Opinion: OPEC Reads Diesel While the Agencies Count Totals

OPEC Reads Diesel While the Agencies Count Totals On Wednesday morning in New York, crude futures were trading at $91.41 per barrel for Brent, up nearly $3 from the previous week, even as the EIA simultaneously published a crude inventory build of 4.4 million barrels, one of the largest single-week additions on record. The American Petroleum Institute had estimated a draw of 328,000 barrels the day before. The agencies' own surveys disagreed by 4.7 million barrels, and the market shrugged and rallied anyway. That divergence is a useful lens through which to read this week's broader forecasting war. The IEA and EIA now project 2026 global oil demand growth below what most analysts would consider statistically distinguishable from zero, a figure Ole Hansen at Saxo Bank described Monday as part of a more-than-two-million-barrel-per-day gap between the three major forecasters, highlighting what he called "extraordinary uncertainty." OPEC maintains demand will grow by 0.6 million barrels per day this year. Wall Street is mostly siding with the agencies. The data sitting in the products market is siding with OPEC. If global oil demand were genuinely contracting in any meaningful structural sense, the barrel fraction most sensitive to freight, industrial activity, and agriculture would be softening first and hardest. Trucking cycles, farm equipment, and locomotive fleets run on diesel. When those slow, distillate demand falls before anything else. Instead, ULSD sits at the 91st percentile of its 52-week range, up 33.8% above its 200-day moving average, with distillate inventories running 13% below the five-year average. Heating Oil closed Friday at $4.49. Managed money in NY Harbor ULSD holds a net long of 14,038 contracts, hardly the positioning of traders who believe industrial demand is rolling over. The agencies would respond that U.S. total products supplied, their own domestic demand proxy, was already running 2.9% below year-ago levels even as Brent rallied nearly $6 on the week to close Friday at $93.70. That is a genuine tension, and it deserves a direct answer. Aggregate demand figures capture everything from jet fuel to petrochemical feedstocks; they are a blunt instrument in a market where the product mix is shifting. What the total-products-supplied figure cannot tell you is whether the softness is concentrated in gasoline, where EV penetration is real and accelerating, or spilling into distillates. The distillate price and inventory data are unambiguous that it is not the latter. A 13%-below-average distillate stock position does not coexist with a demand-contraction story without some reconciliation the agencies have not yet offered in public. The crude inventory picture makes the reconciliation harder, not easier. The 4.4 million barrel EIA build brought commercial stockpiles to 428.8 million barrels, right at the five-year average, after API had suggested a draw. The most plausible explanation is a mismatch in refinery run-rate timing between the two surveys: crude arriving in tank while throughput lagged. Distillate output simultaneously fell to 5.2 million barrels per day in the same reporting period. Refiners appear to be accumulating crude input while cutting distillate yield, which means any crude surplus implied by the demand-contraction narrative will not automatically translate into product relief. The pipeline from crude surplus to diesel surplus requires refinery runs that are currently not happening at the yield ratios that would close the distillate gap. Brent at $93.70 and WTI at $86.64 reflect a market that has, at minimum, priced the supply story as incomplete. The agencies are also publishing their demand forecasts into a structural blind spot. Fingrid, the Finnish grid operator, called for "engine power", diesel generation, to balance data center load this week. This is not an isolated event; it is the visible edge of a demand category that neither IEA nor EIA sector models capture cleanly. Western agency oil demand accounting is organized around transport and industrial end-use. Backup and load-balancing generation for AI data centers sits in a category that barely existed when these models were designed. Bloomberg's own reporting shows that data centers face more intense local opposition than solar farms and nuclear plants, suggesting these facilities cannot readily expand grid connections at the pace their power draw is growing. When grid connections are constrained, operators run what they have. In Europe, that frequently means diesel. A portion of OPEC's demand optimism may be structurally correct for reasons that do not appear in any published agency model. The winter adds a further complication that neither side of this debate has fully priced. NOAA puts the probability of a very strong El Niño this fall and winter above 90%. That pattern historically suppresses Northeast Asian heating demand, freeing LNG cargoes globally, a bearish force for oil substitution. But the EU simultaneously loses Russian LNG supply on January 1, and European gas storage is tracking 61.8% full with Aurora projecting 68–73% by November 1. The revised EU target is 80%. Germany sits at 50.2% full. The Netherlands, a major LNG regasification hub, is at 42.7%. If El Niño suppresses Asian LNG demand and frees cargoes, that flow runs straight into a Europe that needs it; spot price relief for European buyers offsets itself against the structural shortfall. Oil sits in this picture as the swing fuel for power generation when gas tightens, and TTF closed Friday at $65.83. JKM settled at $22.94. The spread between Asian and European LNG prices is not wide enough to assume arbitrage will resolve the EU storage gap without oil-fired generation stepping in. OPEC is reading the same distillate screen as the rest of the market. The group's 0.6 million barrel per day demand-growth forecast may be too high in aggregate, but the directional disagreement with the agencies reflects something real: the product-level signals OPEC weights are pointing in a different direction than the aggregate flow data the IEA and EIA prioritize. Managed money net long in WTI sits at 103,715 contracts. Speculators are not positioned for a demand collapse. The agencies may prove correct on the annual average. Demand growth globally may well print below 0.5 million barrels per day for 2026. But the distillate market, the refinery run-rate data, the emerging diesel-generation load from data centers, and the structural LNG tightness heading into winter together constitute a set of product-level signals inconsistent with the comfortable demand-contraction story. Crude at $93.70 Brent is not the price of a demand collapse. Neither is $4.49 heating oil against inventories 13% below the five-year average. The agencies are counting total barrels. OPEC is counting which barrels, and where they go. That distinction is already visible in the price. The market has made its call.