Complex data center workloads such as training machine learning models and running artificial intelligence (AI) applications would take a very long time if powered by central processing units (CPUs only). To that end, specialized semiconductors are used to accelerate computationally intensive AI tasks.
In that vertical semiconductor, Nvidia(NASDAQ: NVDA) graphics processing units (GPUs) have emerged as the industry standard. In fact, the company has between 80% and 95% market share in AI accelerators, according to analysts. But recently Nvidia shareholders got some worrying news from a rival Broadcom(NASDAQ: AVGO).
Broadcom sells a variety of semiconductor products, including integrated Wi-Fi and Bluetooth chips Apple a Samsung smartphones, as well as networking chips in Arista switches. But Wall Street is particularly enthralled by its leadership in application-specific integrated circuits (ASICs.) ASICs are purpose-built chips for specific use cases such as accelerating artificial intelligence (AI) workloads.
Analysts estimate that Broadcom has about 60% market share in custom AI chips because of its relationship with three hyperscalers, a term that refers to companies with large data center footprints. Although Broadcom has not identified its hyperscale customers, analysts generally believe they are parent Google Alphabet, Meta Platformsand TikTok parent ByteDance.
Broadcom estimates that revenue from its three existing hyperscale customers will range from $60 billion to $90 billion in 2027, up from $12.2 billion in 2024. In other words, the company predicts that normal AI chip sales will will increase by at least 70% annually in the next three. years, but perhaps as fast as 95% annually.
That's disappointing for Nvidia shareholders because it means Broadcom will likely gain market share in AI accelerators. In fact, analysts are Morgan Stanley estimate that ASICs will account for 13% of AI accelerator sales in 2027, up from 11% in 2024. They also think that figure could hit 15% in 2030. But there's more bad news for Nvidia shareholders.
Broadcom CEO Hock Tan told analysts on a fourth-quarter earnings call that Broadcom has selected two new hyperscalers that are likely to be revenue-generating customers by 2027. That means revenue from custom AI chips could grow faster than 95% annually in the next year. a few years. Importantly, although Broadcom did not identify the customers, analysts believe they are the creator of Apple and ChatGPT OpenAI.
Image source: Getty Images.
I mentioned that Broadcom select two additional hyperscalers as potential customers. CEO Hock Tan himself used that word because Broadcom won't develop ASICs for small companies, and small companies wouldn't be interested in using custom AI chips.
There are two reasons for that. First, designing ASICs is expensive, so the customer must have a data center footprint large enough to warrant the cost. Piper Sandler Analyst Harsh Kumar recently told CNBC that each chip costs about $500 million to design, so it wouldn't make financial sense to work with clients who can only order a few thousand. Instead, orders must be in the range of 250,000 to 500,000 units.
Second, custom chips are less flexible because they are designed for specific workloads and lack supporting software development tools. Nvidia offers a robust ecosystem of code libraries and pre-trained models that simplify the development of GPU applications. No such tools exist for ASICs. That means using custom chips requires a high level of technical expertise, which means Broadcom's customer base is limited.
Additionally, companies experimenting with ASICs may ultimately decide that the costs outweigh the benefits. Antoine Chkaiban at New Street Research says that only two companies have used custom AI silicon at scale: Google and Amazon. So Nvidia is well positioned to maintain its leadership in AI accelerators. Bank of America analysts estimate it will hold a 75% share of the market in 2030, down only slightly from 80% in 2024.
Looking ahead, Wall Street thinks Nvidia's adjusted earnings will grow at 34% annually through fiscal 2027, which ends in January 2027. That consensus makes the current valuation of 53 times adjusted earnings looks very reasonable. Potential investors can buy a few shares with confidence, and existing shareholders have good reason to be optimistic going forward.
Have you ever felt like you missed the boat when buying the most successful stocks? Then you'll want to hear this.
On rare occasions, our expert team of analysts discuss a “Double Down” stock. recommendation for companies they think are about to pop. If you are worried that you have already lost your opportunity to invest, now is the best time to buy before it is too late. And the numbers speak for themselves:
Nvidia:if you invested $1,000 when we doubled down in 2009,you would have $362,166!*
Apple: if you invested $1,000 when we doubled down in 2008, you would have $48,344!*
Netflix: if you invested $1,000 when we doubled down in 2004, you would have $491,537!*
Right now, we are issuing “Double Down” alerts for three amazing companies, and there may not be another opportunity like this anytime soon.
John Mackey, former CEO of Amazon subsidiary Whole Foods Market, is a member of The Motley Fool's board of directors. Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Bank of America is an advertising partner of Motley Fool Money. Randi Zuckerberg, former director of market development and spokesperson for Facebook and sister of Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool's board of directors. Trevor Jennewine He has held positions at Amazon, Arista Networks, and Nvidia. The Motley Fool has positions and recommends Alphabet, Amazon, Apple, Arista Networks, Bank of America, Meta Platforms, and Nvidia. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.
