ARM (ARMH) is the biggest pure play technology licensor in the world with a market capitalization in excess of $23 billion. The company's IP is used in a vast majority of graphics and logic processors shipped in the world today. ARM architecture has a near monopoly in the mobile devices market with all the major players such as Qualcomm (QCOM), Nvidia (NVDA), Samsung (GM:SSNLF), Apple (AAPL) etc. using ARM IP to design and produce their mobile chips. The company's design places emphasis on low power consumption which has made it a favorite for use in mobile devices. x86 architecture is not optimized for energy consumption, as a result of which it has almost no presence in the mobile industry. ARM gets a licensing fee for each processor that is shipped using its IP and also a royalty payment from its customers. Mobile device shipments have boomed in the last 3-4 years and ARM's revenues has almost doubled in the last 4 years. The stock price has outrun the profit and revenue growth, returning more than 700%. The stock is highly expensive now with a forward P/E of a ~80x and a P/S of ~ 25x. It is hard to make a decent return by buying stocks trading at such high valuation. However, shorting stocks only due to valuation can be dangerous as such stocks can keep going higher. A potential catalyst is crucial in deciding to short a high valuation stock. The catalyst in my view will be the introduction of Intel's Silvermont chips during late 2013.
Why ARM is a potential Sell candidate
1. Valuation, Valuation and Valuation - ARM is probably the most expensive semiconductor stock with a P/S 25x, P/B of 11.5x and a P/CF of 97.1x. The forward P/E of ~79x implies almost perfect execution by the company in the future. Though the revenue and profit growth in the past have not been bad, they have not been good enough to justify the current lofty valuations. Revenues have grown by 24% and net profits by ~58% in the last three years.
2. x86 Competition - Intel (INTC) is ARM's biggest and perhaps the only large competitor in the technology industry. Almost every other big semiconductor company uses ARM designs in its MPUs and a number of companies also use ARM designs in their GPUs. Intel has been spending a massive amount of money to make chips for the mobile devices market. The company is going to introduce Silvermont chips in December 2013, which could drastically change the landscape for ARM architecture. Intel has never really put too much money into mobile processors as it was too small compared to the PC and server processor market. However, the rapid growth of smartphones and tablets has dramatically altered the semiconductor hierarchy. Qualcomm and Samsung have seen their MPU sales surge even as x86 vendors have languished. Intel's new chips will be manufactured on its leading 22 nm node and will have huge performance gains from the previous generation. I have already written about how Baytrail and Merrifield could see Intel capture a big share of the mobile phone processor market.
3. Foundries fall behind Intel - Intel has been producing chips on the 22 nm node since last year while its major competitors like Samsung, Global Foundries and TSMC (TSM) are still stuck in the 28/32 nm node. Intel is all set to produce the first chips on the 14 nm node in 2014, while TSMC might get to the 20 nm node by end of 2013.
Samsung's 28nm chip production is still lagging far behind rival TSMC, industry observers believe. Samsung currently makes the majority of its advanced chips for 32nm chip contracts placed by Apple, the observers said. Samsung's 28nm chip output remains limited due to insufficient yield rates, in contrast to TSMC's 28nm chip production that already reached commercial levels, the observers noted. TSMC, which does not provide 32nm process technology, has said that the company would triple production and revenues from 28nm wafers in 2013 compared with 2012. The foundry set a record capex of US$9.5-10 billion for 2013, also stepping up efforts to bring 20nm capacity online.
There are doubts whether TSMC or others will be able to get to 20 nm by 2013, as they do not have the process and manufacturing expertise of Intel. If these foundries fail to make the necessary technology advancements, then Intel will sweep the mobile devices market with its Atom chips. ARM stock will crater dramatically if this happens. Intel does not allow foundry services for chips made using the ARM architecture, even though its factory utilization was just 50% during the 4th quarter of 2012.
4. Imagination Technologies competition - Imagination Technologies is a company with a business model similar to ARM Holdings. The company licenses its IP for graphics processors to major mobile device companies such as Apple for a small fee. Imagination directly competes with ARM in graphics processor designs but does not have much of a presence in the logical processor space.
ARM Upside Risks
1. Mobile Device Growth - The sales of smartphones and tablets are showing no signs of stopping despite an exponential growth in the last 5 years. Tablets shipments will exceed 200 million units, while smartphone shipments will be more than 900 million units in 2013. This means that almost a billion mobile devices will use ARM designed processors in 2013. This is a stupendous number though ARM gets a tiny license fee for each processor that is shipped.
2. Server Opportunity - Currently the server market is totally dominated by the x86 chips made by Advanced Micro Devices (AMD) and Intel. This is a huge market where Intel's high performance Xeon products have a ~95% market share. Now ARM is trying to penetrate this market and has lined up a number of big and small companies to make servers using ARM processors. Even AMD has procured an ARM license to make micro servers using ARM processors. Intel has reacted strongly to this move by announcing the release of low power Avoton processors. The success of ARM in this market is not assured given that ARM has been traditionally strong in low power consumption rather than heavy duty processing. However, if ARM manages to get even a small market share in this huge market, then it will expand its TAM. ARM processors are now being used in a wide variety of applications besides mobile devices.
Marvell announcing the deployments of some of their first commercially available ARM-based servers with Baidu, as well as LSI Logic announcing its Axxia 5500 range of communications processors, the high-performance, power-efficient networks.
3. Strong Partners - Except for Intel, every major processor company licenses ARM designs. Intel has held a stranglehold over the computing device MPU industry for a long time now. The rapid growth of mobile devices using ARM processors has broken Intel's hold over the industry. Strong players like Qualcomm, Samsung, Nvidia etc. will do everything possible to decrease Intel's power over the industry. This means that they will continue to support ARM unless it becomes impossible for them to do so. If ARM fails, then Qualcomm might be the biggest loser since it holds more than 50% of the mobile SoC market through its Snapdragon processors. Nvidia is also trying to build a big mobile processor business using Tegra processors. Samsung designs and manufactures its own processors using ARM designs. Apple too designs its own "A chips" based on ARM IP. They will not like to buy x86 chips from Intel as they would have to cede more profits and lose differentiation as well.
ARM has been one of the best performing technology stocks in the last 5 years. The stock has climbed from ~$5 to ~$50 now giving a massive profit to investors. The stock has more than doubled in the last year and currently has a market capitalization of ~$23 billion. This is a bit too excessive in my view. The stock remains very near its all-time high of $50.56 and should cross it given the momentum of the broader market and strong mobile shipments.
Though ARM has revenues of less than a billion dollars, its IP and designs are critical in powering an industry which generates revenues of more than $300 billion a year. However, ARM's valuation has gone completely out of sync with its fundamentals. It's hard to justify the lofty valuations using any rational valuation method. Competition from Intel is going to seriously hurt ARM in 2014 as it starts producing energy efficient, high performance Atom chips using the most advanced fabs on the planet. However, shorting ARM might not be the best idea now. The reason is that Intel is mostly absent from the mobile chip market which is growing at a ferocious pace. Smartphone and tablet shipments are exceeding forecasts regularly and that is giving ARM stock a huge boost. While I am convinced that ARM remains a top short idea, the timing to shorting ARM remains crucial. I think it will be a good idea to build a short position in ARM in the latter half of 2013, when Intel will launch its 22n mobile chips. This might be the catalyst which will mark the decline in the ARM stock.