The verdict starts landing tonight—and Teradyne is the first receipt

The verdict starts landing tonight—and I already have skin in it.

On Monday, I wrote that this week would test whether the AI trade could survive a harder question.

Not whether AI demand exists.

Whether the cash returns can justify the capital being committed.

Tonight, the first major answers arrive.

The Federal Reserve announces its decision at 2:00 p.m. ET.

Then Microsoft and Meta report after the close.

One of my own new positions has already reported.

Teradyne delivered a large earnings beat this morning and the shares rose sharply.

The three events are different, but they belong to the same market argument:

What is already priced in, what new information arrived, and how much should investors pay for the future?

Checking Monday’s thesis

My week-ahead research identified Microsoft as the cleanest AI-capital test of the week.

Alphabet had already established the tension.

Google Cloud grew 82%, Search remained resilient and Cloud profitability improved dramatically. The stock still closed down 7.1% after Alphabet raised its capital-spending forecast and reported negative quarterly free cash flow.

That did not show that AI demand had failed.

It showed that growth no longer receives an unlimited valuation exemption from the cost required to produce it.

Microsoft and Meta now face that test through two different business models.

Microsoft: Azure must explain the bill

Microsoft is primarily an Azure test.

Management guided Azure and other cloud-services growth to 39–40% in constant currency after delivering 39% in the previous quarter. Demand was still exceeding available capacity. Microsoft’s previous earnings call

That creates a demanding setup.

If Azure reaches or exceeds the guide, investors will still ask what was required to produce the growth.

If it misses meaningfully, elevated AI infrastructure spending will become harder to defend.

The important questions are:

Did Azure deliver the 39–40% growth management guided?
Is enterprise AI adoption broadening beyond infrastructure consumption?
Are operating margins absorbing the higher depreciation and operating costs?
Is free cash flow keeping pace with the investment cycle?
How much more capital does Microsoft expect to commit during fiscal 2027?

A headline earnings beat can coexist with an unsatisfying answer to those questions.

Alphabet just demonstrated that.

The stock will react not only to whether Microsoft grew, but to whether the growth changes the expected return on its AI capital.

Meta: can AI strengthen the engine funding it?

Meta faces the same test through advertising.

Its core business must continue generating the cash that funds a rapidly expanding infrastructure programme.

The market will focus on:

Advertising revenue growth
User engagement
AI-driven recommendation quality
Advertising efficiency
Operating margin
Capital-spending guidance

The strongest answer would not be another promise about a future AI product.

It would be evidence that AI is already making the existing business more valuable—better recommendations, higher engagement and more effective advertising.

That would show the investment feeding the engine that pays for it.

The risk is that infrastructure spending accelerates faster than the benefits visible in revenue, margin and cash flow.

Meta is scheduled to release its second-quarter results after the market closes today. Meta Investor Relations

The Fed goes first

Before either company reports, the Federal Reserve decides whether to leave its target range at 3.50%–3.75%.

Holding remains the more likely outcome, but markets entered the meeting pricing roughly a one-in-three chance of a quarter-point increase.

That makes this more than a routine hold.

At the previous meeting, policymakers were evenly divided on whether rates should rise. Chair Kevin Warsh’s communication has also been deliberately less prescriptive than the forward-guidance style markets became accustomed to.

The statement matters.

The press conference may matter more.

Investors will listen for how Warsh balances recent inflation improvement against the risk that oil, tariffs or demand could keep price pressure elevated. Reuters’ Fed preview

The sequence creates an unusual stack:

The discount rate is repriced.
Microsoft and Meta report.
Investors apply the new rate backdrop to two of the market’s largest AI-spending programmes.

A strong earnings report can look different after a hawkish Fed.

A disappointing report can become more damaging if yields rise at the same time.

That interaction is one reason I am not trading into the events.

Teradyne delivers the first receipt

Teradyne reported before the larger tests arrived.

Second-quarter revenue reached $1.329 billion, up 104% from a year earlier and above the roughly $1.22 billion analysts expected.

Adjusted earnings were $2.47 per share, compared with an approximately $2.09 consensus.

The third-quarter outlook was also stronger than expected:

Revenue of $1.20–1.30 billion
Adjusted earnings of $1.85–2.15 per share

Consensus had been near $1.06 billion of revenue and $1.56 of adjusted earnings.

The shares responded with a sharp high-single-digit rise during the session. Teradyne’s Q2 release

The report matters fundamentally.

It strengthens the evidence that demand for semiconductor test equipment remains robust, particularly around AI-related memory and compute investment.

It also provides better information about the earnings path than existed when I entered the position.

But the market reaction does not prove that I made a good decision.

What the TER result proves—and what it does not

I bought TER last Thursday as one of eight new positions in a scheduled deployment cycle.

The stock was already down roughly 4% one day after entry.

It had cleared the quality screen, ranked highly enough for inclusion, fit the portfolio construction rules and entered at partial size while cash remained around 58%.

I did not buy it because I knew this quarter would beat expectations.

I did not have an informational advantage on revenue, earnings or guidance.

The position existed before the result because the process qualified it.

Today’s report is new evidence about the company.

It is not retroactive evidence about the quality of the original decision.

That distinction is easy to lose when a position rises.

A trader can make a reckless earnings gamble and still be rewarded.

A trader can also make a disciplined, well-sized decision and experience an immediate loss.

The outcome changes the account.

It does not rewrite the decision.

This is the worked example for Lesson 21: A winning trade does not prove the decision was good.

The correct way to review the position

TER should be evaluated in two separate passes.

The decision review

Using only information available before the earnings report:

Did the company meet the quality criteria?
Did the opportunity rank justify inclusion?
Did the position improve or weaken portfolio diversification?
Was the initial size appropriate for the uncertainty?
Was the deployment consistent with the scheduled process?

Those questions judge the entry.

The information update

Using the earnings report that arrived afterward:

Did revenue and guidance strengthen the operating thesis?
Has valuation changed enough to affect expected return?
Does the new information alter the intended holding size?
Has the stock become too extended for another tranche?
Does the portfolio now have more correlated AI-infrastructure exposure than intended?

Those questions judge what happens next.

Combining the two reviews creates hindsight.

Separating them creates learning.

What I am doing into tonight

Nothing new.

I am not adding Microsoft or Meta before the close.

Two of the world’s largest companies reporting immediately after a live Fed decision is exactly the kind of binary-event stack I do not trade into.

That is not a bearish call on either company.

It is a statement about where I do and do not have an edge.

I hold what the system has already placed, including TER and the rest of Thursday’s deployment.

I keep the remaining cash available.

I let the events produce information and price.

If Microsoft or Meta disappoints and quality companies are pulled toward valid demand zones while their underlying theses remain intact, the system may eventually have something to do.

If the reports are strong and prices move away, I do not chase them.

The opportunity is not tonight’s headline.

It is the relationship between the new information, the price it creates and the rules that determine whether that price is actionable.

The larger lesson

Tonight will produce winners and losers.

Some will have made disciplined decisions.

Some will simply have guessed correctly.

The P&L alone will not tell us which is which.

Teradyne’s beat is welcome.

It strengthens the company evidence and improves the position outcome.

It does not give me permission to claim that I predicted the print.

The process must be judged without borrowing certainty from a result that had not happened yet.

A good outcome belongs in the account. A good decision must already exist before the outcome arrives.

Cash remains around 58%.

Positions held.

Levels marked.

I do not predict the print.

Ask Teradyne.

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.