October 2026 producer prices expose a difficult combination for earnings and AI investment: final-demand PPI rose 6.52% year over year, energy increased 37.03%, and semiconductor and other electronic component manufacturing prices climbed 30.64%. These are substantial price pressures, although their recent trajectories differ.
The management question is whether customers can absorb higher prices fast enough to support margins and finance the next wave of investment. Energy, freight and computing equipment can raise buyers’ bills while price-sensitive customers constrain sellers’ ability to pass them on. That creates a demanding environment for businesses promising returns from expensive AI infrastructure.
Analysis cutoff: November 15, 2026. Data cover October 2026, released November 13. Sector comparisons below use BLS not seasonally adjusted (NSA) index levels and ATN calculations.

October PPI: high annual inflation, uneven recent momentum
The separate BLS seasonally adjusted release reported a 0.3% October increase, down from September’s revised 1.4%. Energy fell 0.1% in October on that basis. These movements qualify any claim that inflation accelerated uniformly in the latest month.
| Industry or series | Y/Y | Oct. vs Jan.–Sept. average | Aug.–Oct. vs May–July average |
|---|---|---|---|
| Final demand (WPUFD4) | 6.52% | 2.64% | 1.25% |
| Goods (WPUFD41) | 10.24% | 3.98% | 1.52% |
| Foods (WPUFD411) | 1.71% | 1.30% | -0.34% |
| Energy (WPUFD412) | 37.03% | 12.61% | 5.09% |
| Services (WPUFD42) | 4.98% | 2.01% | 1.09% |
| Transportation and warehousing (WPUFD422) | 18.53% | 9.30% | 3.85% |
| Less foods and energy (WPUFD49104) | 5.05% | 2.03% | 1.07% |
Energy remains a major source of annual pressure, but October brought some relief. Its NSA index fell 1.77% from September. Freight-related services show a different pattern: transportation and warehousing prices continued to rise. A large annual increase can coexist with a monthly decline, and an elevated quarterly average can persist after a shock starts to ease.
| Industry or series | Y/Y | Oct. vs Jan.–Sept. average | Aug.–Oct. vs May–July average |
|---|---|---|---|
| Total manufacturing | 13.29% | 5.54% | 2.38% |
| Services less trade, transportation and warehousing | 3.25% | 1.12% | 0.49% |
| Total trade | 7.56% | 3.17% | 2.67% |
| Retail trade | 5.48% | 2.56% | 1.15% |
| Wholesale trade | 8.94% | 3.56% | 3.66% |
| Transportation | 14.98% | 7.75% | 3.30% |
| Transportation and warehousing | 13.99% | 6.84% | 2.98% |
| Selected health care | 2.26% | 0.95% | 0.46% |
Manufacturing and transportation stand out against the more moderate annual readings for selected health care and services excluding trade, transportation and warehousing. However, these are selling-price indexes, not direct measures of production costs or earnings. Higher output prices may benefit some suppliers while squeezing their customers.
Trade indexes generally measure margins between selling and acquisition prices, rather than the total price of merchandise or net profits. Nor can warehousing inflation be recovered simply by subtracting the transportation percentage from the combined transportation-and-warehousing percentage; aggregation requires compatible levels and weights.
Production stages: an inflation pipeline, not a guaranteed timetable
| Industry or series | Y/Y | Oct. vs Jan.–Sept. average | Aug.–Oct. vs May–July average |
|---|---|---|---|
| Stage 4 (WPUID54) | 7.98% | 3.14% | 1.90% |
| Stage 3 (WPUID53) | 8.42% | 3.13% | 1.05% |
| Stage 2 (WPUID52) | 13.30% | 3.59% | 1.86% |
| Stage 1 (WPUID51) | 13.91% | 5.62% | 2.62% |
The annual increases are strongest in stages 1 and 2. These input indexes provide more direct evidence of pressure on businesses buying intermediate goods and services than an industry’s own output-price index does.
Yet the pipeline does not move mechanically or at a fixed speed. Stage 2’s NSA index fell 0.34% in October, while stage 1 rose 0.59%. Contracts, substitution, competition, productivity and demand determine how much reaches finished products. Some businesses may absorb higher costs through margins; others may reduce orders or investment.
Higher replacement costs can also increase working-capital requirements. They do not automatically raise the carrying value of existing inventories: accounting methods, purchase timing and possible write-downs matter.
Utilities: generation prices are the sharper pressure point
| Industry or series | Y/Y | Oct. vs Jan.–Sept. average | Aug.–Oct. vs May–July average |
|---|---|---|---|
| Utilities | 6.69% | 2.83% | 4.61% |
| Electric power generation | 20.21% | 23.38% | 21.73% |
| Bulk power transmission and control | 11.61% | 3.51% | 2.56% |
| Electric power distribution | 1.30% | -2.78% | 0.57% |
| Industrial electric power | -1.96% | -3.14% | 2.41% |
| Natural gas distribution | 7.29% | -1.24% | 0.14% |
| Industrial natural gas | 3.02% | -4.78% | -1.07% |
The divergence is consequential. Electricity-generation selling prices are much stronger than distribution prices, while industrial natural gas is below its January–September average. These indexes measure different activities and should not be treated as interchangeable electricity bills.
For data centers and industrial buyers, exposure depends on local tariffs, power-purchase agreements, fuel mix and contract resets. A high generation index can signal pressure on future procurement, but it does not establish an equal increase in every customer’s bill. The index alone also cannot attribute industrial electricity movements to line upgrades, voltage requirements or AI demand.
Technology: annual price shocks meet a more selective customer
| Industry or series | October MoM | Y/Y | Aug.–Oct. vs May–July average |
|---|---|---|---|
| Computer and electronic product manufacturing | 2.18% | 14.46% | 2.06% |
| Computer and peripheral equipment manufacturing | 0.00% | 21.64% | 6.86% |
| Computer storage device manufacturing | 0.00% | 80.34% | 17.94% |
| Semiconductor and other electronic component manufacturing | 3.91% | 30.64% | 4.63% |
| Bare printed circuit boards | 0.00% | 46.43% | 7.39% |
| Printed circuit assemblies | 11.72% | 150.56% | 15.24% |
The exceptional storage and circuit-assembly readings were reproduced from the specified BLS series. Their interpretation needs care: storage prices were effectively unchanged in October despite the very large annual increase; circuit assemblies continued to rise sharply. These categories are broader than AI-specific hardware, exclude imported production as such, and cannot establish the cost change of a particular GPU, server or data-center project.
The earnings channel is conditional but economically important. Higher equipment purchase prices can increase capital expenditure and, subsequently, depreciation. Electricity raises operating costs. R&D and financing add further demands on cash. Whether earnings deteriorate depends on revenue growth, utilization, efficiency, pricing and the timing of these expenses.
The demand-side concern is that households protecting essential spending may resist premium AI subscriptions. If consumers favor mid-priced goods and services, providers could face pressure to deliver more functionality at a lower price. This is the demand scenario assessed here; PPI itself does not measure disposable income or prove a broad shift toward particular price tiers.
Consumer subscriptions are also only one market. Enterprise software, developer tools and usage-based services can earn revenue when customers see measurable productivity gains. The relevant test is whether revenue and gross profit cover inference, development and infrastructure costs—not whether every household can afford the highest-priced plan.
AI developers already generate revenue. The difficult phase is therefore investment before sustainable profitability and cash generation, rather than investment before any revenue exists. Higher hardware prices may support suppliers’ revenues while making buyers’ returns harder to achieve; the market effect will differ across the chain.
Inflation and rates: pressure persists, but the path remains open
The evidence supports continued vigilance about energy, freight and computing-equipment prices. It does not establish that CPI or PPI must rise faster in every coming month. Sustained energy relief, weaker demand or productivity gains could reduce pass-through; renewed fuel increases could extend it.
Higher policy rates would add financing pressure to a cost-intensive investment cycle, but monetary policy also responds to inflation and employment. Rate increases cannot directly create energy supply, while premature easing could complicate inflation control. The balance is a policy trade-off, not a conclusion that one action is inevitably disastrous.
A 4.25% yield on the 13-week Treasury bill is an author-defined monitoring threshold for tighter short-term financing conditions. Its significance would depend on the speed and persistence of the move, inflation expectations and broader funding markets; it is not an automatic trigger for a Fed decision or a systemic break.
Credit stress could transmit through supplier financing, customer payment delays and refinancing needs. PPI alone cannot establish the probability of an AI company failing or quantify resulting market losses. Receivables quality, financing maturities, credit spreads and cash flow would be the evidence needed to evaluate that risk.
The conclusion from the numbers is a cost-and-return challenge, rather than a uniform acceleration story. Strong annual producer-price increases leave buyers exposed even where monthly inflation eases. For AI, the decisive question is whether paying demand and operating efficiency can grow fast enough to finance increasingly expensive infrastructure. Watch fuel and power procurement alongside customer retention, margins and free cash flow.
Related ATN analysis: September PPI: energy, diesel and technology inflation.
Sources and methodology
ATN recalculated the tables from BLS monthly index levels retrieved through the public API on November 15, 2026. October observations are preliminary; the release incorporates revisions to earlier months. Results use the retrieved vintage and may change after subsequent revisions. Tables are NSA throughout; the separate official monthly headline is SA.
- Y/Y: (October 2026 ÷ October 2025 − 1) × 100.
- October versus January–September average: (October index ÷ arithmetic mean of the nine preceding 2026 monthly indexes − 1) × 100. This is an equally weighted average of monthly levels, not conventional YTD growth and not an annualized rate.
- Three-month-average comparison: (mean August–October ÷ mean May–July − 1) × 100. This is a comparison of adjacent three-month averages, not October versus July.
- Annualization: where used, [(ending level ÷ starting level)4 − 1] × 100 for a three-month interval. Multiplying a quarterly percentage by four is only an approximation. For example, generation’s 21.73% three-month-average increase compounds to about 119.60%, not 86.93%. Short-term NSA annualization can magnify seasonal effects, so the tables retain the period changes.
Industry series are identified by their BLS industry names; commodity and production-stage codes appear where useful. Overlapping aggregates must not be added together, and percentages are not contributions without appropriate weights.
- BLS: October 2026 PPI release, published November 13 — official headline and revision context.
- BLS public API: monthly index observations — underlying series used for ATN calculations.
- BLS: PPI definitions and coverage — domestic producer selling prices and industry classifications.