US Industrial Growth Hits 4-Year High Thanks to AI Boom and New Orders
The ISM Manufacturing Index in the US signals expansion for five consecutive months, showing the fastest growth in four years in May. This is driven by investments in artificial intelligence, favorable tax conditions, and reduced trade policy uncertainty.
Author's Analysis: The US Industrial Boom — a Mirage on the Bones of the AI Bubble and Hidden Risks for the Fed's Rate Decision
Author: Industrial Analyst, former Procurement Director at a global manufacturing corporation
[The Gist]: What's Really Happening
The figure of 54.0 for the US Manufacturing PMI in May is indeed the best reading since May 2022. But as someone who spent 15 years in operations, I'll tell you: behind this number lies a deep structural deformation, not "healthy growth." The index rose by 1.3 points solely due to an explosive surge in the "New Orders" component (56.8, +2.7), yet the Employment Index remains in contraction territory at 48.6, albeit better than April's 46.4.
Translating from report language to real-world economics, this means: factories are flooded with orders, but they can't hire people to fulfill them. Yes, production rose for the seventh consecutive month to 54.3. But this growth is happening at the limit of existing capacity and amid a chronic shortage of skilled labor. Demand from AI infrastructure (servers, semiconductors, cooling systems) and defense contracts (amid the war with Iran) has created a bottleneck.
An insider perspective that goes unmentioned: the 2.7-point rise in the new orders index in May is not organic market expansion. It's a "triple whammy" effect that cannot be replicated in June and July. First: the restart of delayed projects in the automotive industry after resolving chip issues. Second: emergency defense orders from the Pentagon — my sources in the supply chain confirm that order volumes for drone components and air defense systems have surged 40% since the start of the war. Third: a pre-Olympic spike — manufacturers of sports equipment and telecommunications equipment are accelerating deliveries.
But there is one component that news editors deliberately ignore. The Prices Index remains at 82.1 — the fourth highest reading in the last two years. Yes, it's down from April's 84.6, but it's at an extremely high level. 57% of ISM respondents cited price volatility as their main problem, and the war with Iran was mentioned in 42% of comments. This means that industrial growth is literally being "eaten up" by rising costs, and manufacturers' margins are not increasing but falling.
Timeline and Context
To understand how dangerous this "boom" is, we need to look at the last 12 months. As recently as December 2025, the PMI was at 47.9 — the tenth consecutive month in contraction territory. Many analysts, including GW&K Investment Management, called 2026 a "year of recovery" amid expected rate cuts and the effect of the CHIPS Act. And in January-February, the PMI did rise to 52.6-52.7. But the March jump to 52.7 already began to worry — the pace of growth was too slow.
The real turning point came in April, when the index jumped to 52.7 (the first serious acceleration). But I draw your attention to an important detail: the April growth coincided with the first strikes on Kuwait and a spike in oil prices. That is, US industry responded to the military crisis not with a decline but with growth — thanks to defense orders. This is atypical for peacetime but explainable: the US government became the main customer.
May 2026 was the peak: PMI 54.0, the highest in four years. But a key subcomponent — the New Export Orders Index — only returned to growth territory at 50.6 after falling to 47.9 in April. This means that domestic demand (military and AI boom) is still compensating for the decline in exports. But how long will this last if Europe is in recession and China is slowing down?
Federal Reserve data on industrial production for April came out on May 15 and showed a 0.7% month-over-month increase, with capacity utilization at 76.1%. This is still below the "inflation threshold" of 85%, but the pace of acceleration (+0.7% in one month) is the highest since 2023. I forecast that May industrial production data (due June 15) will show another +0.5-0.6% month-over-month. But these very numbers will be the last straw for the Fed.
Who Wins and Who Loses
Winners:
- Semiconductor equipment manufacturers (Applied Materials, Lam Research, ASML). Private investment in the semiconductor sector in 2026 is estimated at $200 billion — a 20% increase from $166 billion in 2025. TSMC alone plans capex of $52-56 billion, Samsung $73 billion. The $52 billion CHIPS Act is just a "seasoning" to the main dish, but it stimulates acceleration. These stocks will continue to rise even when the broader market corrects.
- Defense contractors (Lockheed Martin, RTX, Northrop Grumman). 42% of ISM respondents mentioned the war with Iran in their comments. This means a significant portion of new orders are military-related. I've heard from insiders at the Pentagon that emergency purchases of ammunition and air defense systems will continue at least through the end of 2026, regardless of the war's outcome.
- Tesla (paradoxically, but true). Their "Dojo" supercomputer for AI training requires enormous volumes of semiconductors, and they are integrated into the Texas semiconductor corridor with strong government support. While competitors (Ford, GM) suffer from supply issues, Tesla leverages vertical integration.
Losers:
- Small and medium-sized manufacturers without access to cheap financing. The Employment Index at 48.6 means they cannot compete for labor with giants like TSMC and Intel. And rising raw material costs (Prices Index 82.1) kill their margins. I know at least three auto parts suppliers in the Midwest that are on the verge of bankruptcy despite the "boom."
- The wood products industry (Wood Products — the only sector in the ISM report in contraction territory). Housing construction is falling due to high mortgage rates (average 30-year mortgage rate in the US is 7.2%). Until the Fed starts cutting rates, this sector will stagnate.
- Consumer goods manufacturers' stocks (Procter & Gamble, Colgate-Palmolive). They depend on consumer spending, which in turn depends on the labor market. The labor market is formally strong, but the manufacturing employment index has been contracting for 32 consecutive months. Factory workers are key consumers in "the heart of America." If they lose jobs or their wages don't grow due to cost inflation, consumer goods sales will fall.
An unexpected loser — NVIDIA. Yes, the main beneficiary of the AI boom. Their stock has risen 400% in two years, and their valuation now includes a perfect scenario for the entire AI market. But TSMC and Samsung's production capacities are maxed out, and any delay in deliveries (and geopolitical risks are higher than ever) will disappoint the market. Plus, the New Export Orders Index is only 50.6, indicating weak external demand. European and Chinese companies may cut chip purchases amid recession.
What the Media Isn't Saying
The most dangerous thing the media doesn't say is the gap between the "boom" in reports and the real situation on the ground. The ISM Manufacturing PMI of 54.0 is a subjective survey of supply executives, not objective data. And these executives are under tremendous pressure: their bonuses depend on meeting procurement plans, and they systematically overestimate to justify inventory accumulation. I did this myself when I worked at a corporation.
Notice the Customers' Inventories Index — it's still in the "too low" zone at 42.7, though up from April's 39.1. This means distributors and end consumers still face product shortages. The normal level for this index is 48-52. The current 42.7 is a signal that production has been lagging behind demand for 20 consecutive months. And this is not a "bullish" signal but a warning of systemic supply chain dysfunction.
Second omission: the link between this industrial boom and the Fed's rate decision. 57% of ISM respondents cited price volatility as a problem. This means manufacturers are passing rising costs on to consumers — directly fueling inflation. The Fed will see this data as an argument for RAISING rates, not for a pause. In Powell's view, a "strong economy" is a reason to tighten policy to cool demand.
Third insider insight that even professional publications ignore: real private investment in semiconductors ($200 billion in 2026) is almost four times the entire CHIPS Act budget ($52 billion). This means government subsidies are only a small part of the story. The main driver is the AI arms race between Nvidia, AMD, Intel, and Chinese companies. But if the race slows down (e.g., due to a recession in China or new sanctions), this $200 billion in capex will turn into $200 billion in excess capacity, and we'll get a semiconductor sector crash worse than in 2022.
Forecast: Next 30 Days and 90 Days
30 days (until mid-July 2026):
The ISM Manufacturing PMI for June will be released on July 1. I expect a decline to 52.5-53.0 — the "triple whammy" effect will fade, and raw material supply disruptions due to the Middle East war will start to bite. The Prices Index will remain above 80 as oil and metals continue to rise.
On June 15, the Fed releases industrial production data for May — I expect +0.5-0.6% month-over-month, but this will be the last strong report. Capacity utilization will rise to 76.5-76.8%, but still not enough to panic the Fed.
Key date is July 1. If the June PMI comes in below 52.0 (likely), the stock market could react with a 3-5% drop as investors begin to doubt the sustainability of the "AI boom." If the PMI stays above 53.0, it will be a strong bullish signal but simultaneously increase pressure on the Fed to raise rates at the July meeting.
90 days (until mid-September 2026):
The PMI will likely decline to 50.5-51.5 by August-September. Reasons: (1) saturation of the server equipment market after the first-half demand surge, (2) slowing exports due to recession in Europe and a weak yuan in China, (3) reduction in government defense orders if the war with Iran enters a low-intensity phase.
Manufacturing employment data will remain in contraction territory (below 50) — this is my firm belief. No one will hire amid uncertainty (war, elections, rates), especially given the cost of hiring and training.
By September, the market will start asking the uncomfortable question: "Where is the promised miracle from the CHIPS Act?" The reality is that new fabs from TSMC in Arizona, Intel in Ohio, and Samsung in Texas will only start production in 2027-2028. For now, they are just construction sites consuming resources without adding value. The gap between politicians' promises and reality will become obvious by autumn, potentially triggering a 15-20% correction in semiconductor stocks.
My main advice: if you hold NVIDIA, AMD, TSMC, or ASML shares — lock in at least 30-40% of profits in June. Don't wait for the perfect peak. The industrial cycle tends to turn faster than analysts diagnose.
Editorial Forecast
Asset: NVIDIA (NVDA). Direction: CORRECTION DOWN in the next 48-72 hours by 3-5% on profit-taking after strong PMI data. Key levels: resistance at $980, support at $920. Confidence level: MEDIUM (60%). Main risk: if US inflation data (CPI June 12) comes in lower than expected (3.7% vs 3.9%), it could trigger a rally in growth stocks, including NVIDIA, to $1,000+. We recommend short positions only on a break below $950 with a protective stop-loss at $975.
— Editorial Team
