There’s little doubt that Microsoft Azure has been on a tear, growing 120% year-over-year in the last quarter and mounting the only credible competition for Amazon Web Services. That’s the good news.

The bad news is that this could be costing Microsoft in terms of its profitability.

As noted by Steve Nellis, Microsoft’s continued success with Azure could result in profit margins that are “a lot lower than what Microsoft’s investors have been used to for the past few decades.” Microsoft doesn’t really have a choice, but the transition will be painful, all the same.

The pains of transition

Microsoft is a gargantuan cash machine, generating billions of dollars in profit each quarter from its Windows and Office businesses. Anyone who has worked for a software company with a commanding lead in a given segment knows that these profits can persist for years, even as new license revenue fades.

SEE AWS and Microsoft Azure at the center of IT’s most exciting change in decades (TechRepublic)

Usually, this license revenue gets gobbled up by new competitors. In the case of Microsoft, however, it has been cannibalizing itself. To its credit, Microsoft has begun competing with its tried-and-true Office and Windows businesses, introducing cloud-based alternatives and splitting up its reporting into three groups (Productivity and Business Processes, Intelligent Cloud, and More Personal Computing) to help shield these burgeoning businesses from some of the harsh glare of Wall Street’s eye.

“Intelligent Cloud” includes Microsoft’s successful Azure platform revenue, but it also includes older (and very profitable) server tool businesses. Burying Azure in the midst of more profitable products is critical for giving Azure some breathing room as it grows in strength, but it hasn’t stopped Microsoft’s overall profitability from declining. Consistently. For years.

From 80.16% in June 2010 to 64.70% percent in June 2015, Microsoft’s profitability has declined every year for the past several years. While there are a range of scapegoats, cloud is arguably the most persistent.

Losing by winning

The world is moving to cloud computing, and Microsoft can either opt to sit on its cash cows and wait for oblivion or it can embrace the new normal (and lower profitability) of cloud computing. Microsoft, to its great credit, has chosen the latter course.

But, in so doing, Microsoft has set itself up to lose profitability even as it wins market share and revenue. Cloud, in a way, is the gift that keeps on taking, not giving. As Nellis writes, “As the newer, less-profitable services [within the Intelligent Cloud business] become a bigger portion of its Azure group, it’s likely to end up with margins a lot closer to AWS.”

The problem, however, is that Amazon thrives on the lower margins of cloud, and constantly wants to push those margins even lower.

But wait! It gets worse.

Amazon, for its part, is using this relatively thin profitability to build “fat” infrastructure, as Andreessen Horowitz’s Benedict Evans captures:

Evans then asks, “The question to ask isn’t whether Amazon is some profitless ponzi scheme, but whether you believe Bezos can capture the future.”

SEE AWS profit keeps feeding the Amazon beast (TechRepublic)

Based on operating profits of 36.4%, the answer seems to be “yes,” and a yes that will continue to crimp Microsoft’s own profits even as it successfully competes.

Microsoft’s profitable future

And yet, Microsoft is doing exactly what it should be doing: Winning. No, it isn’t yet displacing AWS, but Microsoft has quickly become a strong competitor to AWS, one that CIOs expect to invest heavily in for years to come.

The difficult thing will be to keep winning even at the cost of profitability. Different business models and levels of profitability will change how the Redmond giant invests in and benefits from innovation.

In short, we should expect Microsoft to look very different in five years than it looks today, precisely because it’s winning in the market. Not many companies survive a transition of this scale, but Microsoft appears to be the exception to the rule.