WordPress must orient itself toward AI within five years or cede the web to platforms that did

WordPress holds a substantial share of the published web today. Jason Calacanis’s argument is that this is exactly the kind of fact that can lull a platform into the wrong strategic posture. His call is blunt: “This company needs to point WordPress in a forward direction, to think about how you take advantage of what’s going on in AI and become a relevant player in the next 5 years. Otherwise, you won’t be.”

That is a falsifiable claim with a named company, a named technology category, and a named time horizon. It is not a general observation that AI matters for the industry. It is a directed forecast about a specific organization’s window for action, and it closes somewhere around 2030.

Calacanis’s argument is not that WordPress is broken today. It is that the current moment represents a directional choice, not a maintenance question. Platforms that orient themselves toward AI’s capabilities while those capabilities are still forming will accumulate structural advantages that latecomers cannot easily replicate. Platforms that treat AI as a feature layer added onto an existing architecture will fall behind organizations that treated it as a premise from the start. In Calacanis’s framing, Automattic has time to make the right move, but that time is finite and running.

The distinction between a product roadmap problem and a strategic identity problem sits at the center of what makes the call substantive. A product roadmap problem gets resolved by shipping features, adding integrations, and publishing statements about the company’s AI vision. A strategic identity problem requires the organization to decide what it is for in a world where content creation, site management, and developer tooling are all being restructured by the same underlying shift. Those two problems have different solutions, different timelines, and different consequences for getting the diagnosis wrong.

Installed base is a trailing indicator. It measures where organizations have been, not where they are heading. The metric that will determine whether WordPress remains the default publishing layer for the web is not how many sites run on it today. It is whether the tooling, the workflows, and the architectural assumptions that define the next generation of web publishing are built around WordPress or built around something else. That competition is happening now, while the standards are still forming, which is precisely why Calacanis’s five-year framing carries weight. Platforms that wait for the new standard to crystallize before repositioning will find that the repositioning cost has become prohibitive.

The call is also worth taking seriously because it targets a company with real resources and a dominant installed base. Calacanis is not predicting the failure of a marginal player. He is arguing that scale and history are insufficient substitutes for strategic direction, and that a platform which powers a large share of the web can still lose the next version of that web if it fails to earn its place in it. The installed base that represents WordPress’s strength today is also the thing that can make the urgency feel less acute than it is, because the consequences of moving slowly will not show up immediately in the numbers that organizations track quarter to quarter.

What makes the forecast testable is precisely its specificity. By the early part of the next decade, the evidence will either support Calacanis’s call or refute it. Developer attention, new site decisions, the tooling that AI agents reach for when they need to publish to the web: these are the leading indicators that will show whether WordPress moved early enough and decisively enough to secure its position in an agentic publishing environment. The five-year clock is when the draft closes, and the choices Automattic makes in the near term are the ones that will determine which way the answer goes.