To The Moon

Over the past week, major asset classes delivered a strikingly uneven set of returns. Crude oil took first place with a 10.64% gain, silver followed with 4.04%, copper and gold posted modest gains of 1.10% and 0.81% respectively, while aluminium fell 0.73% — the only commodity to close lower. Against that broad commodity strength, U.S. equity index futures retreated across the board. Both of the week's commodity narratives point to the Middle East. On crude: renewed U.S.–Iran confrontation, a Houthi strike that shut in 400,000 barrels per day of capacity at Saudi Aramco's Jazan refinery, Red Sea tanker traffic falling to multi-month lows, and OPEC+ preparing to stop raising output targets together pushed up the pricing of supply-disruption risk. On aluminium: according to Reuters, war in the Middle East has damaged Gulf smelting capacity and forced consumers to draw on exchange stocks, sending LME aluminium inventories to their lowest level since 1998. At the same time, Fed policy expectations are shifting from cuts toward hikes, and long-end Treasury yields did not retreat over the same period. But when price is placed alongside inventory, the real story of the week is a set of directional divergences: the biggest gainer, crude, actually saw an official inventory build over the same period; gold, whose physical stocks have bled out by more than 30% in a year, rose just 0.81%; aluminium, with inventories at their lowest since 1998, was the sole decliner; and copper, which saw the largest drop in visible inventories, had a substantial share of that decline amount to nothing more than metal moving to a different warehouse.

As of the close on 24 July 2026, weekly performance across key assets was as follows:

Chart 1　Weekly performance of key assets (red = up, green = down)

In an environment where macro expectations keep swinging back and forth, watching price action alone is no longer enough to grasp the main thread of asset performance. Inventory changes, by comparison, better describe physical supply and demand, while fund flows better reflect allocation preferences. It is therefore worth examining U.S. equities, Treasuries, crude, copper, aluminium and precious metals through the twin lenses of inventory and capital.

1. Equity fund outflows double; the yield spread flips from narrowing to widening

According to the latest data from ICI:

About ICI (Investment Company Institute): founded in 1940, ICI is one of the most central trade associations in the U.S. fund industry. Its statistical coverage spans roughly 98% of the assets held in U.S. funds registered under the 1940 Act, and its flow data is widely regarded as the authoritative source for tracking subscriptions and redemptions in U.S. mutual funds. ICI also publishes long-running statistics on the assets and flows of regulated funds both in the U.S. and globally; its methodology is stable and its coverage broad, which is why it is heavily cited by brokers, research houses and financial media.

Outflows from U.S. equities widened markedly. For the week ended 15 July 2026, U.S. equity mutual funds are estimated to have seen net outflows of $18.104 billion (0.1% of assets as of 31 May), of which domestic equity funds accounted for $14.456 billion and world equity funds for $3.648 billion — the pressure was concentrated in the home market. On a marginal basis, that is close to double the prior week's net outflow of $9.664 billion, indicating a significant intensification of withdrawals.

Inflows into Treasuries slowed. Over the same week, bond mutual funds are estimated to have taken in net inflows of $4.511 billion (0.1% of assets), split between $3.138 billion into taxable bond funds and $1.373 billion into municipal bond funds (0.2%); that is roughly $2.621 billion less than the prior week's net inflow of $7.132 billion. Taken together, the week reads as 「equity funds bleeding faster while bond funds attract less」 — the outflow from equities did not convert into inflows to bonds, which looks more like a contraction of the overall risk budget than a directional reallocation between stocks and bonds.

Chart 2　US Fund Net Flows (equity / bond fund net flows, source: ICI)

$SPDR S&P 500 ETF Trust(SPY)$ $S&P 500(.SPX)$ $E-mini S&P 500 - main 2609(ESmain)$ $Micro E-mini S&P 500 - main 2609(MESmain)$ $Micro E-mini S&P 500 - Sep 2026(MES2609)$ $E-mini Dow Jones - main 2609(YMmain)$ $Micro E-mini Dow Jones - main 2609(MYMmain)$ $Dow Jones(.DJI)$ $SPDR Dow Jones Industrial Average ETF Trust(DIA)$ $NASDAQ 100(NDX)$ $E-mini Nasdaq 100 - main 2609(NQmain)$ $Micro E-Mini Nasdaq 100 - main 2609(MNQmain)$ $Invesco QQQ(QQQ)$ $Micro 10-Year Yield - main 2607(10Ymain)$ $10-YR T-NOTE - main 2609(ZNmain)$

On the yield curve, the U.S. 10-year Treasury yield (blue line) printed a latest reading of 4.61% and the 3-month (orange line) 3.90%, maintaining a positive spread of roughly 71 basis points. Against the prior period's 4.50%, 3.85% and 65 basis points, the long end rose about 11 basis points and the short end about 5, widening the spread by roughly 6 basis points.

Looking at the marginal shape at the far right of the chart, this period marks a clear reversal from the last: previously it was 「short end up, long end down」 that drove the spread narrower, whereas now both ends are rising together with the long end moving materially more. In terms of position, the 10-year is already near the top of its one-year range (it printed 4.71% on 23 July, a one-year high), while the 3-month has lifted only gradually and remains below the midpoint of its range. That tells us the force driving the curve is coming from the long end rather than the short end — the backdrop being inflation readings still clearly above target (core PCE of 3.4% in May), policy expectations shifting from cuts to hikes, and a rising term premium, all acting together. Overall, the current curve reflects a repricing of forward inflation and term compensation, rather than a repricing of near-term liquidity.

Chart 3　U.S. 10-year and 3-month Treasury yields (source: U.S. Treasury)

$SPDR S&P 500 ETF Trust(SPY)$ $S&P 500(.SPX)$ $NASDAQ(.IXIC)$ $Invesco QQQ(QQQ)$ $ProShares UltraPro QQQ(TQQQ)$

2. Crude: inventories break below the five-year floor, but the rally is not inventory-driven

According to the latest data from the EIA and Bloomberg:

For the week ended 17 July 2026, U.S. commercial crude oil inventories (excluding the Strategic Petroleum Reserve) printed 411.7 million barrels, a build of 2.0 million barrels from the prior week; Cushing inventories printed 19.37 million barrels, a draw of 674,000 barrels — the two moved in opposite directions. But position matters far more than the weekly change. The EIA's own framing is that commercial crude inventories sit about 6% below the five-year average for this time of year; recomputing across the same weeks of 2021–2025, the current level is not merely below the average but has broken through the bottom of the five-year range by roughly 7.3 million barrels. Cushing is more extreme still, sitting about 2.53 million barrels below its five-year floor. On both seasonal charts, the marginal shape at the far right shows the 2026 line (orange) running beneath the five-year range and hugging its lower edge — physical supply and demand are at their tightest in five years.

Placing that position next to price gives us the first divergence of the period: crude rose 10.64% on the week, while the marginal signal from the EIA over the same span was a 2.0 million barrel build. Inventories did not get tighter, so this leg of the rally cannot be attributed to a larger-than-expected drawdown. On a marginal basis, the inventory position below the five-year floor provides a floor for price — a slow-moving variable — while the bulk of this week's gain came from a repricing of supply-disruption risk, a fast-moving variable that forms quickly and fades just as quickly. Overall, crude sits in a combination of 「absolute inventories extremely tight, marginal destocking paused, price carrying a sizeable risk premium」: the fundamental foundation is solid, but the price has a visibly larger exposure to headlines.

Chart 4　U.S. commercial crude oil inventories (vs. five-year range; 2026 in orange)

Chart 5　Cushing crude oil inventories (vs. five-year range; 2026 in orange)

$United States Oil Fund LP(USO)$ $WTI Crude Oil - main 2609(CLmain)$ $E-mini Crude Oil - main 2609(QMmain)$ $Micro WTI Crude Oil - main 2609(MCLmain)$

3. Copper: visible inventories draw down fast, but the increments keep flowing to the U.S.

According to the latest data from Bloomberg:

As of late July 2026, total visible copper inventories across the three major exchanges stood at roughly 1.079 million metric tons, down about 122,000 mt (−10.2%) from the 1.201 million mt of the prior period, extending the downtrend in place since March. But breaking the total into its three venues reveals a structure quite different from 「synchronised destocking」. Shanghai copper inventories came in at roughly 69,600 mt, down 74,400 mt (−51.7%) from 144,000 mt and back to a two-year low; LME copper inventories printed roughly 369,000 mt, down 97,000 mt (−20.8%) from 466,000 mt. Running counter to both, CME copper inventories reached roughly 706,000 short tons (about 640,500 mt), up 54,000 short tons (+8.3%) from the prior period and setting yet another record high — on the chart, still a one-way line up with barely a pullback.

On a marginal basis, the defining feature of the copper market this period is not simple destocking but a geographic redistribution of inventory: over the same span in which the three-venue total fell by 122,000 mt, the CME alone accumulated roughly 49,000 mt. According to Reuters and trade media, this configuration is tied to expectations of U.S. tariffs on copper under Section 232: the tariff expectation opens a cross-market price differential, and deliverable metal is shipped to the U.S. for delivery. The implication for price has to be read in two parts. The deep drawdowns in Shanghai and at the LME represent genuine physical tightness — deliverable stocks at the Shanghai Futures Exchange have fallen roughly 82% since early May, and according to Shanghai Metals Market (SMM), China's social copper inventories fell 41,600 mt in a single week while spot premiums rose in step. But the record high on the CME side means that metal has not actually been consumed; it is simply sitting in a different warehouse. High U.S. inventories do not equate to strong U.S. demand. That is why copper's modest 1.10% gain on the week is not inconsistent with the seemingly bullish reading of 「visible inventories down more than 10%」 — the market prices 「inventory relocation」 and 「inventory consumption」 separately.

Chart 6　Total visible copper inventories across the three exchanges (metric tons)

Chart 7　Shanghai copper inventories (metric tons)

Chart 8　LME copper inventories (metric tons)

Chart 9　CME copper inventories (short tons)

$ETFS COPPER(COPA.UK)$ $Copper - main 2609(HGmain)$

4. Aluminium: LME stocks hit their lowest since 1998, yet it was the only decliner

According to the latest data from Bloomberg and Reuters:

As of late July 2026, combined aluminium inventories across the three major exchanges stood at roughly 729,000 metric tons, down about 115,000 mt (−13.6%) from the prior period's 844,000 mt. Unlike last period's split of 「overseas destocking, domestic building」, all three venues drew down in the same direction for the first time. The overseas leg is the most important line in this section: LME aluminium inventories came in at roughly 273,000 mt, down 42,000 mt from 315,000 mt, which according to Reuters is the lowest level on record since 1998 — the direct cause being war in the Middle East damaging Gulf smelting capacity and forcing consumers to draw on exchange stocks to cover the shortfall. Domestically, Shanghai aluminium inventories printed roughly 455,000 mt, down 72,000 mt from 527,000 mt, ending the first-half build that had run from about 100,000 mt up to roughly 520,000 mt; COMEX aluminium inventories fell further to just 751 mt from 1,378 mt.

On a marginal basis, this is the strongest inventory reading of any commodity this period: all three venues moving the same way, a 13.6% drop in the total, and an overseas level at a 28-year low. And yet the price alongside it was the only decline in the field — aluminium fell 0.73% on the week. This constitutes the second divergence of the period, and its direction is precisely the opposite of crude's. The explanation shares the same root as crude's: given that the LME aluminium drawdown was driven mainly by Gulf capacity being hit by war, then once supply recovers faster than expected and the geopolitical risk premium begins to unwind, price reacts ahead of inventory. Overall, aluminium sits in a state of 「visible inventories extremely tight, price yet to reflect it」; what merits tracking is how that divergence converges.

Chart 10　Total aluminium inventories across the three exchanges (metric tons)

Chart 11　LME aluminium inventories (metric tons)

Chart 12　Shanghai aluminium inventories (metric tons)

Chart 13　COMEX aluminium inventories (metric tons)

$ETFS ALUMINIUM(ALUM.UK)$ $ALUMINUM FUTURES - main 2610(ALImain)$

5. Gold and silver: gold inventories bleed out one-way; silver rallies as net longs shrink

According to the latest data from Wind and the CFTC:

Inventory side (physical fundamentals): COMEX gold inventories printed a latest reading of 27.026 million troy ounces, continuing their clear decline from the prior period's 27.883 million (a fall of roughly 856,000 ounces, or −3.07%); COMEX silver inventories printed 331 million troy ounces, a modest increase from the prior period's 323 million. Historically the two are moving in exactly opposite directions: gold inventories are extending an unusually clean one-way downtrend, having fallen from roughly 38.5 million ounces over the past year to the current level — a cumulative loss of more than 30% — with no sign of the drawdown slope flattening at the right-hand edge, so physical support remains firm. Silver inventories, by contrast, have risen modestly and persistently since confirming an interim bottom in March; while still within a tight range by the standards of recent years, the marginal direction has shifted from tightening to easing.

Chart 14　COMEX gold inventories (million troy ounces)

Chart 15　COMEX silver inventories (million troy ounces)

Capital side (positioning): as of the latest data, COMEX gold non-commercial longs stood at 224,800 contracts against 40,900 shorts; silver non-commercial longs at 36,700 against 13,300 shorts. On a marginal basis, gold shows 「longs cut, shorts added」 — longs down 2,525 contracts and shorts up 247, taking the net long from 186,700 to 183,900 contracts and the long/short ratio from 5.59 to 5.50. Silver shows 「both sides added」 but with a sharply asymmetric split: longs up just 495 contracts while shorts rose 2,104 (more than four times the increase in longs), cutting the net long from 25,000 to 23,400 contracts and the long/short ratio markedly from 3.23 to 2.76.

Taken together, this set of data speaks to widening disagreement rather than a strengthening trend: the net long structure remains intact in both metals and the long side has not been overturned, but net longs narrowed in both over the week and the long/short ratio fell in tandem, indicating that speculative money willing to stand on the short side at current prices is increasing. Silver's case deserves particular attention, because it contradicts both price and inventory: silver rose 4.04% and was the second-strongest commodity of the week, yet over the same span its net long shrank by 1,609 contracts, exchange inventories continued to build modestly, and a substantial portion of the increase in open interest sat in spreads rather than directional positions. Rising price alongside a shrinking net long usually means the positioning base beneath that advance is thinning — which is why the short-squeeze narratives circulating around silver warrant caution.

Chart 16　COMEX gold non-commercial long vs short positions (contracts, source: CFTC)

Chart 17　COMEX silver non-commercial long vs short positions (contracts, source: CFTC)

$Gold - main 2612(GCmain)$ $E-Micro Gold - main 2612(MGCmain)$ $1-Ounce Gold - main 2610(1OZmain)$ $E-mini Gold - main 2612(QOmain)$ $SPDR Gold ETF(GLD)$ $Silver - main 2609(SImain)$ $E-mini Silver - main 2609(QImain)$ $iShares Silver Trust(SLV)$ $Micro Silver Futures - main 2609(SILmain)$ $100-Ounce Silver - main 2609(SICmain)$

6. Conclusion: read the flows, know what is in the warehouses

Over the past week, the most striking feature on the inventory side was three directional divergences between price and inventory. In crude, both commercial inventories and Cushing broke below the bottom of their five-year ranges, leaving physical supply and demand at their tightest in five years — yet the marginal signal was a build while price rose more than 10%. The contrast in aluminium is more direct: LME inventories at their lowest since 1998, all three venues drawing down together for the first time, and price nonetheless the sole decliner in the field. Copper is a third structure: visible inventories fell more than 10%, but almost entirely at the LME and in Shanghai, while the CME alone set a record high — a geographic redistribution rather than genuine consumption. In precious metals, gold's physical inventories continue their one-way grind lower while silver's have edged up; the two are moving in opposite directions. On the capital side, equities and bonds weakened together: weekly net outflows from equity funds nearly doubled against the prior week, and while bond funds retained net inflows, their momentum slowed. The change in the Treasury curve was clearly long-end led, with the positive spread widening to 71 basis points — a complete reversal of the narrowing dynamic of the prior period. In precious-metals positioning, gold showed 「longs cut, shorts added」 and silver 「both sides added, with shorts rising more than four times as much as longs」, and net longs narrowed in both.

Overall, the common thread this period is that divergences — between price and inventory, and between price and positioning — are multiplying at the same time, and they do not point the same way. That points neither to a bullish nor to a bearish conclusion, but it does indicate that price is currently driven more by expectations and risk premia, while the signals from physical supply and demand lag behind. In such an environment, the value of inventory and positioning data lies not in supplying a direction, but in flagging which advances rest on the thinnest foundations and which declines are inconsistent with fundamentals. How will markets trade next week? We will look again then.