OpenAI・Anthropic・Google・xAI, Why Trillions Flow into Unprofitable AI Businesse

2026-07-19 · Source: AI on Medium · Field: Business & Management — Corporate Strategy & Leadership, Capital Markets & Investment Management, Economic Analysis & Policy · Depth: Intermediate, long

Summary

Major AI companies like OpenAI, Anthropic, Google, and xAI are attracting trillions in investment despite current unprofitability, driven by unique business models and investor expectations. Their cost structures are dominated by model training (e.g., GPT-5 class models exceeding \$500 million), inference (OpenAI projected \$14.1 billion in 2026), and massive infrastructure projects like OpenAI's \$500 billion "Stargate." Revenue streams include individual/enterprise subscriptions, API billing (80% of Anthropic's revenue), cloud provider resale, advertising integration, and government contracts. These firms rely on massive unlisted funding rounds, with OpenAI valued at \$852 billion and Anthropic at \$965 billion, often involving strategic investors. Google, however, funds AI through its existing high-profit search business and proprietary TPUs. Investment is sustained by a "growth story" narrative, circular investments, patient capital, long-term governance, and anticipated IPOs, all predicated on a ferocious decline in inference costs (e.g., 1,000-fold reduction for GPT-4 class quality by 2026) and unlimited demand growth.

Key takeaway

For investors evaluating AI companies, recognize that current valuations are driven by future market share and declining inference costs, not immediate profitability. You should scrutinize the long-term sustainability of their funding mechanisms, such as patient capital and strategic partnerships, and assess their compute efficiency. Be aware that the "declining cost of intelligence" premise is critical; if usage growth outpaces cost reduction, current investment logic could collapse.

Key insights

The AI industry's massive investments are sustained by future growth expectations, declining compute costs, and diverse funding mechanisms despite current unprofitability.

Principles

• AI company costs are dominated by compute, not personnel.
• Inference costs are projected to drop 1,000-fold by 2026.
• Strategic investors prioritize market share over immediate profit.

In practice

• Diversify AI investments across multiple companies.
• Prioritize compute efficiency for long-term viability.
• Explore outcome-based billing for AI services.

Topics

• AI Business Models
• Generative AI Investment
• Inference Cost Reduction
• AI Infrastructure
• Strategic Partnerships
• Public Benefit Corporation

Best for: Investor, Executive, Director of AI/ML

Related on AIssential

• AI Is Too Expensive — AI's current economic model is unsustainable due to massive hyperscaler investments, high operational costs, and unproven profitability of AI labs.
• Leaked financial docs show OpenAI is losing billions of dollars a year — OpenAI's rapid revenue growth is outpaced by massive R&D and operational costs, leading to significant losses despite high valuation.
• Can someone explain to me if Anthropic is about to become profitable or not like I am five? — The debate over Anthropic's profitability highlights how investment capital can obscure true operational earnings in frontier AI.
• The $600 Billion Just Vanished: Corporate’s AI Delusion — The enterprise AI bubble is financially unsustainable due to a massive revenue-to-cost imbalance.
• A Whimper, Not a Bang — The future of AI will be defined by restricted access and concentrated power, not universal capability.
• AI Is Slowing Down — The AI industry's financial model is unsustainable, requiring trillions in revenue by 2030 that current demand cannot support.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI on Medium.