Summary

• ARM's stock is overvalued, with a 25% growth expectation over 35 years, which is unrealistic given its erratic margins and headwinds.

• Arm Holdings' business model relies on licensing and royalties from chip designs, creating a strong ecosystem but facing competition from RISC-V.

• Financial performance shows solid revenue growth but unstable free cash flow and margins, raising concerns about long-term stability.

• Key risks include competition from RISC-V, too reliant on China, and potential issues with major client Qualcomm, making the stock a sell.

SweetBunFactory

Overview

Arm Holdings (NASDAQ:ARM) is at the heart of the current AI boom. Its designs draw the more efficient and smaller chips in the technology space. On February 12, 2024, the stock price rose by 60% in two days (Figure 1). Since then, the stock price has stalled.

Figure 1: Seeking Alpha

I think the business is excellent in an industry with good prospects: semiconductor design. Its revenue growth is solid but not spectacular, and margins have been erratic. The stock price is priced as its free cash flows will grow 25% in the coming 35 years, which, I think, is overvalued. The company faces some concerning headwinds, so I recommend selling the stock.

Business model and growth drivers

ARM sells intellectual property over chip designs to chip designers like NVIDIA (NVDA), QUALCOMM (QCOM), or Broadcom (AVGO), which saves time when designing a new chip from scratch. Nowadays, current chips, due to AI and edge computing, have to be smaller and have more computation power, meaning that energy optimization is key, and ARM designs specialize in this optimization. Considering that a chip is made of billions of transistors and half the cost of AI is from energy consumption, any savings from the design phase are crucial for chip companies.

I like its economic model based on capital efficiency and recurring revenue streams. The company doesn’t have to invest much CAPEX (just 6.3% of total revenue in the last quarter), and the focus is on keeping the best-qualified engineers in semiconductor design (currently 6,429, 83.4% of total staff).

The company earns from a one-time license agreement, and later, once the equipment with an ARM chip design has been launched to the market, it receives royalties, providing a steady cash flow.

However, I consider that the main characteristic of its business model is the ecosystem generated around its portfolio. Semiconductor manufacturing, OEMs, and foundries are attached to the ARM design platform. The ecosystem advantage is supported by 20 million developers and a complete suite of tools necessary to design chips quickly. That ecosystem creates barriers to imitation, and finding a different approach to achieve optimal design is challenging. Those developers already have the skills and the routine to use ARM tools, and foundries have manufacturing templates customized for ARM designs. The whole value chain is optimized. It is hard to break those relationships. This ecosystem has made that 90% of smartphones work with an ARM design.

Financial Performance

Royalty revenue increased 23% to $514 million in Q2 FY2025, which is solid growth. In the last four quarters, the company has grown 22%. This increase is due to a slightly recovered smartphone market and higher royalty per chip thanks to Armv9 delivery.

Watching the annualized contract value is a good way to better understand the evolution of licenses. The annualized contract value increased 13.1% in Q2 FY2025 from a year ago. The average growth rate in the last four quarters has been 14.2%. I consider royalties and licenses to be growing at a healthy pace.

Those attractive revenue streams come with a gross profit ratio of 96.2% in Q2 FY2025, a two-percentage-point improvement from a year ago. Research and development is 60.1% of total revenue, a decrease from 77.7% a year ago but in line with the latest quarters. Selling, general, and administrative expenses decreased by 17% to $241 million due to share-based compensation and IPO-related costs but offset by marketing and professional service costs. So, EBITDA improved from -$115 million last year to $108 million this quarter. However, if we look at the bigger picture, we can see variability in margin (Figure 2), which is a concern.

Figure 2: Author

In Q2 FY2025 and Q1 FY2025, free cash flow was negative at $47 million and $319 million, respectively. This was mainly due to an increase in net working capital. Accounts receivable rose by $225 million because of a mismatch between revenue recognition and customer payments in royalty revenue. At the same time, contract assets related to licensing agreements increased by $151 million. Besides, there was a $193 million decrease in employee compensation.

Ultimately, the business model is exceptional for me but doesn’t translate into stable results.

Valuation: the Dominance Company Method

In my previous article about ARM (Beyond The Surface: What Led Me To Buy Arm Holdings), I used a nontraditional valuation method: the Dominance Company Method. The technique succeeded (Figure 3) even when most Seeking Alpha analysts recommended selling the stock, and my model and analysis were bullish. Since then, the stock has gained 92% vs 19% upside in the SP500.

Figure 3: Seeking Alpha

The methodology says that if a stock is dominant in a market with high-growth expectations, you shouldn’t use the typical 10-year model with a terminal value of 3%. The historical evolution of companies like Microsoft (MSFT) or Apple (AAPL) has grown by 18% annually over 35 years.

I have used the valuation method to see the growth rate through 35 years implied in the stock price. Figure 4 shows the result of the valuation, and you can see that the implied growth rate is 25%.

Figure 4: Author

That implies that ARM would have been much better than Microsoft and Apple over the years because the growth rate is exponential. I don’t think this is going to be the company's evolution because last quarter’s free cash flow per share didn’t grow steadily (Figure 5), and the company is facing several headwinds.

Figure 5: Author

Headwinds to overcome

As a confirmation of my valuation, Figure 6 shows how stress is reflected in the pricing of ARM in the market. The Price-To-Free Cash Flow multiple is much higher than that of other companies like Nvidia or Broadcom, whose valuation has skyrocketed lately due to the AI boom. That means that any sign of a problem will turn the multiple down. The main sign of the problem is the RISC-V competition. RISC-V is an open-source instruction set architecture that competes directly with ARM’s designs. Membership in RISC-V International, a body that promotes the standard, has increased significantly. Major players like Qualcomm, Meta, and Google are investing in RISC-V development.

China is prone to advance in this technology as it can shelter from occidental countries' potential bans. Alibaba Group Holding Limited (BABA) is developing one of the highest-performing RISC-V processors, and it is currently in production. It is used in its server to compute complex AI calculations. Besides, ARM derives a substantial portion of its revenue from China, which is about $193 million or 22.9% of total revenue in Q2 FY2025. Any deterioration in US-China relations could severely impact ARM's business.

Qualcomm represents about 10% of the ARM business, and by itself, it is a risk. Qualcomm is using RISC-V technology to sell some of its chips for Microsoft devices and has the power to migrate current ARM technology to RISC-V. Both companies face legal litigation that could impact the relationship and threaten its revenue growth.

Figure 6

Risks

The main risk of my bearish thesis is that the adjacent market that ARM is pursuing will grow more rapidly than I expect in my model. My model assumes that total revenue will grow 25% over 35 years. Those new markets must develop at least a 40% growth rate in the following years. Let’s dive into my projections for those markets.

ARM’s revenue growth rate in data centers, automotive, and IoT over the next ten years will be substantial but will not meet my expectations. I see a potential reach of $20-35 billion by 2034, which means an average annual growth rate of 15-20% below the 40% expected.

The data center sector will grow from $8.98 billion in 2024 to $31.35 billion by 2034, representing a CAGR of 13.3% due to cloud providers like Microsoft, Google, and AWS promoting energy-efficient chips over their cloud infrastructure. The automotive sector will gain $5-8 billion annually with 15-20% of the automotive semiconductor market by 2034 because of the development of Automotive Enhanced (AE) processors and upgrading the scope to Compute Subsystems (CSS). ARM has a dominant position in the $10.66 billion microcontroller market and is expected to grow to $23.5 billion and capture $15-20 billion.

Conclusion

I recommend selling the stock. Its price is too high for the company's future prospects. Its expanding markets, data centers, automotive, and IoT won’t give it a 40% growth rate, which the company needs to justify its valuation. Also, many obstacles could damage its performance, such as open-source RISC-V architecture, business in China, and tense relations with key clients like Qualcomm. Those issues could push the stock price down in a high-valuation company such as ARM.