Running Hot Into Scarcity: Why the AI Trade Is About to Hit a Wall

Jordi Visser's latest macro breakdown is one of the most important pieces of analysis we've covered this year. The former CIO of Weiss Multi-Strategy Advisors isn't calling the AI trade a bubble — he's saying something more nuanced and arguably more dangerous: the fundamentals are real, but the physics of the situation are about to change.

The headline: Jordi has exited his entire semiconductor position, including the Micron trade that delivered him a 5–8x return. He's not shorting. He's rotating — into silver, Bitcoin, and commodities. And his reasoning comes down to three converging forces that most of the market is ignoring.

---

The Exhaustion Model Is Flashing Red

Jordi has built a proprietary screening tool he calls the exhaustion model. It classifies stocks into tiers — extreme exhaustion, elevated exhaustion, somewhat exhausted, and low exhaustion — based on the magnitude and speed of their moves, cross-referenced with candlestick pattern analysis.

When he tested it against last Tuesday's market action, the results were striking. The average move across stocks in the extreme exhaustion category came in at 4.1x — almost exactly where the model predicted reversals would cluster. The median lined up across every tier except one.

For options traders, his practical takeaway: find the names sitting in extreme exhaustion, then look for which ones have implied volatility that hasn't caught up to the move yet. That's where the mispricing lives.

The candlestick patterns he's tracking tell the same story. Micron has had large wicks on seven of its last eight trading days — meaning intense battles between buyers and sellers at those levels. He's seeing the same reversal patterns in South Korean semiconductor names, French AI plays, and Japanese tech. This isn't isolated. It's global.

---

The Hormuz Bottleneck Nobody Is Pricing

This is where Jordi's analysis gets uncomfortable. The Strait of Hormuz remains shut. That's not opinion — it's fact. And while markets have priced in the oil impact, they haven't even begun to price the downstream effects on semiconductor manufacturing.

Every component in a data centre rack — cooling systems, advanced packaging, power semiconductors, optical fibres, petrochemicals for chip production — depends on something flowing through the Strait. Sulfuric acid, critical for mineral processing, isn't getting through. Mosaic is curtailing fertiliser production due to sulfur shortages. Exxon and Shell have informed major suppliers including Costco to prepare for empty shelves of motor oil within weeks.

The semiconductor supply chain runs on just-in-time logistics. When 20% of global oil flows are blocked, it's not just an energy story. It's a production story. And production risk is what kills earnings momentum — not demand destruction, but the inability to deliver against demand that's still growing.

Jordi's key insight: Q1 earnings were spectacular because of massive hoarding and front-loading. Companies ordered everything they could get their hands on. The One Big Beautiful Bill's bonus depreciation provisions meant that capex spending flows straight to revenue for the recipients while expenses get stretched out for the spenders. That's an accounting tailwind that flatters earnings in the short term.

But if Hormuz stays shut and production can't keep pace, volumes slow. If volumes slow and pricing normalises simultaneously, you get hit on both sides. That's the risk nobody is modelling.

---

DRAM: The Leading Indicator That Just Rolled Over

Jordi overlaid two charts that every semiconductor investor should see: the six-month rate of change in DRAM pricing against SOX relative to NDX performance, going back to 2017 using Z-scores.

The correlation is tight. When DRAM pricing momentum turns negative, semiconductor outperformance over the Nasdaq follows — every time. Sometimes DRAM leads by months, sometimes they roll together, but the direction is consistent.

The rate of change in DRAM pricing has now turned negative. SOX relative to NDX hasn't caught down yet. That gap is the trade.

As Jordi puts it: once DRAM momentum rolls over, the semi trade becomes acutely sensitive to bad news because the market can no longer lean on accelerating memory prices to justify positioning. And bad news — in the form of Hormuz-driven production constraints — is coming.

---

The Institutional Stampede

Goldman Sachs' trading desk reported that sovereign wealth funds and asset managers have been panic-buying into AI names since the calendar flipped to May. This is a classic late-cycle momentum chase — institutional managers who were structurally underweight AI are now being forced to chase performance to protect their benchmarks.

Jordi has written extensively about this dynamic. His piece "Your Capex is My Opportunity: The Benchmark Arbitrage of the AI Buildout", published by Institutional Investor, lays out why this forced rotation is structural but also why it creates the conditions for violent unwinds when momentum breaks.

The momentum factor has hit 70 on both the weekly and monthly timeframes simultaneously. Historically, when both trigger at the same time, momentum tends to stall and then underperform for roughly a year. We're at a five standard deviation overshoot of price momentum versus nominal. Semiconductors are trading 62% above their 200-day moving average.

These aren't bearish indicators in isolation. They're stretch indicators. And when you combine stretch with a supply chain catalyst that can actually break the earnings trajectory, the risk-reward shifts dramatically.

---

The Amazon Analogy

Jordi makes a critical distinction that separates his view from the permabears. Amazon went from a $400-equivalent price to a 9x return between 2006 and 2020. Nobody called it a bubble because it took 14 years. The AI semiconductor trade has delivered the same magnitude of move in months.

The issue isn't whether semis deserve to be here. They probably do. Global semiconductor revenue is on track to exceed $1.3 trillion in 2026, up over 60% from last year. The capex is real. The demand is real. Agentic AI requiring 1,000x more compute than generative AI — as Jensen Huang said this week — is real.

The issue is time compression. When you compress a 14-year move into six months, you get parabolic price action that is inherently unstable. And parabolic moves end in speed crashes, not gentle consolidations.

---

The Rotation: Where the Money Goes Next

Jordi's positioning tells you where he thinks the next parabola forms: crypto and commodities.

Bitcoin is sitting below its prior highs while parabolic moves are already underway in adjacent markets — tokenisation and stablecoins are going vertical. The Clarity Act is moving through Congress with 72% probability of passing before year-end. The UK is relaxing stablecoin restrictions. Amazon has integrated crypto payments through Coinbase and Stripe via its Bedrock agent platform.

Jordi's signal for when the retail crypto wave begins: Dogecoin. When DOGE breaks above its current consolidation range, that's the sign that retail is re-engaging with crypto. It took roughly a year for Micron to transition from institutional to retail participation — he expects a similar timeline for the crypto rotation.

Silver is his other conviction position. With CPI breaking above three-month bills, we're back in a negative real yield regime. That's historically the strongest environment for precious metals. The scarcity trade — which has been the dominant theme of his thematic portfolio all year — intensifies from here.

---

The Consumer Is Already Collapsing

Buried in Jordi's data is a chart that should alarm anyone still positioned in consumer discretionary names. Equal weight consumer discretionary versus equal weight consumer staples has fallen to its lowest level since liberation day. This isn't cap-weighted — Amazon and Tesla aren't distorting it. This is the actual consumer economy.

Of 26 household durable names, 22 are down year-to-date. Restaurant stocks are getting destroyed — DoorDash down 34%, Wingstop down 50%. Real wages have now turned negative, with CPI running above the Atlanta Fed median wage tracker for the first time in this cycle.

This matters because it's the other side of the bifurcated economy. AI capital spending is booming. Consumer spending is cracking. Both things are true simultaneously, and most portfolios aren't positioned for both.

---

What to Watch

Jordi's checklist for the weeks ahead:

DRAM spot pricing — daily. The leading indicator for semiconductor relative performance. It's already turned. Watch for acceleration.

Hormuz shipping data — weekly. Every week the Strait stays shut, inventory buffers thin further. The production impact is cumulative and non-linear.

Korean semiconductor ADRs — daily. South Korea is the canary in the coal mine for global semiconductor momentum. The candlestick reversals started there.

Dogecoin — weekly. The retail signal for crypto re-engagement. When it breaks out, the rotation accelerates.

Hyperscaler stock performance relative to S&P — weekly. Still down year-to-date on a relative basis despite the AI narrative. If they roll over again, the broader market loses its last support.

Breadth — daily. S&P 500 breadth broke below its 50-day moving average and has been flat since January. Hindenburg Omens have triggered on both NYSE and NASDAQ simultaneously — only the 19th time in history. The S&P win rate over the following five months in those instances is below 50%.

---

The AI buildout is a $90 trillion secular trend. It's not going away. But secular trends don't move in straight lines, and the combination of parabolic price action, rolling supply chain bottlenecks, and institutional capitulation buying creates the conditions for a speed crash — not a bear market, but a violent repricing that creates the next entry point.

Jordi's not telling you to be bearish. He's telling you to be ready.

---

This analysis is based on Jordi Visser's weekly macro update for subscribers, May 2026. Jordi is the former CIO of Weiss Multi-Strategy Advisors and one of the most respected macro strategists in institutional finance. PTL Signal tracks his Scarcity framework and other macro indicators at ptlsignal.com/visser.

For informational purposes only. Not financial advice. Do your own research.