Is Microsoft on sale? and Its Case in the AI Race

(This story appeared on my substack page around a week ago: https://open.substack.com/pub/piproberts25/p/is-microsoft-on-sale-and-its-case?r=6kjdcg&utm_campaign=post-expanded-share&utm_medium=web)

Last October, Microsoft was trading at around US$540 per share. Today it trades closer to US$385. For one of the highest-quality businesses in the world, that represents a meaningful re-rating.

Microsoft possesses some of the strongest competitive moats in global markets. Through Windows and Microsoft Office, it enjoys a level of market dominance that, in many enterprise environments, approaches monopoly-like characteristics. For countless businesses, Microsoft’s products are not merely convenient, they are essential. The cost and disruption involved in migrating away from its ecosystem are often so great that switching simply isn’t a practical option.

Beyond these flagship products, Microsoft has built a remarkably diversified business. Revenue comes from cloud computing, gaming, cybersecurity, advertising, developer tools, enterprise software, LinkedIn, and many other sources. This diversification supports enormous free cash flow, an exceptionally strong balance sheet, and one of the most recession-resistant customer bases in the market.

These strengths give Microsoft significant pricing power. The company has demonstrated its ability to increase prices over time, with most enterprise customers accepting the increases because the value provided far outweighs the cost of switching.

Over the past six months, a number of high-quality companies have experienced valuation compression. Whatever the reason, being able to purchase Microsoft on a price-to-earnings ratio of around 22 times feels unusual. Everyone knows Microsoft is a great business, so it is rarely cheap.

The AI Question

Artificial intelligence is the technology theme dominating markets today, and Microsoft’s primary vehicles in that race are Azure and Microsoft Copilot.

Perhaps Microsoft’s greatest advantage is not Azure itself, but its existing enterprise relationships. Nearly every large organisation already uses Windows, Office, Active Directory, Teams or some combination of Microsoft’s products. Selling AI into an existing customer base is considerably easier than convincing organisations to adopt an entirely new technology platform.

However, history suggests that being an early leader in a technological revolution is no guarantee of long-term success.

Commodore was one of the pioneers of the personal computer revolution. The Commodore 64 became one of the best-selling computers of all time and helped introduce computing into millions of homes. Yet Commodore International entered bankruptcy in 1994.

Nokia dominated the mobile phone market in the early 2000s. The Nokia 3310 remains one of the most iconic mobile phones ever produced, but the smartphone revolution left the company behind. Microsoft eventually acquired Nokia’s handset business in 2014, only to later write off much of the investment.

The same pattern has occurred throughout history. Hundreds of automobile manufacturers emerged during the early twentieth century, yet only a handful survived. Early aviation saw dozens of aircraft manufacturers disappear despite the industry’s enormous long-term success.

The lesson is straightforward: even if AI becomes one of the defining technologies of this century, many of today’s leaders will not necessarily remain leaders decades from now.

Why Microsoft May Be Different

Microsoft’s advantage is that it is not relying solely on winning the AI application race.

The company’s existing software ecosystem naturally places AI products directly in front of hundreds of millions of users. Copilot is already being integrated throughout Microsoft 365, Windows, GitHub and many of the company’s enterprise offerings, making adoption far easier than for a standalone AI company trying to acquire customers from scratch.

More importantly, Microsoft is also one of the world’s largest cloud providers through Azure.

This distinction matters.

Whether customers use Microsoft’s own AI models, OpenAI’s models, or even develop their own, many of those workloads still require cloud infrastructure. In that sense, Microsoft is not only participating in the AI race, it is also selling the “shovels.”

The IBM Case Study

IBM’s experience during the personal computer revolution provides an interesting comparison.

In 1981, IBM launched the IBM PC, establishing what would become the industry standard. Its open architecture allowed dozens of manufacturers to build compatible machines, creating fierce competition and steadily reducing hardware margins.

Ironically, Microsoft became one of the greatest beneficiaries of the PC revolution by supplying the operating system rather than competing directly in hardware. It was, in effect, selling shovels during a gold rush.

By 2005, IBM had exited the personal computer market altogether, selling its PC division to Lenovo. The business it helped create had become commoditised, but IBM survived because it had already shifted its focus toward higher-margin enterprise software, consulting and infrastructure.

Even though IBM’s venture into the PC industry did not lead to the long-term success enjoyed by the likes of Apple, the fact that it was never its only income stream meant that its long term survival didn’t rely on the success of that one venture.

The Investment Case

Microsoft is certainly making an enormous bet on AI. The company is investing tens of billions of dollars into data centres, specialised chips and cloud infrastructure. These investments may take years to generate attractive returns, and if demand for AI ultimately falls short of today’s expectations, those returns could disappoint.

Governments and regulators are also paying increasing attention to AI, competition and data infrastructure, creating the possibility of higher compliance costs and additional regulation.

There are, of course, other risks. Microsoft faces ongoing currency exposure due to its global operations, integration risk from large acquisitions, and its share price is influenced by passive investment flows because it represents such a large weighting in many major indices.

Yet the investment case remains compelling.

If Microsoft’s AI strategy succeeds, shareholders should benefit from continued growth across Azure, Copilot and the broader Microsoft ecosystem.

If AI proves less profitable than markets currently expect, Microsoft’s downside appears more limited than many pure AI companies. Its established software franchises, enterprise relationships and recurring cash flows should continue generating substantial earnings regardless.

The greatest limitation may simply be Microsoft’s size. As one of the world’s largest listed companies, even exceptional business performance requires enormous earnings growth to materially move earnings per share.

That may ultimately be Microsoft’s greatest strength as an investment today. Investors are not relying on one speculative technology to justify the business. They are buying one of the market’s most consistently profitable businesses at a valuation that is considerably more reasonable than it has been for much of the past several years, while receiving a free option on AI success. This may even be underselling how Microsoft is positioning itself, which is as a toll road of enterprise AI. The infrastructure providers often earned extraordinary returns regardless of which applications ultimately succeeded.