How Venture Capital Operates in the Age of AI: Key Insights from Niсk Davidov

What’s happening to venture capital in the age of AI, and how are Silicon Valley funds adapting? The technological revolution fueled by the explosive rise of generative AI is reshaping not only industries but the very mechanics of venture investing. Traditional fund strategies are losing relevance, and the winners are those who can move fast, navigate the technology stack, and stay close to founders. In a closed investor call hosted by Raison, Niсk Davidov — GP of DVC AI Fund I — shared his perspective on what venture capital looks like today, and how it’s set to evolve.

From Adtech to AI Venture: Nick Davidov’s Journey

Nick Davidov began exploring AI back in 2012 — first as a builder of AI-powered advertising products, and later as an investor. It was the heyday of programmatic advertising, when algorithms could decide in milliseconds which ad to serve a user.

We used to joke that soon an algorithm would predict your click even before you thought about it.

Around the same time, AlexNet emerged — the neural network that taught computers to see. Not long after, they began to understand language. That marked the dawn of a new technological era: one in which AI could plug into virtually any part of the economy. Because language, after all, is the interface for everything we’ve built as a civilization.

Why Silicon Valley Is Worth More Than Money

When Nick landed in the Valley, he quickly discovered its greatest asset — proximity to the right people. In thirty minutes at a bar, you can gain insights that might take months to uncover elsewhere.

I spent six months studying the Brazilian market with consultants. Then I went to a party at Facebook’s office, and over a beer, two executives — from Facebook and Baidu Brazil — handed me a complete go-to-market strategy. And then they said, ‘Hey, why don’t we just do it together?

Nvidia, Energy, and the Future Stack: How AI Is Reshaping Infrastructure

AI products are built on foundational layers: electricity, chips, cloud platforms, models, infrastructure software, and applications. Each layer carries its own risks and opportunities.

According to Andrej Karpathy, the world will require a billion times more computation over the next decade. Even if chip performance doubles every nine months, humanity will still need to massively scale energy production and rebuild infrastructure from the ground up.

At the chip level, TSMC dominates manufacturing — but the real strategic game is being played by Nvidia. Unlike Cisco in the early 2000s, Nvidia isn’t stuck at the “plumbing” layer. It’s climbing the stack — into clouds, foundation models, and even end-user products.

Cisco, during the dotcom era, made the best routers and believed it would "own the internet." But the value went to those who built on top of the pipes — the product creators. Nvidia has taken the opposite approach: through vertical integration, it avoids the commoditization trap.

And that’s critical in an era of explosive compute demand: those who control the full stack don’t just move faster — they capture more margin.

Machines Are the New Customers: How AI Is Transforming SaaS and Its Business Model

Building cloud infrastructure is a capital-intensive business that blends finance, development, and deep tech. But even hyperscalers like CoreWeave, Lambda, and Nebius are now facing the realities of commoditization. Profitability often hinges on creative accounting — like depreciating GPUs over a six-year horizon.

Yet two things remain true:

• AI is getting smarter.
• AI is getting cheaper.

Just like in the early 2000s, the real value is created at the application layer — when products solve actual business problems. That’s why the future belongs to AI-native products with clear business models and sticky user bases.

The SaaS sector is undergoing a fundamental shift: traditional “per-seat” licensing is fading, and a new kind of customer is emerging — AI agents. These agents aren’t limited by office hours or headcount. They can spin up thousands of parallel processes on demand. As a result, the SaaS monetization model is shifting from “per user” to pay-per-usage.

AI doesn’t get tired, doesn’t complain, and has infinite time. It can spin up 1,000 virtual machines and do what no analyst ever could.

This marks the birth of a new economy — one built not on seats, but on cycles.

For SaaS businesses, things are only getting better. It’s just that their primary customers are no longer people — they’re machines.

This new paradigm is a direct extension of Nvidia’s strategy: moving up the stack toward value creation. The same applies to the next generation of venture funds. The winners won’t be those betting on infrastructure — but those who can identify and support AI products with scalable, usage-driven economics.

A New Kind of Fund: How RollingFund Connects Engineers and Startups

Niсk and his wife Marina brought their founder experience into the heart of their fund’s philosophy:

Who really helps a startup? It’s not a big fund — it’s a Google engineer who writes a personal $50K check and shows up on a Saturday to help. That’s when we realized: we need to build a fund for people like that.

That’s how RollingFund was born — a platform where most LPs are engineers and developers. What unites them is a single AI-powered system that connects portfolio companies with domain experts.

RollingFund is designed for value-added angels, engineers, and operators — people who don’t just write checks but roll up their sleeves and work alongside founders. It’s not just a community; it’s an AI-native platform that bridges startups with a network of 160+ engineers worldwide.

It doesn’t matter whether I’m working today or not. The machine works. It’s a scalable architecture that ensures a baseline of quality — every time.

Disciplined Focus: Why DVC AI Fund I Chases Returns, Not AUM

DVC AI Fund I may look like a traditional venture vehicle on the surface — $75M in capital, a $100M hard cap, and a four-year deployment window — but its spirit is radically different. At the core is strategic austerity:

• Strictly Seed / Pre-Seed / Series A
• No follow-on investments
• Relentless focus on returns over assets under management (AUM)

This kind of discipline is rare in an industry increasingly obsessed with scale for scale’s sake.  >We don’t want to play the AUM game like Andreessen Horowitz. We’ve got our own money in the fund. We’re not playing for management — we’re playing for returns.

Avoiding follow-ons isn’t ideology — it’s operational clarity:

The moment you start doing follow-ons, you need more capital. Then come larger checks, blurred focus, and falling returns. We’re not going down that road.

DVC’s portfolio already includes 124 companies, among them Perplexity, Kursor, Avoca, and Splitworks. Over just four years, the fund has backed 2 unicorns and 1 decacorn.

The fund’s proprietary AI analyzes 220 variables per startup to predict upcoming rounds. But final decisions remain human.

We’re not guessing. We know a company is about to raise in three months. But it’s always a person who decides — the founder, their motivation, their context. That can’t be automated.

Take Kick, an AI-driven accounting startup:

We saw the signals and offered a check. The next day, the founder said, ‘OpenAI is giving us $1M, you’re in for $1M, valuation is $35M.’ Three months later, they raised a Series A at $120M. We made 2.5x in a single quarter.

This is what places DVC AI Fund I in a rare class of venture firms — built with its own stack, operating at engineering speed, and guided by investor discipline.

Why Raison Invested in DVC AI Fund I

As part of our fund-of-funds strategy, we chose to invest in DVC AI Fund I because it represents exactly the kind of partner we value: those who combine deep technical expertise with a fundamentally human approach to investing.

DVC AI Fund I has built a strong ecosystem, developed its own proprietary AI infrastructure, and—most importantly—respects the early stage for what it truly is: a phase where teams matter more than capital.

We believe that being present at Seed and Pre-Seed requires a different toolkit, a different tempo, and a different kind of attention. DVC AI Fund I delivers on all three—not just as an investor, but as an operator.

Takeaways: Where Venture Is Headed in the Age of AI

• As AI accelerates, so do the demands on infrastructure, strategy, and decision velocity.
• The winners will be those who combine technological access with deep empathy for founders.
• The future of venture belongs to those who help in practice — not just with capital, but with talent, customers, and code.

AI is no longer just a technology. It’s becoming the infrastructure of a new civilization. And those who recognize this early — who can act quickly and build value not just with money but with hands-on support — will lead the next wave.

If you’d like a deeper dive into DVC’s investment logic and Nick Davidov’s vision, watch the full recording of our private session at the link below. Please note: the call is conducted in Russian.

Link