Microsoft Stock Is Down 28% From Its High. Is Azure the Way Back?

Microsoft shares are still about 28% below their 52-week high of above $550, set last October, despite some recent recovery. One analyst at Fishtown Capital argues the market has mispriced Microsoft's AI position and that strong Azure and Copilot numbers in upcoming earnings could shift the narrative back toward "AI winner" and push the stock above $500. The thesis rests on three pillars: a valuation near multi-year lows, heavy capital expenditure that mirrors Amazon's long-term playbook, and accelerating enterprise adoption of Copilot.

What happened

Data point	Detail
52-week high	Above $550, set last October
Current decline from high	Approximately 28%
Analyst price target reference	Back above $500
Comparable historical period	Late 2022, before a significant rally

Microsoft set a 52-week high above $550 last October and has since fallen roughly 28%. According to Fishtown Capital, a family office investor with a disclosed long position in the stock, the current valuation sits near multi-year lows. The analyst draws a parallel to late 2022, when the stock was similarly out of favour before a substantial recovery.

The core argument is that the market is treating Microsoft’s large capital expenditure programme as value destruction. Fishtown Capital disagrees, comparing the approach to Amazon’s years of heavy infrastructure investment that eventually paid off for shareholders.

Why it matters

Two products sit at the centre of the recovery thesis: Azure (Microsoft’s cloud computing platform) and Copilot (its AI assistant embedded across Office, Windows, and developer tools). According to the analysis, Copilot adoption is expected to pick up pace because large enterprises already run deep on Microsoft software and are reluctant to introduce outside AI vendors for security and compliance reasons. That stickiness, the argument goes, gives Microsoft a structural advantage that won’t show up fully in one quarter’s numbers.

For business operators who already pay for Microsoft 365 or Azure, this framing matters practically. If Copilot usage inside enterprises accelerates, Microsoft will likely raise per-seat pricing tiers and push AI features deeper into workflows. Teams that have not yet evaluated what Copilot does inside their existing stack may find the decision made for them by IT or finance departments within the next 12 months.

The broader point about CapEx is worth watching too. Microsoft is spending heavily on data centres and chips. If that spending produces measurable Azure revenue growth, the current price dip could look like a buying window in hindsight. If it does not, the spending becomes a drag on margins with no obvious catalyst to reverse it. Our coverage of Meta’s 550 MW gas plant investment shows just how material these infrastructure bets have become across the industry.

Our take

The thesis is coherent, but readers should weigh two caveats. First, the author holds a long position in MSFT and disclosed it. That does not make the analysis wrong, but it is a reason to verify the numbers independently rather than take the framing at face value.

Second, the “entrenched enterprise relationships” argument for Copilot cuts both ways. Yes, enterprises won’t rip out Microsoft. But that same inertia means they also won’t rush to pay more for AI features they haven’t fully evaluated. Copilot adoption has been slower than Microsoft initially signalled, and a single strong earnings print doesn’t resolve whether enterprises are actually getting ROI from it, a question our coverage of OpenAI’s “useful intelligence per dollar” framing touches on directly.

For businesses considering AI integration inside their own operations, the more useful question is not whether MSFT stock recovers, but whether Copilot actually reduces time or cost in your specific workflows before you commit to higher-tier licensing.

What to do about it

1. Run a Copilot trial on a contained workflow (meeting summaries, email drafting, or code review) before your next Microsoft renewal conversation.
2. Ask your Microsoft account manager what the per-seat cost difference is between your current plan and a Copilot-enabled tier.
3. Track Azure pricing changes at your next renewal; accelerating demand can push costs up even if the headline stock story is bullish.
4. If you are evaluating alternatives, compare Copilot against standalone tools on a task-specific basis rather than as a platform decision.

The stock story may resolve itself with one good earnings report. The product question, whether Copilot earns its seat licence cost inside your business, takes longer to answer and matters more.

Source: Bing News · GitHub Copilot