OpenAI and Anthropic are pouring billions into training AI models like GPT-4 and Claude.
But as competition heats up, their path to profitability is becoming increasingly complex.
Why?
Because one of their primary revenue streams — selling API access to large language models (LLMs) — is rapidly becoming a “zero-margin business.”
At least that’s what Aidan Gomez, CEO of rival AI provider Cohere, argued during a recent podcast appearance.
The Bigger Story
We’re witnessing a classic case of commoditization in the LLM market, fueled by cutthroat competition.
At the model level, it’s a race to the bottom on pricing. Companies are slashing API fees — or offering free access — to gain market share and acquire more usage data.
Meanwhile, at the product level, we’re seeing a deep challenge arise: AI-powered chatbots are prone to what Klue’s CTO, Sarathy Naicker, calls “invisible differentiation.”
Despite different architectures and training methods, GPT-4, Claude, and other LLMs offer strikingly similar user experiences: you type a prompt, you get a text response—rinse and repeat.
The result? A landscape where the nuanced differences between AI models are virtually imperceptible to the average user, making product differentiation a challenge.
Why You Should Care
When products offer similar features and users can’t easily tell them apart, brand becomes king.
Here’s the deal:
- Brand Drives Early Adoption: In emerging markets where products evolve rapidly but differentiation is minimal, brand becomes a catalyst for user adoption. This is why OpenAI is ahead.
- Brand as a Quality Shortcut: When users can’t easily spot the differences between products, they face a choice: spend hours testing tools, wade through reviews, or simply trust a well-known brand. Most will take the path of least resistance.
📌Go deeper: Check out last week’s Coffee & Compete podcast to hear Sarathy’s thoughts on “invisible differentiation” in the AI era.