Biz: The New Game

Nothing signals the arrival of spring quite like the RFPs for tech due diligence assessments—and, to be honest, I’d been eagerly looking for them after the post-Christmas slowdown. These projects are among the most demanding and gratifying work in our field, calling for intense rigor and a nuanced grasp of the ever-evolving technology landscape.¹.

It was one of those early team conversations—the sort where you sift through legacy assets and assess just how much needs to be updated to meet a customer’s demands. That’s when it struck me: I no longer know how to value a tech company’s IP.

It’s not that I’ve lost the skill—rather, the skill no longer anchors to anything concrete. I can evaluate the quality of the tech, the team’s performance, and the soundness of the architecture. But now, that’s as relevant as judging a phone by whether its charger fits a Schuko plug.².

Of course it fits. They all fit. That’s the point.

Just plug it in.

When electricity first arrived, companies that specialized in wiring buildings enjoyed a real competitive edge. Infrastructure was complex, costly, and demanded true expertise—whole industries thrived on that advantage. Now, electricity is a utility, instantly available at the flip of a switch. Nobody judges a factory by its wiring anymore; they judge it by what it produces and who it serves.

Some are slow to accept this shift, but for code, that moment has already passed.³

The technical labor once needed to build a digital product—the heavy lifting of coding, integrating, and deploying—now feels like it comes straight from the wall. While not entirely free, it’s accessible enough that it rarely limits what you can create.⁴ The real barrier has shifted, but our valuation mindset hasn’t kept pace.

There’s plenty of noise about this signaling the death of SaaS. I don’t buy it. SaaS valuations have shifted because the risk profile changed—a perfectly rational reason for stock prices to move. Some SaaS companies will thrive, others will fail, and a few will invent new monetization models. Code becoming a utility no more kills software than electricity killed manufacturing.

But it does surface the question I couldn’t answer in that meeting: if code downloads from the wire as tokens, what actually gives a tech company its value?

The multiplier game of valuation was always part fiction⁵—yet it served a purpose. It offered a common language for negotiation and a baseline that roughly mirrored reality. But that story only held because the code was difficult to build. Now that it isn’t, the multiplier game has lost its anchor.

So here’s my current view on where the real value sits: four patterns I keep seeing in companies that have something durable—assets that can’t be generated by the model.⁶

Velocity

The first pattern is speed. With technical R&D requirements drastically reduced, the traditional multi-year development cycle—the one that once justified massive upfront investments—has collapsed into months or even weeks. This fundamentally changes what matters.

Annual Recurring Revenue tells you where a company is. Velocity tells you how fast it’s getting there. Two companies may have identical ARR, but if one took three years and the other just three months, that speed becomes everything when building the product is nearly free.⁷

The companies poised for premium valuations are those that can swiftly turn new ideas into revenue—and, crucially, sustain that revenue in a landscape where competitors can build just as fast. In a world where the barrier to entry is an internet connection and an Agent harness, speed to market is the ultimate differentiator.

Local, Targeted Data

The second pattern is about what the models can’t do.

Large language models regress to the mean—they’re generalists by design. Recently, I asked a smaller open-source model who the president of Finland was. The result? It spiraled into a loop: Finland’s presidential election had just occurred, and the model’s training data cutoff left it unable to process the new reality. Without up-to-date, domain-specific context, it simply couldn’t answer.

This underscores a crucial point: domain-specific, actively managed data is a genuine moat. None of it appears on a balance sheet—yet in the agentic era, the chasm between what’s recorded and what’s truly valuable has never been wider.

Owning a proprietary dataset—say, ten years of Finnish cosmetics consumer habits—makes you vastly better equipped to serve that audience than any generic LLM. The ability to curate and leverage targeted data, whether through fine-tuning, retrieval-augmented generation, or simply strong data governance, is becoming a primary competitive advantage. If you possess the right data and a living “memory” of your domain’s reality, you hold the cards.

Ecosystem Positioning

This is the one that’s harder to measure, but arguably more important than it appears at first glance.

In the past, securing your place in a digital ecosystem—building an outstanding API, managing complex data conversions—was a prized technical asset. Today, that’s Schuko-plug territory: API integration and code conversion are nearly frictionless. Technical architecture, on its own, is no longer a moat.

What matters now is the ability to shape and anchor ecosystems through business contracts and relationships. Where comparability once meant similar revenue profiles, it now means similar ecosystem positions—who has the contracts, not just the code.⁸

When technology is freely connected, the true asset becomes your contractual integration into real-world, offline, or tightly regulated service ecosystems. This is far harder to track than an API call, but the legal and business rights to operate within a specific ecosystem create a formidable barrier to entry. And unlike code, contracts can’t be downloaded from the wall.

The Human Layer

This is the factor that should be obvious—yet it rarely gets enough attention.

When execution becomes cheap, judgment commands a premium. Companies have always paid for external advisors and consultants—not for better spreadsheets, but for a trusted, independent perspective. Now, when AI can instantly generate endless models and strategies, judgment itself becomes the product.

In an era where AI can devise ten plausible strategies in moments, a company’s most valuable asset is the human who can discern which strategy is truly sound. This isn’t just “soft thinking”—it’s the hardest thing to automate, and arguably the only thing that can’t be sourced from the Schuko plug.

That’s why company culture and employee retention are now central to valuation—not just for startups, but for every business. To win, you need a culture that attracts and keeps people with extraordinary judgment. If your best people leave, they take your one irreplaceable asset with them.

Patterns That Didn’t Make the Cut

As I researched, a few other models surfaced. For now, I’ve left them out of the core patterns, and here’s why:

Physical assets are the most obvious candidate. Owning land, equipment, or heavy infrastructure offers real, measurable, undeniable value. Yet, tech investors consistently discount these. Capital-heavy industries tie up resources, running counter to what high-growth tech valuation rewards—and the agentic era doesn’t change that calculus.

Brand is trickier. When anyone can build the product, brand becomes the trust signal—an enormously valuable differentiator. But brand is notoriously hard to measure on the balance sheet, and I haven’t found a way to operationalize it in due diligence. That doesn’t make the brand less real; it just means it deserves its own exploration.⁹

Bringing It All Together

The best venture investors have always valued companies this way—intuitively, looking for patterns rather than relying on spreadsheets. The difference now is that this lens is no longer optional: everyone needs it, not just VCs with decades of pattern-matching experience.

Building companies isn’t over—the rules have simply changed. If you’re relying on your codebase as your moat, you’re already behind. It’s like selling the quality of your wiring in a world where electricity comes from the wall.

This model isn’t the final word—but the old checklist has had its day. The ground has shifted, and it’s time we shift with it.