Anthropic's Q3 Results Are In — The Costs Are Exploding

SUMMARY

Ed Elson, joined by John Foley, Deirdre Bosa, and Mark Zandi, examines Anthropic's amended S-1 filing, the deepening tech sell-off, and the economic landscape ahead of the U.S. midterms. The discussion covers Anthropic’s explosive growth and spending, the challenges facing tech giants amid rising debt, and the widening gap between the AI-driven stock market and the broader American economy.

MAIN POINTS

• Anthropic's Q3 numbers reveal $17 billion in revenue and nearly $900 billion in future compute commitments, raising concerns about sustainability. [1:01]
• Elon Musk's comments on Grokbot's unsustainable costs highlight industry-wide challenges with AI profitability. [6:00]
• Anthropic's compute costs are rising faster than revenue, with quarter-on-quarter compute expenses up nearly 80%. [9:01]
• Blackstone expects Anthropic to receive a high junk credit rating post-IPO, complicating its fundraising prospects. [12:00]
• Tech giants like Amazon, Google, and Meta are experiencing negative free cash flow and increasing reliance on debt to fund AI buildouts. [16:01]
• Wall Street consensus projects a return to strong free cash flow for hyperscalers by 2030, despite current financial pressures. [21:00]
• Mark Zandi explains the divergence between the AI-driven stock market and the real economy, with most Americans not benefiting from AI gains. [27:02]
• Zandi grades the U.S. economy as a B overall, noting an A++ for the AI sector but a D for the broader population, and predicts potential political shifts in the midterms. [35:01]

DETAILED ANALYSIS

Anthropic’s amended S-1 filing has brought unprecedented transparency to its financial outlook, revealing both extraordinary revenue growth and daunting future obligations. The company reported $17 billion in revenue for the third quarter, representing a twelvefold increase year-over-year, but also acknowledged a slowdown in quarterly growth from 144% to 47%. More striking is the disclosure of nearly $900 billion in compute and data center commitments over the coming years, with $70 billion due within the next twelve months—an amount that exceeds current annualized revenue.

This scale of expenditure underscores the immense capital intensity of the generative AI sector and raises questions about long-term sustainability, especially as price competition intensifies due to new entrants and discounted offerings, particularly from Chinese open-weight models.

John Foley of the Financial Times highlights the inherent unpredictability in projecting Anthropic’s future revenue, with internal guidance pointing to a potential $200 billion in annual revenue by 2028. If achieved, this would justify a valuation in the $1.4 trillion range, comparable to other high-growth tech firms like SpaceX. However, the risks are significant: growth rates are decelerating, and the company is already offering discounts to maintain competitiveness.

The amended filing also introduces new risk factors, including the possibility of reaching a threshold where AI systems begin to self-improve rapidly, necessitating operational slowdowns for safety—an event that could disrupt revenue streams in unpredictable ways.

Industry-wide, profitability remains elusive. Elon Musk’s public remarks about Grokbot, SpaceX’s AI chatbot, encapsulate the sector’s challenges: the cost of serving advanced AI products often exceeds revenue, forcing even well-capitalized firms to scale back or reconsider their strategies. Anthropic’s own compute costs have surged from $3.4 billion in early 2026 to $9.9 billion in the most recent quarter, outpacing revenue growth.

While operating losses had narrowed significantly compared to the previous year, the latest quarter saw a return to negative territory, exacerbated by non-cash losses tied to complex financial instruments. The path to profitability is further complicated by the expectation that Anthropic will receive a high junk credit rating after its IPO, as communicated by Blackstone. This rating will increase borrowing costs and limit access to capital, forcing the company to rely more heavily on equity raises or expensive debt to fund ongoing operations.

The challenges facing Anthropic are mirrored across the technology sector. Deirdre Bosa notes that the recent tech sell-off has erased $5 trillion in market value from leading AI and chip stocks since early October, pushing the Nasdaq 100 into correction territory. Major firms like Amazon, Google, and Meta are now reporting negative free cash flow, a stark departure from their historically robust balance sheets.

The financing of AI infrastructure is increasingly reliant on debt, with AI-linked debt issuance projected to reach $1 trillion next year, up from $655 billion this year. Rising interest rates and credit rating downgrades, such as the S&P’s recent move on Meta, are compounding the risks. Even Microsoft, the only hyperscaler currently maintaining positive free cash flow, has tapped the bond market for $25 billion, signaling sector-wide caution.

Despite these pressures, Wall Street consensus remains optimistic about the long-term prospects of the hyperscalers. Projections suggest that by 2030, Amazon, Google, Meta, and Microsoft could collectively generate over $600 billion in free cash flow, implying that current financial strains may be temporary. However, this outlook is contingent on the realization of aggressive growth forecasts and continued demand for AI services.

Sentiment in San Francisco, according to Bosa, remains bullish, especially among industry leaders like Nvidia, which continues to benefit from surging demand for GPUs. Nevertheless, the fate of smaller cloud providers and newer entrants—the so-called neo-clouds—remains uncertain, as they may struggle to secure the capital needed to compete at scale.

Turning to the broader economy, Mark Zandi of Moody’s Analytics emphasizes the growing disconnect between the AI-driven stock market and the lived experience of most Americans. While the total value of U.S. publicly traded stocks has doubled in four years, largely due to AI, the majority of Americans do not own significant stock holdings or even their own homes. Real wages have stagnated or declined as inflation and interest rates have risen, eroding purchasing power and fueling widespread economic dissatisfaction.

Zandi attributes much of the current economic malaise to policy decisions, particularly the war in Iran, tariffs, and restrictive immigration measures, all of which have contributed to higher prices and slower growth. He grades the overall economy as a B, but notes that this average masks a stark divide: the AI sector is thriving (A++), while the broader population faces deteriorating conditions (D). This polarization is likely to influence the outcome of the midterm elections, with economic grievances driving potential shifts in political power.

In summary, Anthropic’s Q3 results and the broader tech sector’s financial dynamics illustrate both the transformative potential and the acute risks of the current AI investment cycle. The sector’s future hinges on the ability to convert massive capital outlays into sustainable, profitable growth amid intensifying competition, rising costs, and uncertain macroeconomic conditions.

LINKS

• Prof G Markets YouTube channel
• Prof G Markets newsletter subscription
• Order 'Notes On Being A Man'
• Subscribe to No Mercy / No Malice
• Prof G Markets Instagram
• Scott Galloway Instagram
• Ed Elson Instagram
• Ed Elson X (Twitter)
• Ed Elson Substack
• HIMS weight loss and health services

KEYWORDS