Jesse Felder: The AI ‘Bubble of All Bubbles’ Is About to Hit Reality

Key Concepts

• AI Bubble: The central theme of the discussion, characterized by extreme speculation in the technology sector, particularly around Artificial Intelligence.
• Frankenstein of Bubbles: A metaphor used to describe the current market situation, combining elements of past bubbles like the dot-com and housing bubbles.
• Narrative-Driven Speculation: The role of compelling stories and narratives in driving investment decisions, often detached from fundamental value.
• Retail and Institutional Investors: Both individual and large-scale investors are participating in the speculative frenzy.
• NVIDIA Earnings: Seen as a crucial "bellwether" for the tech sector and the AI bubble debate.
• Hyperscalers: Large technology companies that provide cloud computing services (e.g., Microsoft, Google, Amazon).
• OpenAI and Anthropic: Key private companies at the forefront of generative AI development, experiencing significant losses.
• Compute Costs: The expense associated with processing power and data centers, a critical factor in AI business models.
• Overbuilding: The potential for excessive investment in infrastructure (data centers) leading to a surplus of supply.
• Capital-Intensive Business Models: AI businesses require significant investment in hardware and infrastructure, unlike previous "capital-light" tech models.
• Network Effects: The phenomenon where a product or service becomes more valuable as more people use it; absent in many current AI models.
• Financialization of the Economy: The increasing dominance of financial markets and instruments in the overall economy.
• K-Shaped Economy: A metaphor describing an economy where different segments experience vastly different outcomes, with the wealthy benefiting disproportionately.
• Labor Share of Income vs. Corporate Profit Margins: The inverse relationship between the portion of national income going to workers and the profits earned by corporations.
• Commodity Supercycle: A prolonged period of rising commodity prices driven by strong demand and limited supply.
• Real Assets vs. Financial Assets: A shift in investment focus from stocks and bonds to tangible assets like commodities.
• Insider Buying: A signal of confidence in a company's prospects by its own executives and directors.

The AI Bubble: A "Frankenstein of Bubbles"

Jesse Fedler of The Felder Report argues that the current market situation, particularly within the technology sector, can be described as a "Frankenstein of bubbles." This analogy combines elements from past speculative manias, such as the dot-com bubble and the housing bubble, incorporating both financial and technological components. A powerful narrative is driving extreme speculation among both retail and institutional investors, creating what Fedler calls the "bubble of all bubbles."

Lack of Fundamental Underpinning

A key argument is the absence of strong fundamental support for the current valuations. Fedler states that the revenues needed to justify the massive AI spending are simply not present. He highlights OpenAI's ambition for explosive revenue growth to support an estimated $1.5 trillion in data center spending, yet notes that no viable business model has yet been identified to achieve this. This lack of a clear path to profitability is a significant concern.

The "Worst Kept Secret"

Interestingly, Fedler observes that there's a widespread awareness of the potential bubble, leading to a situation where "everybody seems to know that it's a bubble." This collective acknowledgment, however, hasn't curbed the speculative fervor.

The Role of OpenAI and Anthropic in Inflating Big Tech Earnings

A critical factor fueling the current tech rally, according to Fedler, is the financial performance of major hyperscalers being artificially boosted by the losses of private AI companies like OpenAI and Anthropic.

Losses Masking Weakness

Fedler cites a Wall Street Journal article by James Macintosh, which points out that "100% of the earnings growth in these hyperscalers has come from the losses at OpenAI and Anthropic." Specifically, OpenAI incurred a $12 billion loss in the third quarter, which accounted for approximately 65% of the total earnings growth for the "Magnificent 7" companies. Anthropic added another $8 billion in losses. These substantial losses from private entities are being funneled into big tech companies for data center development and hyperscaling services, thereby inflating the reported earnings of these larger corporations in the short term.

OpenAI's Unsustainable Business Model

The core issue with OpenAI's model, as described by Fedler, is that "the cost of serving every query is greater than the revenue generated by every query." This "lose money on every transaction, make it up in volume" approach is inherently unsustainable, leading to exponentially growing losses.

Sam Altman's Strategic Interest in a Compute Cost Crash

Fedler posits a "devious plan" where Sam Altman, as a major buyer of compute power, has a vested interest in seeing compute costs plummet. He argues that the most effective way for OpenAI to survive is through an "overbuilding and then a crash" in compute costs. By encouraging massive investment in data centers, Altman aims to create an oversupply that will drive down prices dramatically, potentially making OpenAI's business model profitable. This scenario, while beneficial for OpenAI in the long run, is detrimental to the hyperscalers who are building this infrastructure.

SoftBank's Financial Strain

The financial strain of funding OpenAI's losses is evident. SoftBank, a key investor, has reportedly sold its Nvidia stake and is leveraging its portfolio through margin loans to meet its commitments to OpenAI. This highlights the difficulty in continuously securing funding for escalating losses.

The Mechanics of a Bubble and its Potential Beneficiaries

The discussion delves into the historical patterns of bubbles and who might emerge as winners.

Historical Parallels: Fiber Optics and Railroads

Fedler draws parallels to past bubbles, such as the fiber optic buildout where an oversupply of dark fiber drove down costs, enabling new internet-based business models. Similarly, historical railroad bubbles also involved massive infrastructure development followed by consolidation and cost reductions.

The "Benefit" of Overbuilding

Even figures like Mark Zuckerberg acknowledge that overbuilding infrastructure can be beneficial in the long run, as it leads to lower costs for users. However, this comes at the expense of the companies that build the infrastructure, many of whom may go bankrupt.

Unpredictable Winners

The lesson from past bubbles, like the dot-com era, is that it's difficult to predict future winners. While the internet revolution led to the rise of companies like Meta, their emergence was not foreseen during the initial speculative phase.

The Shift from Capital-Light to Capital-Intensive AI Models

Fedler contrasts the highly profitable, "capital-light" business models of big tech over the past 15 years (e.g., Meta, Microsoft) with the current AI landscape. AI business models are inherently "hugely capital-intensive" and lack the economies of scale and network effects that drove previous tech successes. Unlike platforms where incremental users generate near-100% profit, generative AI incurs significant incremental costs per user. Furthermore, the absence of strong network effects means users can easily switch between competing AI platforms, preventing the creation of "walled gardens" and monopolies.

Berkshire Hathaway's Investment in Google and Market Sentiment

The conversation touches upon Berkshire Hathaway's recent stake in Google (Alphabet).

Google's Cash Flow Strength

Fedler acknowledges that Google possesses strong cash flows, differentiating it from companies like Meta, which he suggests are currently spending more than their free cash flow net of stock-based compensation.

A "Modified Index Fund" Approach

He characterizes Berkshire Hathaway's position in Google as a "tiny position" and part of a broader strategy to create a "modified index fund" portfolio, focusing on "cash cows" and avoiding higher-risk names.

Net Seller Stance

Despite this investment, Fedler emphasizes that Berkshire Hathaway has been a net seller of assets for eight consecutive quarters, building its largest cash position relative to total assets in its history. This indicates a cautious, rather than bullish, market outlook from Warren Buffett's firm.

The Intertwined Nature of AI and the Financial System

The discussion explores whether the AI sector has become "too big to fail" and its potential impact on the broader economy.

Not "Too Big to Fail" in the Traditional Sense

Fedler does not believe AI is "too big to fail" in the same way the banking system was during the 2008 financial crisis, where a failure could cascade through interconnected liabilities.

Increased Stock Market Importance

However, he argues that the stock market's significance relative to GDP and household wealth has never been greater. Investors across all demographics are heavily invested in the stock market, making an AI bust potentially more damaging than past sector-specific downturns.

Exacerbating the Fiscal Deficit

A significant AI bust could also negatively impact government finances by reducing capital gains tax revenue, thereby exacerbating an already problematic fiscal deficit.

The Tech Revolution vs. Short-Term Hype

Fedler clarifies his stance on technology, distinguishing between long-term revolutionary potential and short-term overhyping.

Long-Term Integration of Technology

He believes in the long-term transformative power of technology, citing the internet as an example. It took over a decade for companies to effectively integrate the internet into their businesses and develop viable models. Similarly, AI's true impact will likely unfold over 10-15 years.

Current Stage: Early and Overhyped

Currently, Fedler sees the AI sector as being in its "very early stages" with significant overhyping. Businesses are struggling to effectively implement AI, with some finding it "functionally useless" despite significant investment. This contrasts with the hype around mass layoffs and AI replacing human workers.

The Unlikely Scenario of a Slow Bubble Deflation

Fedler argues that a slow, orderly deflation of the AI bubble is unlikely due to the high levels of leverage in the market.

Unprecedented Leverage

He points to record-high margin debt, record-low cash accounts among investors, and massive speculation in options markets, particularly call buying. This leverage creates a significant risk when the bubble begins to burst.

The Impact of Leverage on Market Dynamics

When speculators bet on a bubble bursting, they may attempt to unwind leverage, leading to margin calls and liquidations. The massive call buying has created a bid in the market by forcing dealers to buy underlying securities. A shift to put buying could force dealers to sell equities, potentially triggering a sharp downturn.

The Speed of Market Movements

The speed at which market movements can occur today, facilitated by mobile technology, puts immense pressure on policymakers to intervene in a crisis.

The Real Economy and the Risk of Recession

Fedler paints a grim picture of the underlying real economy, suggesting that AI spending has been a key factor propping it up.

Weak Underlying Economy

He believes the underlying economy is "pretty poor," with rising unemployment and limited job creation outside of healthcare and education. A falter in AI spending could quickly lead to a recession.

The Commodity Market's Message: Inflation, Not Deflation

Contrary to expectations of deflation following a bubble burst, Fedler believes the commodity market signals a resurgence in inflation.

Gold as a Leading Indicator

The strong performance of gold, a leading indicator for commodities, suggests that deflation is not on the horizon. Copper prices are also strengthening, and natural gas prices are nearing multi-year highs.

The Fed's Role and Fiscal Deficits

Fedler anticipates that the Federal Reserve may need to re-engage in quantitative easing to support the Treasury market, especially given the massive fiscal deficits. This monetary stimulus, combined with the existing deficit, makes a deflationary outcome unlikely.

Political Ramifications and the Shift in Investor Sentiment

The discussion highlights the growing political implications of economic inequality and the potential shift in investor focus.

Affordability as a Political Driver

Recent elections in the US have shown affordability to be a primary concern for voters, suggesting that policymakers may be constrained by public sentiment, especially in an inflationary environment.

The "K-Shaped Economy" and Corporate Profit Margins

Fedler reiterates the concept of the K-shaped economy, where the ultra-wealthy are increasingly detached from the struggles of the majority. He links corporate profit margins to labor's share of income, suggesting that a narrowing of the "K" would necessitate a reduction in profit margins to increase labor's share.

Gen Z's Growing Influence

He identifies Gen Z as a potential source of political pressure for policies that benefit their interests, which could come at the expense of corporate profit margins. The persistent low labor share of income and high youth unemployment are seen as dangerous indicators.

The Decline of Aspirational Asset Ownership

The traditional model of individuals aspiring to be asset owners through programs like 401(k)s is eroding with the rise of the gig economy and contractual work. This lack of "skin in the game" for a growing segment of the population fuels a desire for change.

Structural Inflationary Pressures

Fedler argues that the trend of rising corporate profit margins and falling labor share has been a driver of disinflation for decades. A reversal of this trend, coupled with demographic shifts leading to a smaller workforce supporting a larger population, will likely create structurally inflationary pressures.

Investable Opportunities in a Risky Environment

Despite the pervasive risks, Fedler identifies potential investment opportunities.

A Shift to Real Assets

He believes there's a paradigm shift from investing in financial assets (stocks and bonds) to owning "real assets."

Energy Stocks as an Attractive Opportunity

Energy stocks are highlighted as particularly attractive, trading at low valuations (6-7 times earnings) with strong cash flow. The demand for energy is soaring due to AI and LNG exports, while investment in new oil and gas development has declined significantly over the past decade.

Leading Indicators and Insider Buying

Fedler points to the strong performance of gold as a leading indicator for commodities and the potential for oil prices to follow suit. He prefers investing in energy stocks over direct commodity ownership, citing significant insider buying in the sector as a compelling signal.

Underowned Energy Sector

Despite being the best-performing sector over the last five years, the energy sector remains significantly underowned by investors, presenting a unique opportunity.

Critical Minerals: A Cautionary Note

Fedler expresses less conviction in critical minerals due to a lack of compelling insider activity in that space, though he acknowledges it hasn't been a primary focus of his research.

Contrasting Valuations: AI vs. Energy

He contrasts the extremely high valuations in the AI space with the cheap valuations of energy companies, noting significant insider selling in AI-related companies like Coreweave and Amazon, while energy companies are actively buying back stock and experiencing insider purchases.