Stop Buying Platforms. Start Building Ecosystems.

There was a moment in Karim’s talk that nailed something I’ve been thinking about for a while. He described AI’s impact on development as an hourglass: the middle part—actually writing the code—has gotten dramatically faster. But requirements on one side and testing on the other haven’t shrunk at all.

His math was blunt: you’re not saving ten hours down to one. You’re saving ten hours down to six. The coding might take one or two hours now instead of four. But the remaining time is still requirements, testing, scaling—the work that AI can’t compress because it’s fundamentally about understanding your business, not generating syntax.

For the enterprise, this means the promise of “AI makes everything faster” doesn’t hold up that cleanly. If you can nail your business requirements and your technical requirements, you can speed up development significantly. But rushing something live still doesn’t work for enterprise organizations. Anyone who has been in this space long enough knows that.

The part of Karim’s keynote that hit hardest was about data ownership. In his words, the SaaS verticalization of the last twenty years has created “dead ends of access to data.” Your data lives in silos—each vendor’s silo—and you can’t get to it in any meaningful way.

The most forward-thinking enterprise clients—and Karim referenced Crowd Favorite’s own work with a large technology client in California—are thinking differently. They want to create their own data lakes. They want a source of truth that spans their systems, not one that’s trapped inside a vendor’s product.

The next step is what makes this real: private, custom LLMs that can mine that data and find patterns across all of it. You can’t do that when every tool in your stack is trying to verticalize your data into its own ecosystem. You can’t do that with legacy systems that have APIs bolted on as an afterthought.

This isn’t a future vision. This is what’s happening right now at the organizations that are thinking past the current hype cycle.

Karim’s word for what enterprise teams actually need was simple: foundation. A solid, reliable foundation that you build on top of—not a platform that tries to do everything and locks you in.

On open source, he was direct: some projects have tried to over-innovate in the last few years and created problems for their users. Others have stayed stagnant, clinging to backward compatibility without being willing to take the risks necessary to move forward.

TYPO3 v14 LTS, which Benni Mack presented right after Karim’s keynote, represents that middle path—a foundation built on more than two decades of refinement that has already solved problems like internationalization, scalability, and multi-site management that newer platforms are still working through.

But the metaphor that actually landed was this: think of a content management system not just as a content system—think of it as an operating system. One that lets you plug in fast-moving innovations and move quickly but safely. An evergreen platform that doesn’t need to be constantly migrated but can be built upon, extended, and adapted.

The pressure isn’t just on commercial vendors. Open source has the same problem now. As the CEO of the agency that brought WordPress to Fortune 50 companies, Karim has standing to say it—open source platforms that have taken private equity in the last decade are starting to show the same patterns as commercial vendors. Rushing features to market before they’re ready. Chasing growth metrics instead of stability.

TYPO3’s governance model—how releases are organized, how the project is funded, how decisions are made—is different. I’ll cover that in more detail in a separate post based on Jam’s talk at the Summit. But the short version: not all open source is governed the same way, and the governance model is what determines whether a project stays aligned with its users or starts drifting toward other interests.

Karim’s keynote was about time horizon. Not what platform should we pick for 2026, but what foundation will still work in 2036. What he called an “evergreen platform”—something you don’t have to keep migrating, but can keep building on.

At b13, we’ve built on TYPO3 for over 20 years—and Karim’s framing is exactly how we think about enterprise projects. The best investments aren’t in platforms; they’re in foundations that give you the freedom to move.

If you’re thinking about your content management strategy with a longer time horizon, we should talk. And if you want to understand how TYPO3 fits into the enterprise ecosystem conversation—the summit in Atlanta was just the beginning.