(UBS)
Anthropic said it will limit access to its new “Mythos” model, which it claims can identify and exploit new vulnerabilities across major operating systems and web browsers. The company said the model has already found thousands of severe vulnerabilities in existing software, prompting the firm to extend limited access only to a small group of critical technology and infrastructure partners in order to patch up security problems now.
Our view: Without taking any single name views, we think this episode reinforces two points for investors. First, frontier AI capabilities are advancing quickly, which may help stoke continued demand growth too. At the moment, measures of AI revenue, token usage, and compute consumption all point to rising demand. Second, AI is moving beyond a discretionary productivity tool and closer to an embedded operating layer for functions like cybersecurity, where reinforcing software against attacks and detecting vulnerabilities could become increasingly essential rather than optional. Within the big AI labs, we think the most durable investment opportunities are likely to sit with companies that can scale, secure, distribute, and monetize AI tools over time. We continue to favor diversified exposure across the AI value chain and geographies, including exposure to quality AI leaders and ecosystem enablers.
Original report - Inflation, Iran talks to put market optimism to test,10 April 2026.
