While it’s hard to imagine the chip and technology sector repeating a great year like 2023, both financial and technology analyst groups are pretty much in agreement that 2024 will be just that. There is.
With the rise of AI and the need for ever-greater computing power, semiconductor companies will likely see another strong year in 2024, albeit with varying levels of success.
First, let’s take a look at some of the key players in the data center and IT infrastructure market. This segment fueled his 2023 AI demand, driving up valuations and consuming nearly all the chips these players could produce. There’s no reason why we shouldn’t see an additional 20% to 50% increase in silicon and infrastructure opportunities, allowing nearly every player to benefit to some degree.
Nvidia NVDA,
is the clear leader in Clubhouse, with its stock up more than 200% in the past year thanks to GPU data center dominance, the most powerful and complex platform for training AI models that will revolutionize computing. workloads. Other companies are trying to catch up with him in 2023 and will continue to do so in 2024, but we can expect to see the rise of competing chip options from Advanced Micro Devices AMD.
in-house developed silicon such as Microsoft MSFT;
alphabet google,
Amazon.com AMZN,
Start making some progress.
Nvidia’s data center division’s revenue will grow as the market expands, but it seems inevitable that market share will inch back as competition increases. The company hopes to offset that with new product and service offerings. If Nvidia truly becomes more than just a chip supplier to the market, offering what CEO Jensen Huang has dubbed an “AI factory,” or other AI and computing microservices, it could be the same or even better by 2024. It is possible to see more revenue growth.
Nvidia’s main competition will be AMD and its MI300 family of GPUs, which will be announced in late 2023. CEO Lisa Su is confident about the new chip, making bold statements about expected revenue, and customers like Microsoft are buying early. As a true second-tier supplier of high-performance, high-memory-bandwidth chips capable of handling the most demanding AI workloads, AMD should stand to benefit greatly on its own. However, AMD will also continue to see its market share improve in the data center CPU space, with its EPYC processor series gradually cornering the dominant Intel Xeon family.
Intel INTC,
In the data center sector, 2024 could be the cloudiest yet most interesting year. Last December, the company announced its next-generation Xeon processors, codenamed Emerald Rapids, and teased Gaudi 3, a dedicated AI processor that uses a unique architecture compared to GPUs from Nvidia and AMD. Gaudi 2 is currently on sale, but not much is known about its success in 2023 or what his 2024 customer pipeline will look like. This custom AI architecture offers unique power and performance benefits for some specific workloads based on the testing I’ve seen, but for large-scale AI customers, We need to convince them that this is a long-term solution. Intel will see some benefit simply because his Nvidia has limited GPU inventory of its own, but CEO Pat Gelsinger has much bigger ambitions than just his Nvidia leftovers.
Finally, in the data center sector, it is worth mentioning cloud service providers and the expected growth there. Microsoft Azure, Google Cloud, and Amazon AWS are the three biggest players expected to see a significant increase in computing usage by AI companies large and small, with off-the-shelf solutions from AMD, Nvidia, and Intel as well as proprietary solutions. We also offer Custom chips in a fully verticalized solution stack. All three of these companies and the meta platform META
Companies such as have verbally committed to development roadmaps for these custom silicon and software, but it remains to be seen how much of their cloud service offerings will actually utilize them over the next 12 months. is not clear.
Another risk spot for cloud providers is the need for data security and privacy in enterprises deploying AI applications and solutions. In this case, it may make more sense to handle some of the AI training and inference workloads in on-premises data centers, which could be a win for Dell. technology dell,
HPE HPE,
To fill that gap.
The customer field for semiconductor companies may be more interesting than the data center field, even though the growth rate is much lower. The impact of AI PCs and the rise in general client computing is likely to create dramatic winners and losers over the next 12 months, with either incumbent leaders outpacing the competition or the same competition gaining significant market share. You can either take it away.
Microsoft’s potential game changer
The unique opportunity here lies with Microsoft. While the push for AI PCs is still vaguely defined and more of a marketing term than something consumers can understand or trust, for Microsoft it’s about devices, operating systems, content management, generative AI tools, and more. This is an opportunity to redefine what personal computing means across the globe. more.
Just as the smartphone revolution changed the way we think about accessing and creating data, the next big change will be how we integrate it with AI. This could usher in a renaissance for the PC, taking back mindshare from smartphones as great devices for getting things done and organizing large collections of data and information. What better way to fight back against Apple AAPL?
More than a Mac with features that Apple hasn’t really tackled yet? The windows may be small, but they’re certainly there.
The three major companies surrounding “AI PC” are Intel, AMD, and Qualcomm QCOM.
to this point. Intel launched its Meteor Lake chips, now called Intel Core Ultra, last December, promising to bring faster AI computations to all generations of laptops in 2024. The ability to integrate a graphics engine on the chip allows Intel to strike a balance between performance and battery life. Intel Core Ultra gives Intel the opportunity to scale up the AI PC space, shipping millions of processors to market this year, creating thought leadership (Intel = AI on PCs), and possibly driving up prices. Drive revenue by offering these chips to customers who want to take advantage of this.
AMD has been equipping some of its notebook PC chips with dedicated AI accelerators since the middle of last year, but sales and market saturation status are unclear. In December, the company announced its next two generations of Ryzen laptop chips with improved AI performance in 2024 in hopes of taking more of this chip market from Intel. For AMD, there is a possibility that it will continue to provide his OEM on behalf of Intel, and as long as it has a competitive and reliable AI PC strategy, it can take advantage of any marketing that Intel and Microsoft push. can.
Qualcomm is also in the fray, with its Snapdragon X-series chips announced last October expected to launch in mid-2024. These processors are based on the Arm architecture rather than x86 and have a much more powerful NPU for AI processing than Intel or AMD options. This is part of Qualcomm’s strategy to transition from a communications company to a computing company, and we expect to announce significant wins with key system partners in the coming months. Qualcomm is also the only chip company that appears to be investing heavily in brand marketing in the customer space, taking advantage of the stagnation of its chip competitors to make Qualcomm and Snapdragon the biggest player in consumer AI computing. We hope that this will lead to progress in the field.
Although it doesn’t come up often in the AI PC conversation, Nvidia has an interesting position that it plans to highlight in 2024. It’s a GPU, the same one used in the data center AI segment, but with less memory, making it ideal for high-quality applications. High-performance, high-throughput AI processing on your PC. For high-end content creators, game developers, video and 3D animators, the power of a discrete GPU to run AI on a local machine far exceeds that of an integrated NPU from Intel, AMD, or Qualcomm. The top GeForce products have over 800 TOPS (Tera Operations per Second), while his NPU on Intel Core Ultra is only 10 TOPS. This comes at the cost of power and, in the case of the laptop’s Nvidia GPU, battery life, but for many AI workloads pure performance is all you need. Needless to say, Nvidia’s software stack is second to none, and almost every AI application developer has been using his Nvidia GPU since day one, giving you another potential advantage to leverage.
3 unknowns
There are three big questions I’m tracking through 2024. These influence the forecasts and analysis above and broadly set the tone for what 2025 will look like.
-
Where will most of the AI processing take place? In the cloud, at the edge, or on the consumer’s local device? This will make the chip war in the data center or client sector more difficult for the companies mentioned in this article. The degree of influence will be determined to some extent.
-
How popular will custom silicon options from Microsoft, Alphabet, Amazon, and Meta be compared to traditional silicon providers from Intel, AMD, and Nvidia? These are big investments and require hardware and software to get right. Both require billions of dollars and thousands of engineers. Will verticalization benefits, or performance and power efficiency benefits, provide enough benefits to offset the momentum of traditional chip company products?
-
This is a bit wild, but will Intel continue as an integrated integrated device manufacturer, or will it eventually split into a manufacturing company (making chips) and a product company (designing chips)? There continue to be signs that Intel is exiting non-core businesses such as Mobileye, its FPGA group, and even its AI software business. The changes here provide an opportunity for other chip companies to expand their supplier choices and give Intel products the opportunity to forge new partnerships.
Mr. Ryan Shrout signal 65 with the founder of shrout research. Follow him on X @Ryan Shrout. Shrout has provided consulting services to AMD, Qualcomm, Intel, Arm Holdings, Micron Technology, Nvidia, and others. Shrout owns Intel stock.
more: Apple stock looks expensive compared to the rest of the Magnificent Seven
Also read: Nvidia stock breaks new ground ahead of CES after healthcare partnership
