Every year, Oxx brings together founders, investors, operators, and advisors from across the European B2B software ecosystem for Scope – an afternoon dedicated to the questions that matter most to the people building and funding the next generation of software companies. This year, 100+ guests gathered at the St Pancras Hotel in London for Scope 2026, and the scope was the collision between AI and software: what it’s creating, what it’s disrupting, and what it demands of the people leading through it.
We called it The Cambrian Moment. 540 million years ago, the Cambrian explosion produced almost every body format alive today – eyes, spines, limbs, exoskeletons – in what geologists describe as an “explosion.” Most of the species that emerged are extinct. But the ones that survived became the progenitors of everything alive now. We think a similar tipping point is playing out in software. Here’s a summary of what the day surfaced.
AI is software. Software is AI. And there’s about to be a lot more of it.
The narrative that AI is killing software doesn’t survive contact with the data. Public software valuations are down roughly 58% below the ten-year median, but the picture is far more uneven than the headlines suggest. Vertical and infrastructure software is holding up well, protected by proprietary data, regulatory complexity, and deep process embedding. It’s the shallow horizontal workflow tools that face a genuine reckoning. The market’s instinct to differentiate is right, its instinct to dump all software into one basket is not.
Because AI is software, and software is AI. We are yet to see AI implemented through any means other than software, and no one building software today would do so without AI tools. What’s actually happening is an explosion. Coding agents are collapsing the cost and time to build, expanding the pool of software entrepreneurs far beyond today’s 50 million professional developers. Autonomous agents are becoming the new software users, consuming applications round the clock. And entirely new categories are emerging — most notably the “context graph,” which captures not just what happened in a business process but why, giving agents the contextual intelligence they need to make autonomous decisions. A category that didn’t exist two years ago, and one that represents a trillion-dollar market opportunity.
More builders. More products. More users. Entirely new categories. That’s the Cambrian Moment.
The real bottleneck isn’t the technology. It’s the humans.
But if the technology is moving that fast, what’s actually holding things back?
The washing machine was commercialised in the early 1900s. It took until the 1970s for household ownership to reach 90% – held back not by the technology but by infrastructure gaps, distrust of the machine’s cleaning ability, and the entrenched social systems around the old way of doing things. AI faces strikingly similar constraints. Only 12% of workers today fully trust AI to understand their specific work context. 60% fear their colleagues will think less of them for using it. BCG’s 10/20/70 framework captures the split: 10% of the value from AI comes from the AI itself, 20% from the technology and data infrastructure, and 70% from the workforce change that has to happen around it. We are asking two fundamentally different questions when we talk about invention and innovation – “can we build it?” and “will people actually adopt it?”
The operators at Scope confirmed this from the inside. At one company, token spend grew 62% in two months – but the harder shift was getting twenty-year veteran developers to accept that the way they’d always worked was no longer the standard. At another, the transformation was framed as moving every department from “human as the builder” to “human as the architect.” Across the financial institutions that panellist Didem Ün Ateş advises on AI implementation, the biggest blockers are almost never technical — they’re governance, trust, and the gap between what leadership says it wants and what the organisation actually rewards. And as Alex Depledge, serial entrepreneur and government advisor, put it in a fireside: it’s managing the transition that is the real problem, not the technology itself.
The consensus across every session was clear. The companies making real progress haven’t just adopted AI tools: they’ve redesigned their organisations around them, with leadership personally involved and a culture where it’s safe to experiment and fail.
When AI raises the floor, the difference is humans.
So if technology is commoditising and adoption is the real constraint, what separates the companies that will survive this Cambrian window from the ones that won’t?
People. The data now challenge the repeat-founder orthodoxy – previously successful serial entrepreneurs were the strongest predictor of venture success, and that’s no longer the case. The best founders are getting younger. They’re attending top research institutions at higher rates. And the dial that’s gone up most in how investors assess them isn’t solely on technical depth – it encompasses intellectual honesty, adaptability, and the willingness to relearn.
This isn’t abstract. One founder at Scope described hiring an AI agent as his chief of staff: given the same context as human advisors, the agent’s strategic recommendations were comparable. His cognitive load dropped. His decision-making improved. But the takeaway wasn’t that AI replaces the founder, it’s that it amplifies them. When emotional ownership meets AI-augmented capability, the gap between founder-led and non-founder-led companies widens. Founders matter more in this era, not less.
The selective pressure is on.
The Cambrian explosion produced extraordinary diversity, and then most of it went extinct. The same pressure is arriving now. The companies that emerge from this moment will be deeply embedded in core business processes, will compound value through proprietary data and context, and will be led by people who combine a builder’s instinct with the honesty to keep adapting.
The next generation of global B2B software leaders is being built right now, and a meaningful number of them will be European.
Deeper write-ups from each Scope 2026 session are coming over the next few weeks.
Check out more pictures and information around Scope 2026 here: www.oxx.vc/scope-2026
