Katie Wood: open AI models are not a threat, but a growth catalyst for OpenAI and Anthropic
Contrary to the widespread fear that publishing open weights of AI models undermines the business of leading labs, ARK Invest CEO Cathie Wood offers a radically different view. In her conviction, open source is not a "time bomb" for OpenAI and Anthropic, but a powerful lever that will only strengthen their market dominance.
Wood argues that the main danger of open models lies not in competition at all. When a developer publishes all model weights, anyone can download it, run it on their own hardware, and modify it without restrictions. However, it is precisely this accessibility that creates fertile ground for malicious actors. The more powerful open models from Meta, Mistral, and DeepSeek become, the wider the attack surface for cyberattacks.
This logic flips traditional investor thinking. The rise in threats forces the corporate sector to more actively purchase the most advanced paid AI solutions for protection. Thus, each new open algorithm is not a loss of customers, but an additional argument for businesses to invest in expensive yet reliable protection based on closed models.
The AI Security Arms Race
Research confirms this point of view. Specialists at the UK AI Safety Institute found that in terms of cyberattack capabilities, open models have already come very close to leading proprietary developments, lagging by only four to seven months. This means the threat is not hypothetical, but quite real and rapidly growing. The ARK AI primetime report calls the rise in AI infrastructure spending the main trend of the decade.
Wood directly names OpenAI, Anthropic, and the promising SpaceXAI as the main beneficiaries of this dynamic. Paradoxically, it is precisely open models that will be the reason these companies capture the lion's share of revenue in the AI sphere. Notably, ARK Invest already invests in all three organizations.
Market Prospects and IPO
Against the backdrop of these statements, it is worth noting that Anthropic has already filed a confidential S-1 application for listing, and its valuation is approaching $1 trillion. OpenAI, in turn, could go public as early as September 2026. In Wood's view, both giants are not victims of the open-source revolution, but its main beneficiaries.
My analysis: Wood's thesis looks convincing, especially in the context of growing cyber threats. However, it is important to remember that ARK Invest has a direct financial interest in the growth of these companies. Nevertheless, the fundamental logic is strong: in a world where everyone can arm themselves with AI tools, demand for premium protection will only grow. This creates sustainable "defensive" demand for industry leaders, making their positions even stronger than in the absence of open alternatives.