Arista Networks CEO Jayshree Ullal said memory has become the new gold due to the rapid expansion of artificial intelligence [1].
The statement highlights a critical shift in the hardware supply chain. As AI models grow in complexity, the physical infrastructure required to store and process data is becoming a primary bottleneck for global tech scaling.
Speaking during an interview on CNBC’s "Mad Money" program, Ullal said that AI is increasing the demand for memory [1]. She said that this surge in necessity has made memory as valuable as gold [1].
This demand is reflected in the company's financial outlook. Arista Networks expects to reach $11 billion in revenue in 2026 [3]. This figure represents an increase of about $2 billion over the revenue generated in 2025 [3].
Ullal said the company's quarterly results are linked to the broader trend of AI-driven growth [1]. The reliance on high-capacity memory is now a central component of the infrastructure needed to support large-scale AI deployments, a trend that continues to push valuation and demand higher across the sector.
As companies race to build more powerful AI clusters, the ability to secure and scale memory capacity has moved from a standard procurement task to a strategic priority for the industry [1].
“Memory is the new gold.”
The comparison of memory to gold suggests that semiconductor components are no longer just commodities but strategic assets. For the broader tech industry, this indicates that the pace of AI adoption is currently limited by hardware availability, meaning companies that control the memory supply chain will hold significant leverage over the development of next-generation AI.
