From release to refill, the loop has closed. Before the crisis Japan's government held about 263 million barrels of crude, the world's third-largest stockpile. It released roughly 80 million barrels in March and a further 36.5 million barrels - about 20 days of consumption - in July; by end-July 2026 government holdings had fallen to about 182 million barrels. In days-of-cover terms, the national petroleum reserve fell from 146 days at end-2025 to 103 days at end-July 2026. In August a METI committee approved restoring government stocks to roughly 90 days of cover, and against the FY2027 target of meeting 90 days of crude import requirements Tokyo needs to add about 48 million barrels. On the preliminary basis published in early October 2026, Japan's total domestic petroleum stocks equate to about 200 days (102 days government, 93 days private, plus 5 days held jointly with producing countries inside Japan). Both the release and the refill follow published schedules: in the fourth quarter Japan returns to the Middle East and US crude markets as a buyer that looks at delivery, not price.

The tender terms reveal the buyer's real preference. JOGMEC's 28 August tender sought 2 million barrels of Murban crude for the Shibushi storage base in south-west Japan, priced at a USD 20/bbl discount to the Murban official selling price and inclusive of freight, demurrage and insurance, with delivery between 15 October and 14 December 2026. Three details matter. First, choosing a grade close to Japan's import structure shows the refill is about composition, not just volume. Second, folding freight, demurrage and insurance into a single discount means the buyer has bought out all delivery uncertainty in one number. Third, a two-month delivery window shows it prioritises certainty of arrival over hitting a particular price window. This buyer's willingness to pay for delivery certainty is clearly higher than its price sensitivity.

The cost lands on freight and source substitution. As of end-February 2026, 94.2% of Japan's crude imports came from the Middle East, with about 93% of that moving through Hormuz. Rerouting options run either through Saudi Arabia's Red Sea terminal at Yanbu and the UAE's Fujairah, or via Suez and then around the Cape of Good Hope. Idemitsu confirmed in August that the ongoing rerouting has stretched some Saudi-to-Japan voyages from 20 days to 50-60 days, and Argus estimates that shipping Saudi crude to East Asia around the Cape costs 64% more than the shorter route through Bab el-Mandeb. The supply base is also being reshuffled: US crude rose from under 4% of Japanese imports in 2025 to more than 30% by August 2026, while S&P Global Platts reported in early June that a US crude cargo bought by Idemitsu priced more than USD 20/bbl above the Dubai benchmark. Japan's cabinet recently approved three crude supply-security bills covering rerouting support, shipping insurance and investment in producer-country pipelines - policy is underwriting high-cost supply security.

What it means for SNSUC. First, restocking demand is price-insensitive and delivery-sensitive: a quote should compete on vessel type, load port, insurance and demurrage allocation rather than on price alone. In this round the party that can write guaranteed arrival into the contract beats the party that shaves two dollars off the price. Second, Japanese buyers have very low tolerance for changes in discharge port; rerouting and war-risk premiums must be itemised and verifiable, and quotes should state delivery windows rather than vague months. Third, source substitution is becoming structural (US crude above 30%), so cross-basin arbitrage windows will keep opening and closing: quotes into Japan need two benchmarks - Middle East (Dubai/Murban) and US - with an explicit rule for switching between them. Fourth, restocking demand is stacking on top of winter demand, so Middle East-Japan freight retains support through the fourth quarter; any cost model premised on freight falling back needs a margin of safety. The above is a market observation based on public data and does not constitute investment advice.