MCP and the Next Land Grab

Every few decades, something comes along that completely resets the board.

Not just a new product — a new platform. A new surface area where the early movers build empires and everyone else spends the next ten years trying to catch up.

We've seen it before.

The Pattern

When the internet went mainstream in the 90s, a wave of people saw it coming and built fast. Some of them built garbage. Some of them built Google. The point isn't that every early mover won — it's that every major winner was an early mover.

Then came the App Store. Apple opened it in July 2008 with around 500 apps on day one. Facebook was one of them. They didn't have to be — they were already a dominant website — but they showed up anyway, and ended the year as the most downloaded app on the platform. They understood what was happening before most people did. Get there early, learn the terrain, and figure it out.

That's the pattern.

So What's MCP?

In November 2024, Anthropic released the Model Context Protocol — MCP — as an open standard. If that sentence didn't make your ears perk up, let me try again.

The best analogy I've seen is USB-C. Before USB-C, every device had its own cable, its own port, its own headache. USB-C came along and standardized it — one connection type, everything works together.

MCP does the same thing for AI. Before MCP, if you wanted an AI assistant to interact with your calendar, your Slack, your database, your anything — you had to build custom integrations from scratch for every single combination. It was a mess.

MCP standardizes the connection. Build once, plug in anywhere. An AI agent can now read your emails, update a task in Asana, pull data from your CRM, and respond in Slack — all through the same protocol.

By February 2025 — three months after launch — over 1,000 open-source MCP connectors had already been built. OpenAI and Google DeepMind both adopted it. This thing moved fast.

The Shift Nobody's Talking About Loudly Enough

Here's the part I keep coming back to.

For the last two years, the race has been: who can build a product with AI in it? Slap a chatbot on your app. Add a "summarize this" button. Ship it, call it AI-powered, and see what happens.

That race is almost over. That bar is the floor now.

The next race is different. It's not about who built a tool with AI — it's about who made their tool work with AI.

Think about what that actually means. Instead of building a standalone AI product, the winning move is making your existing platform readable, controllable, and usable by an AI agent. You become part of someone's AI-powered workflow. You're not competing with AI — you're a node in it.

The companies getting there first are building MCP servers for their platforms right now. They're not waiting to see how this plays out. They're the Facebook-on-day-one crowd.

What This Looks Like in Practice

Picture this: someone sits down and opens Claude, or GPT, or whatever AI they use. They type: "Summarize everything from my email and Notion today, reschedule my afternoon meeting, and draft a reply to the client thread."

That's not science fiction anymore. That's what MCP enables — an all-in-one command layer where you talk to one AI and it reaches into all your tools.

The companies whose platforms support this workflow are the ones that survive the transition. The ones who don't? They become friction.

Nobody wants friction in a world where everything else just works.

Where This Goes

I don't think most people have fully internalized how significant this moment is. Not because MCP is complicated — it isn't — but because the implications are.

We're not just talking about smarter apps. We're talking about a wholesale reorganization of how people interact with software. The AI becomes the interface. The tools become the backend.

We've been here before. The internet. The App Store. The early movers built the map while everyone else was still asking for directions.

MCP is the new surface area. The connectors are being built right now. The question isn't whether this becomes the standard — it already is.

The question is whether you're building on it yet.

And here's the thing — we're still in the early innings. MCP is vastly underutilized right now. Most people haven't even heard of it, and the use cases being built today barely scratch the surface of what's coming. Industries, workflows, and entire business models are going to be rebuilt around this protocol, and most of the world is still asking "wait, what's MCP?"

That gap between where we are and where this is going? That's the opportunity.

Feel free to reach out — I love talking about this stuff.