cd ~/
Agentic AIMulti-Agent SystemsMCPPlatform ArchitectureInfrastructure

Every Platform Is an MCP Server Now

MCP hit 97 million monthly downloads and Salesforce just went fully agent-native. The integration layer for agents is decided.

Evyatar Bluzer
3 min read

The integration problem for AI agents is effectively solved. MCP won.

The Numbers

Model Context Protocol hit 97 million monthly SDK downloads in March. That's up from 2 million at launch in November 2024. One of the fastest adoption curves in open-source history.

But raw downloads don't tell the full story. What matters is who's building MCP servers - and as of last week, the answer is everyone.

Salesforce Just Went Headless

At TDX on April 15, Salesforce announced Headless 360. The entire platform - every data object, workflow, and business logic block - is now exposed as APIs, CLI commands, and 60+ MCP tools. Coding agents like Claude Code, Cursor, and Codex get native Salesforce access through the protocol.

This isn't an integration layer bolted on top. Salesforce rebuilt itself to be operated by agents, not humans clicking through a browser.

They're not alone. Databricks rebranded its AI Gateway as Unity AI Gateway the same week, wrapping its existing governance model around agent access to LLMs and MCP servers. The A2A Protocol crossed 150 supporting organizations in April. The Linux Foundation's Agentic AI Foundation now stewards both MCP and A2A.

What This Means for Builders

I've spent the last year wiring agents to enterprise systems. The integration tax was brutal - custom API wrappers, auth flows, brittle glue code for every platform. That era is ending.

The practical shift:

  • Tool discovery is standardized. Your agent discovers MCP tools the same way regardless of whether it's talking to Salesforce, Databricks, or your internal services. No custom code per platform.
  • Governance follows the protocol. Permissions, auditing, and policy enforcement wrap around MCP calls at the gateway level. You don't build security per-integration anymore.
  • Composability is real. An agent can chain tools across platforms in a single execution - query CRM data, run a model, write results back - all through the same protocol layer.

The second-order effect is bigger. When every platform speaks MCP, your agent's capability surface expands with every new MCP server that ships. You don't build integrations. You gain them.

The Deeper Pattern

Every major platform shift follows the same arc: fragmentation, then protocol convergence, then explosion of higher-level capability. HTTP did it for the web. REST did it for services. MCP is doing it for agents.

We're past the convergence point now. The debate about which protocol wins is over. The interesting work - the part I'm focused on - is what you build when the plumbing is no longer the problem.

Comments