmaps unrealistic.

“I keep writing six-month visions, and then a month later, we’ve shipped almost everything on the list. This has happened three separate times in the last couple of months.”

He prefers short, iterative planning cycles and constant course correction — a mindset reinforced by two books he often recommends:

• Algorithms to Live By by Brian Christian and Tom Griffiths
• Scale by Geoffrey West

Both explore how systems and organizations can make better decisions at scale. Zeb says the lessons from these books “sit in the back of your brain and help you make meaningfully better decisions.” (He first heard both authors speak through The Long Now Foundation, whose talks he credits for shaping his thinking.)

Lessons from Vercel’s Second Act

Vercel’s second act demonstrates how aligning pricing and product decisions around experimentation can accelerate learning and scale. For Zeb Hermann and v0, every release is a new hypothesis, every pricing change a design decision, and every experiment a step toward learning faster.

Listen to the full episode

Even more insights

Discover lessons from the field on using pricing to drive adoption, build sticky habits, and create an unfair advantage in AI markets.

Get the Pricing for AI Distribution Playbook →