Nvidia(Nasdaq: nvda) He said he was waiting lately Artificial Intelligence (AI) The infrastructure expenses to jump between 3 trillion and $ 4 trillion at the end of the decade. This is just a big number. Cloud computing And other major technology companies continue to compete to create the AI capacity, which brings chip manufacturers to a jealous position.
Nvidia has been the winner so far, but that’s not the only one. Let’s look at the three chipsets that will benefit the most of these $ 3 trillion.
Nvidia works in the center of AI. Graphic Processing Units (GPU) has become a standard to educate large language models from strengthening video games, and the company managed to make it a wide trench. The Cuda Software Platform was the key to this. Nvidia enabled developers to learn to program GPUs in Cuda. When this happened, companies were locked in the software ecosystem.
Nvidia is also smart on the network side. The registered nvlink connection allows GPUs to work together as a single unit, providing great benefit with AI workloads. Meanwhile, the acquisition of Mellanox gave even more power to networking and enabled its chips to support larger AI clusters. The power of the network portfolio emerged in the last quarter, and the network data center income doubled almost $ 7.3 billion.
With both software and network advantages, NVIDIA is positioned in a way that continues to be the leader in AI infrastructure construction. Although 90 % GPU does not have a market share, it is still a company to be defeated.
Advanced Micro Devices(NASDAQ: AMD) He lives in the shadow of Nvidia, but is changing in a way to play in the powerful aspects of Sunday. Education dominated the first AI wave and gave Nvidia’s Cuda edge. However, my inference is now the fastest demand, and AMD has already gained some important jobs here. Many of the largest AI companies use GPUs for inference and count seven of the top 10 AI players as customers.
AMD is also part of Nvidia’s Ualink consortium that tries to create an open intermediate connection standard to compete to Nvlink. If this, it may offer more options to create clusters to data centers and cut one of NVIDIA’s advantages. This still begins, but shows how AMD works with others to narrow Nvidia’s trenches.
The company is not just about the GPUs. EPYC Central Processing Units (CPU) gain a share in data centers and still have a strong PC and game chip business. AMD does not need to cross Nvidia to win. Just getting a larger demand for a larger inference can make the CPU business one of the larger long -term beneficiaries of AI Builyour.
Image Source: Getty Images.
Broadcom(Nasdaq: Avgo) It adopts a different approach when it comes to artificial intelligence production, but it also saw explosive data center growth. While Nvidia and AMD fought on GPUs, Broadcom has established a strong position in the data center network area. Ethernet switches, optical intermediate connections and digital signal processors are critical in moving large amounts of data, and as AI clusters grow, network needs grow with them. Therefore, AI network revenue increased by 70% in the last quarter.
However, it may have an even greater opportunity with special AI chips. Broadcom has long been the leader in the design of integrated circuits specific to the application for a long time, and recently started to work with hyper scales (companies operating large data centers) who want to increase performance and increase costs by developing special chips for AI workloads.
Helped Alphabet Create tensor processing units and now work with a few other large customers in new designs. The management says that three remote customers can place 1 million clusters in 2027 financially, which will represent an opportunity of $ 60 billion to 90 billion dollars. This number does not even contain new relationships such as established. Apple.
In addition to all these, Broadcom has VMware. The unit is switched to subscriptions and helps enterprises run AI in hybrid and very Bloud environments and provide another growth way to Broadcom. Bring everything together and you have another chip company that will benefit very much as AI infrastructure expenditures increase.
Imagine this before buying stock at Nvidia:
. Motley Fool Stock Advisor Analyst team determined what they believed Top 10 stocks For investors to buy now… And Nvidia was not one of them. 10 shares that make the cut can produce monster returns in the coming years.
When think Netflix It made this list on December 17, 2004 … If you invested $ 1,000 during our advice, You have $ 651,599!* Or when Nvidia It made this list on April 15, 2005 … If you invested $ 1,000 during our advice, You have $ 1,067,639!*
Now worth drawing attention Stock consultant Total average return 1,049A performance that breaks the market compared to 185% for -S & P 500. Don’t miss the last 10 lists, it can be used when you join Stock consultant.
Geoffrey seiler There are positions in the alphabet. Motley Fool positions and recommends in advanced micro devices, alphabet, apple and nvidia. Motley recommends Fool Broadcom. Motley Fool’s Explanation policy.
