Intel: The Dark Horse In AI Computing For 2024
Contributor since 2023 / 594 Followers
Summary:
- Intel’s 5th Gen Xeon Scalable processors, Emerald Rapids, improve AI performance and efficiency, positioning Intel as a strong player in the AI market.
- The processors offer higher performance, lower cost, and increased power efficiency compared to previous models, making them ideal for high-performance computing and AI technology.
- Intel’s focus on edge computing and their ability to handle both small and large AI models give them (potentially) an advantage over competitors like Nvidia.
FinkAvenue
Investment Thesis
With Intel (NASDAQ:INTC)’s release of Intel’s 5th Gen Xeon Scalable processors, also known as Emerald Rapids, I believe their future in the AI market is strong as these new processors help work towards the trend of smaller LLMs and running AI models more efficiently.
Launched on December 14th 2023, these processors feature up to 64 cores and boast a 40% performance improvement over their predecessors, allowing them beat out their competition when it comes to the increasing demands of AI and other high-performance applications while operating at a lower cost (this part is big). Emerald Rapids also offers a 42% higher performance on AI inference and 21% general compute performance gains compared to their previous model, Sapphire Rapids, along with a 36% increase in performance per watt across various workloads. This improvement is particularly relevant in an era where data centers account for a significant portion of global power consumption.
Moreover, Emerald Rapids signifies Intel’s progress in refining its design methodologies, even though it hasn’t incorporated the latest process and packaging technologies. This incremental yet impactful release improves performance by double digits and reduces manufacturing costs, marking a positive turn for Intel after a few challenging years. The ramp-up of Intel’s foundry business and the emphasis on manufacturing execution, processor design improvements, and workload optimizations suggest a solid growth path for the company. In my opinion, I think the stock is a buy going into 2024. I think they are the dark horse in the race for AI computing.
Background: What Has Changed In The Last Two Months?
While I have recently written about Intel post Q3, I think the release of Intel’s 5th Generation Xeon Scalable processors, known as Emerald Rapids, represent a pivotal moment in the high-performance computing front and a material increase in the upside potential. As I mentioned above, these processors are a fairly new release, therefore their effects are soon going to be seen in Intel’s journey towards reasserting their dominance in the CPU market. Emerald Rapids is the latest model, coming after Intel’s 4th Gen Sapphire Rapids series and is developed on the sophisticated ‘Intel 7’ process node. A substantial improvement between the two is that this new model features up 64 cores. This increase will be very good for high-performance computing and AI technology as it will lead to higher processing power.
As the world of AI technology is developing, the demand for more powerful yet energy-efficient solutions is vastly growing. Intel’s new processors show an improved performance per watt ratio. With the release of Emerald Rapids, Intel is expected to gain a leading position in the high-end CPU sector. Another separating feature of Emerald Rapids is the delivery of enhanced core densities and a larger L3 cache. With features like this, performance and efficiency are on the rise.
What Makes Intel’s New Chip So Special?
In recent years, the development of AI technology has significantly accelerated, resulting in a more competitive CPU market for Intel. But with the release of Emerald Rapid, Intel has bumped up in the ranks due to their new processor demonstrating significant advancements in performance, power efficiency, and security. The three main focuses of this processor include performance, power, and security. Despite the natural gains from the process shrink (from 10nm to 7nm), Intel is still consistently able to provide improvements in these categories. In order to enhance performance, Emerald Rapids introduces a two-tile, 32-core architecture. This decreases the reliance on applications and VMs to access resources on other dies. Not only that but the significant speed boost in the UltraPath Interconnect (UPI) and support for 5.6 GHz DDR5 memory, allows for faster data movement and processing.
Another improvement that has been made is an upgrade in L3 cache to 320 MB, allowing for further increase in performance. Emerald Rapids achieves about 1.34x the performance per watt relative to Sapphire Rapids (a previous model), along with the reduction of approximately 100 watts when idle.
As mentioned above, one of Intel’s main focuses when developing this new technology is security. Intel has integrated application-level security through SGX with virtual machine/environmental security enabled by trusted domain extensions (TDX). This combination provides comprehensive security while maintaining minimal latency, guaranteeing protection for diverse computing environments.
Emerald Rapids’ design specifically targets business-critical workloads, including AI, analytics, database, and cloud-native applications. By integrating acceleration engines there has been a performance increase of up to 1.4x compared to Sapphire Rapids in AI applications. This allows their technology to meet the evolving needs of enterprise IT organizations and cloud providers.
Better for Local AI Compared To Competitors
What makes Intel stand out from their competitors is how these processors handle local AI models. They perform especially well in the context of edge computing, the computation of networks close to the user, compared to one of their main competitors, Nvidia. Intel’s approach includes reducing the need for complex build outs. They do so by integrating AI into an already existing data center, which is opposite of what Nvidia does. By doing so they are able to provide a smooth increase in AI performance without increasing overhaul; therefore making it more accessible and less disruptive for companies
Intel’s 5th Gen Xeon processors aren’t just specialized to handle small AI models but can also handle large AI models, such as those with up to 20 billion parameters thanks to their built-in AI acceleration. This has been proven through real life cases such as the use of their technology by IBM. Intel’s new processor was used by IBM to improve performance on its WatsonX.data platform. On top of this, these chips are 77% cheaper in terms of their total cost of ownership over a 5 year lifecycle. Wow. This is the key to AI using edge computing networks.
As I mentioned above, what makes these processors stand out is their focus on edge computing. Taking a deeper look into this, Intel positions their processors as AI workhorses for edge use cases, arguing that AI inference, a crucial component of AI models, is increasingly happening at the edge rather than centralized data centers. This allows for faster, more efficient processing, along with lower latency. Intel’s focus of edge computing is consistent with the increasing trend of decentralized AI processing, which offers companies an advantage when it comes to real-time offering and an advantage in data processing and decision-making.
Valuation
While these new chips have only been publicly unveiled for the 3 weeks, I think the upside is starting to come into focus.
The global AI chip market hit $53 billion this year and is expected to double to $119.4 billion by 2027. I think Intel can capture 20% of this market share given how much future AI computation resources will be allocated to edge computing (where Intel is trying to be a differentiator).
Given this, Intel could capture a marginal $23.88 billion in revenue. At a forward price to sales multiple of 3.93, this implies an additional $93.84 billion in market cap upside, or 44.1% upside in share price potential (assuming the company does not issue more stock).
Keep in mind, this price to sales ratio is far lower than the 20.81 times forward sales that Nvidia (NVDA) trades at. If the market begins to see the innovative potential of these chips for edge computing I think the price to sales multiple could get closer to the ones NVDA trades at. Then things could get even more interesting. I am sticking with the price to sales multiple of 3.93 here as a more conservative estimate.
Risk to thesis
Although the future for Intel’s 5th Gen Xeon processors looks promising, it won’t be entirely smooth sailing for them. The AI market is highly competitive therefore they face serious competitors. Some examples are AMD, Ampere Computing, and Nvidia, who are also making big developments in their technologies such as AMD’s 4th-gen Epyc and Instinct accelerators, Ampere’s success with Arm-compatible processors in the cloud, and Nvidia’s AI-optimized Superchips.
Another risk of Intel stems from their recent history of their Sapphire Rapids processors. During the launch in 2021, Sapphire Rapids faced repeated delays, decreasing confidence in Intel’s abilities. After these issues, Intel has worked towards structural changes, more specifically I’m referring to the changes in their approach to CPU design and pre-silicon simulation.
While Intel’s Emerald Rapids Xeons offer a 20% performance uplift and simplified architecture, they fall when it comes to core count compared to some of their competitors. This is a crucial issue for Intel as core count is becoming more and more important in cloud computing. Intel’s focus on per-core performance over core count may prevent them from being a leading force in high-core-count markets. Another difference between Intel and their competitors is their cloud based processors. This may present an issue for Intel as a lot of major cloud providers like AWS and Microsoft are shifting towards custom silicon and Arm-based designs.
Bottom Line
Intel’s 5th Gen Xeon Scalable processors, Emerald Rapids, represent a major opportunity for Intel to gain a leading position within the CPU market. These new processors are able to meet the increasing demand for powerful, yet energy-efficient Ai solutions. Key features such as enhanced core densities and a larger L3 cache contribute heavily to this. In order to be successful in this journey, Intel will need to outperform intense competition from rivals like AMD and Nvidia. The changing market leaning towards high-core-count and custom silicon solutions may also develop as a roadblock for Intel. Along with that, Intel will need to use what they have learned from past delays to succeed. Intel’s ability to innovate and consistently execute its strategies will be crucial in overcoming these challenges and leveraging its technological advancements in the rapidly evolving CPU market. However, I think they have a real shot of pulling this off. I am optimistic. I think the stock is a strong buy heading into 2024.
Analyst’s Disclosure: I/we have no stock, option or similar derivative position in any of the companies mentioned, but may initiate a beneficial Long position through a purchase of the stock, or the purchase of call options or similar derivatives in INTC over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Noah Cox (account author) is the Co-Managing partner of Noahs' Arc Capital Management. His views in this article are not necessarily reflective of the firms. Nothing contained in this note is intended as investment advice. It is solely for informational purposes. Invest at your own risk.
Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.