Summary
- Kevin Warsh, the Chair of the Federal Reserve, emphasized the significance of artificial intelligence (AI) in his keynote speech at Jackson Hole, describing this moment as a "hinge point in history."
- He noted that annual sales from the top two AI labs have surpassed $100 billion, showing a staggering increase of over 500% from the previous year, and indicated that the Fed is now considering AI as "potentially a new factor of production."
- Warsh is particularly focused on observing the growth rate of capital spending related to AI, which he refers to as its "second derivative."
During his inaugural keynote at Jackson Hole as Federal Reserve chair, Kevin Warsh drew considerable attention for his comments on artificial intelligence, even as he remained tight-lipped about interest rates. Nestled in the section titled "Preparing for Future Policy Conjunctures" lies crucial information for those interested in the flow of investments into AI.
Warsh began by reflecting on the economic climate prior to the AI boom, where economists warned of "secular stagnation" following the 2008 financial crisis. He pointed out that this notion of excess capital chasing insufficient investment opportunities is now outdated.
Myriad: What will the Fed decide in September? Make your prediction.He supported this assertion with data, revealing that business capital expenditures—termed "the seed corn of future economic growth"—are increasing at their fastest rate since 2021, with a growth of about 9% over the last year, over half of which is associated with AI investments. Looking ahead, he stated that he would be monitoring not just the level of spending but also its growth rate, or "the second derivative."
1. AI Development Surpasses Expectations
According to Warsh, the advancements in artificial intelligence have outstripped the expectations of even its most ardent supporters. He remarked, "The potential for substantially higher growth is on the rise."
He elaborated further, stating, "Endless streams of capital are flowing into various AI-related infrastructures, creating a sort of hyper-Moore's law effect."
Moore's law, which observes that computing power tends to double approximately every two years, is being outpaced by what Warsh terms a "hyper" version, suggesting that AI capabilities are advancing even more rapidly. This perspective is a notable shift for a central banker traditionally viewed as hawkish, indicating that he sees AI as a significant driver justifying the recent influx of capital into the market.
2. The AI Economy is Valued at $100 Billion
Warsh explained that the convergence of capital and labor has given rise to the large language models that are central to AI technology. He noted, "Users purchase tokens to access these models."
He quantified this by stating, "Annualized sales of tokens for the two leading labs are reported to exceed $100 billion, marking a growth of more than 500% compared to last year."
In this context, a token represents a fundamental unit of access sold by AI companies, essentially a segment of text that the model processes. By providing a specific dollar figure, Warsh emphasizes that the Fed now considers token revenue a measurable part of the economy, rather than merely a niche tech statistic.
3. AI Recognized as a Factor of Production
Warsh indicated that the Federal Reserve is closely monitoring the market dynamics within the AI sphere. He stated, "We acknowledge that AI is a new variable—potentially a new factor of production—that will impact both the economy and monetary policy."
Factors of production typically include labor, capital, and land. By categorizing AI in this way, it signifies that the Fed views token consumption as something that could influence the economy's production capacity without causing inflation—essential for determining appropriate interest rates. Misjudging this could lead to errors in monetary policy.
Myriad: When will OpenAI launch GPT-6? Make your prediction.Warsh posed critical questions regarding AI's potential to drive a meaningful and sustained increase in productivity. He also questioned whether token usage would complement or compete with labor, to which he suggested that neither answer is clear. A Fed task force is currently investigating these issues, but he candidly noted that their findings would not influence immediate policy decisions.
This acknowledgment reveals a significant gap: the Fed recognizes AI's macroeconomic relevance but admits to lacking a comprehensive framework for it, while still making rate decisions based on models predating this spending surge. Bill Gates has further explored the labor aspect of this discussion, advocating for a robot tax and protections for jobs that cannot be automated.
4. Uncertainty Surrounds Value Distribution
Warsh did not claim to have definitive answers regarding the implications of AI. He expressed uncertainty about whether AI would truly enhance global productivity and when such changes might manifest.
He also pondered whether the economic benefits would be distributed globally or within specific sectors. "How much of the surplus will go to owners of scarce resources—AI labs, chip manufacturers, energy producers, and cloud service providers?" he questioned.
A Fed task force is examining these topics, but he reiterated that its conclusions would not impact current policy decisions. Warsh's inquiry regarding the distribution of value gained a real-world context just days before his speech, as Nvidia reported record revenues of $96.2 billion and announced $366 billion in future AI infrastructure investments, alongside plans to acquire Hugging Face for approximately $12.9 billion. Conversely, a previous report indicated that 95% of generative AI companies are struggling to succeed.
This suggests that value may be consolidating rather than being distributed evenly. If a single chipmaker can capture such a substantial portion of the surplus generated by AI advancements, Warsh's questions about market structure appear to be finding concrete answers.