This week, the market had decided Microsoft was the AI trade gone wrong.

Nearly $190 billion in planned capex. Shrinking free cash flow. A stock down more than 30% from its highs. That was the story heading into Tuesday.

Then Microsoft reported, and the market flipped overnight.

Azure accelerated instead of slowing. Copilot crossed 30 million paid seats. The stock ripped roughly 15% on Thursday and added about $260 billion in value in a single session, its best day since 2008.

The easy read is relief. The real read is more interesting, and it changes how you should think about this stock for the next year, not just this week.

That's what The Pragmatic Playbook digs into today: what Thursday's rally is actually pricing in, and whether Microsoft still has real room to run.

Missed Wednesday's deep dive on Visa's earnings?

The options market is pricing Apple and Amazon for meaningful post-earnings moves, but the setup is very different for each. Apple’s implied move is around 4%, which is high compared with its recent earnings reactions, while Amazon’s implied move is closer to 6% to 7%, reflecting bigger uncertainty around AWS growth and AI capex. In plain English, traders are not just watching the numbers. They are betting this earnings night could reset expectations for two of Big Tech’s most important stories.

The Fool’s point is that Fed Chair Kevin Warsh captured public frustration with inflation in three words: “Deliver it, already!” The Fed kept rates unchanged at 3.5% to 3.75% in a 9-3 vote, but Warsh offered few specifics on when action might come, even as inflation remains above target. The takeaway is that markets may be losing patience with Fed talk, especially with three officials already voting for a hike.

Broadcom’s five-year, over $200 billion Samsung agreement is really about locking down scarce AI supply. The deal gives Broadcom access to Samsung memory, including HBM, plus potential use of Samsung’s 2-nanometer foundry and advanced packaging capabilities. The bigger message is that AI chip winners are no longer just competing on design. They now need guaranteed memory and packaging capacity, and that security is getting very expensive.

Radar transcribes 115,000+ podcasts where executives, officials, and analysts talk candidly and publicly, transcribed and queryable within minutes of airing.

Watch a company, a person, or a theme with Radar and get alerts whenever they’re mentioned.

Radar is built by former Twitter and Tesla engineers, using an AI-native transcription pipeline that delivers high accuracy and extensive data enrichment.

Try Radar for free today

Microsoft just had its best trading day in nearly two decades. Shares ripped about 15% on Thursday after fiscal fourth-quarter results, adding roughly $260 billion in market value in a single session.

The easy read is that Microsoft "proved the bears wrong." Azure grew faster, not slower. Capex fears eased. Case closed.

Pull it apart and the story gets more interesting. Microsoft is still down about 7% for the year, still sitting roughly 19% below its 52-week high, and still spending close to $190 billion this year on the same AI infrastructure buildout that terrified investors two months ago.

The Pragmatic Playbook is here to answer one question. Is this rally the start of a real re-rating, or a relief bounce wearing a turnaround's clothes?

Start with what the numbers actually say.

Azure revenue growth accelerated to 43% at constant currency, up from 40% the quarter before. That is not deceleration. That is a business getting faster while carrying a much bigger base.

Commercial remaining performance obligations, essentially Microsoft's contracted backlog, jumped 84% to $678 billion. Copilot crossed 30 million paid seats, putting Microsoft's AI product line on roughly a $10 billion annual run rate.

Revenue for the quarter hit $90 billion, up 18% year over year. Adjusted EPS of $4.74 beat estimates of $4.24 by one of the widest margins Microsoft has posted in years.

That's not a business that limped over a lowered bar. That's a business that cleared the old one.

None of that changes the fact that Microsoft is still one of the biggest capital spenders on the planet.

Capex is tracking toward roughly $175 to $190 billion this year, almost entirely funneled into data centers and GPUs. Free cash flow already felt it, falling to $15.8 billion in an earlier quarter this year, down from $20.3 billion a year prior, even as reported net income kept climbing.

One extraordinary earnings report doesn't undo that math. It just bought the market's patience for another quarter.

Not a resolved risk. A postponed one.

There's a second layer here worth sitting with.

A meaningful share of the enthusiasm around Azure and Copilot is downstream of a broader AI infrastructure buildout that increasingly involves the same handful of companies investing in each other's growth. Microsoft, OpenAI, and the chipmakers underneath them are becoming more financially intertwined by the quarter.

That's not a scandal. It's a concentration of risk that doesn't show up cleanly in any single earnings report.

Because if AI infrastructure demand ever cools even modestly, Microsoft isn't just exposed through its own capex. It's exposed through the health of the partners it's increasingly betting alongside.

Microsoft spent the first half of this year sliding from an October high near $555 down toward $349, a drawdown of roughly 34% that made this the stock's worst start to a year since the dot-com era. By early July it was trading in the high $380s, a stock the market had essentially stopped believing in.

Thursday changed the picture violently. Shares ripped from the $390s to roughly $448 in a single session, the kind of move that usually takes months, not hours.

Even after that pop, this isn't a stock making new highs. It's sitting about 19% below its 52-week peak and still down roughly 7% for the year, which means Thursday's move closed a gap more than it opened a new one.

The next real test is whether $448 holds as a floor or gets sold into like every other bounce attempt did earlier this year. If it holds and grinds toward the $480 to $500 zone where the stock traded before the capex panic started, that's the market saying the turnaround is real. If it fails and slides back under $420, that's the market telling you Thursday was relief, not conviction.

Management guided to 45% constant-currency Azure growth for fiscal Q1 2027, above what analysts had modeled. If Microsoft delivers anywhere near that number next quarter, this rally gets a second leg. If it comes in even a few points light, expect the capex conversation to reopen immediately.

Free cash flow, not revenue, was the actual source of this year's selloff. Watch whether it stabilizes or keeps compressing as capex ramps toward $190 billion. A flattening free cash flow trend would do more to validate this rally than another strong Azure headline.

I covered this exact tension back in June, when Microsoft was down 33% while the business kept growing 18%. The setup has changed since then, but the underlying question hasn't. How much of this growth depends on a small circle of AI partners investing in each other, and is that circle getting more or less concentrated.

Citi, Wells Fargo, and Bernstein all raised targets into the $600 to $650 range within a day of the print. Watch whether more conservative shops, like Evercore near $528, move up to meet them or whether the Street stays split. A split Street after a move this big usually means the rally isn't fully priced in yet.

Thirty million paid seats sounds impressive until you ask how fast that number needs to grow to justify the AI infrastructure spend behind it. The next disclosure on seat growth and revenue per seat will tell you whether this is a real product cycle or a big number still searching for scale.

I didn't expect Microsoft to erase four months of fear in a single session, and honestly, neither did most of Wall Street.

But the number that matters isn't the 15%. It's that the business grew into its valuation instead of the valuation collapsing to meet the business, which is the opposite of what the bears were betting on.

This doesn't mean the capex risk is gone. It means the market just got real evidence that Microsoft can spend historically large sums on AI infrastructure and still accelerate its highest-margin business at the same time.

Stocks don't need new highs to prove a thesis. They need to stop confirming the bear case, and this week, Microsoft finally did.

Love him or hate him, Musk moves markets. His next launch hits July 22, and the smart money is already positioning. Our analyst found 3 stocks set to ride it — with entry points and a buy/sell playbook.

Show Me the 3 Stocks

This week is a good reminder of what actually moves stocks versus what just moves headlines.

Microsoft didn't get cheaper on Thursday. It got proven.

The capex fear that hammered this stock all year was never wrong to consider. It was wrong to treat as permanent.

The investors who sold in June on the scary capex number and the investors chasing this week on the exciting Azure number are often making the same mistake in opposite directions.

Both are reacting to a single data point instead of watching whether the underlying story stayed intact.

The edge was never in predicting Thursday's number. It was in not panicking in June and not chasing in July.

If this week's Playbook saved you from a bad entry, send it to one friend who needed it too.