Quality Compound

Jamie Dimon and Stan Druckenmiller both flagged US debt & fiscal trajectory as unsustainable within 10 years — back in April 2016. That was 10.5 years ago.

Time's up. The math hasn't improved.

Nothing a significant USD devaluation vs. gold can't fix — but that's the kind of reset nobody wants to admit is already in motion.

Quality investing comes down to a simple filter: would you own this company for 5-10 years if the market went dark tomorrow? If the answer's no, move on.

We're hunting businesses the market — or even a rough macro cycle — can't break. Durable earnings power, pricing leverage, strong balance sheets, rational capital allocation. The kind of names that compound through any environment.

Great stock = great company + great valuation. Both matter. Overpaying for quality still hurts returns. But start with the business — if it can't stand on its own fundamentals for a decade, valuation is irrelevant.

That's the bar.

As an investor, not an economist — the difference matters. My views don't stay on a blackboard. They go into real portfolios, real capital allocation.

None of us know the future. But the odds we'll look back at today's rates as "the good old days" just went up. Cost of capital matters. If you're building a portfolio for the next cycle, this is the environment you're pricing in now.

Clear signal here: if this trajectory holds, inflation's heading significantly higher over the medium term.

For portfolios: long-duration bonds remain a fade. The risk/reward on long-end Treasuries continues to deteriorate. Real assets, inflation-linked instruments, and quality equities with pricing power look better positioned.

This isn't a short-term trade call — it's a structural shift in how we think about fixed income allocation and real return generation. Duration risk is real.

Counterintuitive dollar dynamics worth watching: When USD strengthens too aggressively, foreign holders don't get to choose what they sell — they dump what they *can* sell to defend home currencies.

So if DC pushes real fiscal austerity to cut deficits, you'd expect Treasury supply to tighten. Instead, you get the opposite: effective UST supply surges as foreign central banks and sovereigns are forced sellers.

It's a reflexive loop. Stronger dollar → FX stress abroad → forced Treasury selling → higher yields → more dollar strength. The very policy meant to stabilize debt markets can trigger the unwind.

Keep an eye on foreign Treasury holdings and currency volatility. If austerity talk heats up and the dollar rips higher, watch for disorderly moves in the long end.

Dollar has no real successor — and that's the point.

Europe? Worse off politically and economically. China? Not a democracy, currency controls, roach motel — you get in, can't get out. Economy's a black box with its own issues. Switzerland? Too small.

When capital wants out of the dollar, there's nowhere else to park. That's why it flows into the one reserve asset no government prints: gold.

This structural reality underpins both dollar strength and gold's role as the ultimate hedge. Not bullish or bearish — just how the system works.

Interesting structural point on US tech infrastructure: if the hyperscalers ($MSFT, $GOOGL, $AMZN, $META) are effectively too-big-to-fail with implicit government backing — especially given their AI buildout and national security importance — why would anyone buy long-dated Treasuries at lower yields?

You'd just buy hyperscaler bonds at slightly higher yields with the same implicit backstop. That dynamic could keep pushing long-end Treasury yields higher until something breaks or policy shifts. Worth watching as AI capex scales and these companies become even more systemically critical.

Violent short squeeze overnight — $58M in crypto liquidations in an hour as $BTC reclaimed $80K. Fed Governor Waller's comments just shifted September rate cut odds to a true toss-up. Yen strengthening again, which matters for carry trades and risk appetite.

Beneath the surface of equity indices, some interesting divergences forming. Not chasing here — waiting for cleaner setups. Long weekend ahead, so positioning matters.

Staying patient. React to what the market gives you, don't front-run narratives.

AI infrastructure valuations have stretched hard — we've seen this movie before with every major tech cycle. The question isn't whether we're in bubble territory (we clearly are in pockets), but whether the underlying business fundamentals can catch up to the multiples.

Right now, earnings quality from the hyperscalers and semiconductor plays is still solid. CapEx is real, deployment is accelerating, and monetization is starting to show up in results. That's different from pure speculation.

But yeah, when retail is piling into anything with "AI" in the name and you're seeing 50x+ forward multiples on companies with no clear path to profitability? That's froth. We've probably hit peak fear around the bubble narrative itself — which often marks a near-term bottom for sentiment.

Still long quality names with real revenue tied to AI infrastructure. Trimmed some of the more speculative stuff. This shakes out the weak hands, then we go again.

Energy looking like the most attractive sector right now. Healthcare second. Utilities just got downgraded.

Four warning signs building beneath the surface:

1. Semis underperforming — notable given AI narrative
2. Industrials breaking six-month support
3. Crude oil nearing breakout from 3.5-month base
4. Defensive inflationary leaders outperforming — historically not a bullish setup

Still constructive on the market, but tracking these pressure points. Risk management matters when sector rotation gets this pronounced.

S&P 500 sitting just 2% off all-time highs, but under the hood it's a different story — only 39% of components above their 20-day moving average.

I typically watch the 50DMA for real pullback setups, but this kind of short-term breadth deterioration while we're still near record levels? Worth noting.

Narrow leadership can persist for a while, but when internal momentum weakens this much, it often precedes either a healthy consolidation or something more meaningful. Watching how this evolves — quality names holding up better than the average stock right now.

Breadth is quietly deteriorating under the surface of the $SPY.

The Advance/Decline line is rolling over, and now more S&P 500 names are trading below their 20-day MA than above it. We saw similar breadth weakness right before the sell-off earlier this year.

Does this guarantee another leg down? No. But as we head into September and October — historically choppier months — I'm watching internals closely. The index can mask weakness for a while, but breadth often gives you the early signal.

Breadth matters. It's not a perfect timer, but when participation narrows this much, you pay attention.

Classic bubble cycle playing out again — rational early investment morphs into euphoria, overinvestment, then collapse. Most investors ride it up and give it all back. Only a handful of companies survive the shakeout.

Same pattern every time: railroads in the 1840s, dot-com in 2000, now AI. Early fundamentals get drowned out by imagination and momentum. The technology is real, the winners will be massive — but the crowd always overdoes it.

Stay disciplined. Focus on balance sheets, cash flow, and who's actually monetizing. The survivors will compound for decades.

Treasury Secretary Bessent says AI firms need to explain how they benefit the public.

The public already gets it. AI will "benefit" workers the same way offshoring to China "benefited" US manufacturing — which is still down ~27% from 2000 levels.

We've seen this movie before. Tech displacement without a plan for those displaced. The AI infrastructure build is real, the productivity gains are real, but let's not pretend there isn't a labor market reckoning coming for white-collar work just like there was for blue-collar.

As a PM, I'm long the AI winners — $NVDA, hyperscalers, the infrastructure plays. But from a macro perspective, the labor and political fallout from this transition is underpriced. History doesn't repeat, but it rhymes.

Andy nails the AI capex vs. inflation timing problem.

Bessent's bet: massive AI buildout = productivity gains = disinflationary within six months.

Reality check: productivity is still theoretical. Inflation from the buildout is already here.

Dallas Fed: data center electricity demand alone could add ~0.05pp to PCE in 2026. Allianz shows regions with heavy data center exposure already seeing +0.19pp on CPI from power costs.

Meanwhile: electricity prices +4.2% YoY, IT goods +3.6%, computers/peripherals +5.1%.

Why? Hundreds of billions competing for the same scarce inputs simultaneously: power, natural gas, transformers, copper, land, construction labor, GPUs, HBM, DRAM, advanced packaging. Companies borrowing heavily to finance it.

That's a demand shock before it's a productivity shock.

AI may eventually raise productivity and lower unit costs — but only after infrastructure is built, tech diffuses, and firms convert spend into measurable output.

Dismissing current inflation in energy, hardware, infrastructure, and credit with a future productivity story isn't serious analysis. It's narrative.

Bottom line: AI's disinflationary benefits haven't arrived. The inflationary bill is already being paid.

Private construction spending on manufacturing still down >20% YoY — even with the AI infrastructure build-out in full swing. That's the real tell.

Two things:

1) We're pouring tens of billions into geopolitical spending (Iran, etc.) instead of domestic productive capacity. Misallocation at scale.

2) If AI capex rolls over — and it will eventually — this chart gets uglier fast. The infrastructure spend is masking serious weakness in broader industrial investment.

We're not building enough real capacity. AI data centers aren't enough to carry the whole manufacturing sector. If that theme cracks, we've got a problem.

Oil calls moving toward $100 — supply tightness or demand optimism? Worth watching energy names if this holds.

SoftBank looking at another $10–20B bond raise to back OpenAI. Aggressive move, but fits the AI infrastructure theme we've been constructive on.

Goldman pushing AI capex peak out to 2028. If they're right, the hyperscalers, datacenter REITs, and semiconductor plays still have runway. Quality names in this space remain interesting on pullbacks.

Stay patient. React to what's real, don't chase predictions.

Kicking off the week with a look at how equities are digesting higher crude prices — $SPX holding up better than you'd expect given the energy move. Also running through positioning in gold, silver, and $BTC as we head into a data-heavy stretch. Market's showing some resilience here, but watching how sustained this is if oil stays elevated. Quality names still working.

Japan and China trimming Treasury holdings to fund domestic needs — but who steps in as buyers when US deficit financing needs are at all-time highs? Real structural question for rates and the dollar.

Patrick Oddoux walked through the creditor retreat dynamic: fewer natural buyers, bigger deficits, higher term premium likely. Also flagged France's fiscal stress ahead of elections — he's buying downside protection there.

European exposure? Defense and power grid infrastructure only. Real assets over duration. Short the long end makes sense in this setup.

Fundamental repricing of sovereign risk and funding dynamics playing out in real time.

September setup getting interesting. Korean traders unwinding leveraged chip ETFs fast. NAAIM exposure back near extremes. Warsh just said he'd be "hard pressed" to call financial conditions restrictive — that's a signal. NFP Friday, $SPX entering historically weakest month.

Lots of price action, data, and flows converging. Patience here. React, don't predict.