Cathy Wood: Open AI code is not a threat, but a catalyst for superprofits for OpenAI and Anthropic
Contrary to the dominant narrative in the industry, ARK Invest CEO Cathie Wood argues that the proliferation of open models with public weights does not undermine the positions of market leaders, but rather creates new revenue streams for them. In her logic, the more powerful free alternatives become, the higher the demand for premium protection that only closed developments can offer.
Wood is flipping conventional investor logic on its head. It is widely believed that the progress of Meta (recognized as extremist and banned in Russia), Mistral, and DeepSeek will inevitably lead to a collapse in subscription models from OpenAI and Anthropic. However, the head of ARK sees a different pattern here: the growth of open weights expands the attack surface for cyber threats, meaning businesses will be forced to more actively purchase the most advanced paid AI solutions for their own security. Paradoxically, the threat from open-source becomes the main commercial argument for closed ecosystems.
Escalation of Cyber Threats as a Demand Driver
Wood states outright: the stronger the open model, the more compelling the case for expensive protection based on cutting-edge AI. Research from the UK AI Safety Institute confirms this alarming dynamic—in terms of cyberattack capabilities, leading open models are already comparable to developments from market leaders from four to seven months ago. This is not just a theoretical discussion: in the latest ARK AI Primetime report, the rise in AI infrastructure costs is called a decade-long trend, and it is security that becomes the key catalyst for these investments.
Beneficiaries of the New Reality
Wood predicts that OpenAI, Anthropic, and, in the future, SpaceXAI will capture the majority of revenue in the AI sector. ARK Invest has already invested in all three entities. Notably, Anthropic has filed an S-1 application for a listing, with its valuation approaching $1 trillion, while OpenAI could go public as early as September 2026.
"Paradoxically, open models are becoming the main reason why OpenAI, Anthropic, and, in the future, SpaceXAI will capture the majority of revenue in the AI sector," Wood emphasizes.
My take: This logic looks convincing but requires caution. Demand for protection will indeed grow, but in the long term, the margins of leaders could suffer if regulators begin to restrict the monopolization of security. Nevertheless, over the next 2-3 years, Wood's scenario looks at least justified—the market will pay for reliability, not openness.