Satya Nadella's Microsoft is not the Microsoft of Bill Gates and Steve Ballmer.

In some ways, that's because Gates and Ballmer missed some of the big trends in computing like, say, smartphones. Microsoft had such a dominant position in the PC era that it could fight competitors using Windows like a castle with high walls and many cannons. Their most infamous tactic in those days was dubbed "embrace, extend, and extinguish," whereby they incorporated into Windows the new features and new applications developed by rivals but then altered them in ways the rivals couldn't match. But playing rough led to lawsuits and eventually the big 1998 antitrust case.

In the mobile era, however, Microsoft has no such castle. Google and Apple control the operating system software on billions of phones and the high walls and many cannons belong to them (and they're raising the same kinds of competition and antitrust worries that got Microsoft in trouble). Nadella got the top job in part because he wasn't the exec who lost mobile and he wasn't tied to the fading power of the Windows division. He oversaw Azure, Microsoft's cloud computing platform. And Azure had to compete against Amazon Web Services, the OG of the cloud.

So as we start to see the outlines of the post-smartphone era take shape, how will Nadella compete?

I got a window into his strategy last week when Microsoft sent me one of its HoloLens 2 headsets and invited me to a virtual meeting of sorts.

Each of the three meeting participants was in a different city and wearing a HoloLens, and we were all connected over the Internet via Microsoft's just-made-public Mesh mixed reality service. Looking around my office, I could see cartoonish figures of the other two participants, or avatars, that moved around in real time to match how those people were moving back in each of their own offices. Before me was a virtual table with a couple of digital sea creatures floating in the air. I could reach out and move them around, examine them more closely, or stretch them bigger.

The amazing part was that as I did that, the other people in the meeting could see exactly what I was doing with the virtual jellyfish and shark. Everything looked a bit cartoonish, as I said, but the location tracking and the real-time movements were pretty spot on (at least until someone hit a wrong menu item and accidentally dropped out).

The key here isn't really the HoloLens hardware. In fact, Microsoft is making Mesh compatible with as much gear as possible, including Apple's Mac computers and Facebook's Oculus VR headsets. The Mesh service itself isn't what consumers or business users will use. It's more of a plumbing service that any app developer can use to build in these kind of virtual, real-time collaboration capabilities. The real magic is happening out on Azure's cloud servers, which are tracking everything and everyone in the virtual meeting space and calculating how to display them for each participant without any delays.

“It’s time for us to reflect on how the cloud will change over the next decade, and the innovation our changing world will require from the cloud,” Nadella said in his keynote address at March 2's unveiling of Mesh at Microsoft's Ignite conference.

That's smart for a couple of reasons. First, no one thinks Microsoft is going to win the race to develop the best mixed reality and virtual reality hardware. On March 2, after Microsoft publicly demoed Mesh with the HoloLens, PCMag's Sascha Segan compared it to Microsoft's ill-fated "Pocket PC" effort from 20 years ago. Analyst Michael Gartenberg was even more negative, running down the list of hardware categories where Microsoft was first but lost to Apple, including music players, phones, and smartwatches. And Apple is already working on its own VR devices.

Second, and more importantly, Microsoft's strength remains in developing software platforms and tools, particularly ones that can leverage the popularity of its still popular business software like Office and Teams (Teams is getting Mesh features pronto, Microsoft says). Amazon or Google may try to build something better than Mesh and Apple may try to ignore it or lock it out. The service only wins if Nadella and his team outmaneuver and out-innovate the pack.

Competing on product quality? How quaint.