A blowout jobs report erased the rate-cut hopes. A dovish Fed chair walks into his first meeting with the market almost certain he must do the opposite of what he promised. And the artificial intelligence boom everyone is celebrating turns out to be the very thing keeping inflation alive.
Five numbers that frame the week: a strong economy became the market’s biggest fear.
“The cruelest trap in markets is a good number that arrives at the wrong time.”— Anthony Rosenthal, Weekly Market Pulse, June 2026
TL;DR
Markets in brief. The S&P 500 closed at 7,383.74, up 7.9 percent year-to-date but down 2.6 percent on Friday. The 10-year Treasury yield sat at 4.48 percent, near the high of the cycle, as the jobs report pushed investors to price higher-for-longer rates. WTI crude ended near 90 dollars, up 43 percent for the year. Gold held close to 4,485 dollars. The fear gauge, the VIX, jumped into the low twenties from the mid-teens.
The week’s hinge variables. Three conditions will decide where this goes. First, what Kevin Warsh’s first Federal Open Market Committee meeting on 16 to 17 June reveals through its “dot plot”, the chart in which each official marks where they expect rates to go. Second, whether the 2-year-to-10-year yield gap keeps widening as long-term borrowing costs climb. Third, whether the VIX pushes above the high-twenties into the meeting. Each is taken up in the Analytical Takeaway.
The crash gauge is still benign at 27, but for the first time this spring one signal has turned red.
The interest rate on a typical thirty-year US mortgage tracks the 10-year Treasury yield, and on Friday that yield sat at 4.48 percent, near the highest of the cycle. The reason is a single number, the May payrolls blowout set out in the summary above: good news for workers became bad news for markets, because a hot jobs market is exactly what stops a central bank from cutting rates. The irony of the year is that Kevin Warsh, who took the Fed chair signalling openness to earlier cuts, now walks into his first meeting on 16 to 17 June with the market roughly 85 percent certain he will be forced to raise rates instead.
The deeper point connects the jobs report to the technology boom. Torsten Slok of Apollo argued this week that the artificial-intelligence build-out is itself an inflation engine: the half-trillion dollars a year that the big cloud companies are spending on data centres is not an abstraction, it is real demand for power, copper, transformers, concrete and skilled labour, all arriving into an economy already running at 3.8 percent inflation. In plain terms, the thing the stock market is celebrating, the capital-spending super-cycle, is the same thing keeping the Fed pinned. Slok also noted there is, so far, “zero evidence” of jobs being lost to AI; if anything the spending is creating them. So the boom is not deflationary as the optimists assumed. It is the opposite.
For a saver or a borrower, the so-what is concrete. If Warsh’s dot plot shows a 2026 hike, the 2-year yield, which moves with rate expectations, climbs toward 4.3 percent, the 10-year tests 4.70 percent, and mortgage and business-loan costs stay higher for longer. If instead he signals patience, the 10-year could rally back toward 4.25 percent and the equity market exhales. The whole year now hinges on a single chart released at one meeting.
The build-out is not an abstraction: it is a bid for four physical things in finite supply.
| Input | What the AI build-out needs | The supply problem |
|---|---|---|
| Electricity | Hyperscaler capital spending running at roughly 500 to 600 billion dollars a year, much of it new data centres | Grid interconnection queues stretch to around 2,600 gigawatts with multi-year waits; new firm power cannot arrive in time |
| Copper | Every data centre, transformer and motor is copper-intensive | A projected supply shortfall of about 31 percent against 2035 requirements; a new mine takes 10 to 16 years |
| Transformers & grid kit | Medium- and high-voltage equipment to connect the load | Lead times measured in years; a global bottleneck that more money cannot quickly clear |
| Skilled labour | Electricians, linesmen and construction trades to build it | A tight labour market, the same one that just printed 172,000 jobs |
The takeaway a non-specialist can carry away: when half a trillion dollars a year chases power, copper, equipment and tradespeople that cannot be conjured at speed, the price of those inputs rises. That is the mechanism by which an investment boom turns into an inflation problem, and why it keeps the Fed pinned.
The Taylor Rule is a simple formula central banks use as a sense-check: it estimates where the policy rate should sit given inflation and the health of the economy. Plug in this week’s numbers — a neutral real rate of about half a percent, core PCE inflation at 3.3 percent (1.3 points above the 2 percent target), and an economy at roughly full employment — and the rule points to a policy rate of about 4.4 percent. The Fed funds rate today is 3.50 to 3.75 percent, a midpoint of 3.625 percent. In plain terms, policy is running roughly three-quarters of a percentage point easier than the rule prescribes. That gap is the cage: it is why a chair who wanted to cut may instead be told by his own framework to tighten. (An estimate from the rule, not a forecast; r-star is itself uncertain.)
The May jobs blowout, detailed in the summary, nearly doubled consensus with wages firm. The 10-year yield is at 4.48 percent. Market-implied year-end rate-hike odds have jumped to roughly 85 percent. Broadcom’s AI-guidance miss triggered a roughly one-trillion-dollar semiconductor selloff.
The AI capital-spending cycle and the inflation problem are one story, not two. A dovish chair is likely to be forced into hawkish guidance simply to ratify what the data demands. The equity market near records and the bond market at cycle-high yields are pricing incompatible outcomes; the 17 June meeting is where they reconcile.
A softer-than-expected dot plot would force a violent unwind of the hike trade and rally both bonds and rate-sensitive shares. A renewed Hormuz flare-up that pushes oil back to the mid-90s would harden the inflation case and remove the patience option entirely. A genuine consumer rollover, hinted at by Lululemon’s weak guidance, would reopen the cut debate from the demand side.
The problem nobody could manufacture. A stellarator is one of the two main designs for a fusion reactor, the kind of machine that tries to recreate the power source of the sun by holding a superheated gas, called a plasma, inside a magnetic cage. The stellarator’s great advantage is stability; its great curse is the magnets. To hold the plasma it needs coils bent into fiendishly complex three-dimensional shapes, each one hand-wound to tolerances finer than a human hair, each one effectively bespoke. For sixty years that manufacturing nightmare is a large part of why fusion has stayed twenty years away.
Thea Energy, spun out of the Princeton Plasma Physics Laboratory in 2022 and based in Kearny, New Jersey, was founded to attack exactly that bottleneck. Its bet is deceptively simple: replace the one impossible coil with dozens of small, identical, flat magnets, and shape the magnetic field not by bending metal but by adjusting the current in each magnet in software. The hard three-dimensional machining problem becomes a manufacturing-and-control problem, the kind of problem modern industry is good at.
Where Thea stands. On 27 May the company announced an oversubscribed 100 million dollar funding round, taking total private investment to roughly 130 million dollars. The lead investor is Thomas Tull’s US Innovative Technology Fund, alongside Idemitsu Kosan, a Japanese energy major whose presence hints at a future power-buyer relationship. The chief technology officer, Dr David Gates, is a stellarator authority from Princeton, which is the academic-to-commercial signal in its strongest form. The roadmap runs through a demonstration device called Eos, with construction targeted around 2027, to a commercial plant, Helios, targeted around 2035.
The strategic uncertainty. Here is the moment of genuine doubt that every honest case study needs. Thea’s architecture is elegant on paper, but no one has yet shown that an array of simple planar magnets can confine a plasma well enough to clear the net-energy-gain bar, the point at which a reactor produces more energy than it consumes. That is the bar that has defeated every fusion programme for half a century. And Thea is not alone: Commonwealth Fusion Systems and Helion are far better capitalised and are racing for the same prize, including the same small pool of data-centre customers willing to sign long-term power contracts. The company that locks up those buyers first may define the market before Thea’s first demonstration even runs.
The resolution, and the transferable lesson. What the oversubscribed round bought Thea is not certainty but time and a specific kind of credibility: a specialist, strategic syndicate rather than generalist momentum money, and a Japanese energy partner that signals a path to a real customer. The transferable lesson is the one that runs through every great hard-technology story: the winning move is often not a better version of the hard thing, but a way to turn the hard thing into an easy thing. Thea is not trying to build a better impossible magnet. It is trying to make the magnet ordinary. Whether physics co-operates is the open question, and it is genuinely open.
On our investment framework this scores as a watch-list name, not a buy: a clean, never-covered sector for our readers, a top-tier founder, and a perfect fit with this week’s theme that the AI build-out’s scarcest input is firm, clean power, but with pre-commercial physics risk and a 2030-plus timeline that keeps it firmly in the “watch and learn” category.
Who owns civilisation-scale power? If a handful of private firms commercialise fusion first, they will hold something no company has ever held: the ability to manufacture effectively unlimited, clean, firm electricity at a location of their choosing. The same data-centre operators funding fusion today would be among the first to buy that power tomorrow, which raises the prospect of the most fundamental resource in an economy being priced and allocated by a small number of private actors before any regulator has decided how it should be governed.
There is a second, quieter risk. Fusion is not fission, but it is not free of governance questions: the fuel cycle, the handling of tritium, and the export-control status of the underlying plasma-physics and high-field-magnet expertise all sit in a regulatory grey zone that has not kept pace with the science. An investor or analyst evaluating Thea, or any fusion start-up, should treat the concentration of energy supply and the immaturity of the governance framework as category risks, not company quirks.
Friday, 8:30am, Washington. The Bureau of Labor Statistics released a single line of data, and within minutes it rewrote the market’s view of the rest of the year. The payroll figure was nearly double what forecasters expected, and that was precisely the problem. By the close the major indices had fallen hard, the Nasdaq posting its worst day since April 2025. Nothing broke. The economy simply did too well, and a market that had spent the spring betting on rate cuts had to tear up the bet.
The chip rout. The selling started a day earlier. Broadcom reported a record quarter on Thursday evening, with revenue up 48 percent and AI-chip sales up 143 percent, and the shares fell about 15 percent anyway, because the forecast was merely very good rather than spectacular. When the most important AI chipmaker after Nvidia disappoints an elevated bar, the whole sector pays: Marvell and Micron fell around 16 and 13 percent, Intel and AMD around 11. Roughly a trillion dollars of share value evaporated. It was a useful reminder that in a momentum market, the punishment is not for missing, it is for not beating by enough.
The consumer tell. Beneath the macro drama, Lululemon offered a clean read on the median household. The athletic-wear retailer beat on profit but cut its outlook, with gross margin down 410 basis points, most of it tariffs, and North American same-store sales down 6 percent. Tariffs are now visibly passing through to prices, and the discretionary shopper is starting to flinch. That is the demand-side counterweight to the inflation story: the build-out is hot, but the household is cooling.
Oil’s round trip. Crude spent the week reversing itself. WTI ran from the high-80s back above 95 dollars mid-week on renewed Iranian missile activity, then settled near 90 by Friday. The round trip, and what it says about reading oil against the diplomatic headlines, is covered in Geopolitical Watch.
Crypto’s quiet exit. While shares sat near records, Bitcoin fell about 13 percent on the week to roughly 62,000 dollars, on a record streak of withdrawals from Bitcoin funds. The point worth noticing: this happened without a broad risk-off panic. Money is leaving digital assets even when other risk markets are calm, which looks like a structural rotation rather than a mood swing. Ethereum fared worse, down toward 1,580 dollars and now off 47 percent for the year.
Most risk gauges are still calm; the one that moved is the yield curve, and it moved the wrong way.
A composite score of 27 says the same thing in plain language: the plumbing of the market is calm even though shares had a rough Friday. Credit spreads are tight, company-default fear is low, and factory orders are rising. The single warning is the yield curve, and a curve steepening because long-term rates are climbing on inflation worries is a slower-burning risk than a credit-market seizure, not an imminent one. For a long-term investor the message is not to sell, it is to notice that the cost of being wrong about the Fed has gone up, and to make sure portfolio risk is sized for a higher-for-longer rate path rather than for the cuts the market spent the spring expecting.
Two years ago an AI model could hold a useful thought for a few minutes before losing the thread. This week Jack Clark, in his Import AI newsletter, reported that the length of time the newest systems can work on their own, without a human stepping in, has reached about twelve hours. Researchers at METR think hundred-hour autonomous tasks are plausible by the end of this year, and OpenAI has openly said it wants an “automated AI research intern” running by September. The significance is not any one figure; it is the slope. A machine that can work unsupervised for twelve hours is a different economic object from one that needs a minder every few minutes.
That is why the jobs debate in the Analytical Takeaway is not settled. The same week that Torsten Slok showed no AI-driven job losses in the hard data, Clark showed the capability that could eventually cause them accelerating. Both are true at once, which is the honest state of the argument.
This takes about four minutes. Paste the following into Claude: “You are an economist. In plain English, explain how the half-trillion dollars a year being spent on AI data centres could be feeding inflation rather than reducing it. Then give me the single strongest counter-argument.”
What Anthony found: the model produced a clean two-channel answer, that data centres bid directly for power, copper and skilled trades while their productivity payoff arrives only later, and then offered the sharpest rebuttal, that if the productivity gains land they are disinflationary with a lag. That is exactly the Slok-versus-the-build-out tension this edition is built around, reconstructed by a model in under a minute. The wider observation: most competitors are still describing the AI revolution. The ones pulling ahead are using it to stress-test their own arguments before they publish them.
For the sixteenth consecutive week, the Strait of Hormuz refuses to resolve. The “largely negotiated” framing that drove oil down almost ten percent in late May reversed inside a week: WTI ran back above 95 dollars mid-week as Iran launched ballistic missiles toward neighbouring states and US forces struck Qeshm Island in response, before settling near 90 by Friday. No formal signature, no stand-down by the Islamic Revolutionary Guard Corps, and the 60-day window on the memorandum runs only to early July. Iranian sources said on 1 June there would be no dialogue until the fighting in Lebanon ends.
The lesson the WMP has logged for four months holds: oil prices the reality of the Strait more honestly than the diplomatic headline does. The structural floor is unchanged, with roughly 12.8 million tonnes a year of Qatari liquefied-natural-gas capacity offline and a three-to-five-year repair timeline that is independent of any ceasefire.
The currency split. The strong dollar that followed the jobs report told two different stories in emerging markets. The Turkish lira slid further, with the dollar buying about 46 lira, as Turkey’s own inflation and credit strains continue. The South African rand, by contrast, has held its ground for the year, with the dollar near 16.45 rand, supported by the country’s commodity exports into a firm metals market. The same dollar move helps a commodity exporter and hurts a deficit-running importer; the two are not interchangeable, and reading them as one number misses the mechanism.
While the tape spent the week obsessing over a single jobs report and a single dot plot, here is a number that no central bank meeting will move. In 1820, only about one person in ten on Earth could read and write. Today the figure is roughly nine in ten. In the span of two centuries, humanity went from a world that was overwhelmingly illiterate to one that is overwhelmingly literate, and most of that gain came in the last seventy years.
The share of the world that can read has gone from roughly 12% to about 87% — most of it within living memory.
Source: Our World in Data, “Literacy rate” (UNESCO; van Zanden et al.). Long-run series.
The reason this matters to an investor is compounding of a different kind. Every percentage point of literacy is a larger pool of people who can be trained, can start a business, can read a contract, can use the very AI tools this publication writes about each week. The build-out in the Analytical Takeaway needs power and copper; it also needs minds capable of using what it builds, and the supply of those minds has never been larger. That is the deeper answer to this week’s hawkish trap: the same forces straining power and copper are also enlarging the human capability that ultimately pays the build-out back. Bad news, a hot jobs report read as a threat, makes the headlines. The quiet, century-long expansion of human capability does not, and it is the more durable trend.
The instinct to worry about the week’s rate scare is understandable; it is also incomplete. The data, over any horizon longer than a news cycle, points up.
On 3 June, SpaceX filed an amended prospectus to sell shares to the public for the first time. The numbers are without precedent: a raise of about 75 billion dollars at a valuation near 1.75 trillion dollars, with a Nasdaq debut targeted around 12 June under the ticker SPCX and Morgan Stanley as lead bank. If it prices at the top of the range, it will be the largest initial public offering ever, surpassing Saudi Aramco.
It is worth explaining the mechanics in plain terms, because they are the whole story. An initial public offering, or IPO, is the moment a private company first sells shares to ordinary investors. Three features of this one matter. First, the float, the slice of the company actually being sold, is tiny relative to the whole: raising 75 billion against a 1.75 trillion valuation means only a few percent of the company changes hands. Scarcity is engineered in. Second, the lock-up, the period during which existing insiders are barred from selling, means even fewer shares will trade in the early months, which tends to amplify price moves in both directions. Third, the structure keeps voting control firmly with the founder, so public shareholders are buying the economics, not the steering wheel.
The analytical frame, rather than a recommendation, is this: SpaceX is being brought to market as a scarce trophy asset, a fixed and tightly controlled supply of equity in a business with no real public-market comparable, sold into enormous demand. That combination, scarce supply meeting deep demand, is what reprices an entire category, and it is the same logic that drives prices for sports franchises and rare collectables. The risk it imports is the mirror image: when almost nothing trades freely, the price can detach from the fundamentals in either direction, and the first lock-up expiry, when insiders can finally sell, is the moment that detachment is tested. For a reader, the lesson is to separate the quality of the company, which is real, from the mechanics of the listing, which are designed to make the shares scarce before anyone has agreed what they are worth.
Follow the money. There is a second-order effect worth naming. Vacuuming roughly 75 billion dollars of institutional cash into one listing does not create that money, it relocates it, and it lands in the same week that Bitcoin saw a record run of withdrawals from its exchange-traded funds, falling about 17 percent to around 61,000 dollars, with Ethereum down 47 percent for the year. The cleanest reading is a structural rotation: scarce, prestige mega-cap equity pulling capital out of the assets investors hold when they have nothing scarcer to buy. It is a hypothesis, not a measured flow, but if the SpaceX book prices several times oversubscribed while digital-asset funds keep bleeding, the migration is the story rather than the coincidence.
Each week the WMP scores the central economic argument about artificial intelligence as a contest between two forces. The Displacer is the case that AI substitutes for human labour, concentrating the gains in capital and eventually destroying the spending power that the economy runs on. The Augmenter is the case that AI raises human productivity, expands output, and spreads the gains broadly. Both sides agree AI is powerful. They disagree about whether it is a demand shock or a supply shock.
This week, four pillars, one point each:
Running total — the Augmenter 7, the Displacer 1. This week the Augmenter takes three pillars to the Displacer’s one. The honest reading is not that the debate is settled: the automation horizon keeps the Displacer firmly on the board, but every hard number that landed this week, jobs, wages, financing costs, favoured the Augmenter. A single month of genuine white-collar displacement data would change that immediately.
This week’s deep dive looks at the first of the twelve income strategies the WMP tracks: an actively managed income fund, geared with a Lombard loan. The terms first. An income fund holds a diversified basket of dividend-paying shares, bonds and other yielding assets. A Lombard loan is simply borrowing against that portfolio, the way a homeowner borrows against a house, using the investments themselves as collateral. The idea is to earn the fund’s yield on both your own money and the borrowed money, keeping the difference.
Suppose an investor holds 1,000,000 dollars in a diversified active income fund yielding 7.0 percent, paying 70,000 dollars a year. A private bank offers a Lombard facility at 50 percent loan-to-value, priced at the short-term rate plus 1.5 percentage points. With the short-term rate near 3.6 percent, the borrowing cost is about 5.1 percent. The investor draws 400,000 dollars, well inside the limit, and buys more of the same fund. That borrowed slice earns 7.0 percent, or 28,000 dollars, and costs 5.1 percent, or 20,400 dollars, for a net carry of 7,600 dollars. Total income rises from 70,000 to 77,600 dollars, lifting the yield on the original capital from 7.0 to 7.76 percent.
The danger is the same one that bites every geared strategy. If the fund’s value falls far enough, the loan-to-value breaches the bank’s limit and triggers a margin call, a demand to repay part of the loan immediately, often forcing a sale at the worst possible moment. With short-term rates now more likely to rise than fall, the cost of the loan can also climb, squeezing that 7,600 dollar carry toward zero. The strategy works in calm, high-yield conditions and turns vicious in a drawdown.
Covenant note: negotiate a fixed margin spread and a notice period on margin calls rather than same-day repayment, and confirm whether the facility is callable at the bank’s discretion. The difference between a 48-hour notice and a same-day call is the difference between an inconvenience and a forced sale.
| Strategy | Indicative Yield | Spread vs IG | WoW | Action |
|---|---|---|---|---|
| 1. Active income fund + Lombard | 7.0 to 7.8% | +300 to 380bps | Cost up on hike risk | Watch |
| 2. IG / split-rated CLO tranches | 6.4% | +220 to 260bps | Stable | Hold |
| 3. Listed infrastructure debt/equity | 5.9% | +180bps | Firm on power demand | Add |
| 4. Business development companies | 10.8% | +560bps | Watch credit quality | Hold |
| 5. Agency mortgage REITs | 13.5% | +150bps (asset OAS) | Hurt by higher rates | Reduce |
| 6. Senior secured leveraged loans | 8.6% | +420bps | Floating-rate, resilient | Add |
| 7. Preferred shares / hybrids | 7.4% | +300bps | Soft on rate move | Hold |
| 8. Real asset royalties | 6.8% | +260bps | Oil round-trip neutral | Hold |
| 9. EM hard-currency sovereign carry | 8.1% | +390bps | Strong-dollar headwind | Watch |
| 10. High-yield municipal bonds | 6.2% (tax-free) | +275bps | Stable | Hold |
| 11. Private credit direct lending | 11.1% | +575bps | Spreads holding | Hold |
| 12. Trade & supply-chain finance | 9.0% | +450bps | Short-tenor, defensive | Add |
The week’s rate move tilts the whole book toward floating-rate and short-tenor strategies, six and twelve, and away from the rate-sensitive ones, five in particular. Geared strategies like the one above earn their keep only if the cost of the gearing stays contained, which is precisely what the 17 June meeting will decide.
What I am watching and why, not a recommendation to buy or sell. Two names carry a genuine new development this week; the rest sit in Portfolio Watch below, where MP Materials’ four-week score this weekend is noted.
What the market may be missing. The dominant theme this week is that the AI build-out is real demand for physical inputs, and rare earths sit at the heart of the magnets that go into motors, turbines and defence hardware. USAR is the rare American name trying to build a complete chain from mine to finished magnet, and this week the thesis kept playing out: the Department of Energy grant is funding the processing step, the part of the chain China dominates. The market still treats USAR partly as a speculative miner; the variant view is that the dual federal backing, a Defense Department purchase floor plus DOE funding, is closer to a utility contract than a mining gamble.
The historical parallel. The cautionary rhyme is Molycorp in 2010 and 2011, the last time a rare-earth monopoly narrative met strategic-necessity framing, then collapsed on production execution and a weak balance sheet. The key difference this time is the cost-plus government purchase structure, which removes the revenue risk that sank Molycorp. The parallel is a warning about execution, not about demand.
What would change the thesis. Brazil’s competition regulator, CADE, is reviewing the Serra Verde acquisition; a block there, or any withdrawal of the federal purchase commitment, would remove the spine of the case. That review is the live risk into the 13 June score date.
What the market may be missing. FCX returns to the radar this week with a fresh catalyst that ties directly to the edition’s theme: copper rose about 2 percent on the week, and Freeport itself rallied around 12 percent, the equity’s leverage to the metal on full display, exactly as Torsten Slok was arguing that the AI build-out is raw demand for copper, power and metal. FCX is the largest listed pure play on that demand. The market still prices it on the near-term Grasberg production ramp, where a 2025 mud-rush slowed output; the variant view is that the supply story dwarfs the mine-level noise, with roughly a 31 percent projected copper shortfall against 2035 requirements and new mines taking ten to sixteen years to build.
The historical parallel. The question is whether 2026 copper is 2004, the early innings of a multi-year supercycle, or 2007, the blow-off top. The honest answer is that the 2004 case rests on a supply deficit that physically cannot be closed quickly, whereas 2007 was a demand story that a recession ended. This is a supply story.
What would change the thesis. The US-China tariff pause, which expires in mid-August, is the swing factor; a breakdown there would hit copper demand sentiment and the shares with it. FCX was formally scored incorrect at its four-week mark in May on price, and this remains a structural, multi-quarter view, not a four-week call.
Spare a thought for the economists. For two years they have been forecasting the jobs market into a gentle slowdown that would let the Fed cut, and for two years the American worker has politely declined to co-operate. This month they predicted around 80,000 new jobs. They got 172,000. There is a particular kind of professional indignity in being wrong because the thing you study did better than you said it would, like a doctor disappointed to find the patient has gone for a run.
The wider comedy of the week was the sight of a market frightened by good news, selling shares because too many people found work. Somewhere in this is a lesson about what happens when an entire asset class is positioned for one outcome and the data insists on another. Mostly, though, it is a reminder that the market is not the economy, and on Friday the two were barely on speaking terms.
A jobs print came in far too strong,
And equities sang the wrong song;
The dove, now ensnared,
Must hike, it appears,
— the trap that he’d set all along.
2026 Scoreboard
25 assets ranked by year-to-date return · Baselines locked 1 January 2026 · Week 21 · 6 June 2026 · Friday 5 June verified closes (direct API) · Baltic Dry carry-forward, MSCI EM flagged (EEM/index baseline)
Five of the Scoreboard’s asset classes are still pointing in different directions — the Repricing thesis is intact.
2026 YTD Performance — All 25 Assets — Week 21
| Rank | Asset | 1 Jan 2026 Baseline | Week 21 Close | YTD % |
|---|---|---|---|---|
| 1 | Baltic Dry Index | 1,877 | 3,222* | +71.66%* |
| 2 | WTI Crude | $63.20 | $90.54 | +43.26% |
| 3 | USD/TRY | 35.40 | 46.06 | +30.13% |
| 4 | Nikkei 225 | 51,830 | 66,588 | +28.47% |
| 5 | DAX | 20,073.90 | 24,759 | +23.34% |
| 6 | Nasdaq 100 | 25,200.50 | 28,958 | +14.91% |
| 7 | Russell 2000 | 2,481.91 | 2,834 | +14.17% |
| 8 | Copper | $5.682 | $6.26 | +10.23% |
| 9 | S&P 500 | 6,845.50 | 7,383.74 | +7.86% |
| 10 | Euro Stoxx 50 | 5,740.15 | 6,062 | +5.61% |
| 11 | FTSE 100 | 9,948.30 | 10,368 | +4.22% |
| 12 | HYG | $78.15 | $79.43 | +1.64% |
| 13 | Swiss SMI | 13,248.10 | 13,388 | +1.06% |
| 14 | MSCI EM | 1,595.20 | 1,602* | +0.43%* |
| 15 | Gold | $4,341.10 | $4,337.10 | -0.09% |
| 16 | LQD | $109.02 | $108.17 | -0.78% |
| 17 | Silver | $70.61 | $68.94 | -2.35% |
| 18 | AGG | $102.20 | $98.17 | -3.94% |
| 19 | Nifty 50 | 24,415 | 23,367 | -4.29% |
| 20 | Hang Seng | 26,340 | 24,962 | -5.23% |
| 21 | Natural Gas | $3.514 | $3.23 | -8.11% |
| 22 | TLT | $94.27 | $85.06 | -9.77% |
| 23 | USD/ZAR | 18.63 | 16.31 | -12.47% |
| 24 | Bitcoin | $87,850 | $60,923 | -30.65% |
| 25 | Ethereum | $2,967 | $1,580.86 | -46.72% |
* Prior-period carry or derived value — Friday close not separately verified from a named source at production time. See exception report. The chart and table refresh live from the verified Supabase record for edition 21.
Every company that has appeared in On the Radar is tracked here until its formal four-week score date. A company moves from On the Radar to this appendix when there is no fresh catalyst that week — the analytical call is intact, but there is nothing new to add.
| Company | Entry | Week | Current Close | Return | Original Thesis | Score Date |
|---|---|---|---|---|---|---|
| Bloom Energy (NYSE: BE) | $207.10 | Wk 14 | ~$291 | +40.7% | On-site power for AI data centres; deployable in months versus multi-year grid queues | Scored Wk 18: CORRECT. 3-month track continues. |
| Arista Networks (NYSE: ANET) | ~$147.00 | Wk 20 | ~$156.58 | +6.5% | EOS software lock-in as AI clusters shift to Ethernet; networking at a hardware multiple | 27 June 2026. Fell ~6% on the week in the chip rout; thesis intact. |
| Talen Energy (NASDAQ: TLN) | $361.01 | Wk 15 | ~$370 | +2.5% | AWS nuclear PPA worth close to the whole enterprise; PJM spot optionality unpriced | Scored Wk 19: PARTIAL (+3.2%). Resolves on Cornerstone catalyst. |
| Venture Global LNG (NYSE: VG) | $13.08 | Wk 16 | ~$12.96 | -0.9% | Structural Qatar LNG gap independent of Hormuz diplomacy; highest spot exposure among US exporters | Scored Wk 20: INCORRECT on price. Structural thesis intact. |
| MP Materials (NYSE: MP) | $67.21 | Wk 17 | ~$65.46 | -2.6% | Only US rare-earth mine with a Defense Department cost-plus price floor; strategic-utility re-rating | 7 June 2026 — scores this weekend. |
This week: FCX and USAR carry fresh catalysts and appear in On the Radar above. MP Materials reaches its formal four-week score this weekend and will be finalised at Monday’s scoring run; on Friday’s close it is tracking a small loss against a bullish call.
| Indicator | Latest | Prior | Direction |
|---|---|---|---|
| Nonfarm Payrolls (May 2026) | +172k | ~+80k cons. | Blowout; rate-cut hopes erased |
| Unemployment Rate (May 2026) | 4.3% | 4.3% | Steady, low |
| Average Hourly Earnings YoY | +3.4% | +3.3% | Wages firm, inflationary |
| Core PCE YoY (Apr 2026) | 3.3% | 3.3% | No new print; sticky |
| Headline PCE YoY (Apr 2026) | 3.8% | 3.8% | No new print; energy-driven |
| ISM Manufacturing (May 2026) | 49.2 | 48.7 | Sub-50 headline; new orders 56.8 |
| Michigan Consumer Sentiment | 44.8 | 44.8 | No June print yet; near record low |
| Fed Funds Rate (current) | 3.50-3.75% | 3.50-3.75% | Hold; first Warsh decision 16-17 Jun |
The whole curve is near cycle highs, and the 2s10s gap has re-steepened to plus 43 basis points.
| Tenor | Yield | WoW Change |
|---|---|---|
| 2-Year Treasury | 4.05% | -11bps |
| 5-Year Treasury | 4.45% | flat |
| 10-Year Treasury | 4.48% | +0bps, cycle-high area |
| 30-Year Treasury | 5.12% | +2bps |
| 2Y-10Y Spread | +43bps | Re-steepening |
| HY OAS | 285bps | Tight |
| IG OAS | ~92bps | Stable |
Copper led the week as the AI-build-out demand story reasserted; oil round-tripped.
| Commodity | Close | WoW % | YTD % |
|---|---|---|---|
| WTI Crude Oil | $90.40/bbl | +4.4% | +43.0% |
| Gold | $4,337.10/oz | -5.4% | -0.1% |
| Silver | $73.50/oz | -3.1% | +4.1% |
| Copper | $6.52/lb | +2.2% | +14.8% |
| Natural Gas (Henry Hub) | $3.23/MMBtu | -2.6% | -8.1% |
| Baltic Dry Index | 3,222* | Prior-week carry | +71.7%* |
| Date | Event | Relevance |
|---|---|---|
| 7 Jun 2026 | MP Materials 4-week score date | On the Radar accountability |
| ~12 Jun 2026 | SpaceX (SPCX) Nasdaq debut | Largest IPO in history, ~$75B raise |
| 13 Jun 2026 | USAR 4-week score date | CADE review is the live risk |
| 16-17 Jun 2026 | FOMC — first Warsh decision + dot plot | Highest-impact catalyst of Q2 |
| ~13 Jun 2026 | Michigan sentiment (June preliminary) | Consumer-rollover watch |
| 27 Jun 2026 | ANET 4-week score date | On the Radar accountability |
| Wk 22 (13 Jun) | WMP 2026 Thesis Check-In | Repricing thesis review |
| Early Jul 2026 | Iran 60-day MoU window expires | WTI direction, LNG spread |
| Mid-Aug 2026 | US-China tariff pause expires | Copper, EM, global trade |
| Pair | Rate | YTD % | Driver |
|---|---|---|---|
| USD/TRY | 46.05 | +30.08% | Lira weak; domestic inflation, EM credit pressure |
| USD/ZAR | 16.45 | -11.70% | Rand resilient on firm metals; softer on Friday risk-off |
| DXY (US Dollar Index) | ~100.6 | ~flat | Firmer on the hawkish jobs repricing |
| EUR/USD | ~1.092 | +~2% | Softer euro as dollar firms |
| Indicator | Level | Signal |
|---|---|---|
| VIX | ~21.0 | Spiked on jobs-report selloff (estimate, flagged) |
| MOVE Index (bond volatility) | ~86 | Below 110 caution line |
| HY Credit Spread (OAS) | 285bps | Below 350bps danger zone |
| S&P % above 200dma | 58.4% | Below 60%; breadth slipping |
| 2Y-10Y Yield Spread | +43bps | Re-steepening — recession-precursor signal |
| Crash Probability Score | 27/100 | Lowest band, but up from 2 |
| Company/Event | Data Point | Relevance |
|---|---|---|
| Broadcom (AVGO) | Q2 revenue $22.2B (+48%); AI chips $10.8B (+143%); -15% on guidance | AI-infra bar now very high; ~$1T sector selloff |
| SpaceX (SPCX) | S-1/A filed 3 Jun; ~$75B raise at $1.75T; Nasdaq ~12 Jun | Largest IPO in history; scarcity-engineered float |
| Import AI / METR | Independent work-horizon ~12h (Opus 4.6); ~100h plausible by year-end | Recursive acceleration (Displacer, Pillar 1) |
| Apollo / Torsten Slok | “Zero evidence” of AI job losses; build-out is inflationary | Augmenter Pillar 2; links AI to the Fed trap |
| Lululemon (LULU) | Beat EPS; weak guide; GM -410bps (tariff); NA comps -6% | Tariff pass-through; softening consumer |
| Flashpoint | Status | WMP Assessment |
|---|---|---|
| US-Iran / Hormuz | MoU unsigned, 16th week; Iran missiles + US Qeshm strike mid-week | Oil round-tripped to ~$90. 60-day window runs to early July. |
| Qatar LNG / Ras Laffan | 12.8 mtpa offline; 3-5yr repair | Structural supply gap unchanged, independent of diplomacy. |
| US-China Tariffs | Pause expires mid-August | Swing factor for copper and EM. Watch for extension. |
| Lebanon | Fighting blocks Iran dialogue (Iranian sources, 1 Jun) | A new gate on the Hormuz diplomatic track. |
Scoreboard closes: Friday 5 June 2026 market closes. S&P 500, Nasdaq direction, WTI, gold, Bitcoin and Ethereum verified via web search (TheStreet, CNBC, Bloomberg, Yahoo Finance). Several closes are derived from verified Thursday 4 June closes (automated daily fetch, Yahoo Finance, in the Supabase wmp_price_daily table) plus the stated Friday market moves; these and the prior-period carries (Baltic Dry, MSCI EM) are marked with an asterisk and listed in the exception report. Baselines locked 1 January 2026.
Market Probability Dashboard signals: high-yield OAS, MOVE and VIX from market data; ISM new orders from the ISM May 2026 report; percentage of the S&P above its 200-day average from MacroMicro/StreetStats; insider clusters from OpenInsider/Vickers. The VIX level is an estimate consistent with Friday’s selloff and is flagged.
Analytical calls are logged at entry and scored at the four-week mark in the Supabase intelligence database (project cxldftvilyhhxcuynhfs). On the Radar entries are framed as “what I am watching and why”, not investment recommendations. Full disclaimer in the footer.