Issue 009 is here — the weekly business newsletter brought to you by Pretus. The free platform that helps you master the finance recruiting cycle with AI-driven interview prep for IB, CRE, WM, AM, & more + networking, behaviorals, and an exclusive ecosystem with 1-on-1 coaching and insights you can’t get anywhere else.

Big Tech earnings week is supposed to create clarity.

This one delivered three different verdicts at once.

Microsoft and Amazon showed that the AI infrastructure spending wave is producing real, accelerating revenue. Apple showed that spending big on AI to sell hardware works until it doesn't, and that consumer AI monetization is still a theory. The Fed held rates for the fifth consecutive meeting, but three regional presidents broke ranks in the most visible internal disagreement the FOMC has seen in years. And while all of that was happening, South Korea's stock market triggered its ninth circuit breaker in 40 days.

A lot moved this week. Here is what it means.

Starting in South Korea, the KOSPI triggered its ninth circuit breaker of 2026 this week, including back-to-back halts that have no precedent in exchange history. The index is down 41% in 40 days. To put that in scale: the 2008 financial crisis took the S&P 500 roughly a year to fall 57%. South Korea just lost a comparable magnitude in six weeks.

Samsung Electronics and SK Hynix, which together make up a significant portion of the index, are being destroyed by sustained foreign selling. The semiconductor trade unwind that started with Kimi K3 in Issue 008 is not a US market story anymore. It is a systemic shock to the Korean economy. The ₩2.5 quadrillion in market cap that has evaporated is not an abstraction. Pension funds, retail investors, and institutional money are all under water. The government has been intervening repeatedly and it has not held.

In the US, the S&P 500 ended the week roughly flat at 7,413. Earnings beats from Microsoft and Amazon did not produce a sustained rally. Instead, institutional money rotated hard out of mega-cap tech and into healthcare and financials, both of which hit record highs. That rotation is meaningful. It tells you professional investors are not convinced AI equity valuations are fully justified by the revenue numbers, even the good ones.

Boeing also reported. Revenue came in at $24.6 billion, ahead of estimates. But EPS landed at -$0.76 versus -$0.29 expected, driven by a $280 million charge on the Air Force One presidential aircraft program. The stock fell 4%.

Set the loss aside for a second. Free cash flow turned positive at $631 million. That is the first positive FCF print Boeing has posted in years. And 171 deliveries in the quarter is the best rate since 2018. Boeing is not fixed. But it is measurably less broken than the version of the company that existed 18 months ago. For recruiting purposes: this is what a turnaround thesis looks like at the data level. Revenue up, FCF positive, delivery cadence improving, but charges still widening the loss. The story depends entirely on which metric you weight.

The Federal Reserve held rates at 3.50%-3.75% on Tuesday. That part was fully expected by markets.

The vote was not.

9-3.

Three regional presidents voted against the hold. Beth Hammack of Cleveland, Neel Kashkari of Minneapolis, and Lorie Logan of Dallas all dissented in favor of an immediate 25 basis point rate hike. That is the most overt public disagreement inside the FOMC in years, and arguably the most significant display of internal friction since Chair Kevin Warsh took over.

To understand why this matters, you have to understand Warsh's communication strategy. His July statement ran 130 words. Jerome Powell's average under his tenure was over 300. Warsh has deliberately compressed public guidance. Less press conference theater. Less forward guidance. The idea is that markets should price in uncertainty rather than front-run Fed language.

Under that framework, three dissenters is the loudest signal available. The Fed cannot shout. Warsh will not shout. But Hammack, Kashkari, and Logan just did, through the vote itself.

The statement cited Middle East geopolitical uncertainty as a rationale for holding. That framing matters. It tells you the Fed is now treating geopolitical risk as a rate decision input, not just an equity market concern. When geopolitics is on the short list of reasons to hold, the risk calculus inside the FOMC has expanded beyond the dual mandate.

Markets are now pricing two hikes in 2026. September odds sit at 82%. For a committee that has been pausing for five consecutive meetings, that repricing is significant.

The takeaway for how you think about this sector: a divided Fed with a hawkish tail has real second-order effects. Higher-for-longer compresses LBO multiples, extends duration risk in fixed income, and keeps cap rates elevated in commercial real estate. These are not abstract macro positions. They are the conditions your future employer is underwriting deals against right now.

Heading into this week, the market had a binary question: is $725 billion in AI infrastructure spending generating real revenue, or is it the most expensive forward bet in technology history?

Microsoft and Amazon answered. Apple complicated it.

Microsoft

Revenue: $90.0 billion. Up 17.75% year over year. That beat estimates by more than $1 billion.

Azure: up 43%. Microsoft Cloud crossed $100 billion in annual revenue for the first time in history. Copilot commercial seats surpassed 30 million. But the number Wall Street spent the most time on was this: $678 billion in commercial remaining performance obligations, up 84% year over year.

That number is contracted future revenue. Companies have signed agreements with Microsoft and the cash is coming. The enterprise AI thesis is not a forecast. It is an order book.

Amazon

Total revenue crossed $200 billion in a single quarter for the first time in the company's 30-year history. AWS grew 37% year over year, its fastest rate in 18 quarters. AWS revenue hit $42.2 billion. Operating income rose 43%, with a 39.4% operating margin.

CEO Andy Jassy disclosed that Amazon's AI business and its custom silicon business are each running at more than $25 billion annually. Two new lines of business, both at $25B+ run rates, both accelerating.

The infrastructure AI thesis is not just working. It is compounding.

Apple

Apple beat on EPS at $2.02 versus $1.89 expected. iPhone revenue was up 22%. Those are strong numbers.

But services revenue missed. $30.74 billion versus the $31.22 billion consensus estimate. And guidance came in below expectations, with management citing supply constraints as the limiting factor. The stock fell more than 6% after hours.

The read: Apple's AI strategy is working as a hardware driver. The iPhone upgrade cycle has been real, and AI features have contributed to it. But the services layer, which is where the recurring margin lives, is not accelerating the way the market needed it to. Consumer AI monetization, whether through subscriptions, in-app AI tools, or intelligence-driven services growth, is still not a clean story at the revenue line.

AI is producing real, measurable, accelerating revenue at the infrastructure layer. Microsoft and Amazon proved it. At the consumer monetization layer, the story is still being written, and Apple is the proof.

That distinction is going to matter in interviews this fall. When someone asks you about AI and equities, knowing the infrastructure vs. consumer split is not a detail. It is the whole thesis.

Forward to someone who needs it. More next week.

— Clip’d by Pretus