For more than a decade, Satya Nadella’s tenure at Microsoft has been held up as one of the great corporate turnarounds of the modern technology era. He rescued the company from the strategic drift and cultural sclerosis of the Steve Ballmer years, refocused it on cloud computing, embraced open source and transformed Microsoft into a growth juggernaut once again.
The numbers speak for themselves. Under Nadella, Microsoft created extraordinary shareholder wealth, with its market capitalisation soaring and its relevance restored across enterprise IT. Azure became a genuine rival to Amazon Web Services. Microsoft rediscovered its ability to ship software people actually wanted to use. For years, it felt like Nadella could do no wrong.
But recent events raise an uncomfortable question: has the Nadella era peaked?
Microsoft’s second-quarter financial results last week, which rattled investors and sent the share price sharply lower, suggest that the market is becoming uneasy.
In the quarter, Microsoft beat market expectations with revenue of US$81.3-billion and net income of $38.5-billion, but investors focused less on the earnings surge and more on the cost of getting there.
Azure cloud revenue grew 39%, in line with estimates but slightly slower than the previous quarter, while capital expenditure linked to AI and data centre expansion jumped to $37.5-billion – well above analyst expectations. As a result, Microsoft’s shares fell sharply, dropping about 10% and marking their steepest one-day decline since March 2020.
The sell-off reflects growing concern that Microsoft may be overspending on AI infrastructure before demand and profitability are fully proven, the Wall Street Journal reported (paywall) last week. Executives acknowledged that limited AI hardware availability is constraining Azure’s growth, even as demand continues to exceed supply.
Running ahead
Even so, investors are increasingly wary of how much of Microsoft’s cloud growth depends on AI – and whether the massive data centre buildout can be justified if AI adoption spreads more slowly than anticipated.
Nadella himself has cautioned that AI’s real-world impact is still lagging its technical capabilities, reinforcing fears that the sector could be running ahead of sustainable returns.
Microsoft is no longer the world’s most valuable company. In fact, it is no longer even in the top three. That alone does not define success or failure (Microsoft continues to be a highly successful enterprise, as the latest earnings attest), but it does signal a shift in sentiment – and sentiment matters for a company that has spent aggressively on a promise about the future.
Read: Big Microsoft price increases coming in 2026
That promise is artificial intelligence. And increasingly, it looks like Microsoft may be overplaying its hand.
There is no doubt that AI is a transformative technology. Large language models, generative tools and AI-assisted software development will change how work gets done. Microsoft was smart to partner early, and to integrate AI capabilities across its stack.
But the problem is not that Microsoft is investing in AI. The problem is that it is injecting AI into absolutely everything, whether users want it or not.
Copilot is now in everything, from Word, Excel and Outlook to Teams, Windows, Edge and seemingly any surface Microsoft can find. It is starting to feel less like thoughtful product design and more like a frantic attempt to justify the billions being poured into AI infrastructure. We do not need Copilot in everything from SwiftKey to Minesweeper. AI should be available when you want it, not shoved relentlessly down our throats.
Indeed, forced adoption is not adoption at all. If customers are not pulling these tools into their workflows because they genuinely add value – and are instead being nudged, nagged or defaulted into them – Microsoft risks inflating usage metrics without building durable demand. That is a dangerous game when you are spending tens of billions of dollars on data centres to support those services.
Which brings us to capital expenditure.
Microsoft’s AI-driven spending spree has been breathtaking. Data centres, GPUs, power contracts, networking infrastructure – the company is spending like a business convinced that exponential demand is inevitable. Investors, however, are beginning to ask the more difficult questions: what if that demand does not materialise at the scale or speed being assumed? The future trajectory of the US stock market depends on the answer to that question.
The market reaction to Microsoft’s recent results suggests that investor concerns are growing. Slowing growth in some areas, combined with relentless capital intensity, is a toxic mix for valuation. Overspending on infrastructure ahead of proven adoption is how tech giants turn visionary bets into margin-destroying mistakes. Witness the dot-com boom and bust for evidence of this.
Windows
Then there’s Windows. Nowhere is Microsoft’s strategic confusion more visible than in its operating system using by hundreds of million of people daily.
Once the beating heart of the company, Windows today feels like more of an unloved afterthought. Windows 11 has been a particular disappointment. Compared to the polish of Apple’s macOS or the flexibility and performance of modern Linux distributions, Windows 11 often feels cluttered, inconsistent and bloated.
Instead of fixing the basics, Microsoft’s answer appears to be: add more AI!
Read: Valve’s Linux console takes aim at Microsoft’s gaming empire
This is exactly the wrong approach. Injecting Copilot into a flawed operating system does not make it better; it makes the flaws more irritating. Users do not want AI panels and pop-ups layered on top of an experience that already struggles with coherence. Windows needs a top-to-bottom rethink – a genuine overhaul of performance, UI consistency, reliability and system services.
Why is Microsoft not doing that?
The cynical – and probably correct – answer is that Windows is no longer where the profits are. Azure and enterprise services drive Microsoft’s financial engine. Windows is a distribution channel, not a growth business. Spending the kind of money required to reinvent Windows would not move the revenue needle in the way another hyperscale data centre might.
That logic makes sense on a spreadsheet. It makes less sense for the long-term health of the ecosystem. Windows is still the primary computing interface for office workers around the world. Neglecting its quality while stuffing it with AI features risks alienating the very base Microsoft depends on.
None of this criticism seeks to undermine Nadella’s achievements. Compared to the Ballmer years – a period marked by missed mobile opportunities, strategic arrogance and cultural stagnation – Nadella’s leadership has been extraordinary. He modernised Microsoft’s culture, embraced developers and open ecosystems, and made the company relevant again.
But great leadership eras do end. The most dangerous moment for any successful CEO is when yesterday’s winning strategy becomes tomorrow’s overreach.
Microsoft today feels like a company spending first and asking questions later. It feels like a company that has forgotten that restraint, focus and user trust matter as much as ambition.
The biggest gains of the Nadella era may already be behind us. That does not mean decline is inevitable – but it does mean Microsoft needs to slow down a bit and listen more carefully to its users.
Read: Vukani Mngxati appointed as CEO of Microsoft South Africa
There is little doubt that AI will shape the future of computing. But the future does not belong to companies that force technology onto their users to fit their business models. It belongs to those that deploy it where it truly belongs – and have the discipline to hold back where it does not. – © 2026 NewsCentral Media
Get breaking news from TechCentral on WhatsApp. Sign up here.
