GSX Weekly Commodity Comment

Key Takeaways:

The most interesting trade in softs this week may not be the one making the most noise. Coffee is sliding as Brazil’s harvest story gains weight, but cotton is quietly building the cleaner upside case: tighter balance sheets, weather risk, improving export flow and an oil market that could yet change the fibre equation.

Cotton has support, but not a free pass. U.S. export sales for the week ended May 21 were firmer, with shipments also improving, though cumulative movement still needs to do more work before the demand story can be called convincing. West Texas rain has taken some of the urgency out of the drought headline, but it has not removed production risk. Add in crude near the top of the recent range and a possible lift to polyester costs, and cotton has the better forward setup, provided buyers keep showing up.

Coffee is carrying the opposite problem. Arabica is near its lowest level since November 2024 as the market leans harder into a bigger Brazil crop and a smoother harvest outlook. But the sell-off is not without upside risks. Certified and destination stocks remain thin enough that any delay in Brazilian flows, renewed weather risk or stronger nearby demand can still turn a comfortable paper balance into a physical squeeze.

The macro layer is what makes this week different. Brent rose again on June 3 as Gulf hostilities flared and U.S.-Iran talks stalled, while Reuters reported that Hormuz transits remain well below pre-conflict levels. Bond markets are already watching the inflation risk, with the U.S. 10-year yield around 4.46% and traders again pricing some chance of rate hikes. When will the pressure become too great that will we see a TACO? There is always a pressure point when this is possible, it seems that this time it isn’t so simple and the one thing to watch is that Crude inventories around the world are been drawn so there is only so much longer this can last for before this becomes a much bigger problem….

Cotton has moved from a breakout retest into a deeper correction, with only a tentative recovery attempt showing so far. The key change is that the old 82–84 c/lb breakout area remains clearly lost and is now overhead resistance rather than support. July 2026 cotton closed Friday, 29 May, at 76.15 c/lb, while December 2026 closed at 79.59 c/lb and March 2027 at 80.72 c/lb. Compared with the 22 May closes used last week, that leaves July down 127 points, while December was up 26 points and March up roughly 53 points. The recovery has therefore been uneven: nearby July remained under pressure, while deferred contracts stabilized modestly.

By Tuesday, 2 June, the market had bounced but had not repaired the technical break. July 2026 cotton closed at 77.04 c/lb, December 2026 at 80.54 c/lb and March 2027 at 81.72 c/lb, keeping the board below the old 82–84 c/lb breakout zone even after the rebound. The broader cotton benchmark was quoted around 76.99 c/lb on 2 June, up on the day but still down more than 7% over the month, leaving prices well below the mid-May upper-80s highs.

Spot indicators weakened again. USDA’s Weekly Cotton Market Review showed seven-market spot quotations averaging 72.12 c/lb for the week ending 28 May, down 433 points from 76.45 c/lb the previous week, though still above 62.61 c/lb a year earlier. Daily spot quotations ranged from a high of 72.63 c/lb on 22 May to a low of 71.37 c/lb on 27 May. Liquidity improved from the prior week’s near standstill but remained thin: spot transactions totaled 4,493 bales, versus just 6 bales the previous week and 1,499 bales a year earlier. ICE July ended the USDA review week at 76.77 c/lb, down from 77.98 c/lb the week before. That means the cash market is still confirming near-term futures weakness, even though season-to-date spot transactions remain well ahead of last year at 1.50 million bales versus 958,561 bales.

Cotton futures began to retrace from the 89 resistance area throughout May, and have continued lower into the 0.382 retracement at 77. We are counting the downside as wave 2 through an Elliott Wave lens.

It’s worth noting that wave 2 is generally the sharpest retracement in an Elliott count, commonly terminating in the 0.618 retracement area. However, it can still overshoot as deep as the origin of wave one. So far, price action is adhering nicely to standard wave behaviour.

Last week we outlined the 76 area of Fibonacci confluence highlighted by the green box, forecasting a bounce to emerge from that region. The green box has held, and the upside is firmly underway this week.

We continue counting this as a (b)-wave rally for now, with another downleg to follow in wave (c) of 2 towards 70. This target area can be further refined when the (b)-wave top is in.

The latest CFTC futures-and-options data, for 26 May, showed managed money holding 74,828 long contracts and 20,628 short contracts, leaving a net long of 54,200 contracts. That is down 7,845 contracts from the prior week’s 62,045-contract net long. The market therefore remains fund-long, but the fund length has been cut back materially. This was not fresh bearish short-building: managed money actually cut 8,021 longs, while shorts fell slightly by 176 contracts, meaning the decline in the net long came overwhelmingly from long liquidation, only marginally offset by short-covering.

Commercial positioning has eased again, but it remains a major caution point. Producer, merchant, processor and user accounts were 79,975 contracts long and 209,017 contracts short, leaving them net short 129,042 contracts. That is less short than the previous week’s 133,307-contract net short, because commercial shorts fell by about 4,861 contracts, while longs were broadly flat/slightly lower, down about 594 contracts on the CFTC change line. So this is still not a fresh escalation in commercial selling; it is another modest reduction in hedge pressure. Even so, the outright commercial short remains large, equal to 41.6% of open interest, and still shows heavy producer and merchant selling into the rally.

USDA’s U.S. cotton balance remains heavy. For 2025/26, production is pegged at 13.90 million bales, exports at 12.00 million, domestic use at 1.60 million and ending stocks at 4.40 million. That leaves stocks-to-use near 32%.

The 2026/27 outlook is modestly tighter. USDA projects 13.30 million bales of production, 12.30 million of exports and 3.90 million in ending stocks, cutting stocks-to-use to 28.1%. That is supportive, but not tight enough to drive prices without help from weather or demand.

Some abandonment risk is already in USDA’s forecast. The May balance sheet uses March planting intentions and adjusts harvested area using regional abandonment averages, with the Southwest adjusted for moisture. West Texas dryness alone may not be enough; the market needs evidence conditions are worsening beyond USDA’s assumptions.

The global balance is more constructive. USDA sees 2026/27 world production at 116.04 million bales, consumption at 121.69 million and ending stocks falling to 71.84 million from 77.27 million. Still, Brazil remains a major competitor, with exports projected at 15.00 million bales in 2026/27.

Demand is improving but not decisive. Latest U.S. old-crop upland sales were 131,800 running bales, while new-crop sales reached 216,000, led by Pakistan. Shipments, however, were 289,400 running bales, unchanged on the week and below the four-week average.

USDA/NASS’s latest Crop Progress report shows U.S. cotton planting largely back on schedule. As of the week ended May 31, the crop was 66% planted, up from 53% a week earlier, slightly ahead of last year’s 64% but just behind the five-year average of 67%. Texas reached 56% planted, up from 42%, though still behind both last year and average at 59%.

The bigger issue is no longer acreage getting seeded, but whether stands establish well enough to limit abandonment. Texas soil moisture improved but remains stressed: 51% of topsoil and 57% of subsoil were rated short or very short. Nationally, moisture also improved, with topsoil short/very short easing to 36% and subsoil to 39%.

Drought coverage in Texas has eased but remains broad. Drought.gov shows 57.4% of the state in D1-D4 drought as of May 26, down from 65.3% a week earlier. That statewide number still understates the cotton risk because production is concentrated in drier western areas.

Weather has become more helpful for West Texas, but not cleanly bearish. Recent storms brought useful moisture across parts of the Panhandle, Permian Basin and Rolling Plains, while also raising risks from hail, high winds, flooding and planting delays. CPC’s medium-range outlook continues to favor above-normal rainfall for Texas and the southern Plains.

U.S. cotton export demand improved in the latest USDA/FAS report, with current-crop upland net sales rising to 153,600 RB for the week ending May 21, up 17% from the prior week and 32% above the four-week average. Buying was led by Vietnam, Pakistan, China and Turkey, giving the report a broader and less bearish tone than last week’s Pakistan-led rebound.

Shipments were the more supportive feature. Upland exports increased to 317,700 RB, up 10% week over week, with Vietnam, Pakistan and Turkey the main destinations. On an all-cotton basis, shipments reached 334,900 bales, keeping USDA’s 12.0 million-bale old-crop export target within reach, although cumulative sales and shipments remain slightly behind last year.

New-crop demand eased from the prior week but remained respectable, with 112,000 RB of 2026/27 upland sales. Pakistan again led the book, followed by Mexico, South Korea and Turkey. The forward sales picture is improving, but demand is still too concentrated to call it broadly bullish.

The export outlook is now better than last week. We would lift the working full-year U.S. export view to around 11.5–11.7 million bales, with upside toward USDA’s 12.0 million if weekly shipments continue near or above 300,000 bales. A shortfall versus USDA would still add carry-in to next season’s balance sheet, but the risk now looks closer to 300,000–500,000 bales, rather than last week’s larger downside case.

Weather has turned less price-supportive as rain chances remain active across parts of West Texas and nearby cotton areas. Verified moisture would reduce abandonment risk and pressure weather premium. But the forecast remains uneven, so missed or poorly placed rains would keep support under cotton futures.

Coffee’s front end is still trading scarcity, but the board has turned back toward harvest pressure. July arabica settled Tuesday, June 2, at 259.20 c/lb, down 1.40 c/lb, after touching a 1½-year nearest-futures low, while July robusta settled at $3,462/t, up $24/t but still below last week’s mid-$3,500s test. The bullish case remains exchange stocks: ICE arabica certified stocks have slipped to 432,781 bags, a 3½-month low, and robusta inventories remain historically tight. But thin stocks are now a squeeze risk rather than a trend-setter, as the market increasingly discounts heavier second-half supply.

Brazil is the cap on any rally. Conab sees the 2026 crop at a record 66.7 million bags, up 18%, while private estimates are higher and point to near-record exports in 2026/27. Vietnam is also leaning bearish for robusta, with January-April exports up strongly year on year, even as weather worries linger in the Central Highlands. Colombia’s softer exports still support washed arabica differentials, but not enough to alter the broader balance. Net-net: coffee has moved from an inventory trap into a transition trade. Tight nearby stocks can still punish shorts, but unless arabica reclaims the mid-260s and then the 270s—and robusta regains $3,500-$3,550—rallies look tactical and vulnerable to renewed producer selling.

The weekly chart exhibits a downside continuation candle, consistent with the bearish macro outlook for coffee futures. The 1.618x Fibonacci extension at 254 was flagged last week as this downleg’s target region. Also note that the red 200WMA is converging in the same area, increasing the strength of support.

Notably, price/MACD divergence is beginning to develop on the weekly timeframe. This signal of exhaustion in the current impulse increases conviction in the anticipated wave (iv) rally.

Drilling into the 1-hour timeframe, the late-May rally exhibited the hallmarks of corrective rather than impulsive price action: messy and overlapping. While the market extended slightly beyond our initial expectation in the (c)-wave, this did not alter the structure. KC has since broken down as anticipated, sharply, and has resumed trending toward the wave (iii) target region of 254.

Arabica’s positioning backdrop remains thin, though the latest move was less aggressively bearish than the prior washout. CFTC futures-and-options data for May 26 showed managed money in Coffee C net long 17,434 contracts, versus roughly 17,900 a week earlier. Funds cut 1,254 longs, but also covered 782 shorts, meaning the decline in net length was driven more by continued long liquidation than by a fresh wave of short selling.

Commercial positioning offered little fresh comfort. Producer, merchant, processor and user accounts were net short 19,963 contracts, slightly wider than the 19,862 net short seen the previous week. Commercial longs rose 997 contracts, suggesting some industry buying on weakness, but commercial shorts increased a little more, by 1,098 contracts, pointing to renewed origin hedging as prices tried to steady.

The report is already stale, and subsequent price action has made the setup harder to call supportive. July Coffee C fell to 259.20 cents a pound on June 2, with arabica as at a 1.5-year nearest-futures low. The broader message is unchanged: arabica’s fund cushion has not rebuilt. Net length has slipped again, to barely 17,400 contracts, leaving the market exposed to further liquidation unless prices can reclaim the 280/285 c/lb area and, eventually, rebuild momentum toward 300 c/lb.

Robusta looks firmer by comparison. The latest London positioning data showed speculative managed money raising its net long to 13,447 lots, equal to about 2.24 million bags, from 11,097 lots in the prior report. That rebuilding fits robusta’s tighter physical tone, but it is not enough on its own to validate rallies. With July robusta at $3,462 a ton on June 2, confirmation still needs to come from nearby spreads and physical demand, not just from a friendlier fund slate.