7 Stocks to Solve AI's Data Transfer Bottleneck
As AI focuses on power and memory, a new bottleneck emerges: data transfer speed between chips. A report highlights 7 stocks that could provide the solution.
While the AI industry scrambles to secure electricity, land, and high-bandwidth memory (HBM), another critical constraint is emerging: moving data between chips fast enough. According to a report from 24/7 Wall St., 7 stocks are positioned to solve this bottleneck.
Details
The new bottleneck is the speed of interconnects between AI processors in data centers. As large language models grow, the need for high bandwidth between chips becomes crucial. The report highlights companies developing interconnect technologies, including NVIDIA (NVDA) and Marvell Technology (MRVL), along with others specializing in switches and optical cables.
Context
These challenges come as the AI industry experiences massive growth, with companies investing billions in new data centers. Data transfer issues are not new but are exacerbated by increasing model sizes and chip counts. Proposed solutions include NVIDIA's NVLink and silicon photonics technologies.
What It Means for Investors
Investors seeking opportunities in the AI supply chain may find these stocks offer exposure to solving one of the most pressing bottlenecks. However, these technologies are still in development and may face technical and regulatory hurdles.
Frequently Asked Questions
Found this useful? Share it