The Sunday Brief — July 26, 2026: Two Tariff Regimes, Two Different Rulebooks

What defined the week ending July 26, 2026?

Two tariff actions landed four days apart, and they operate under different rules. Confusing the two will cost you money in both directions.

On July 24 at 12:01 a.m. Eastern, USTR enacted a Section 301 framework across 60 economies at 10% or 12.5% ad valorem, concluding forced-labor enforcement investigations opened in March. The action reaches 99.4% of US imports by value. Free trade agreement preference does not automatically defeat the duty, but goods entered free of duty under USMCA or CAFTA-DR are exempt outright, alongside a long product exemption list.

On July 20, four days earlier, three presidential proclamations under Section 338 of the Tariff Act of 1930 imposed additional 50% duties on Canadian dairy, alcoholic beverages and motor vehicle goods, effective August 19. Those duties apply whether or not a good originates under USMCA. Section 338 carries no time limit.

The rate differential is the story. Section 338 sits four to five times higher than the Section 301 layer, applies to a narrower product set, and offers no origin defence.

Preliminary flash PMI data released July 24 showed the S&P Global US Manufacturing PMI easing to 53.8 in July from 53.9 in June, below a 54.3 consensus. Production continues to expand. Supplier delivery times lengthened at the fastest pace since August 2022, extending worsening lead times to 11 consecutive months. Input price inflation re-accelerated to a 14-month high. S&P Global cites Middle East geopolitical tension as a contributing factor, particularly on delivery times.

For CPOs and CFOs, the mandate before Monday is to price both regimes correctly, and to stop treating origin as a single binary variable.

Key developments this week

How does the new Section 301 forced-labor regime actually price?

What happened: USTR published a Notice of Action on July 23 concluding its Section 301 investigation into foreign forced-labor import prohibitions. Trading partners committed to adopt and effectively enforce a prohibition face a 10% additional duty. Partners with no prohibition in place face 12.5%. Of the 60 economies investigated, 54 failed to impose and effectively enforce a prohibition, and six failed to effectively enforce one.

Five partners work differently. The European Union, Japan, South Korea, Switzerland and Taiwan receive a combined-rate structure. Where a product's column 1 rate already equals or exceeds the applicable threshold, no additional duty is assessed. Where the column 1 rate falls below, duty is assessed so the combined rate equals the threshold. Thresholds are 10% for the EU and Taiwan, and 12.5% for Japan, South Korea and Switzerland.

The exemptions are extensive. Products of Canada and Mexico entered free of duty under USMCA are exempt under headings 9903.05.93 and 9903.05.94. Textile and apparel goods of Costa Rica, the Dominican Republic, El Salvador, Guatemala, Honduras and Nicaragua entered free of duty under CAFTA-DR are exempt under 9903.05.95. Section 232 articles are exempt under 9903.05.90, covering aluminium, steel, copper and derivatives, passenger vehicles, light trucks, medium and heavy duty vehicles and their parts, wood products, and semiconductor articles. Civil aircraft and parts, pharmaceutical-use articles, donations and informational materials each carry their own heading. Chapter 98 entries are generally excepted, with 9802 provisions dutiable on foreign value added. USTR expanded the final exemption list by 471 HTSUS subheadings after receiving public comment.

Why this matters: the operative test is not whether a preference program exists. The test is whether the specific goods enter free of duty under one. Eligibility on paper shields nothing. A shipment claiming USMCA preference and clearing duty-free escapes the Section 301 layer. The same part number entering under a non-preferential claim does not.

Executive implications: rebuild COGS models line by line against the annexes rather than by country. Confirm your broker reports the correct Chapter 99 heading in the correct trade remedy sequence per CBP guidance in CSMS #69326983, with Section 301 first, followed by Section 122, Section 232 and Section 201. Audit whether your USMCA claims are being filed and substantiated, because the claim is now worth 10% rather than a rounding error.

Why is Section 338 the regime with no origin defence?

What happened: on July 20, three proclamations under Section 338 imposed additional 50% ad valorem duties on specified Canadian goods, effective 12:01 a.m. ET on August 19, 2026. The three cover dairy, alcoholic beverages and motor vehicles. Annexes span products from wine to hockey sticks to cement. Exclusions include energy, potash, goods subject to Section 232, fish and certain critical minerals. The White House stated the duties apply to all covered goods regardless of whether a good originates under USMCA.

USTR framed the action as offsetting Canadian discrimination against US exports, citing provincial removal of US alcohol products, preferential dairy access for the EU, and caps on US vehicle exports to Canada. Canadian imports of US motor vehicles fell roughly 22%, or $5.6bn, from April 2025 through March 2026 against the prior period.

Why this matters: Section 338 has sat largely unused for decades. Revival gives the administration an authority with no rate ceiling below 50%, no sunset, and no FTA carve-out. The statutory 30-day delay leaves a negotiating window before mid-August, so the duties may be adjusted or withdrawn.

Executive implications: if you buy Canadian dairy ingredients, alcohol inputs or automotive components, model August 19 at 50% now and build the negotiation contingency into your Q4 forecast. Check the annexes against your part numbers directly. The motor vehicle annex reaches well beyond automotive.

What does the July flash PMI signal about lead times and input inflation?

What happened: the S&P Global Flash US Manufacturing PMI slipped to 53.8 in July from 53.9 in June, below the 54.3 consensus and above the 50.0 expansion threshold. Supplier delivery times lengthened at the sharpest pace since August 2022, an 11th consecutive month of deterioration. Input cost inflation accelerated to a 14-month high, its highest since May 2025. Selling price inflation reached its steepest pace since August 2022, indicating manufacturers are passing costs through rather than absorbing them.

Two caveats belong on the reading. S&P Global links the delivery-time deterioration substantially to Middle East shipping disruption alongside tariffs. Chris Williamson also flagged one-time spending around the FIFA World Cup and the USA 250 bicentennial, both inside the July 9 to July 23 survey window, as a temporary lift to growth signals.

Why this matters: the shipping picture is worse than the tariff picture. Brent moved from below $70 on July 1 to the mid-90s by late July on renewed strikes, Houthi attacks on Saudi tankers in the Red Sea, and a Caspian Pipeline Consortium suspension affecting roughly 80% of Kazakh exports. Two chokepoints under pressure at once compounds the tariff reset rather than running parallel to it.

Executive implications: stop uncoordinated buffer buying. With new-order momentum softening and holding costs rising, broad safety-stock accumulation locks up liquidity in low-risk categories while leaving specialised, single-sourced components exposed. Target buffers by lead-time variance and sole-source status, not by spend.

Who qualifies for the Phase 3 IEEPA refund order?

What happened: on July 17, 2026, Court of International Trade Senior Judge Richard Eaton issued an order, made available July 21, directing CBP to reliquidate without regard to IEEPA duties any of the plaintiffs' entries liquidated more than 80 days earlier. The order follows the transfer of more than 3,700 refund cases to Judge Eaton's docket. CBP must report on Phase 3 progress by August 4.

The critical limit: the order applies only to companies with cases already filed at the CIT. Non-litigants remain the contested category. DOJ appealed the broader refund order on June 2, arguing the CIT lacks authority to grant relief beyond parties before the court. Government exposure on finally liquidated entries is estimated above $30bn.

What is open to every importer: CAPE Phase 1, live since April 20 inside the ACE portal, covering most unliquidated entries and certain recently liquidated entries. More than $95bn has been queued and more than $40bn disbursed through end June. Phase 2 covers Reconciliation Program entries.

Executive implications: determine your filing status first. If you filed at the CIT, task trade compliance with the entry-summary matching and importer of record submission the order requires. If you did not file, work CAPE Phase 1 for unliquidated and recently liquidated entries, and take legal advice on whether filing at the CIT is worth the cost for your finally liquidated exposure. Either route delivers non-dilutive cash to offset the new July 24 duty layer.

Why is enterprise AI delivering insights rather than the promised cost cuts?

What happened: an SAP survey of enterprise technology leaders, reported by CIO Dive in July, found enterprises seeing measurable AI value in generating business insights and improving customer interactions. Cost reduction and time savings, the two outcomes written into most original business cases, remain difficult to demonstrate. Separate CIO Dive reporting shows agentic workloads driving a sharp rise in total AI usage and spend, with OpenAI advising CIOs to establish visibility into demand, spend and risk before scaling.

Why this matters: many 2025 and 2026 technology budgets were approved on efficiency grounds. The measurable wins are landing in analytics, scenario planning and decision quality. The mismatch is a measurement problem rather than a technology failure.

Executive implications: rebase procurement technology ROI on margin protected, duty exposure avoided, refunds captured and inventory velocity improvement. Those variables align with the current tariff and lead-time environment. Administrative headcount reduction does not.

What does this mean for category strategy?

Two regimes with different origin rules expose the structural weakness of static category management. Annual RFP cycles evaluating suppliers on unit price and a single country-of-origin field cannot price this.

Landed cost now depends on which instrument applies to the specific part number. For a flat-tier origin, total landed cost equals unit price plus freight and insurance plus the column 1 duty plus 10% or 12.5%. For the EU, Taiwan, Japan, South Korea or Switzerland, the calculation is unit price plus freight and insurance plus the greater of the column 1 rate or the applicable threshold, because the Section 301 layer tops up rather than stacks. For USMCA-qualifying goods entering duty-free from Canada or Mexico, no Section 301 layer applies at all, and from August 19 a Canadian-origin good on a Section 338 annex carries 50% whether USMCA applies or not.

Three implications follow.

1. Origin is no longer one field. Model instrument-by-instrument at the part-number level. A single country code cannot carry Section 301 tier, Section 232 overlap, Section 338 annex status, and USMCA claim substantiation at once.
2. Preference claim discipline now has direct P&L value. A USMCA claim you fail to substantiate costs 10%. Sub-tier bill-of-materials traceability moves from a compliance exercise to a margin control.
3. Contracts need duty language, not price language. Build dynamic, index-linked clauses mapping duty liability explicitly between buyer and seller, bound by symmetric risk-sharing collars. Specify which party carries a new instrument, not merely a new rate.

How do leading organizations compress decision velocity?

Two tariff actions four days apart, on different statutes, with different origin rules, is a test of decision velocity. Velocity here means the elapsed time from a regulatory signal to a quantified total cost impact to an executed sourcing decision.

Legacy operating model: wait for the customs broker statement, identify a cost variance, run a manual bill-of-materials audit, shift vendors weeks later. Margin already lost.

Continuous intelligence model: ingest the Federal Register notice and the annexes, map every part number against exemption headings and rate tiers automatically, quantify margin exposure by SKU, and surface pre-qualified alternatives and refund eligibility in the same view. Margin defended.

The difference is not analytical horsepower. The difference is whether your part-number data already carries the attributes the new rules ask for. Organizations with clean sub-tier origin data repriced their book on July 24. Organizations without it are still building spreadsheets.

The Kodiact perspective

World-class manufacturers treat procurement as an engine of balance-sheet resilience rather than a purchasing function. In an era of continuous regulatory change, predicting the next pivot is unreliable. Designing for adaptability is not.

Resilience requires removing the silo between finance and procurement execution. Finance operating alone mandates blanket inventory cuts to optimise working capital. With supplier lead times worsening for 11 consecutive months, stripping buffers during extended delivery delays causes line stoppages far exceeding the holding-cost savings. Procurement operating alone builds buffers everywhere and locks up cash.

Alignment happens when sourcing allocations, capital deployment and inventory buffers are evaluated through one lens covering total landed cost, regulatory defensibility, operational capacity and working capital. Enterprises reaching that state insulate earnings before shocks reach the quarter.

Boardroom questions to ask this week

1. Instrument mapping: for our top five product lines, which parts fall under the July 24 Section 301 tiers, which fall under a combined-rate origin, which are exempt via USMCA or Section 232, and which appear on a Section 338 annex effective August 19?
2. Preference claim integrity: are our USMCA and CAFTA-DR claims substantiated to audit standard, now that a failed claim costs 10% rather than nothing?
3. Refund position: have we established whether we are a CIT plaintiff, and are we working CAPE Phase 1 for unliquidated and recently liquidated entries?
4. Buffer targeting: with lead times worsening for 11 straight months, are our safety stocks concentrated on single-sourced, long-lead components rather than spread across low-risk categories?
5. Decision lead time: how many hours does recalculating total landed cost across our global bill of materials take when a new trade instrument lands at midnight?

Conclusion: structural trade friction is now a permanent operating feature

The week ending July 26 proves the point twice. A Section 301 regime covering 99.4% of US imports with a detailed exemption architecture, and a Section 338 regime at 50% with no origin defence at all, arrived four days apart under different statutes.

Neither regime rewards guessing. Both reward part-number-level precision about which instrument applies. Enterprises relying on periodic reviews and country-level origin fields will absorb margin erosion in both directions, overpaying where exemptions apply and underpricing where they do not. Long-term profitability belongs to leadership teams treating trade compliance as a core financial variable and building the decision velocity to act on it within hours.