Hello, dear readers!
This is our weekly brief on remarkable AI topics, so you can keep up with what the machines — and their investors — are plotting next.
Today's focus — Meta's quiet bet on the "agentic web." The company has acquired Moltbook, a social network built not for humans but for AI agents. If AI systems start acting on our behalf across the internet, will they also need places to hang out?
Also in this week's edition:
Coca-Cola is weaving AI deeper into marketing and product development, using data and generative tools to shape campaigns — and even flavor ideas.
And new data from Andreessen Horowitz suggests the global AI app ecosystem is increasingly clustering around three major hubs: the United States, China and Russia.
Meta Thinks AI Agents Will Need a Social Network
Meta's acquisition of Moltbook — a small social network designed for AI agents — initially raised eyebrows. Why would one of the world's largest social media companies buy a platform where the users aren't people?
The answer may lie in Meta's longer-term vision of the "agentic web." As AI systems become capable of acting on behalf of users — booking services, negotiating transactions or coordinating tasks — the internet may start to host not just human users, but large numbers of software agents interacting with one another.
In that world, agents will need infrastructure. They will need ways to discover each other, exchange information and coordinate actions across platforms — and plot our eventual demise, of course. A network designed specifically for AI systems could become the connective tissue that makes those interactions possible.
For Meta, this could extend far beyond an experimental side project. If agents begin to mediate large portions of online activity, the platform that hosts and organizes those interactions could sit at a powerful new layer of the internet — one that shapes commerce, services and, inevitably, advertising.
In other words, the next social network might not be for people at all.
Is It Real or Is It Coke
Coca-Cola is expanding its use of AI across marketing and product development, embedding the technology into everything from campaign design to consumer analysis.
Executives say the goal is to move toward what they describe as "persuasion-led growth" — whatever that means. Instead of relying mainly on traditional advertising, the company increasingly uses data and AI tools to analyze consumer behavior, test marketing content and adapt campaigns across markets.
The approach is beginning to influence product ideas as well. The limited-edition Coca-Cola Y3000 Zero Sugar, for example, was partly shaped by AI-driven insights gathered from online consumer discussions.
The shift signals how generative AI is moving from experimentation into everyday decision-making inside large consumer businesses — for better or for worse.
The Global AI App Race Is Splitting Into Three Ecosystems
New research from venture firm Andreessen Horowitz highlights just how quickly the generative AI app ecosystem is evolving. The firm's latest ranking of top AI products shows explosive growth across chatbots, image tools, coding assistants and AI-native productivity apps.
But the data also hints at a broader structural shift: the global AI landscape is increasingly clustering around three major ecosystems — the United States (OpenAI, Google), China (DeepSeek, Baidu) and Russia (Yandex).
Each region is building its own constellation of AI products, platforms and distribution channels, often shaped by local infrastructure and technology companies. While many applications still compete globally, the underlying ecosystems are becoming more distinct.
If that trend continues, the future AI market may look less like a single global platform race — and more like several parallel technology ecosystems evolving side by side.
Thanks for reading AIport. Until next Monday — by then, AI will almost certainly have launched three new startups, five new buzzwords and at least one product nobody quite understands yet. 🚀
