Behind the April Producer Price Report Lies the AI Shock
In April, producer prices went beyond what many had anticipated. The main takeaway? The producer price index (PPI) surged by 1.4%. Sales were nearly three times what Wall Street had predicted and saw a 6% increase year over year. Service prices climbed by 1.2%, while goods prices rose by 2%.
Energy prices, unsurprisingly, stole the spotlight. Final demand energy prices surged by 7.8% in April, with gasoline prices soaring by 15.6%. Other fuels like diesel and jet fuel also saw increases. Energy-related services, particularly trucking, experienced notable hikes as well. This situation was primarily an energy supply shock.
However, there’s more to the story. Beneath these soaring energy prices, the April PPI indicated a different type of inflation emerging elsewhere in the economy. This one is demand-driven rather than a result of supply issues, and it’s impacting businesses more than consumers. What we’re observing here is the boom in artificial intelligence investment.
To clarify, this does not seem to dampen consumer sentiment. Monetary policymakers consider price stability in terms of what consumers spend, which likely means the Fed won’t focus too much on these developments. In fact, prices for data processing and internet services actually fell. They rose by 0.4% in April but dropped by the same percentage compared to last year. Additionally, wireless phone prices have dropped nearly 3.7% over the year and remained steady in April. Basically, the PPI isn’t indicating that “cloud” services are becoming more expensive.
The point being made is that, in the physical world, prices are climbing sharply while the “cloud” persists.
The Rising Cost of Physical AI
AI is frequently treated as if it’s weightless, populated by software, algorithms, and models, seemingly above the economy. Yet, the AI boom is also driving growth in capital goods. It demands chips, circuit boards, servers, cooling systems, electrical infrastructure, warehouses, and yes, even construction workers. While AI may operate in a digital realm, it fundamentally relies on the tangible world.
This shift can be easy to overlook. The private capital equipment index appears relatively stable. In April, it increased by only 0.3%, up 4% from the previous year. However, this broad category doesn’t serve as a precise indicator for AI since it includes many items unrelated to AI like industrial machinery, medical devices, and vehicles.
Digging deeper into specific categories offers clearer indications.
Certain segments like printed circuit assembly saw extraordinary increases. In April, it jumped by 26.5%, reflecting a staggering year-over-year hike of 156.5%. These components are crucial for accelerator boards, servers, and other data center equipment.
The prices for electronic components and accessories rose by 8.1% in April and climbed 27.6% compared to the year prior. Storage for computers increased 2% for the month and over 20% for the year. These trends illustrate the growing demand amid the expansion of data centers and high-performance computing.
Moreover, the electrical components are equally pivotal. Electromechanical equipment rose by 2.8% in April and was up 13.3% year over year. Switchgear and industrial control equipment increased by 3.8% month-to-month and 12.1% year-over-year. These elements are fundamental to managing the substantial energy flows necessary for data centers.
The same upward trajectory shows up in infrastructure costs. Inputs for power and communication structures increased by 2.1% in April and 7.4% year over year, with goods inputs climbing by 2.9% in the past month and 9.8% over the year. Additionally, transportation and warehouse costs related to these buildings increased by 8.2% in April and 15.3% year-on-year.
It’s a prime moment for businesses involved in selling to the AI investment wave. The profit margins for capital equipment vendors grew by 3.8% in April and 12.3% for the year. Wholesale margins in machinery and parts rose by 3.5% month over month and 15.1% year over year. Equipment sales also saw a significant increase in margins, indicating AI may be playing a role in this growth.
Signals Not as Obvious
This situation also sheds light on why certain signals might be easily overlooked. AI isn’t neatly categorized in the PPI. The impact can be seen across electronic components, circuit assemblies, and a host of other related areas. To fully grasp AI’s influence, you need to look beyond the standard 34-page PPI report. A more detailed examination is available in the “Detailed report”. In April, this expanded to 329 pages.
The Bureau of Labor Statistics (BLS) might also be underreporting the true extent of AI’s reach because of how it weighs different components in the PPI. Many items contributing to PPIs often lag behind economic realities. The weight structure changes infrequently and tends to reflect economic conditions from 2017. This means that while current AI investments in diverse sectors might lead to price changes, those shifts may not be enough to significantly shift the entire index.
In the long run, AI is likely to create a deflationary effect through heightened productivity. However, before such advancements lead to lower prices, someone has to construct an AI system. And, the costs associated with its development are climbing.
