Arm Holdings (NASDAQ:ARM) is reaching a major paradigm shift in their target market as the firm further develops their CSS architecture. ARM has recently made a few major announcements involving the adoption of their ARM Neoverse across the hyperscalers, including Microsoft (MSFT), Oracle (ORCL), and Amazon (AMZN). As each of these firms further develop their high-performance processors to bolster the speed, reliability, and power reduction, I believe ARM will realize a significant growth opportunity that will take the firm well beyond their smartphone-entrenched market. I am upgrading my recommendation for ARM shares to a BUY rating with a price target of $213.78/share at 54.43x eFY26 price/sales.
Be sure to review my previous coverage of Arm Holdings here:
Arm Will Have A Tough First Year Back In The Public Arena
Arm’s Share Price Has Gotten Ahead Of Operating Growth
ARM Operations
ARM is positioning itself for significant growth in the coming years as “AI everything” starts to take form. Though I believe AI-enabled smartphones and AI-enabled PCs are going to experience a slow adoption in the next two years, CY24/CY25, I do anticipate adoption to begin taking hold as consumers begin upgrading and migrating to the latest devices. To make my thesis clear with respect to this theme, I do not anticipate consumers to upgrade their devices in mass for the sake of upgrading. I believe upgrades will be driven by the necessity to refresh hardware as opposed to seeking out the newest device. Despite my anticipation for a relatively flat consumer device market going into CY25, I do believe there is some value to be extracted within the segment as OEMs migrate from ARMv8 to ARMv9. As adoption accelerates to ARMv9, royalty revenue will follow as more CPUs fit inside the chip. This factor will drive significant value across both the top-line and bottom-line as these higher-powered chips will likely be very margin accretive. ARMv9 is expected to be utilized across multiple verticals, including automotive ADAS, handheld devices, and industrial IoT for enhanced automation, amongst others. Management expects ARMv9 adoption and CSS chips to start ramping up in e2h25.
Bright Spot For ARM Architecture
Though I anticipate a relatively lackluster year for handheld devices, PCs, and automotive sales, there are some bright spots for ARM in the back-half of eFY25 and beyond. As discussed in my initial coverage of Nvidia, Nvidia Doesn’t Need China For Growth In AI Chips, the hyperscalers were in the process of designing their own chipsets for AI factory purposes. These chips are beginning to come to fruition and are broadly adopting ARM architecture for its power efficiency, data retrieval, and integration. I believe this shift away from the x86 architecture into ARM architecture can drive significant growth beyond ARM’s smartphone business in the coming years and will significantly shift the next generation of compute. Though this wasn’t a part of my core thesis for Intel (INTC), a shift to ARM architecture may create a major headwind for their x86 business.
One point that stuck out to me in the q4’24 earnings call was that management is seeking an avenue to enter into infrastructure and take some of the highly sought-after market share from the x86 server chip. On the call, management discussed how the ARM CPU and the NVIDIA (NVDA) GPU on the Grace Blackwell can integrate and allow for higher access to memory bandwidth when compared to the x86. This can be a major shift in server hardware if ARM architecture is utilized in the next generation of AI-enabled infrastructure and has the ability to bring ARM significant growth, if adopted. Though I believe this may be a long-shot expectation as this shift would be the biggest since the separation of hardware and software with the introduction of the x86, it can very well occur as AI factories are somewhat more standalone vs. traditional general compute server farms. I personally wouldn’t bet the farm on this shift to occur; however, if ARM CPUs are adopted in server hardware, I expect this to first occur across hyperscalers before going into enterprise data centers.
ARM announced their partnership with Microsoft in November 2023 to develop the Neoverse CSS that can create more custom, specialized solutions for AI infrastructure and networking devices. Management mentioned in the q4’24 earnings call that their first Neoverse/Microsoft (MSFT) Cobalt chip is now ramping up. This is likely to bleed into Oracle’s (ORCL) AI factories being built out as the firm seeks to take on market share. Oracle is currently working on accelerating the ARM developer ecosystem to transition applications from x86 to ARM architecture. As management at ARM described the CPU+GPU model with Grace Hopper and Grace Blackwell, Oracle will likely be leveraging this model to accelerate AI and HPC applications.
ARM also announced their partnership with Google (GOOGL, GOOG) for their Axion processor. Google announced that the Axion Processors will be available for GCP customers in late CY24 and will provide faster data retrieval, faster video distribution, and GenAI. These chips will provide for better energy efficiency across Google’s AI factories. Google expects 30% better performance than the fastest general-purpose Arm-based instance with up to 50% better performance and up to 60% better energy-efficiency when compared to the current generation x86-based instance.
ARM Financials
Forecasting financials for ARM, I anticipate the firm to begin realizing tailwinds in growth and margin expansion in e2h25 and into eFY26. Given the early ramp of the ARM Neoverse, I do not expect this form factor to be accretive to growth until eFY26, as the ramp up with Microsoft isn’t likely to occur until eFY2h25. I believe this will be the catalyst that will drive ARM to the next level of growth and profitability. In the meantime, I anticipate a relatively flat outcome as a result of the soft smartphone and automotive market, net of new adoption in automotive ADAS.
Valuation & Shareholder Value
One of the biggest challenges ARM faces is their valuation. Shares are currently trading at 53x price/sales, a relatively high valuation for the industry. As a value investor, I believe this premium makes the shares un-investable as shares are priced for perfection.
As a growth investor, I believe ARM shares will realize significant upside given the adoption of ARM Neoverse across the hyperscalers. To add some perspective, Oracle is constructing massive 200MW AI factories and a 1GW AI factory. Given the relationship with ARM, I believe that Oracle will likely utilize ARM CPUs paired with Nvidia’s GPUs to accelerate performance.
Given these growth factors, I am upgrading my recommendation for ARM to a BUY rating with a price target of $213.78/share at 54.43x eFY26 price/sales.
Additional Risks to Consider
There are some risks to take into consideration as we navigate through my investment thesis. To state the obvious first, ARM currently has a very rich valuation, and any minor setbacks can result in a significant decline in the share price. Another risk to consider is that Neoverse may not be as broadly adopted as laid out in my thesis, as this would require significant software upgrades to migrate from the x86 to ARM architecture. In addition to this risk, my expectation is for Neoverse to begin impacting the top-line towards the end of eFY25 and more prominently in eFY26. ARM will likely need to bridge the growth gap with stronger handheld sales in order to maintain their premium valuation before Neoverse ramps up. Any one of these events may lead to my gray-sky scenario and send shares down -34%.
Be the first to comment