In early 2026, when Microsoft announced it would invest $80 billion in AI infrastructure in a single fiscal year, many people read it as a conclusion — the story of a company that had already decided where to place its bets. But listen to the long interview Satya Nadella gave Ben Thompson of Stratechery around the same time, and a different scene emerges. Throughout the conversation, Nadella kept returning to one question from multiple angles: "What are we actually good at?" Investment scale aside, Microsoft was still refining its answer to where it should sit in the AI ecosystem.

A CEO deploying tens of billions of dollars while still refining his company's core competency — to me, that read not as a sign of anxiety but as a fairly honest admission. And inside that honesty is a clue that solo founders and one-person product managers should take for themselves.

Caught Between Partner and Competitor

The topic Nadella handled most carefully in the interview was Microsoft's relationship with OpenAI. Microsoft has invested billions in OpenAI, and OpenAI's models run on Azure. On the surface, it's a partnership. But as OpenAI signs more enterprise customers directly, it increasingly overlaps with Microsoft's Copilot and the Microsoft 365 product line.

Nadella didn't dodge the tension. He pointed to a direction — "we're not in the model layer; we're in the infrastructure and application layers" — but he was cautious about how stable that boundary really is. Microsoft's surest bet right now is inference infrastructure. The math is simple: the more enterprises use AI, the more demand grows for the data centers that handle that compute, and that flows into Azure revenue. Microsoft's current position is to lay the floor that software and AI agents run on.

Where Nadella sounded far less certain was on the concept of the "agentic platform." If AI evolves beyond a simple assistant into something that performs work directly, the role that existing software like Windows and Office plays inside that world changes. Nadella said this would reshape the competitive landscape of the next decade — and yet, on what it would concretely look like, he said they were "still experimenting." When the CEO of the company with more AI resources than anyone on earth says he doesn't yet know the shape of the future platform, that's a signal: this territory is still open.

The Question That Arrives After the Tools Pile Up

One way to read this interview is as a big-company strategy story and move on. By that logic, solo founders and one-person PMs should just grab tools like Notion AI, Perplexity, and Claude and start using them — no strategic agonizing required. And there's real merit to the argument that as AI tools become more accessible, execution speed matters more than strategy. The experience of downloading one tool and seeing your productivity change that same day is real.

But Nadella's struggle points at the problem that arrives after that fast-execution phase has run its course. When the tools multiply and every tool can handle the same tasks, the question "why am I actually good at this work?" takes on a different weight. Microsoft has Azure for infrastructure, Office for productivity, GitHub Copilot for development assistance — and it is still asking where it should stand on top of them.

Once AI routinely drafts your content, organizes your data, writes your emails, and summarizes your meetings, the part a human contributes on top has to be defined far more narrowly and sharply. The same terrain opens up for solo PMs and one-person creators.

There's a story that veteran salespeople often tell. They worked diligently, but the deals didn't close. Looking back later, the problem wasn't a lack of effort. It was that they didn't know where they actually added value. Some people contribute by building relationships, others by interpreting information, others by reading the timing. When AI takes over a large share of those contributions, the person who doesn't know their real point of contribution won't convert new tools into results.

What to Do Before Layering On AI Tools

Pull Nadella's interview down to the level of a solo business, and it yields questions you can examine right now.

Which parts of your work is AI already handling? As AI adoption spreads, the line sharpens between tasks that can be processed quickly and tasks that demand human judgment, trust, and experience. Locating that line is the starting point.

Next, ask where your results actually come from. If you're a creator, you should be able to distinguish whether readers respond to your topic selection, your information synthesis, your voice, or your publishing timing. Which of those AI can already replace — and which it still can't — is what defines your point of contribution.

One thing Microsoft stays conscious of in its relationship with OpenAI is that binding yourself deeply to a single partner narrows your strategic flexibility. The same goes for a solo business: the moment your entire workflow is locked into one tool, that tool's policy change or price hike affects your whole operation. Whatever tools you use, staying mindful of substitutability is a preparation that holds regardless of scale.

The question Nadella is still refining while deploying $80 billion is, in fact, the most basic one. Stacking up tools first and figuring out what you're good at first are two different sequences of work. When Nadella says they're "still experimenting," it's a signal that the question remains open. The preparation a solo PM can make is to write their own answer to that open question — before anyone else answers it for them.