Cocoa Market Review July 2026
OVERVIEW
July saw the market break sharply higher to new highs before retracing the entire move, September London settling at £4,011 and September New York at $5,397, higher on the month but well below the highs of the 9th. The move up was driven by managed money short covering rather than fresh buying, and by the end of the month the net speculative short had begun to rebuild.
The July pod counts offered only marginal relief, while El Niño strengthened over the month. Q2 grind was strong, up close to 10% year on year, though against a weak base and with manufacturer sales volumes disappointing. Against that, the physical evidence of the 2025/26 surplus continued to build, with US warehouse stocks posting their largest monthly increase since May 2021 and a build up of butter and liquor stocks in both Europe and the US.
FLAT PRICE & POSITIONING
July saw volatility remain elevated as the market broke sharply higher following a short respite at the end of June. The initial move to the upside was triggered on 6th July, with London gaining 12.2% and New York 13.1%. September London traded from a low of £3,689 on 3rd July to a high of £4,827 on the 9th before settling the month at £4,011, up £194. September New York closed at $5,397, up $319, having reached a high of $6,478. The move was accompanied by short covering, the market breaking through the resistance that had capped momentum at the end of June.
In New York the managed money gross short fell from 32,562 lots on 30th June to 26,335 by 14th July, a reduction of 6,227 lots, while in London the gross short fell from 14,707 lots to 7,988 by 21st July, a reduction of 6,719. London’s Managed Money net short reached just 956 lots on 14th July, its smallest since September 2025. Origin sold into the rally but, with the CCC and CMC having sold or mostly sold the 2026/27 main crop, there was little forward origin selling to slow or cap the move on the upside. The move on the upside seemed to stall once the managed money position in London went flat.
The retracement then ran through the second half of the month. Over this period we saw managed money increase its net short position. In the report to 28th July the New York net short increased back out to 14,127 lots and London to 4,358, taking the combined net short to 18,485 lots. This came from both directions, with shorts added in both markets, 1,345 lots in New York and 1,159 in London, while longs were liquidated, down 1,178 and 1,584 respectively. Despite reports earlier in the month there has been no fresh managed money long interest, despite the 2026/27 crop uncertainty, strengthening El Niño, and increasing uptrend. Managed money’s long position in London now stands at 4,789 lots, a level seen only a handful of times in the past decade. The lack of fresh managed money participation is partly due to the risk in the market, with volatility and price at these levels, risk per lot is around three times higher than in February, and this lack of participation is itself adding to the volatility we have seen.
It appears as if industry was largely inactive as the market retraced. We had anticipated more involvement as the market came back to £3,800 and $5,000, the levels at which it held at the end of June. Given good cover acquired mostly below where we are now, and slightly better pod counts, industry may be content to wait for lower levels before participating. That said, there is enough uncertainty around the size of the 2026/27 crop and the potential impact of a strong to very strong El Niño that we would not expect them to be far below these levels should the market test them again, and that should limit the downside for now. August pod counts may well shape the next move on flat price.
2025/26 SUPPLY: SURPLUS, ARRIVALS & GRIND
Our view is that the 2025/26 surplus has marginally increased again this month, with market estimates now ranging from 400,000 to 650,000 tonnes. The published forecasts are the more conservative, with the larger surplus estimates coming from within the trade. StoneX raised its projection to 422,000 tonnes from 247,000, driven mainly by a roughly 160,000 tonne increase to the Ivory Coast crop, taking it to around 2 million tonnes, while Transgraph put the current season at 415,000 tonnes.
Ivory Coast arrivals continued to run ahead of last year. July arrivals reached 71,000 tonnes against 56,000 tonnes in July 2025, taking cumulative arrivals to 1.98 million tonnes as at 26th July. The mid-crop has delivered broadly as expected, and the surplus we anticipated becoming visible through this period has duly appeared.
Ecuador bean equivalent exports for June came in at 41,900 tonnes against 33,000 tonnes last year, with total exports for the year now expected at 590,000 to 600,000 tonnes. Nigeria cocoa bean exports for June rose 30% year on year to 18,922 tonnes, a second consecutive strong month after May’s 28% increase. Brazil Q2 production came in at 66,500 tonnes against 40,400 tonnes, taking the crop to date to 154,800 tonnes, up 37% on last year. The ICCO estimated Indonesian production could rise to 220,000 tonnes for 2025/26 from 200,000 tonnes last year.
US warehouse stocks reached 3,365,663 bags at the end of July, a build of 417,777 bags on the month. This is the largest monthly increase in stocks we have seen since May 2021, and this is normally a period where stocks tend to draw down. This has largely been driven by the arbitrage, with New York maintaining its premium and cheap differentials from both West Africa and South America being attracted to New York. Stocks have now risen 1,734,417 bags since the end of December. The level nonetheless remains historically low at just under 3.4 million bags against a ten year July average of around 4.4 million.
On grind, European Q2 fell 4.64% to 316,366 tonnes, the weakest second quarter in six years and, outside the pandemic-affected 2019/20, the lowest since 2013/14. On the nine months to date Europe stands at 946,731 tonnes, its lowest since 2004/05 and a third consecutive annual decline. North America rose 7.65% to 109,659 tonnes and Asia 25.07% to 224,646 tonnes, while at origin Ivory Coast grind rose 27.57% to 158,529 tonnes and Brazil 8.61% to 49,711 tonnes. Ivory Coast June grind was 56,182 tonnes, up 21.66% on the year, taking cumulative grind to 493,300 tonnes against 475,757 last year. Taken together, reported world Q2 grind rose 9.66% to 858,911 tonnes, and on the nine months to date world grind is fractionally lower at 2,557,307 tonnes against 2,564,402 tonnes.
The 9.66% increase was against a very weak base which had itself fallen 8.36%. A meaningful part of Q2’s strength reflects the weak market structure, which allowed cocoa butter to be carried and so encouraged grinding, alongside better uptake in cocoa powder. This increase in grinding has seen butter and liquor stocks build up and as a result butter and liquor has weakened, with ratios trading around the February lows. A notable reversal: ratios bottomed at ten year lows in February, then recovered and appeared to have found a floor, which we read at the time as tentatively constructive for demand, but which now appears to have been structurally led.
On demand, the manufacturers’ results published through July suggested consumption remained poor. Hershey reported second quarter sales volumes down 8% and Mondelez down 1.7%. Lindt, reporting the first half, saw volume and mix fall 7.5%. Barry Callebaut, whose fiscal year ends in August, reported nine month volumes to 31st May down 2.8% at 1,557,239 tonnes, with the third quarter turning positive at 5.7%. Barry Callebaut attributes he quarterly results to elevated demand in its global cocoa business following the market correction earlier in the year, alongside momentum in Asia and progress restoring service levels in North America. Global Chocolate volumes were down 2.3% across the nine months and up only 3.2% in the quarter. Nielsen data cited in the same release, covering 26 countries and excluding e-commerce, showed the chocolate confectionery market down
5.6% over the nine months and 4.4% in the third quarter. Barry Callebaut has, however, improved its full year outlook for 2025/26 to a volume decline of around 1%.
2026/27 OUTLOOK: POD COUNTS, EL NIÑO & DEMAND
Market estimates for the 2026/27 crop range from a small surplus to a large deficit. As with the current season, the published forecasts are the more conservative. This is off the back of the poor early pod counts and the likely impact El Niño will have on production, whilst these high prices are starting to dampen the expected demand recovery for 2026/27.
The July pod counts improved slightly on June but still point to a drop in production from 2025/26. The market is expecting a drop in the Ivory Coast crop of between 10 and 20%, whilst in Ghana, COCOBOD has announced it is expecting Ghanaian cocoa production to fall by at least 16% in the 2026/27 season, as a response to El Niño and the crop’s natural fruit bearing cycle.
Among the published forecasts, StoneX cut its projected 2026/27 global surplus to just 25,000 tonnes from around 149,000 tonnes, after adjusting for the impact on production of a potentially strong El Niño and the poor early pod counts, alongside a modest increase in demand. It expects global production to fall around 6% to 4.83 million tonnes against demand of around 4.75 million tonnes, with Ivory Coast at approximately 1.78 million tonnes, down 11%, Ghana at around 586,000 tonnes and Ecuador near 600,000 tonnes. Transgraph are forecasting a surplus of 80,000 tonnes against 415,000 tonnes in 2025/26, with production down 4.6% to around 4.8 million tonnes from 5.1 million, Ivory Coast down approximately 16%, and demand up around 3% to roughly 4.79 million tonnes, expecting prices to trade in a broad $4,800 to $6,800 range over the next six months. Citi is forecasting an outright deficit, at 56,000 tonnes.
El Niño strengthened over the month, and the coupled ocean-atmosphere system now clearly reflects a strengthening event. The latest Niño-3.4 index reads +1.2°C, and Niño-1+2 at +2.7°C. NOAA now sees an 81% chance of a very strong El Niño during October to December, up from the 63% for November to January cited last month, and a 97% chance the event persists through early spring 2027. An event that would rank among the largest in the record since 1950.
The latest ECMWF seasonal forecast, run from 1st August, shows a clear El Niño signature for December to February. Coastal Ecuador carries a 70 to 100% probability of rainfall in the highest 20% of precipitation, consistent with Niño-1+2 now at +2.7°C against +2.1°C last month, and Ecuador’s flood risk rises once that index pushes above +2°C during the October to February wet season. Indonesia shows the same 70 to 100% probability of falling in the lowest 20%. West Africa now points to a 50 to 70% probability of below normal rainfall across the cocoa belt, with a similar probability of falling into the lowest 20% of precipitation. That period coincides with the Harmattan.
Looking ahead to demand in 2026/27, the market is estimating a 2 to 4% improvement, though outlooks vary as the market interprets how price could dampen the recovery. Lindt is looking for volume stabilisation in the second half and a return to positive momentum in 2027, which rests on pricing actions already taken and on cocoa not running away again. Guan Chong expects grinding to increase in 2026 as demand recovers, but warned the 2027 outlook could weaken as a result of El Niño reducing production and the resulting higher prices. Our own view is that the improvement in demand will be held back by these high prices and the threat of El Niño, and that manufacturers will be more hesitant to move back from reformulated products than the market anticipates. We do still expect to see an improvement, but a more modest one than the 5% recovery the market had been looking for earlier in the year.
We maintain the longer term view that price at this level is only justified if we see a large deficit for 2026/27. As we have mentioned in past reports, we are coming into 2026/27 off a multi year high in the stock to grind ratio, at over 40%. The underlying fundamentals for 2025/26 remain bearish, and
that high stock to grind ratio provides a cushion such that even a small to moderate deficit for 2026/27 would leave the market comfortable. In that scenario we would expect the stock to grind ratio to reassert itself and prices to come off these levels.
MARKET STRUCTURE
The structure continued to reflect the split running through the whole market, with the nearby positions pricing the 2025/26 surplus and the 2027 positions pricing the threat to the 2026/27 crop, alongside EUDR and the ICE position limits.
LONDON
The July London expiry saw 37,730 tonnes delivered. Nigeria and Ecuador dominated at 15,870 and 15,080 tonnes respectively. Both sides were concentrated: FIM tendered 18,280 tonnes and PFU 13,040, with ADM, ITL and MCQ making up most of the balance. On the receiving side just two members took delivery, SCD at 19,010 tonnes and SLM at 18,720.
On grading, July saw 28,360 tonnes graded, of which 24,480 tonnes passed and 3,880 tonnes failed. Of the passed grading, 14,360 tonnes was initial grading, which goes against the December expiry, taking the cumulative total against that expiry to 23,310 tonnes, with the balance of 10,120 tonnes re-grades. Nigeria was the dominant origin in the initial grading at 7,060 tonnes, roughly half the total, followed by Ecuador at 3,190 tonnes and Cameroon at 2,000 tonnes. Ecuador dominated the re-grades at 8,460 tonnes. Certified stocks ended July at 86,070 tonnes, a build of 17,010 tonnes on the month.
The September/December spread (UZ) settled at -£54, firming £8 on the month. Our view here is unchanged: with cheap differentials any strengthening in September is likely to attract more grading and, with 23,000 tonnes expiring in September, this should keep a lid on the September structure. The July grading activity supports this.
December remains the last expiry before EUDR takes effect. The December/March spread (ZH) settled at -£85, weakening £24 on the month. Non-compliant cocoa carries a £100 per tonne discount when delivered against 2027 contracts, so holders have reason to deliver to December 2026 unless the spread pays them to carry into 2027, and this should continue to weigh heavily on the December 2026 structure along with the expiries.
Within 2027 the front spreads were little changed, with March/May 2027 settling at -£3 and May/July 2027 at +£3, slightly lower on the month. Further out the spreads strengthened, with September/December 2027 settling at +£30, up £12, and March/May 2028 at +£17, up £20.
The £100 discount does not prevent non-compliant cocoa being delivered against the 2027 contracts. If holders carry cocoa rather than delivering to December 2026, that will weigh heavily on the 2027 structure until the non-compliant cocoa is delisted. If, on the other hand, non-compliant cocoa dries up in 2027, we could expect those positions to become extremely tight, with the increased position limits adding to that.
NEW YORK
Just 423 lots were delivered against July New York. Availability against September 2026 stands at around 2,206 lots, led by Ecuador at 1,466 lots, followed by Colombia at 264 and Peru at 229. Of the 2,166 lots graded against September, only 229 lots were initial gradings, with the remaining 89% re-grades.
The nearby structure remained weak, with the September/December spread (UZ) settling at -$139, weakening $29 on the month, and December/March (ZH) at -$92, down $14. The 2027 spreads
strengthened, with September/December 2027 at +$37 and March/May 2028 at +$25, up $32. With over 2,200 lots available to September, cheap differentials and stocks rising rapidly in the US, we expect the nearby structure to remain weak.
If EUDR is enforced then New York is likely to attract non-compliant cocoa, which will again weigh on the structure and could keep it weak into 2027. That should in turn cause the arbitrage to weaken, though as of this moment New York is at a premium to London across the 2027 positions, March 2027 at +£25 and May 2027 at +£39.
