Artificial intelligence is entering a new phase in which the technology is no longer defined only by increasingly capable models. The infrastructure supporting those models is becoming just as important.
While companies developing the hardware and software backbone of AI are emerging as some of the stock market’s biggest winners. Anthropic’s Claude is part of another important development.
The company has started watermarking everything it writes, reflecting growing concerns about identifying AI-generated content. As AI-generated text becomes increasingly difficult to distinguish from human writing.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Watermarking could provide a mechanism for tracing or identifying content produced by artificial intelligence. The move highlights a broader challenge facing the AI industry. Generative AI is spreading rapidly across journalism, education, software development, business communication and social media.
While this expansion creates enormous opportunities, it also raises questions about authenticity, attribution and accountability. Watermarking could eventually become one component of a larger ecosystem designed to identify AI-generated material.
Investors are focusing heavily on the physical infrastructure required to power this AI revolution. Twelve companies in the Nasdaq 100 have more than doubled in value this year, and notably, none belongs to the so-called Magnificent Seven. The performance suggests that the market’s AI trade is broadening beyond the largest technology companies.
SanDisk has emerged as the standout performer, gaining approximately 406%. Micron has followed with a rise of about 207%, while Intel has advanced roughly 175%. Other major winners include Arm, Marvell, Western Digital, AMD and Applied Materials.
Their common characteristic is exposure to the infrastructure required to build, operate and expand advanced computing systems. This distinction is important. The first wave of AI enthusiasm concentrated on companies developing consumer-facing applications and the models themselves.
The latest phase is increasingly focused on the companies supplying the chips, memory, networking equipment, manufacturing technology and other components necessary to operate massive AI data centers.
Memory manufacturers, for example, are benefiting from the enormous amounts of high-bandwidth memory and storage required by modern AI systems. Semiconductor designers and equipment manufacturers are also positioned to benefit as hyperscalers and other technology companies continue expanding their computing capacity.
The remarkable stock performances therefore reflect more than speculative enthusiasm. They demonstrate how AI spending is filtering through the technology supply chain.
Every new AI model requires computing power, and that computing power depends on a complex network of semiconductor manufacturers, equipment suppliers and infrastructure providers.
However, investors should recognize that extraordinary gains create elevated expectations. A stock that has risen several hundred percent can become vulnerable to profit-taking, valuation concerns and any slowdown in AI capital expenditure.
The sustainability of these gains will ultimately depend on whether demand for AI infrastructure continues growing at the pace markets currently anticipate. Claude’s watermarking initiative and the surge in AI infrastructure stocks illustrate two sides of the same transformation.
One concerns how society will identify and manage AI-generated information; the other concerns the enormous industrial ecosystem being built to make AI possible. The AI revolution is therefore moving beyond chatbots and headline-grabbing model launches.
It is becoming a fundamental technology cycle, reshaping both the digital information landscape and the semiconductor industry. The biggest opportunities may increasingly lie not only with the companies building AI, but with those supplying the machines that make it possible.