The artificial intelligence sector is experiencing significant volatility, with some companies seeing 20-30% drops and 10% daily fluctuations. This contrasts with the relative stability of Bitcoin and other cryptocurrencies, which have held up better than expected. A key focus for the market is the potential Initial Public Offering (IPO) of Anthropic, a major AI company.

Anthropic's IPO and Market Concerns

Anthropic is reportedly preparing for an IPO, with preliminary discussions and a leaked quarterly Annual Recurring Revenue (ARR) figure of approximately $16.3 million. This news has impacted related companies, with memory stocks and cloud providers experiencing declines. Concerns are mounting that open-source AI models are gaining market share at the expense of closed-source models, potentially affecting revenue for companies like Anthropic and OpenAI.

Jensen Huang, CEO of Nvidia, has projected significant capital expenditure in AI, estimating it could reach $4 trillion by 2030. However, a primary concern is the potential slowdown in growth for leading AI models and the increasing adoption of open-source alternatives. This shift could impact the financial commitments of hyperscalers like CoreWeave, which rely heavily on large clients such as Anthropic and OpenAI.

Hyperscalers and Debt Markets

The financial health of hyperscalers is under scrutiny. There are growing concerns about the unit economics, monetization, and return on investment for these companies. The reliance on commitments from major AI players like Anthropic and OpenAI raises questions about their sustainability if these clients experience a slowdown.

This concern is reflected in the bond markets. Bond yields have increased, indicating market apprehension. New debt issuances from hyperscalers are now being priced at significantly higher rates, with some as high as 13-15% with warrant coverage, compared to 4-5% a year ago. Credit Default Swaps (CDS) for companies like CoreWeave have suggested a substantial probability of default in the coming years.

The situation suggests that many hyperscalers may be struggling to finance their growth initiatives while simultaneously servicing existing debt and interest payments without raising additional capital. The market appears saturated with new debt and equity offerings, potentially forcing these companies into painful restructurings.

Unit Economics and Open-Source AI

The unit economics of AI companies are a significant point of discussion. Reports suggest that some AI companies, including Anthropic, have historically spent a disproportionate amount on computing relative to their revenue. While some have reported positive free cash flow, the language used around these figures has been cautious, with adjustments for various factors.

The rise of open-source AI models presents a challenge to the closed-source model's market share. This dynamic creates a race for companies like OpenAI and Anthropic to continuously invest in developing new models, despite potential revenue pressures and market share erosion. The economic viability of this continuous investment, especially when open-source alternatives are becoming more competitive, is a key question.

The bankruptcy of Character AI has highlighted these challenges. Even after shifting to lower-cost, open-source models, the company reportedly spent $9 for every$1 of revenue. This suggests that many AI startups may face similar economic difficulties, even with price reductions in AI services.

The Role of APIs and Enterprise Clients

While consumer-facing AI subscriptions may not be the primary revenue driver, companies like Anthropic and OpenAI are reportedly generating substantial revenue through their APIs, serving enterprise clients. These API services are estimated to have high gross margins, potentially between 80-90%. Subscriptions may be used as a market penetration strategy, similar to Slack's freemium model, to drive adoption and eventual conversion to lucrative API business.

However, questions remain about customer loyalty and the potential for clients to switch to open-source models if they offer comparable performance at a lower cost. The long-term sustainability of this model depends on the perceived value and stickiness of these AI services.

Market Outlook and Investment Strategy

The market is expected to remain volatile for AI stocks until companies release their S-1 filings, providing clearer financial insights. The potential for Anthropic to achieve positive cash flow is a significant factor for investors to consider. Until then, volatility in AI stocks is likely to exceed that of cryptocurrencies. Investors are advised to avoid excessive leverage and prepare for significant daily price movements.

There is a consensus that the massive capital expenditure in AI infrastructure is unprecedented. However, a significant "duration mismatch" exists, where substantial capital is required to meet demand, but the available capital, largely in the form of debt, may not be sufficient or sustainable in the long term. This could lead to failures in the interim, even if the long-term AI thesis holds.

Tokenization and Institutional Interest in Crypto

At the Token2049 event, a notable trend was increased institutional interest, with fewer retail participants than usual. The focus was on infrastructure and institutional use cases, particularly stablecoins and tokenization. Robinhood's entry into tokenization and its implications for other companies building ecosystems were discussed.

Pearl emerged as a significant talking point, with some analysts suggesting it could be to Zcash what Zcash was to Bitcoin. Pearl is a Layer 1 blockchain derived from Bitcoin, incorporating matrix multiplication (used in AI inference) into its proof-of-work mechanism. This "useful proof-of-work" aims to incentivize miners to perform AI inference tasks while securing the blockchain, potentially addressing concerns about Bitcoin's security budget as miners shift to AI-related ventures.

The narrative around Pearl suggests that Bitcoin's security budget may be at risk as miners are drawn to more lucrative AI workloads. Pearl's model aims to create synergy between AI inference and blockchain security, offering a potentially more economically valuable form of proof-of-work.

Other Crypto Developments

Other notable discussions at Token2049 included:

• Backpack: Highlighted for its shared liquidity layer, aiming to create a global liquidity pool across different regulatory jurisdictions.
• Hyperliquid: Gained attention for its focus on options and their complementarity to perpetual futures, with renewed interest in on-chain options protocols.
• Tokenized Securities: Significant growth was observed in tokenized securities, with a substantial increase in wallet holders and trading volume, particularly outside of US market hours. This suggests a strong demand for tokenized assets as an access story for non-US investors.
• Nubank and Stablecoins: David Vélez, CEO of Nubank, discussed the potential for dollar-denominated stablecoins to become a significant global financial tool, democratizing access to the US dollar and potentially replacing less stable local currencies. Nubank's own crypto wallet has seen substantial adoption, with stablecoins forming the majority of its volume.

The Future of Currencies and Decentralization

The discussion touched upon the potential decline in the number of global currencies, with a focus on the stability of the US dollar and the increasing adoption of dollar-denominated stablecoins. While national sovereignty and monetary policy are factors, the ease of access to more stable currencies like the US dollar through stablecoins could lead to a consolidation of global currencies.

The concept of "Ponzi schemes" was raised in relation to national currencies, suggesting that governments may maintain their own currencies to manage unbalanced budgets and print money. However, the increasing accessibility of superior financial products like stablecoins could challenge this model. The role of state enforcement and violence in maintaining national currencies was also acknowledged.

The internet has facilitated the free flow of information, and blockchain technology is seen as enabling a similar free flow of money. While some countries may attempt to block this, the broader trend suggests a breakdown of global financial barriers.

Content Recommendations

• Book: "Genghis Khan and the Making of the Modern World" and its sequel, "The Conquerors: How Two Hundred Years of War Transformed the World" by Jack Weatherford.
• Documentary: "You Can See Everything," a documentary about Elizabeth Holmes and Theranos, noted for its potential to be a significant social conversation piece.
• Podcast Episode: An upcoming episode of Empire featuring Michael Dempsey of Compound.