Week ending Friday, August 28, 2026
Grow Your Pile is educational and is not investment advice or a recommendation to buy or sell any security. Options involve risk and are not suitable for every investor.
This Week’s Market Intelligence
I read this week’s coverage with one filter: what actually matters for positioning into September?
Two questions that had been hanging over this market got answered in the same week, and they didn’t answer in the same direction. NVIDIA confirmed the AI spending boom is still enormous. Fed Chair Kevin Warsh made it clear that sticky inflation could still mean higher rates. That’s an unusual setup — exceptional corporate fundamentals colliding with a less friendly Fed.
The fundamentals answered first, and loudly
NVIDIA guided to roughly $108 billion of third-quarter revenue, ahead of Wall Street, and projected about 70% revenue growth for fiscal 2028. NVIDIA and AWS are planning another 2 million GPUs across 2027 and 2028. Whatever you believe about an AI bubble, demand for computing infrastructure is not collapsing. (Reuters)
amit flagged the scale of the reaction — NVIDIA adding roughly $453 billion of market cap in a single session as Wall Street rushed to reset targets. We haven’t independently verified that as an all-time record, so take it as the order of magnitude rather than a league table. Either way, that is not a stock reacting to a beat. That’s an entire thesis being re-underwritten in a day.
Doug Kass is taking the other side, and he did it the same week. His argument this week: the AI funding boom has peaked — and his tell is the rush to the exit window, that AI companies are hurrying to go public. He’s not disputing that NVIDIA sold the chips. He’s questioning how long the money that buys them keeps showing up on these terms. Worth sitting with on a week when the tape voted hard the other way.
Then the Fed answered, and the answer was harder
Warsh’s first Jackson Hole speech was the more important event, and it went the other way. He said the Fed would have “more work to do” if policymakers can’t gain confidence inflation is moving far enough toward the 2% target. The two-year yield rose. Odds of another hike went up.
amit caught the part most people missed — the market’s first reaction was wrong: “they really pumped everything 5 min after Warsh’s speech as if he wasn’t the most hawkish we’ve ever seen. 40 min later everything reacted the way it probably should have right after his speech.” His read on the motive is worth holding loosely but hearing: “kinda feels like he was hawkish to ensure credibility vs actually signaling.” A new Fed chair has to buy credibility before he can spend it.
Warsh’s own words, on the inflation data: “While this summer’s PCE and CPI readings were better than expected, they do not tell me that underlying trends have meaningfully improved.” Read that twice. He’s saying good prints aren’t enough. He wants the trend, and he hasn’t got it.
The data underneath is doing both things at once
Liz Ann Sonders posted the August Kansas City Fed manufacturing survey, and it’s the whole conflict in one release. The headline came in at +10, matching estimates and up from +9 prior. New orders came in strong at +16 against +10 prior. But look at the two numbers at the ends: prices paid rose to +55 from +52, and employment fell to 0 from +2.
Costs going up while hiring goes to zero. That is precisely the box Warsh is in, and it’s why this doesn’t resolve cleanly in September.
Trader’s lens. Two answers, opposite directions, and the volatility sits in the gap between them. That’s our window. Inflation prints and the jobs number are volatility events again, which is exactly when short-duration premium is worth selling — but the discipline is not getting long delta at the highs just because earnings are good. Watch the 2-year as closely as the S&P.
Investor’s lens. Don’t read a hawkish Fed as a reason to leave a bull market. Read it as a reason to be pickier inside one. Higher-for-longer rewards real earnings, real free cash flow and manageable debt, and it punishes businesses that need cheap money to function. That’s not a market call. That’s a quality filter.
Where The Year Actually Stands
Small caps still lead the year, and they’re still the most exposed to the rate question for the same reason they always are — the Russell carries the most floating-rate debt, so it wins hardest when cuts are coming and hurts hardest when a hike comes back on the table. Long bonds still haven’t found a floor.
Gold is the one worth pausing on. GLD has given back a good part of its year and sits at +2.66%. We bought into that weakness on Tuesday, restarting the Portfolio 1 gold long by selling the Dec 18 395 put for $1,080. Rick Rule made the case for that side of the trade this week, in conversation with Keith McCullough at Hedgeye: “The dangers of sitting out the gold market are much more probable than the risks of being in it.” His argument is asymmetry rather than prediction — as U.S. debt climbs and purchasing power erodes, being absent costs you more than being early. That’s roughly why we sold a put instead of buying the metal outright: we get paid to wait, and we own it lower if it comes to us.
The Five Stories That Moved Markets
1 · NVIDIA proved the AI boom is still very much alive
The week’s biggest corporate event delivered. Roughly $108 billion of Q3 revenue guided, above expectations, with about 70% revenue growth projected for fiscal 2028, and another 2 million GPUs planned with AWS across 2027 and 2028. Despite the constant bubble debate, demand for computing infrastructure remains enormous. (Reuters) The dissent worth reading is Doug Kass’s — not that the chips aren’t selling, but that the funding behind them has peaked.
Trader Take: The AI trade still has momentum, but don’t assume every AI stock deserves to rally on it. NVIDIA’s results raised the bar for everyone else in the sector. Look for relative strength in semiconductors, networking, memory, power and data-center infrastructure — and treat pullbacks as more attractive than chasing gaps higher.
Investor Take: This report strengthened the secular thesis. The more interesting opportunity may increasingly sit beyond NVIDIA itself, in the companies supplying the chips, memory, networking, power and physical infrastructure needed to build AI capacity.
2 · Powell’s gone, and Warsh just handed Wall Street a more hawkish Fed
Warsh’s first Jackson Hole speech was arguably the week’s most important macro event. He said the Fed would have “more work to do” if policymakers can’t gain confidence that inflation is moving sufficiently toward 2%. Markets took it hawkishly — the two-year yield rose and expectations for another hike increased. (Reuters)
That matters because this market has spent much of the bull run assuming the next major move in rates would eventually be down. Warsh is reminding everyone that isn’t guaranteed.
Trader Take: Keep watching the 2-year and 10-year. If yields break higher again, expensive technology gets vulnerable even when the fundamentals are fine. Financials may benefit from a higher-rate environment.
Investor Take: Don’t fight the Fed. Higher-for-longer favors companies with real earnings, strong free cash flow and manageable debt over speculative businesses dependent on cheap capital.
3 · Inflation is refusing to cooperate
The reason Warsh can stay hawkish got clearer this week. July PCE came in hotter than expected, with the headline running at roughly 3.7% year over year against a 3.6% consensus, and that is the number that matters even though Warsh called the summer’s readings overall better than expected — a decent average doesn’t fix a trend, and July is the direction of travel. Markets promptly raised the odds of another hike. (Reuters)
That sets up the central conflict facing this market:
AI and earnings say bullish. Inflation and rates say caution.
Trader Take: Inflation reports are major volatility events again. If inflation stays sticky, selling short-duration premium around elevated volatility gets more attractive — but I’d stay careful about becoming excessively long delta at market highs.
Investor Take: Inflation near 3.7% isn’t disastrous, but it makes the path toward easier policy much harder. Pricing power, margins and balance-sheet strength become increasingly valuable.
4 · Under the surface, investors quietly pulled a lot of money from stocks
This may be the week’s most overlooked story. Global equity funds broke a 13-week streak of inflows, with roughly $5.9 billion withdrawn. More importantly, U.S. equity funds saw about $22.3 billion of net redemptions — their largest weekly outflow since March. Large-cap funds alone lost roughly $24.7 billion, while mid- and small-cap funds attracted money and bond funds kept taking inflows. (Reuters)
That doesn’t mean a correction is coming. It does tell us institutional positioning has turned more cautious near the highs.
Trader Take: Watch market breadth. If the S&P and Nasdaq sit near highs while fewer stocks participate, that’s a warning. Conversely, continued rotation toward small and mid caps could produce attractive relative-value opportunities.
Investor Take: This isn’t a reason to abandon equities. It’s a reason to rebalance oversized winners and keep some dry powder rather than getting more aggressive after a double-digit year to date.
5 · September’s first big test is already here: jobs
Next Friday’s employment report could be unusually important. July unexpectedly showed a decline in jobs, and economists surveyed by Reuters are looking for only a modest rebound of around 45,000 in August. (Reuters) With inflation still elevated, the Fed now has to balance two competing risks: persistent inflation against a deteriorating labor market. The Kansas City Fed’s employment reading falling to zero this week is a small piece of the same picture.
That creates an interesting paradox. A moderately weak report could actually be bullish, because it reduces the need for another hike. An extremely weak one could reignite recession concerns. A surprisingly strong one could push yields higher, because it gives Warsh more room to tighten.
Trader Take: Friday could produce a substantial move in bonds, QQQ and SPY. Don’t just watch the headline payroll number — watch wages, unemployment, and especially how the 2-year reacts.
Investor Take: We’re entering a phase where “good economic news equals good stock-market news” may no longer always apply. The market wants growth, but not so much growth that inflation and rates accelerate again.
What We Actually Did
We publish every trade, so here’s the week in numbers. Fourteen tickets across Portfolios 1 and 2, roughly $6,714 realized, and no new directional risk added.
Portfolio 2 spent the week managing the QQQ ladder, not growing it. Six rolls in two sessions. On Thursday the 719 went down to the 716 and the 696 and 699 both went up to the 705s, plus the SPY 745s rolled out to October for $2,555 realized on the closed leg. On Friday we did two more: the 716 down to the 712 for a 91-cent debit — paying for four points of distance rather than waiting to see what it cost later — and the 705s up to the 710 for a $156 credit, +$343 realized. Every one of those rolls stayed inside the same expiry or moved out a few days. No new size.
Portfolio 1 took profits and restarted a long. The BSH factory’s financing leg was retired for +$396 with tail coverage still in place, SPCX closed for +$300 once the volatility was gone, and the SPY Oct 16 715 put closed for +$1,209. Then we put capital back to work: sold the SPY Oct 16 750 put for $1,004 and restarted the gold long by selling the Dec 18 395 put for $1,080.
Portfolio 2 also ran a quick AMD trade — an Oct 16 jade lizard opened Monday for $1,032 with no upside risk, closed Tuesday for +$100.
Investor’s lens. Look at what this week wasn’t. With NVIDIA and Jackson Hole both landing, we didn’t add directional exposure into either. We rolled what we had, took profits where the trade had done its work, and kept powder dry. Two of Friday’s QQQ legs are underwater as I write this — QQQ closed at 716.43 with our short strikes at 712 and 710, five days out. That’s the trade we chose, at the size we chose, and you’ll see how it ends either way.
Trusted Voices
The people we read, in their own words.
amit (@amitisinvesting) — The most useful read of the week, twice. On NVIDIA, the scale: $453 billion of market cap added in a single day, “the largest one-day market-cap gain ever recorded in the U.S.” On Warsh, the timing: “they really pumped everything 5 min after Warsh’s speech as if he wasn’t the most hawkish we’ve ever seen. 40 min later everything reacted the way it probably should have right after his speech.” And the motive, offered as a question rather than a claim — “kinda feels like he was hawkish to ensure credibility vs actually signaling.” He also carried Warsh’s own line, which is the sentence to keep from this week: “While this summer’s PCE and CPI readings were better than expected, they do not tell me that underlying trends have meaningfully improved.”
Liz Ann Sonders (Schwab) — The data, with the tension left in. Midweek she posted the hard numbers straight: July durable goods orders +1.1% month over month against +0.5% expected, and the second print of Q2 GDP at +1.5%, in line, with personal consumption at +3.4% beating +3.2%. That is not an economy rolling over. August Kansas City Fed manufacturing at +10 versus +10 expected and +9 prior, new orders strong at +16 against +10 prior, shipments easing to +17 from +20. The two that matter for the Fed’s problem: prices paid up to +55 from +52, and employment down to 0 from +2. She also put out a new On Investing episode on markets, the Fed versus Treasury and Warsh’s speech. We’re flagging it, not summarising it — we haven’t heard it yet.
Keith McCullough (Hedgeye) — Constructive all week, and building the case from different evidence each day rather than repeating himself. Tuesday his chart was “High Yield HYG is voting for Quad 1” — credit confirming what breadth had said the day before. Thursday it was his software book: longs up 32%, 43% and 78% since May 26 against IGV at +9%, with the note that he was “pruning and planting” rather than adding. Friday he walked through risk-managing Amazon and reiterated the Affirm long. That posture is worth borrowing regardless of his book — a constructive process that spends a hot week trimming and rotating instead of pressing is doing the same thing we did.
Rick Rule (via Hedgeye) — The clearest statement of the gold case we saw this week, from his conversation with Keith McCullough: “The dangers of sitting out the gold market are much more probable than the risks of being in it.” He frames it as asymmetry against a backdrop of climbing U.S. debt and eroding purchasing power — not a target, a posture. It lands the same week gold gave back a good part of its year-to-date gain and the week we restarted our own gold position.
Doug Kass (Seabreeze) — The contrarian, and the only voice we read who pushed back on the week’s dominant story. He argues the AI funding boom has peaked, and reads the rush of AI companies toward public markets as the tell. He isn’t disputing that NVIDIA sold the chips — he’s asking who keeps funding the buyers, and how long on these terms. Agree or not, he publishes the argument and the positioning together, which is the standard we hold ourselves to.
Bottom Line
The fundamentals are bullish, but the cost of money is becoming the risk.
NVIDIA removed one major concern this week: AI demand is not collapsing. Corporate earnings remain exceptionally strong — second-quarter S&P 500 earnings growth ran around 33.5% year over year, the fastest since 2021. (Reuters)
But the Fed put another concern firmly back on the table. Rates may still have to go higher.
For traders, I’d stay constructive but disciplined. Keep positive exposure, take advantage of the volatility, and watch yields almost as closely as you watch the S&P.
For investors, I wouldn’t read this week as a reason to exit the bull market. I’d read it as a reason to get more selective. Favor quality, earnings and cash flow. Rebalance the positions that have grown oversized. Keep some buying power for the volatility.
September is setting up as a fight between two enormously powerful forces: AI-driven earnings growth, and a Fed that isn’t yet convinced inflation is beaten.
One more thing, coming separately. We’re doing the technical work in its own note this weekend — a full chart breakdown of the Japanese yen, which has been quietly one of the more important macro stories nobody is watching closely enough. Look for it in your inbox.
Grow Your Pile is educational and is not investment advice or a recommendation to buy or sell any security. Options involve risk and are not suitable for all investors. Selling puts obligates you to purchase 100 shares per contract at the strike price regardless of how far the underlying has fallen; the maximum loss is the breakeven price times the multiplier and can far exceed the premium collected. Rolling a position extends the time you are exposed and does not reduce the obligation. Past performance is not indicative of future results, and results shown are those of our own accounts and are not representative of any subscriber’s results. Read the Characteristics and Risks of Standardized Options before trading.
Tony Battista and Tony Rihan Grow Your Pile
Every options trade in all three portfolios is published, win or lose, with entry, exit and running P&L at growyourpile.com.