tiprankstipranks
Trending News
More News >
Advertisement
Advertisement

How Nvidia’s (NVDA) Tech Will Turn Data Centers into a Huge GPU

Story Highlights

Many data centers, especially older ones, face power limits that restrict how much computing power they can support.

How Nvidia’s (NVDA) Tech Will Turn Data Centers into a Huge GPU

Tech companies are rushing to meet the growing demand for AI by building massive data centers to power GPUs and servers. However, many of these data centers, especially older ones, face power limits that restrict how much computing power they can support. As a result, chipmaker Nvidia (NVDA) has introduced a solution to this challenge with its new Spectrum-XGS network switches. Indeed, rather than adding more physical power to a single site, Spectrum-XGS combines the computing power of multiple data centers. This essentially turns them into a huge GPU that can handle more demanding AI tasks.

Elevate Your Investing Strategy:

  • Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence.

This innovation builds on Nvidia’s existing Spectrum-X technology, which already allows GPU servers within a single data center to work together. The new “GS” version, short for “gigascale,” goes even further by allowing several physically separate data centers to work as one. Interestingly, Nvidia clarified that Spectrum-XGS isn’t brand-new hardware. Instead, it uses the same hardware with improved algorithms that make it possible to transfer data efficiently over longer distances between data centers.

In addition, Nvidia’s Director of Accelerated Computing Products, Dave Salvator, explained that many data centers are “power-capped,” which means that they can’t draw more electricity beyond a certain point. Nevertheless, by linking data centers together, companies can bypass this limit and increase their AI capabilities. Over time, this approach is expected to help developers create even more advanced AI tools. It is worth noting that Nvidia’s announcement also comes just ahead of its upcoming earnings report, which could give investors a clearer picture of how the company is capitalizing on the AI boom.

What Is a Good Price for NVDA?

Turning to Wall Street, analysts have a Strong Buy consensus rating on NVDA stock based on 35 Buys, three Holds, and one Sell assigned in the past three months, as indicated by the graphic below. Furthermore, the average NVDA price target of $198.57 per share implies 11.4% upside potential.

See more NVDA analyst ratings

Disclaimer & DisclosureReport an Issue

1