Everyone’s watching Wyoming. Almost nobody’s watching the cash flow.
This week the market has its eyes on Jackson Hole. But the real signal isn’t in Wyoming — it’s in the balance sheets of the tech companies funding the AI race. On Friday the Dow rose 517.80 points (+0.98%), the S&P 500 gained 0.43%, and the Nasdaq added another 0.43% — yet all three closed the full week in the red, hit by bond-market volatility. The underlying reason isn’t one data point, it’s a wait: the FOMC went into its July meeting split 9-3, the widest gap in two decades, and the market still has no clear framework for reading the new Fed chair. Kevin Warsh, confirmed by the narrowest margin in the role’s history (54-45), gives his first Jackson Hole speech as chair next Friday, August 28. Until then, the Fed’s “reaction function” — which data point actually moves the needle — remains unknown, which is exactly why institutional investors are trimming directional exposure instead of betting on either side of the dual mandate. For anyone managing a portfolio, the practical takeaway is simple: this isn’t the week for directional conviction on rates — it’s the week to manage position and wait for the data, not the narrative.
Nvidia reports August 26 — but the number that matters isn’t quarterly revenue
Nvidia reports earnings on August 26, and Wall Street consensus projects revenue near $28.7 billion (+50% year over year), with the data-center segment contributing over $25 billion. What actually matters is what’s backing that number: hyperscaler backlogs are now verifiable — Google Cloud has accumulated $514 billion in pending contracts (+$52 billion this quarter), and AWS reached $496 billion, 2.5 times its level a year ago. That turns Nvidia’s report into the real thermometer for the AI cycle: if compute demand holds firm, it validates that these backlogs are real future revenue, not promises. If it disappoints, the market starts asking how fast all that capex is actually being monetized. If you’re curious how to govern AI-assisted decisions in your own financial analysis, we covered that in this column on AI in finance.
The idea that the “AI bubble” is about to burst is making the rounds
The noise: the dominant narrative across social media and financial press is that the “AI bubble” is about to burst, fueled by a striking figure — Big Tech generated $599 billion in net income over the past twelve months, but free cash flow was just $169 billion. A $430 billion gap that many read as a sign of unsustainability.
The signal: that gap isn’t inflated profit or an accounting hole — it’s deliberate leverage to fund infrastructure that’s already under contract, and hyperscalers issued roughly $220 billion in debt in 2026 specifically for that. The real risk isn’t that AI demand is fictional; it’s a timing risk: if those contracts take longer than expected to convert into recognized revenue, the correction comes from a financing mismatch, not from a lack of demand. As strategist José M. Echagüe puts it: the income statement tells you if the company is making money; cash flow tells you how long it can sustain the bet without asking permission. That’s the variable to watch before Nvidia’s report — not the “bubble” headline.
“Be fearful when others are greedy, and be greedy when others are fearful.”
— Warren Buffett, “Buy American. I Am.” op-ed, The New York Times, October 16, 2008. With the Fed more divided than it’s been in two decades and a $430 billion cash-flow gap fueling the “AI bubble” headline, Buffett’s discipline applies to the letter: the moment to watch real leverage is before consensus gets scared, not after.