Microsoft (MSFT) Keeps Dropping: Buy the Dip Now, or Wait for Clarity?

BNP Paribas just flagged Copilot upgrades and Azure catalysts as key tailwinds — but MSFT is still down roughly 33% from its peak. Here is what long-term investors need to think through before acting.

Why Is Microsoft Down So Much?

If you own MSFT, you do not need a chart to feel the pain. The stock hit an all-time high of $555.45 in July 2025. As of this week, it is trading around $373 — a drawdown of more than 33% in under a year. Year-to-date, it is down roughly 23%, putting it on pace for what could be its worst annual performance in nearly two decades.

Three forces have combined to drag the stock lower, and it is important to understand all of them before deciding what to do:

First, the capital expenditure shock. Microsoft spent $37.5 billion on capex in Q2 FY2026 alone — a 66% jump year-over-year. The company is building data centres at a pace not seen since the dot-com era. That spending is compressing free cash flow in the short term, which made investors nervous, especially those who bought the stock at premium valuations expecting expanding margins.

Second, Azure growth deceleration concerns. Even though Azure grew 39% year-over-year in the most recent quarter — a genuinely impressive number for a platform of that size — the market had priced in even faster growth. When cloud growth slows even slightly from a high base, institutional investors who modelled 42–45% growth tend to rotate out quickly. The mismatch between expectation and reality hit the stock hard.

Third, the Copilot monetisation question. Microsoft 365 Copilot has reached approximately 16 million paid seats, up 160% year-over-year. But its commercial install base sits between 415 and 450 million users. That means penetration is still below 4%. Investors who expected faster enterprise adoption have grown impatient, and competitors from Anthropic, Google, and others are adding noise to the debate about whether Microsoft can own enterprise AI the way it once owned enterprise productivity.

What Did BNP Paribas Actually Say?

On April 10, 2026, analysts at BNP Paribas held an investor call specifically designed to address the concerns swirling around Microsoft. The headline takeaway: they remain constructive on the stock and see several concrete catalysts ahead that the market may be underweighting.

The key points from BNP’s discussion are worth walking through carefully:

1. The Copilot Product Pipeline Is Accelerating

BNP flagged the May 1, 2026 launch of the Microsoft 365 E7 suite as a near-term inflection point. This new premium tier is priced at $99 per user per month — more than three times the standard Copilot price of $30 — and comes with expanded AI capabilities. Following that launch, BNP expects a broader rollout of tools including Agent Mode, Copilot Cowork, Critique, Council, and Agent 365.

Critically, BNP noted that early product feedback is improving. This matters because a lot of the bear case on Copilot has been anecdotal: enterprise IT teams finding it clunky, usage rates disappointing after initial pilots. If the product quality is genuinely improving — and the E7 launch represents a step-up in capability — adoption curves could surprise to the upside in the second half of 2026.

2. Azure Is More Resilient Than the Bears Suggest

One concern investors have raised is that some of Azure’s capacity is being consumed internally by Microsoft itself — for Copilot infrastructure and large language model development — rather than by external paying customers. BNP addressed this directly, arguing that even internal consumption is a tailwind: heavier token usage and higher GPU rental prices support Azure pricing power over time.

This is a nuanced but important point for long-term investors to absorb. Microsoft’s AI buildout is not a cost centre that competes with its cloud business. The two are structurally linked. Every token Copilot processes runs on Azure infrastructure. The more Copilot scales, the more Azure earns — even if the customer is Microsoft itself.

Beyond that, the commercial backlog tells a compelling story. Microsoft’s remaining performance obligation reached $625 billion, up 110% year-over-year. That is an extraordinary number. It means even if Azure signed no new deals tomorrow, the company has locked-in future revenue that many competitors would take years to accumulate.

3. Investor Fears May Be Overstated

BNP’s investor call was partly structured to directly address the anxiety points the market has focused on: Copilot traction, competition from Anthropic and other AI rivals, and Azure market share erosion. Their view, in plain terms, is that these fears are real but probably exaggerated relative to Microsoft’s actual competitive position.

The enterprise software business is not like the consumer internet. Copilot is not competing with ChatGPT for casual users. It is competing for multiyear contracts with Fortune 500 IT departments, embedded in tools that tens of millions of workers already use every day. Switching costs in enterprise software are enormous. That dynamic tends to favour Microsoft even when the product is not the flashiest option on the market.

The Bull Case in Plain Language

Bank of America, which reinstated coverage with a Buy rating and a $500 price target, laid out the logic that most bulls are working from. Here is how to think about it:

Azure is growing at 39% year-over-year on a base that already generates over $32 billion per quarter in the Intelligent Cloud segment. OpenAI has committed $250 billion in incremental Azure services. That is not a rounding error — it is a structural revenue stream with years of runway. Meanwhile, Azure’s AI workloads are contributing an estimated 13 to 16 percentage points of that 39% growth rate, meaning the AI premium is already showing up in the numbers, not just in management presentations.

Microsoft 365 Copilot’s penetration rate is below 4% of a 430-million-user base. At $30 per user per month, even moving to 10% penetration — which would still be conservative for a product embedded in tools people use every working day — implies a multi-billion dollar revenue uplift that is not yet in the stock price. The new E7 tier at $99 per month could accelerate that dollar expansion even before seat counts dramatically widen.

Net income surged nearly 60% year-over-year in the most recent quarter. The company returned $12.7 billion to shareholders through dividends and buybacks in that same period. These are not the financials of a company in distress. They are the financials of a company investing heavily in its future while continuing to compound returns for long-term owners.

Should You Buy Now or Wait?

This is the question every investor looking at the drop is wrestling with. There is no clean answer that works for everyone, but here are the frameworks worth applying:

If You Are a Long-Term, Buy-and-Hold Investor

The case for buying MSFT gradually at current levels is straightforward. The stock is trading at roughly 22 times forward earnings, compared to a 10-year average closer to 30 times. The business is not impaired — it is accelerating in its most important segment. The selloff has been driven by macro fear, valuation compression in tech broadly, and disappointment relative to sky-high expectations, not by any fundamental deterioration in Microsoft’s competitive position.

Long-term investors who believe that AI will be embedded into every enterprise workflow within the next five years should ask themselves: is there a more logical company to own through that transition than the one that already sits in the operating system, the office suite, the development environment, and the cloud infrastructure of most large organisations on earth? That moat does not disappear because Copilot’s adoption curve is S-shaped rather than vertical.

Buying in tranches — spreading purchases over several months rather than going all-in at once — is a sensible approach given that macro uncertainty (tariff impacts, potential recession odds near 40% on prediction markets) could push the stock lower before it stabilises. Dollar-cost averaging removes the pressure of trying to call the exact bottom.

If You Prefer to Wait for Confirmation

The May 2026 developer conference and the E7 suite launch on May 1 are the two near-term events that could shift sentiment. If early data on E7 uptake is strong, or if Microsoft’s next earnings report shows Copilot seat expansion accelerating materially, that would be a more concrete signal to act on.

Waiting also makes sense if you are already well-exposed to tech and do not need to add more concentration right now. There is no shame in watching the catalyst play out before committing capital. In a market where recession fears are real, patience has its own value.

If You Are Considering Selling

Selling MSFT now at a 33% discount from its peak, after a fundamental business that grew net income by nearly 60% in its most recent quarter, would generally be considered selling into fear rather than responding to a genuine change in business quality. The reasons investors are selling — capex levels, Azure growth expectations, Copilot pace — are all addressable over 12 to 24 months. If your thesis for owning Microsoft is intact, a drawdown is not automatically a reason to exit.

The exception: if your position is larger than you are comfortable holding through further volatility, trimming to a size that lets you sleep at night is always reasonable. Emotional comfort with your position size matters more than being technically right about a stock.

The Bottom Line

Microsoft has fallen a long way from its peak — but the business case for owning it over the long run has not. Azure is a $32 billion per quarter engine growing at nearly 40%. The Copilot platform is in its early innings with less than 4% penetration of a 430-million-user base. The E7 suite launching in May could be the first real test of whether enterprises will pay a genuine premium for deeper AI integration. And BNP Paribas joining Bank of America in flagging these catalysts suggests the institutional view is shifting from caution to selective accumulation.

The risks are real. Capex is enormous. Macro headwinds are persistent. Competition from Anthropic, Google, and others is not going away. These are not trivial concerns, and they deserve weight in any honest analysis.

But at 22 times forward earnings — its cheapest valuation in a decade — Microsoft is no longer asking investors to pay for perfection. It is asking them to believe that the world’s most embedded enterprise software company will find a way to monetise AI at scale over the next three to five years. Given everything it has built, that is not a leap of faith. It is a calculated bet on a business with extraordinary structural advantages.

That is the kind of bet InvestNotBet was built to think through carefully — not to gamble, but to invest with conviction, context, and a clear understanding of the risk you are taking on.