Last week was the warning. This week is the verdict.

Last week was the warning.

This week is the verdict.

Alphabet and Tesla reported, and the market delivered a clear message:

AI growth alone is no longer enough.

Alphabet grew Google Cloud revenue 82% to $24.8 billion. Search remained resilient. Cloud profitability improved dramatically.

The shares still closed down 7.1% after Alphabet increased its 2026 capital-spending forecast to approximately $195–205 billion and reported negative quarterly free cash flow.

The business passed the growth test.

The stock failed the expectations test.

My post-earnings review explained why: the market is no longer asking only whether companies are investing successfully in AI.

It is increasingly asking:

When does that investment turn into sustainable cash flow?

This week, the four megacaps still left to report face versions of that same question.

Microsoft and Meta report Wednesday.

Apple and Amazon report Thursday.

Around those four reports sit a Federal Reserve decision, Chair Kevin Warsh’s press conference, June core PCE and the first estimate of second-quarter GDP.

The events are different.

The market question connecting them is the same:

How much should investors pay for future growth when the capital bill, inflation path and discount rate are all uncertain?

Microsoft: the cleanest AI test

For Microsoft, this is primarily an Azure test.

Management guided Azure growth to 39–40% in constant currency after the business grew 39% on that basis last quarter. Demand remained above available capacity, which makes the coming report a test of both execution and investment efficiency. Microsoft’s previous earnings call

A meaningful miss against that range would matter.

So would guidance suggesting that growth is slowing while infrastructure spending remains elevated.

The headline EPS number may be less important than four questions:

• Did Azure deliver the growth management guided?
• Is AI demand still exceeding capacity?
• How much capital is required to serve that demand?
• Is the spending producing operating leverage, or only more depreciation and cash pressure?

Alphabet proved that exceptional cloud growth does not guarantee a positive stock reaction.

Microsoft now faces the cleanest version of the same test.

Meta: can advertising keep funding the build-out?

Meta’s advertising engine must continue funding its AI investment.

The market will focus on revenue growth, engagement, operating margin, capital spending and whether AI is improving recommendation quality and advertising efficiency.

That distinction matters.

If AI makes the existing advertising system more effective, the investment is already strengthening the core business.

If management asks investors to fund a much larger build-out mainly for products whose economics remain distant, the market may apply the same skepticism it applied to Alphabet.

Meta reports after Wednesday’s close. Meta investor relations

Investors do not need another broad promise that AI will matter.

They need evidence that the spending is improving the business that pays for it.

Apple: the outlier

Apple is the outlier.

This is less about constructing enormous AI data centres and more about the health of the ecosystem that already exists.

The key questions are:

• Are consumers still buying iPhones?
• Is Services maintaining its growth?
• Is Greater China stabilising?
• What does management say about the September quarter?
• Can Apple defend a premium valuation without matching the capital intensity of its peers?

At Apple’s valuation, guidance may matter more than the quarter that has already ended.

The company does not face exactly the same AI-capex test as Microsoft, Meta or Amazon.

But it faces the same expectations problem.

A strong quarter can still disappoint if the market has already paid for stronger guidance.

Amazon: growth versus the infrastructure bill

For Amazon, the focus is AWS growth, cloud margins, capital spending and retail profitability.

AWS is the main bridge between AI investment and current operating profit.

The question is not whether Amazon can find uses for more infrastructure.

It is whether the growth and margin produced by that infrastructure can justify the amount of capital being committed.

That is the same return-on-invested-capital question Alphabet brought to the surface.

Amazon reports Thursday after the close. Amazon’s earnings announcement

If AWS accelerates and margins hold, the market may treat the investment as productive capacity.

If spending rises faster than the evidence of returns, the Alphabet reaction provides the warning.

The 48-hour macro stack

The earnings reports are only part of the week.

The Federal Reserve announces its decision Wednesday.

Holding rates steady remains the base case, but markets are pricing roughly a one-in-three probability of a hike. That makes this a live meeting rather than a routine hold. Reuters’ Fed preview

Warsh’s tone may matter as much as the decision.

A hold accompanied by a clear willingness to tighten in September is different from a hold that treats the oil shock as temporary.

A hike would immediately change the discount rate applied to long-duration technology earnings.

Then Thursday at 8:30 a.m. ET, June core PCE arrives alongside the advance estimate of second-quarter GDP. The releases will add fresh information about both inflation and economic growth before Apple and Amazon report after the close. BEA release schedule

That puts six major events inside roughly 48 hours:

1. Microsoft earnings
2. Meta earnings
3. The Fed decision
4. Warsh’s press conference
5. Core PCE and GDP
6. Apple and Amazon earnings

Each can change a different part of the valuation equation.

Together, they can change the market’s view of growth, margins, inflation, rates and AI capital intensity almost at once.

Oil is helping—for now

One pressure point eased Monday.

The United States and Iran paused attacks, and Brent fell sharply from last week’s close above $100. It traded as low as $87.55 before moving back toward $90. Reuters oil update

Gold traded near $4,093 as the dollar and yields eased. Reuters gold update

That removes some immediate inflation pressure.

It does not resolve the conflict, reopen every supply route or guarantee that oil remains lower.

The correct word is relief.

Not resolution.

A temporary fall in oil makes a Fed hold easier to defend and reduces one source of valuation pressure.

It does not answer the AI-return question facing the four companies this week.

Checking the portfolio after Thursday

Thursday, I deployed capital during the selloff.

I opened eight new positions and added to GLD and SPY. The entries were deliberately spread across businesses that are not all dependent on the same mega-cap platform outcome.

The full reasoning is in my Thursday deployment review.

Seven of the ten positions remain underwater.

That is acceptable within the original design.

They were partial entries made during a falling market. They did not need to identify the exact bottom on day one.

The relevant questions are not whether every position turned green immediately.

They are:

• Did each company qualify under the screen?
• Was each entry partial rather than a claim that the low was in?
• Did the combined exposure remain within the portfolio rules?
• Is the original business thesis still intact?
• Is enough cash available for the next decision?

Around 58% of the account remains in cash.

That is important this week.

The portfolio can absorb new information without needing Wednesday’s or Thursday’s reports to produce a particular result.

What I am doing before the verdict

I am not adding ahead of this event stack.

That is not a forecast that the reports will be bad.

It is recognition that four of the largest companies in the world are reporting around a live Fed decision and the most important inflation release of the week.

Entering immediately before several binary events is not where I believe I have an edge.

I do not know whether Azure will print 38%, 40% or 42%.

I do not know which sentence in Meta’s capital-spending outlook the market will choose to trade.

I do not know whether Warsh will hold, hike or deliver a hold that sounds like a future hike.

I do not need to pretend that I know.

My advantage comes after the information changes price.

I have levels marked on companies I would want to own if disappointing earnings or a hawkish Fed pushes quality names lower.

But a lower price is not enough.

The thesis must remain intact.

That distinction is the subject of Lesson 20: Why the best trades often feel uncomfortable.

Discomfort because price reached a planned level is part of execution.

Discomfort because new information broke the thesis is a reason to stop.

Never confuse the two.

What would make me act

If a company misses expectations and the entire complex sells off, I will not buy the first red candle.

I will ask:

1. Did the report damage the business thesis, or only reset the valuation?
2. Has price reached a demand zone defined before the reaction?
3. Is the invalidation point still defensible?
4. Does the position fit the portfolio’s existing theme exposure?
5. Can I enter at partial size without needing an immediate rebound?

If those conditions are present, discomfort does not disqualify the trade.

If the thesis is broken, the old zone does not rescue it.

That is the decision architecture.

Not prediction.

Not courage.

Classification, then execution.

The verdict

Last week established the warning.

AI demand can be extraordinary while the stock falls because capital intensity and cash flow matter more than the headline growth rate.

This week tells us whether Alphabet was an isolated repricing or the template for the rest of mega-cap technology.

Microsoft tests Azure.

Meta tests the advertising-funded AI model.

Apple tests the durability of its ecosystem.

Amazon tests the economics of AWS and infrastructure spending.

The Fed and PCE determine the rate environment in which every one of those future cash flows is valued.

I do not need to predict what Microsoft says Wednesday.

I need to be ready for what it does to price.

Cash ready.

Levels marked.

Watching the tape.

This is a record of my process and opinions, not investment advice. I hold positions in securities mentioned or related. Copy trading involves risk, including loss of capital. Past performance is not an indication of future results.