Nansen CEO Alex Svanevik commented on the competitive dynamics around AI models, according to a June 14 report from TechFlow. He said open-weights models could become a stronger source of pressure on Anthropic and OpenAI in the future. His point centered on the relationship between model capability, task requirements and cost, rather than on whether every use case needs the most advanced frontier model available.
Not every task requires a “150 IQ-level” model
Svanevik said that not all tasks require a “150 IQ-level” frontier model. In many scenarios, he said, a model with roughly 115 IQ-level capability that costs about 90% less is already “completely sufficient.” Under that framing, the competitive advantage of open-weights models comes from cost efficiency: they do not need to match the highest-end systems in every respect if they can complete many tasks at a far lower cost.
This comparison places pressure on the commercial logic of companies that rely on premium frontier models. If many customers or use cases can be served by cheaper models with adequate capability, the revenue expectations attached to the most advanced models face a more difficult test. Svanevik’s comments specifically named Anthropic and OpenAI as companies that could face greater competition from open-weights models.
Regulation adds another variable to the business model
The AI industry has widely treated the most advanced frontier models as the likely source of future profits. Svanevik’s view challenges that assumption by emphasizing that capability alone may not determine commercial value. Cost, deployment context and actual task needs also shape whether a high-end model is the right product for a given situation.
He also pointed to government restrictions as an important factor. If governments restrict frontier models or block access to them, the revenue expectations of related companies could be affected. When regulation limits the capability or deployment of frontier models, the business path of earning profits from high-end models has to be reassessed by the industry. In Svanevik’s remarks, the cost advantage of open-weights models and the regulatory risk facing frontier models are both key variables in that reassessment.