Napier & Son was the most successful British manufacturer of aircraft engines in the 1920s and 30s with their 12-cylinder Napier Lion powering 163 different types of aircraft between 1918 and 1935. Over that 17 year period the Lion grew from 450 to 1350 horsepower and was, for awhile, the most powerful aircraft, boat, and car engine in the world, holding world speed records in all three venues at the same time. And then the Napier Lion was suddenly gone — a lesson from which Microsoft’s Steve Ballmer could benefit if he and his company don’t repeat it.

Napier perfected their Lion engine over those 17 years, improving it in every way until it was the best and most efficient engine of its class in the world. Then, seemingly overnight, the class changed as air forces and record setters alike suddenly needed more than the 1,350 horsepower a finely-tuned Lion could deliver. Napier’s Lion gave way to Rolls-Royce’s larger and innately more powerful Merlin and Griffon engines and Napier, for all intents and purposes, was gone.

Napier milked its technical and market advantages for a little too long.

What does this have to do with Steve Ballmer and Microsoft? They are Napier, circa 1935 and their Lion is called Windows.

Windows 8 shipped last week to mixed reviews. Ballmer himself called it “a bold re-imagining” of Windows. It’s bold alright, but not bold enough. Windows is doomed.

We can argue all day about whether Windows 8 is better or worse than Windows 7 or even Windows 9, but the real issue here isn’t the software at all but the platform, by which I mean the desktop PC. Companies, governments, families, schools, and individuals are all buying fewer desktop PCs than they used to. Desktop growth has reversed and international desktop expansion is slowing as even that market matures. This year will probably mark Microsoft’s highest desktop sales ever in dollar volume, which sounds good, except that next year sales will be less as they will again the year after and every year past that.

Six years from now (four hardware generations) Windows will be dead. Or free.

And for all his bold re-imagining in New York last week, Steve Ballmer knows this, and that’s his dilemma.

Desktops are fading now, notebooks will be fading soon, both to be replaced by tablets and smart phones where Microsoft not only doesn’t dominate, they aren’t even among the major players.

Death of the desktop is clear not because Windows desktop sales are declining but because Macintosh desktop sales are declining. When Mercedes (Apple) begins to suffer declining unit sales, what does it mean for GM (Microsoft)? Not good.

The only option is to invent the future, which Ballmer and Microsoft are attempting to do by entering the tablet hardware business (again emulating Apple) and cutting bold smart phone deals with outfits like Nokia. But Microsoft, for all its posturing and $1 billion marketing budgets, isn’t any good at inventing the future and knows it. Ballmer lacks confidence that Redmond can invent it’s way out of the current hole. And because he lacks confidence, as does nearly everyone else at Microsoft, of course it won’t happen.

Microsoft didn’t invent the PC but benefited from its invention. Microsoft didn’t invent BASIC, they didn’t invent the PC operating system, they didn’t invent word processor, spreadsheet, or presentation applications, they didn’t invent PC games, they didn’t invent the graphical user interface, they didn’t invent the notebook or the tablet, they didn’t invent the Internet, they didn’t invent the music player or the video game, but they benefited from all these things.

Like Blanche DuBois, Microsoft has relied on the kindness of strangers.

Microsoft may have invented the smart phone. More on that below.

Having not invented any of the products it is known for, why should we expect Microsoft to invent its way out of declining markets? We shouldn’t.

Even video games are in decline and we now see Microsoft trying to turn its 30 million-strong xBox installed base into something like a cable TV network in order to milk that franchise beyond what would otherwise be its death.

Ballmer knows all this. And like Napier, he can keep building his old product line with a twist or two until the market drops out from under him or he can do some real re-imagining and turn Microsoft into a completely different company.

I don’t think he will do it, though, because I don’t think he can do it. Even if Steve Ballmer could envision a better future for Microsoft built on true technical leadership, I don’t think he or his company could follow-through. They are just making too much money doing the old stuff to truly embrace anything new.

Until it’s too late.

This does not mean Microsoft is going away. Their smart phone patents score them $15 for every new Android license of which there are 1.3 million every day. That’s $20 million per day ($7.3 billion per year) to Microsoft for doing, well, nothing.

What Steve Ballmer and Microsoft need to do is clean up their act, quietly trim expenses, maybe even sell a few product lines, and start to seriously stash away cash toward the post-Windows, post-Office world of 2018.

Yes, post-Office. What else can be meant by bundling Office with Windows RT than its value is headed to zero?

If Microsoft can continue to pretend it is big while actually becoming small, they might end up in 2018 with a small residual product line sitting atop $100 billion in cash. Then Ballmer can hand that money to Warren Buffett or to Buffett’s successor and let them manage Microsoft as a mutual fund rather than a technology company.

This is the only future I see for Microsoft because I think Steve Ballmer is a rational man, he understands this, and he sees himself as the only plausible steward for such a sneaky transition. Otherwise, simply as a huge Microsoft shareholder he would have long ago fired himself.

I think this is exactly what has been happening at Microsoft for at least the last 2-3 years, ever since the plan to buy Yahoo cratered.

Ballmer isn’t stupid and he isn’t deluded, he’s a man with a plan — a plan we’re just not supposed to know about yet.

Nothing else makes sense to me.