Nansen CEO Alex Svanevik said on June 14 that open-weights models could place greater competitive pressure on Anthropic and OpenAI in the future. His comments focused on a central tension in the AI sector: the most advanced frontier models may not be necessary for every task, especially when lower-cost alternatives can deliver enough capability for many practical use cases.
Not Every Task Needs a “150 IQ” Frontier Model
Svanevik said the market does not always need a “150 IQ” level frontier model. In many scenarios, he argued, a model with roughly “115 IQ” capability but about 90% lower cost is already “completely sufficient.” That cost-performance balance gives open-weights models a clearer value proposition in situations where users are not optimizing for the highest possible model capability.
The point challenges a widely held assumption in the AI industry: that profits will primarily come from the most advanced frontier models. If many tasks can be handled by models that are cheaper while still capable enough, then the premium-revenue logic behind frontier-model businesses becomes more difficult to sustain across all use cases.
Regulation and Access Limits as Key Business Variables
Svanevik also noted that government restrictions on frontier models, including limits or blocked access, could affect revenue expectations for companies built around such models. In that scenario, model capability is not only a technical question. It also becomes tied to regulatory boundaries and deployment rules that determine what can actually be offered to customers.
For companies such as Anthropic and OpenAI, whose commercial paths rely on high-end model capabilities, the question is whether that path remains intact when regulators begin limiting the capability or deployment of frontier models. Svanevik’s comments frame the issue as a reassessment of AI business models, where the lower cost of open-weights models and the regulatory risks facing frontier models both become decisive factors.