Evaluating the business model of LLMs

The rapid growth and massive valuations of Large Language Model (LLM) companies like OpenAI have sparked intense debate about their long-term business viability, despite their technological impact.

Market dynamics and historical parallels: The LLM industry shows concerning similarities to historically unprofitable sectors like airlines, where technological innovation didn’t translate to sustainable business success.

• Much like the airline industry of the 1960s, LLMs represent cutting-edge technology but face severe structural challenges to profitability
• The airline industry demonstrates how even essential services can struggle financially due to unfavorable market conditions
• In contrast, seemingly simple businesses like Coca-Cola consistently achieve high profitability due to favorable industry structure

Critical industry structure analysis: The five forces framework reveals significant challenges for LLM companies that could limit their long-term profitability.

• NVIDIA holds unprecedented supplier power as the sole provider of essential training chips
• Buyers show minimal brand loyalty, readily switching between different LLM providers
• Direct competition is intense, with multiple vendors offering similar services
• Low barriers to entry allow new competitors to emerge regularly
• Alternative solutions, including human workers and metadata-based approaches, provide viable substitutes

Investment paradox: Despite structural challenges, LLM companies continue to attract massive investment, raising questions about their strategy and investor expectations.

• OpenAI’s recent $6.6 billion funding round at a $157 billion valuation stands as one of the largest VC investments ever
• Potential strategies like developing proprietary chips or building brand loyalty face significant obstacles
• The WeWork cautionary tale shows how even well-funded companies can fail when fundamental business models are flawed

Alternative paths to AI success: While LLM development faces profitability challenges, other AI-related business models show more promise.

• Companies leveraging existing models rather than building their own may find more sustainable paths to profitability
• The success of the underlying technology doesn’t necessarily correlate with the financial success of its creators
• Software companies with unique products and limited supplier dependence typically achieve better margins than hardware-dependent businesses

The NVIDIA factor: The overwhelming dependence on NVIDIA’s hardware creates a fundamental weakness in the LLM business model that could prove difficult to overcome.

• NVIDIA’s monopolistic position in AI chips gives it extraordinary pricing power
• Unlike traditional software companies, LLM developers face substantial ongoing infrastructure costs
• The relationship mirrors the aerospace industry’s dependence on Boeing and Airbus, which has historically limited airline profitability

Strategic implications: Future success in the AI industry may depend more on how companies position themselves within the value chain rather than on direct model development.

• The most profitable opportunities may lie in applications built on top of existing LLMs rather than in model development itself
• Companies should carefully evaluate their position in the industry structure before committing to LLM development
• The long-term winners in AI might actually be the infrastructure providers rather than the model developers themselves