Aspirational CPU manufacturer Arm Holdings (ARM) stock plummeted earlier this week as the DeepSeek saga wiped billions from Western tech firm valuations. As the dust settles, the question is: should investors buy the dip in ARM stock?
Join thousands of data-driven investors – Build your Smart Portfolio for personalized insights.
Regardless of the macro-political shadow, the events unfolding on our screens do not alter my broader thesis about the tech revolution. I see the tech stock selloff as broadly overdone. However, that doesn’t mean I expect a recovery in all futuristic tech stocks across the board. There will be winners and losers from the cornucopia of events unfolding in the tech/AI sector right about now, and I don’t see ARM as one of the winners.
I see shareholder value to be found elsewhere in the tech space rather than in Arm Holdings. Despite this week’s 10% share price correction, I would counsel restraint for tech-focused buy-the-dip bargain hunters eyeing a long position in ARM.
Given its hard-to-justify valuation, I’ve been bearish on ARM for months, with this week’s events further corroborating my bearish position on this dippy stock.
ARM Caught in DeepSeek Crossfire
ARM stock slumped on Monday as the DeepSeek saga piled pressure onto tech stocks, with shares in the British chip designer tumbling 10%, before recovering somewhat in after-hours trading. This decline came amid a broader wave of selling in the artificial intelligence (AI) sector, triggered by the success of Chinese startup DeepSeek’s language model.
The selloff occurred just days after the Stargate Project was first announced — a $500 billion, four-year initiative to build AI infrastructure in the U.S. Importantly for ARM investors, the Stargate Project involves Arm Holdings and its parent company, SoftBank (SFTBY), as key collaborators and funding partners.
The sell-off in ARM stock was nowhere near as pronounced as in other parts of the market. Personal stock favorites of mine, Celestica (CLS) and Credo Technology (CRDO) shed nearly 30% as investors fled from AI infrastructure en masse.
If one were looking to buy some dips, companies like NVDA, CLS, and CRDO would be far better options than ARM, in my opinion.
What Does Cheaper AI Mean for ARM?
The DeepSeek revelation, if taken at face value, presents significant challenges for ARM. As a chip designer, ARM’s business model relies heavily on licensing its designs to manufacturers for various devices, including those used in AI applications. DeepSeek’s achievement of creating a highly competitive AI model at 2% of the regular cost and with fewer chips due to U.S. export restrictions, threatens to disrupt the status quo in AI development.
Cheaper manufacturing methods reduce demand for high-performance, specialized AI chips, potentially impacting ARM’s sales and licensing revenue. The industry may shift towards more efficient, cost-effective solutions, forcing ARM to adapt its designs to meet new requirements. Furthermore, Nvidia’s (NVDA) endorsement of DeepSeek’s approach lends credibility to this new direction, potentially accelerating the trend toward cheaper development.
With this in mind, ARM may need to pivot its strategy, focusing on designing chips optimized for efficiency and cost-effectiveness rather than raw power. This shift could require significant R&D re-investment and a reevaluation of ARM’s product roadmap to remain competitive in the rapidly evolving AI chip market.
DeepSeek Questions Remain Unanswered
Specific questions remain unanswered concerning DeepSeek’s breakthrough development. First among those is whether the Chinese company could access more advanced chips. CEO of Scale AI, Alexandr Wang, told Bloomberg TV at Davos that DeepSeek likely had access to around 50,000 H100 chips. The H100, which has not been approved for sale in China by the U.S. government, reportedly costs around $27,000-40,000.
This suggests that Chinese entities are evading sanctions. Moreover, commentators have pointed out that Singapore is one of Nvidia’s largest clients, with 22% of billings coming from the country in Q3. This seems inexplicably large for a small, gas-dependent city-state, and these Singaporean orders may represent a backdoor to China.
Why is this important for investors? If DeepSeek built its AI model on copious numbers of H100 chips, it would reinforce the thesis that the world needs more high-powered chips and data centers to advance AI.
Lofty ARM Valuation is Unsustainable
My bearish stance is based on ARM’s lofty valuation from key metrics. The company’s forward price-to-earnings (P/E) ratio of 93.5x is significantly higher than the IT sector median of 25.2x, representing a 271.8% premium. Similarly, ARM’s price-to-earnings-to-growth (PEG) ratio of 2.84 is 55.5% above the sector median. These elevated multiples suggest investors are pricing in substantial long-term growth expectations. To top it off, its current P/E ratio is an eye-watering 269.
Compared to historical averages, ARM’s current valuation is also stretched. The company’s forward P/E ratio has fluctuated, reaching as high as 108.7x in June 2024. While the current forward P/E is lower than recent peaks, it remains considerably above sector norms, indicating a potentially overvalued status in the current market. Even when accounting for the firm’s ultra-impressive margins, the valuation is hard to justify.
What is the Future Price of ARM Stock?
On TipRanks, ARM stock carries a Moderate Buy rating based on 14 Buy, four Hold, and one Sell ratings assigned by analysts over the past three months. The average ARM stock price target is $157.20, implying approximately 5% upside potential over the next 12 months.
Strong ARM Insufficient to Warrant Heavy Portfolio Lift
Despite ARM’s leading market position, early-mover advantage, and pioneering market aspirations, I remain strongly bearish on the stock given that its inflated valuation metrics far exceed sector norms, implying substantial long-term growth may be more challenging to achieve than first thought.
While the company is integral to AI infrastructure development, disruptive competition from DeepSeek and potential shifts toward cost-efficient solutions will likely restrict meaningful future growth. At current multiples, the stock is overvalued and vulnerable to a growing number of risks, including from the political sphere of influence, which further muddies the waters in a sector already riddled with asymmetric market risks.
I consider ARM a decent company, but considering its sector and the technology it has developed, other market players perform better from a shareholder perspective. Therefore, I remain bullish on the big tech themes, actively changing life as we know it, but somewhat bearish on ARM Holdings.
