The Opportuna Assembler #3

Welcome to The Opportuna Assembler, a collaboration between 3dot Insights and Opportuna, where we break down the most important shifts in software in under 1000 words.

In this edition we cover:

1. Model Context Protocol: The Game-Changer for Agentic AI – examining why MCP is becoming the “USB-C” of AI tools

2. AI Rails and Revenue Signals – looking at what Q1 earnings say about infrastructure, enterprise spend, and public sector tailwinds

You can sign up for the newsletter here and get regular updates of where we are focusing our investing here. Want more insights between issues? Follow us on LinkedIn.

Model Context Protocol: The Game-Changer for Agentic AI

Remember when every device had a different charger? That’s the headache agentic AI faces when trying to connect to tools and data. Enter the Model Context Protocol (MCP)—the “USB-C” of AI integration, —which is designed to let agents plug seamlessly into GitHub, Slack, calendars, and beyond.

What is MCP?

MCP is an open standard that allows AI agents to interact with tools through a shared interface. Think of Star Trek’s universal translator. MCP Servers “live” inside apps and generate natural language summaries; while MCP Clients embedded in agents interpret those summaries and take action.

Brittle custom integrations become a thing of a past, you —just “plug and play”.

Why It Matters

MCP unlocks three core superpowers for agentic AI:

• Live Context: Agents tap real-time data instead of stale training snapshots. A sales agent, for example, can adjust strategy on the fly using live CRM inputs.

• Tool Chaining: Agents can orchestrate multi-step workflows. In one Anthropic demo, an agent spun up a GitHub repo, committed code, and notified the team via Slack—autonomously.

• Actionable Language: Instead of raw data, agents receive summarized insights (“Your 2 PM meeting clashes with a client call. Reschedule?”).

But Is It Secure?

Yes, —MCP comes with enterprise-grade guardrails:

• Role-based access allows IT teams to define who (or rather what) can do what.

• Audit trails log every action an agent takes, useful for compliance and debugging.

• Rate limits prevent runaway API usage and abuse.

Backers include Microsoft and OpenAI, who are both building MCP support into their ecosystems.

What’s Next

The protocol is evolving fast. Expect:

• An explosion of connectors across SaaS and system tools

• Auto-discovery, where agents proactively suggest new tools

• Multimodal support, enabling interaction with video, images, and sensors

The Bottom Line

MCP turns AI agents into real teammates. With a standardized, secure interface, they can navigate tools, workflows, and environments—bridging the gap between models and the real world.

AI Rails and Revenue Signals: Key Takeaways from Enterprise Software Earnings

AI Infrastructure Build-Out Continues

Hyperscalers are still in capex mode. Meta raised 2024 guidance to $64–72 billion (up from $60–65B), citing internal demand: “Even with the capacity we’re bringing online in 2025, we are having a hard time meeting compute demand.” Microsoft flagged similar pressures. Google and Amazon also noted AI-driven capacity constraints and plans to add infrastructure in H2.

Enterprise IT Spend: Platforms Favored

These earnings calls offered a window into post “Liberation Day” enterprise sentiment. Infrastructure spend remains strong, but software vendors are navigating tighter budgets. Sales cycles are elongating. CIOs are drawing clearer lines between strategic and discretionary spending. That favors platform vendors with embedded usage and quantifiable ROI.

Microsoft had one of its best post-earnings stock reactions in a decade.

SAP reiterated full-year guidance.

ServiceNow nudged its subscription forecast slightly higher—and jumped 20% but, case-specific vendors like Confluent saw sharp drawdowns (-18%).

Efficiency is the new north star. As Satya Nadella, CEO of Microsoft, put it:

Public Sector: Signals, Not a Surge

Public sector demand is real but measured. U.S. federal budgets continue to rise—especially in defense, cybersecurity, and infrastructure. But procurement cycles are uneven, with a mix of bridge contracts and scrutiny. AI is seen as an enabler rather than a budget line item. Managed services tied to outcomes—especially in modernization, data, and security—continue to see traction.

Infrastructure investment is expanding, particularly in the U.S. and Middle East. These projects are long-dated but offer revenue visibility for well-positioned contractors. The opportunity favors incumbents aligned to strategic priorities, not breakout players.

The Bottom Line

AI infrastructure investment is ahead of plan. Software demand is diverging: platforms are winning on urgency and operating leverage, while point solutions face pressure.

The signals are clear: agentic AI is shifting from concept to capability, and the infrastructure race is intensifying. This is something we are seeing across both the public and the private sectors.

Protocols like MCP are not just technical plumbing. They are the connective tissue powering a new generation of adaptive, context-aware systems.

Meanwhile, enterprise software is entering a phase of selective momentum, where platforms with embedded value are breaking away from the pack.

As always, the shape of the future is being assembled one signal at a time.

— The Opportuna Assembler

If you are looking for some deeper dives into the convergence of private and public tech investing, take a look at our Klarna series.

 Was this email forwarded to you? Sign up here.