Lower-cost AI models put pressure on frontier lab pricing

Lower-cost AI models are putting pressure on the pricing assumptions behind OpenAI and Anthropic’s expected public-market valuations, CNBC reported.

Several large companies, including Meta, Shopify, Spotify, and Pinterest, have recently flagged rising AI or inference costs as a business concern. Shopify said economies of scale were partly offset by higher LLM costs.

Enterprise AI spending has increased quickly. CNBC cited CloudZero survey data showing that 45 percent of companies surveyed spent more than $100,000 per month on AI in 2025, up from 20 percent the year before.

Cheaper models are also becoming more capable. CNBC cited Artificial Analysis data showing a wide cost range for comparable workloads across leading models. The report listed Anthropic’s Claude at $4,811, OpenAI’s ChatGPT at $3,357, DeepSeek at $1,071, Kimi at $948, and Zhipu’s GLM at $544 for the same benchmarked workload.

Companies are using routing strategies to control cost. Databricks CEO Ali Ghodsi described an advisor model pattern, where a cheaper model handles most tasks and calls a frontier model only when needed.

Google has also promoted cheaper model options. Sundar Pichai said some companies are already exceeding annual token budgets and said large Google Cloud customers could save more than $1 billion a year by moving most workloads from frontier models to Gemini 3.5 Flash.

Source: https://www.cnbc.com/2026/05/20/cheap-ai-could-derail-openai-and-anthropics-ipos.html