To The Moon

Microsoft and Meta reported earnings on the same night. Both are spending heavily on AI infrastructure, both raised or maintained aggressive investment plans, and both delivered strong revenue growth.

The market still gave them opposite verdicts.

$Microsoft(MSFT)$ rose more than 8% after hours as Azure growth, Copilot adoption and a strong outlook convinced investors that its AI spending is already generating measurable returns.

$Meta Platforms(META)$ fell roughly 9% in premarket trading after free cash flow collapsed and capital expenditure remained close to record levels. Meta’s advertising business is still growing quickly, but investors want a clearer answer on how its enormous compute buildout will create revenue beyond advertising. (Reuters)

The message from the market is becoming clearer:

Investors are no longer rewarding the company that spends the most on AI. They are rewarding the company that can collect the money fastest.

Microsoft finally showed the receipts

Microsoft reported quarterly revenue of $90 billion, up 18% year over year, while Azure revenue grew 43%, beating analysts’ expectation of roughly 40%. The company also forecast Azure growth of about 45% for the current quarter, again ahead of expectations. (Reuters)

The important part is that Microsoft now has several visible AI revenue channels working at the same time.

Azure sells computing capacity to enterprises and AI developers.

Microsoft 365 Copilot adds a paid AI layer to workplace software.

GitHub Copilot monetizes coding assistance.

Security, database and business-software customers can also purchase additional AI services.

Microsoft said paid Microsoft 365 Copilot seats exceeded 30 million, up from 20 million in the previous quarter and above analysts’ estimates. Its commercial cloud backlog reached $678 billion, with the latest sequential increase driven by customers outside the largest U.S. frontier-model developers. (Reuters)

That last point matters.

It suggests Microsoft’s AI demand is becoming broader than a handful of companies such as OpenAI. More traditional enterprises are committing money to cloud and AI services.

Microsoft is spending heavily—but the cash engine still works

Microsoft’s quarterly capital expenditure reached approximately $41 billion, more than 70% higher than a year earlier.

Its free cash flow declined 23% year over year to $19.6 billion, yet that still came in well above market expectations. Microsoft expects reported capital expenditure of about $50 billion in the current quarter and roughly $175 billion for calendar 2026 after changing the accounting treatment of long-term data-center leases. The company said its underlying construction and capacity plans had not changed. (Reuters)

The market accepted that spending because Microsoft delivered three forms of evidence:

• Azure growth accelerated.

• Paid Copilot adoption expanded.

• Free cash flow remained strongly positive.

Microsoft is still spending at an extraordinary rate. Investors can already see where part of the return is coming from.

That is the difference.

Meta’s advertising business remains strong

Meta’s headline revenue was also impressive.

Second-quarter revenue rose 28% to $60.8 billion. Daily active people across its apps reached 3.6 billion, up 3%. Ad impressions grew 14%, while the average price per ad increased 12%. (Meta)

Those numbers confirm that AI is improving Meta’s existing business.

Better recommendation systems can increase the amount of time users spend on Facebook and Instagram.

More accurate ad models can improve targeting and conversion.

Generative tools can help advertisers produce more content and run campaigns more efficiently.

Meta’s core advertising engine is still performing well. Reuters described it as the main business currently financing the company’s wider AI ambitions. (Reuters)

The problem appeared further down the financial statements.

Meta’s free cash flow nearly disappeared

Meta generated only $784 million in free cash flow, down 91% from $8.55 billion a year earlier. Quarterly capital expenditure reached $31.08 billion, and the company narrowed its 2026 capex forecast to $130 billion–$145 billion, raising the lower end from $125 billion. (Reuters)

Its operating margin fell from 43% to 31%, while operating income declined 8%.

Part of that decline came from $2.4 billion in legal charges and $1.18 billion in severance expenses, so the quarter was not purely an AI-spending story. Even so, the cash-flow pressure showed how quickly data centers, servers and chips are absorbing Meta’s operating cash. (Meta)

Meta is building toward roughly 7 gigawatts of computing capacity this year and 14 gigawatts next year. Management believes that compute will support personal AI agents, improve the core advertising business and eventually allow Meta to serve large external customers. (Reuters)

Investors are asking for a more concrete bridge between that capacity and future revenue.

Why Microsoft gets more credit than Meta

Both companies are investing in chips, servers, networking, power and data centers.

Their monetization structures are different.

Microsoft can charge directly for:

• Azure compute usage

• Copilot subscriptions

• GitHub AI tools

• Enterprise security and database services

• AI models and application programming interfaces

Meta’s most proven AI return is still indirect.

AI improves engagement and ad performance, which strengthens its existing advertising business. Meta has discussed personal agents, business agents, smart glasses, model services and potentially selling external compute capacity, but those businesses have not yet produced the same visible revenue stream as Azure and Copilot.

Microsoft’s AI spending is appearing in cloud growth and paid seats.

Meta’s AI spending is currently appearing more clearly in capital expenditure and reduced free cash flow.

That is why similar investment plans produced such different stock reactions.

The Fed held rates—but the decision was not dovish

The Federal Reserve kept its benchmark rate at 3.50%–3.75%, as expected.

The vote was unusually divided: nine officials supported holding rates steady, while three preferred a 25-basis-point increase. The statement said inflation remained elevated relative to the Fed’s 2% goal, partly because of supply shocks affecting sectors including energy. (联邦储备系统)

That dissent matters for technology valuations.

The Fed did not raise rates, but three votes for a hike signal that tighter policy remains a live possibility.

Long-term Treasury yields moved higher after the announcement, with the 30-year yield reaching its highest level since 2007. Higher long-term yields increase financing costs and reduce the present value of profits expected many years in the future. (Reuters)

The Fed has therefore added another filter to the AI trade.

When capital remains expensive, the market becomes less patient with companies whose revenue arrives years after the spending.

The AI trade is moving into a “payback” phase

The Microsoft–Meta split does not mean the AI infrastructure cycle is ending.

Both companies are still spending aggressively, which supports demand across the supply chain.

Chips and manufacturing:
$NVIDIA(NVDA)$, $Broadcom(AVGO)$, $Taiwan Semiconductor(TSM)$

Networking:
$Arista Networks(ANET)$, $Marvell Technology(MRVL)$, $Credo Technology(CRDO)$

Memory and storage:
$Micron Technology(MU)$, $SanDisk(SNDK)$

Power and cooling:
$Vertiv(VRT)$, $Eaton(ETN)$

The change is happening in how investors rank the companies funding that expansion.

The market increasingly prefers businesses that can demonstrate:

• Real customers

• A clear pricing model

• Accelerating revenue

• Positive free cash flow

• A manageable relationship between capex and operating income

AI infrastructure demand remains strong.

The debate is shifting toward who captures the economic value.

Amazon and Apple are next

$Amazon(AMZN)$ and $Apple(AAPL)$ report after the U.S. market closes today. Both calls are scheduled for July 30. (亚马逊)

Amazon: another test of the AI cloud model

Amazon sits closer to Microsoft’s side of the comparison.

Investors will focus on:

• AWS revenue growth

• AI-related contracts and backlog

• Trainium custom-chip adoption

• Capital expenditure

• Free cash flow

If AWS accelerates while cash flow remains resilient, Amazon could reinforce the view that cloud platforms are beginning to earn back their AI investments.

If spending remains elevated without a clear acceleration in AWS, the same capital-return concerns affecting Alphabet and Meta could spread to Amazon.

Apple: the lower-capex comparison

Apple represents a different AI strategy.

Its approach relies more on devices, custom silicon, operating systems, services and external partnerships than on constructing hyperscale AI infrastructure at the same pace as Microsoft, Meta, Alphabet or Amazon.

The market will watch:

• iPhone demand

• Services growth

• China performance

• The possibility of an AI-driven upgrade cycle

• Apple’s long-term AI strategy

• Whether lower capex supports stronger cash-flow resilience

During a period when investors are questioning data-center spending, Apple’s more capital-efficient model may appear attractive.

It still needs to show that its AI approach can create product growth.

TigerComments Take

The most important lesson from Microsoft and Meta is not that one company understands AI and the other does not.

Both have strong businesses and substantial AI advantages.

The market is distinguishing between two stages of monetization.

Microsoft can already point to Azure growth, paid Copilot seats, cloud backlog and positive free cash flow.

Meta can point to stronger engagement and advertising performance, while the revenue model for its broader AI infrastructure remains less visible.

At the same time, the Fed’s divided decision keeps the cost of capital high.

That combination creates a new rule for the AI market:

Companies can keep spending—but the evidence of payback must arrive faster.

Which AI strategy looks strongest now?

A. AI cloud platforms: MSFT / AMZN
B. AI hardware suppliers: NVDA / AVGO / TSM
C. Lower-capex ecosystem: AAPL
D. Wait for more free-cash-flow proof

Do you think Microsoft has already proved that AI spending is paying off, or will the true capital burden only become visible later?

Disclaimer: This post is for market discussion only and does not constitute investment advice. Investing involves risk. Premarket and after-hours share-price moves may change before the regular session.