by Grego
I want to be clear before I start: this is not a nostalgic lament for the web of old. I’m not here to defend blue links or mourn the death of PageRank. What happened at Google I/O 2026 on May 19th is something more structural, and it deserves a strategic reading with a clear head — not a philosophical one.
Google just announced that it’s building a closed abstraction layer over infrastructure it doesn’t own. And the developer and publisher community in our region is treating it, for the most part, like another product launch.
It isn’t.
What really happened at I/O 2026
Three announcements, read together, tell the real story.
AI Mode and the new search engine. Google described it as “the biggest change to its search engine in 25 years.” The redesigned interface doesn’t return links — it immerses users in interactive experiences generated by AI with personalized visualizations, tools, and background agents. Search director Elizabeth Reid called it “end-to-end AI search.” AI Overviews already reaches 2.5 billion monthly users. AI Mode surpassed 1 billion monthly users, with queries doubling every quarter. Clickless searches now represent 60% of all queries. Of every 10 people who used to click through to your site, 6 now never leave Google’s interface.
Antigravity 2.0 and background agents. Google’s agentic development platform isn’t just a coding tool — it’s a work orchestration layer that can launch specialized sub-agents, execute tasks across browsers and terminals, and verify its own results. More relevant to this discussion: “information agents” will continuously scan the web on behalf of the user and deliver synthesized updates without the user ever opening a tab. The web becomes a data source to be consumed, not a destination to visit.
WebMCP. This is the one to watch most closely. Google proposed an “open web standard” that allows sites to expose structured tools — JavaScript functions, HTML forms — directly to browser-based agents. The pitch is openness. The reality is that Google proposes the standard, integrates it into Chrome, and controls the harness (Antigravity) that most developers will use to interact with it. That’s not an open standard. It’s a platform play disguised as standards language.
The numbers behind the thesis
The traffic collapse isn’t a prediction. It’s already happening.
Google search traffic to publishers dropped approximately a third according to Chartbeat data cited by the Reuters Institute. HubSpot estimated it lost between 70 and 80% of its organic traffic. Chegg reported a 49% decline. DMG Media documented drops of up to 89% on some queries. NPR called it “an extinction event” for online news publishers. The organic click-through rate at position one for queries where AI features appear dropped from 27% to just 11%, according to SISTRIX data from March 2026.
These numbers are from before I/O. The new Search — generative UI, information agents, the redesigned search engine — rolls out this summer.
The argument I’m not making
I want to be precise about what the thesis is and isn’t.
This isn’t about AI being bad. Antigravity is technically impressive. Gemini 3.5 Flash outperformed larger models on most benchmarks. The developer tools announced are genuinely useful for those building software.
The argument is narrower: Google is using its control of the browser (Chrome), the default search (AI Mode), and the agent harness (Antigravity) to insert itself as a mandatory intermediary between users and the web. Your content, your product, your documentation — feeds the system as raw material. Whether you receive credit, traffic, or revenue for it is Google’s exclusive decision.
The web was built on a simple contract: publish something useful and people can find it and reach you directly. Google is renegotiating that contract unilaterally. Not through negotiation. Through infrastructure.
What this means if you’re a CTO or engineering leader in Iberoamerica
First: your SEO strategy is obsolete. Not in decline — obsolete. Any content strategy built around ranking for keywords and capturing organic traffic operates on a model that Google is actively dismantling. The benchmark resets. Teams still measuring against 2023 traffic numbers are misinterpreting their own results.
Second: your documentation strategy needs to change. If you’re building developer tools, a SaaS product, or any technical platform — your documentation, your API references, your how-to guides will increasingly be consumed by agents rather than humans. That changes how you structure information, how you version it, and how you think about discoverability. WebMCP is Google’s answer to this; it may not be the only answer, but ignoring it is not a strategy.
Third: dependency risk is real. For years, “don’t build on another platform” was good advice that most companies ignored anyway. The I/O announcements are a reminder of why that advice existed. If your distribution runs through Google — and for most web properties in Iberoamerica, it does — you’re more exposed today than you were a week ago. Direct audience relationships (email subscribers, registered users, repeat visitors) are no longer a nice-to-have. They’re the hedge.
Fourth: this is a governance question for developers. WebMCP being proposed as an open standard by the company that controls Chrome and the dominant default search isn’t the same as being an open standard. The developer community in our region has influence — collectively — over how these standards are shaped. That requires paying attention and participating, not just adopting.
The uncomfortable conclusion
Google built its business on the web. It indexed it, made it navigable, and generated enormous value for itself and, incidentally, for publishers and developers. For a long time, that was a reasonably fair trade.
What’s changing is the direction of value extraction. The web is being repositioned as the training set and data source for an AI layer that Google controls. Your work still matters — as raw material. Whether it matters as something you can share, monetize, or build a business on is another question, and the I/O 2026 announcements suggest the answer is, increasingly: not necessarily.
The engineer in me respects the technical ambition. The strategist recognizes a consolidation play when he sees one.
The question for every tech leader in Iberoamerica reading this isn’t whether Google’s direction is good or bad for the web. It’s whether you’ve already adjusted your strategy to reflect a world where Google is no longer a discovery mechanism that brings you traffic — but a destination that consumes your content to serve its own users.
Because that world already started. I/O 2026 just made it official.
Sources: tante.cc (May 20, 2026), The Next Web, TechCrunch, Winbuzzer, Google Developers Blog, SISTRIX, Reuters Institute, Chartbeat.