Earlier this year, I published our short argument about Arm Holdings (ARMH). With the recent headlines created by the lowered guidance and Deutsche Bank's downgrade of ARMH, as well as the one-day 9% drop that followed, I thought it would be a good time to publish our more recent update. For those of you that have read the past article, you will notice that some of the facts remain the same, while many new ones have been added. (If you are an ARMH bull and hate my view, please make sure to leave a relevant comment before bashing my input with insults.)
Arm Holdings is a European Intellectual Property firm that designs low powered microprocessors that are used in the majority of the mobile devices around the world. ARMH is a fabless company, meaning it does not actually manufacture chips, but rather it licenses its technology to companies like Apple (AAPL), Qualcomm (QCOM), Nvidia (NVDA), Samsung (SSNLF.PK) and many others. Unlike other well-known semiconductors such as Intel (INTC) or AMD (AMD), ARMH's two sources of revenue come from the Royalties and Licensing (R&L) fees it receives for each device that is using its low power chip technology. ARMH's stock has had a major run over the last several years, and it has become another company priced to perfection.
ARMH is currently the U.K.'s most valuable technology company, with a market capitalization of $11.5 billion, even though it's expecting approximately $900 million in revenues, with a cash generation from operations of nearly $200 million. Its respectable growth of more than 20% over the last few years, along with banner contracts with the biggest tech names in the industry, has been applauded by investors to the point where its valuation exceeds companies that are generating nearly 10 times more sales, profits, and everything in between. Investors have created enough noise from their own applause to silence the rationale we believe makes our short argument easy to become comfortable with. Here are some of the key details as to why ARMH is one of our largest short positions:
• Although we agree that ARMH has built a substantial business over the years, we believe that the market has priced in a huge $11.5B premium into a billion dollar a year business because of the company's relationship with AAPL. This relationship clearly does not yield anywhere near as much in revenues or profits as it did in additional market capitalization. AAPL is about 10% of ARMH's revenues, yet each time AAPL beat earnings expectations, we have seen ARMH go up by about $1 per share ($450M).
As an example of how insignificant AAPL's real contribution to ARMH's business is, look no further than AAPL's best quarter ever -- Q4 2011. Even after the monstrous beat of expectations by AAPL, ARMH still missed its earnings expectations. It seems like the bulls believe that by buying into ARMH, they are getting AAPL for a discount. This couldn't be more ridiculous, as the two companies have nothing in common.
In our opinion, the contribution ARMH's market capitalization gets from AAPL is dramatically higher than it will ever get as a business. This significant dependence on AAPL actually creates an additional risk factor for ARMH shareholders. In the event that it loses a part of, or even the entire contract in the future to one of its competitors, ARMH would be left in a circumstance it has never experienced before.
• There has been an extraordinary amount of attention (and increased market capitalization) and hype generated by the expectation of the new Windows 8 based phones using ARMH technology, in addition to the ones using the more traditional X86 (Intel & AMD). We know that this hype is practically guaranteed to disappoint investors, and we know this by simply using grade school math. The value of ARMH has increased by nearly $4 billion since the first announcement. This is slightly more than ARMH will ever make throughout the lifetime of the deal. Hypothetically speaking, even if Microsoft (MSFT) were to sell a billion ARMH mobile devices (dramatically more than Apple has ever sold since the first iPhone came out) in the next 12 months, it would at most translate to somewhere between $75 - $85 million in "revenues" for ARMH. With the ASP of its IP going down approximately 10% per year since 2005, it is actually more likely that the revenues would be even lower if MSFT became that big of an ARMH customer, since they would surely renegotiate better rates than even the mighty AAPL.
• The use of ARMH in servers is another example of a hyped up expectation that has hit the wires with ridiculous expectations. Even if these expectations are met, the new business will only start affecting ARMH in two to three years and generate $5 to $10 million in total royalty revenues by 2015. Without going into too much detail, anyone familiar with technology would immediately tell you that they would not choose the power efficient, yet powerless ARMH chip over X86 technology in their server. It's nice for email, but that's about it. If analysts are putting a $1 per share value on the future server business ($450 in market cap), it would take investors over 100 years to break even on their investment, assuming they continue selling it past the 1st year.
• When one thinks of successful long term businesses, they immediately think of the "brands" created by those businesses they are familiar with. Whether that brand is Coca Cola, IBM, Intel, or General Electric, the brand itself usually only defines a fraction of the services the company at large offers, yet it still represents to us their entire existence.
IP businesses have no brand recognition. A simple way to prove that is by asking yourself if you knew (or cared) that your current smartphone was likely powered by an ARMH IP chip. If you're a techie, it's possible that you knew, but if you're everyone else, it's highly unlikely. This lack of brand recognition means that licensees like Qualcomm, Apple, Samsung and others will have very little hesitation to use any other supplier as soon as ARMH's energy efficiency idea is no longer the most efficient.
The likelihood that ARMH IP will not stay Np. 1 forever is 100%. This major risk of that happening sooner rather than later makes it a bad business model that must change in order to allow the company to survive, let alone thrive. Successful tech companies let you know that you're using their technology even if many users don't know what it means. You'll always know if you're using Microsoft's (MSFT) product, but while you may have likely used ARMH, chances are you haven't noticed.
• While the overall fabless business model does have some cost and overhead benefits, ARMH's Royalty & Licensing fees model is completely flawed over the long run. Our research shows us that since it does not actually manufacture, package or sell its products, its business has already become commoditized, with average licensing fees per unit dropping nearly 10% per year since 2005 ($0.117 to $0.037 per unit). At this rate, it would become almost worthless in just a few years.
Royalties are not faring much better, dropping by almost 50% per unit since 2005 ($0.099 to 0.051 per unit). Although we don't actually believe that ARMH will give its chips away for free, it is very possible that it will be forced to change its business plan and increase R&D costs drastically if it plans on finding a way to increase its ASP's and brand value. This increased R&D is not being calculated by the company, or by any of the sell-side analysts covering the stock.
*Please note that our averages may be different than the ones reported by ARMH, since licensing is a one-time fee, and Royalties are paid per unit. We decided that calculating a weighted average for all still delivers the points.
• The relentless argument by bulls is that ARMH will sell more devices in the future, as well as more chips per device. As smartphones become the new standard, the average core per phone has increased from one to over two in the last 12 years. While this argument may be true, it is flawed for a few reasons. First, it has already reached near-capacity in the mobile phone industry by having its IP in approximately 95% of the smartphones. Despite the growth projections of the smartphone industry (which to some extent we agree with), this overwhelming market share leads us to believe that there is more of a downside than an upside for ARMH's smartphone future. No tree grows straight up forever.
Second, although rapid growth is anticipated in the rest of the mobile industry, ARMH's commodity-type business model has miniscule R&L fees that continue to decline each year. So even when smartphones become the "only" phones and eventually reach four chips per device, the existing business model will not allow shareholders to reap the same benefits from those chips as they do from today's. As the existing patents become older, they will continue to lose value at a rapid pace, eventually becoming obsolete and virtually worthless well before its patent expiration. To compete with new threats from Intel and other competitors, ARMH will have to increase R&D at an exponential rate, which we believe will lead to lower profits (or even losses) for even higher revenues in the next few years. This domino effect would also culminate with the elimination of its dividend.
• Since ARMH is fabless, it owns close to nothing outside of its patents. Nearly half of ARMH's book value is comprised of goodwill (approx $875M worth of goodwill out of $1.78B). This makes the company nothing more than a big team of (approx 2,000) very intelligent people, with patents that will depreciate in value over time as competitors get into the market and make current technology obsolete.
Unlike the AOLs of the world that managed to sell their patents for a price that's nearly half their "recent" market cap, ARMH doesn't have that same option if it plans on staying in business. Just like AOL, Kodak used to value its patents at much higher values several years ago, until it got a wakeup call at a recent bankruptcy auction. Another small player in the sector, MIPS Technologies Inc., has not had as much success and recent rumors indicate the company may have put itself up for sale in the market. This sale is actually as critical for ARMH shareholders as it is for MIPS, because it would not only quantify how much an acquiring company really believes the IP business model is worth, but it would also tell ARMH who it's going to be competing with in the next couple of years. Apple, Samsung, Google (GOOG) or Qualcomm would not be very good for business.
• Mobile competition is coming for the first time in ARMH history, from both INTC, and inevitably AMD, with the real possibility of others. After successfully implementing its low power chip strategy, ARMH has had a clear road to ruling the market with no major company/competition to stand in its way. This is currently ending as you're reading this. With Intel recently announcing the production of a mobile phone chip, as well as a partnership with Integrated Device Technology Inc (IDTI) to develop an integrated transmitter/receiver chipset for wireless charging technology, it is quite apparent that it's committing to take this mobile market away from ARMH. Say what you will about who has the better, more energy-efficient chip, having a business with no competition is better than one that competes with the 800-lb gorilla and its lesser, albeit hefty nephews (AMD and others).
• To first warn anyone looking at the company's financials, make sure to have a currency converter and a sharp eye for small print to make sure you're translating the numbers correctly. We view ARMH's "Normalized Operating Expenses" as the equivalent to the U.S.'s "Non-GAAP," since they exclude acquisition related, share based payments, and disposal or impairment of investments. We also believe ARMH also plays with the currency conversion on its financial reports to magnify, or enhance financial numbers. Only the pence sterling is used for EPS (because it appears larger than dollars). Only the british pound is used in Expenses (because it appears smaller than dollars). The U.S. dollar is mainly used for Revenues (because it looks bigger than pence or BP). Although ARMH reports revenues in BP as well as U.S. dollars, the other financial numbers are only reported in the currencies I listed above, forcing you to constantly convert everything. Aside from being annoying, this can be considered manipulative behavior, which leads us to wonder what else the company is willing to manipulate if given both the necessity, as well as the opportunity.
• In a recent Analyst Day, ARMH's management stated that although licensing CAGR has been 14% for the last 8 years, it is now projecting a CAGR of 5-10% moving forward. Royalties had a 22% CAGR, but are now projected to drop to 10-15%.
• Current 1st half revenue Run-Rate is $844 million, despite $885 million estimates for 2012.
• Margins are beginning to shrink. Recent example: during ARMH Q1 Earnings, the stated Cash generation was $58.3 million, which is $4.6 million LESS than Q1 2011, despite having $24 million more sales. Cash generation disclosure is very strange and continues to change every quarter or two.
• Market Share penetration capacity limitation is already showing in recent numbers, as the recent growth rates are dramatically lower than past. In Q1, 1.1 billion chips were shipped for mobile phones/computers (100mm less than Q4 2011, same as Q4 2010 and Q1 2011).
• Increased expenses and lower margin already began to some extent as Q1 2012 Normalized operating expenses were 66.1£(British pound 1.61x1 dollar) ($106mm) which was HIGHER than Q4 2011 by a few hundred thousand despite more sales in Q4, and HIGHER than Q1 2011 by 6mm£ (10%) or $10mm. The blame was put on increased R&D.
ARMH also projected normalized operating expenses in Q2 2012 to be £67-69 million, which is £10mm higher ($16mm) than Q2 11 or 17% higher. It also used the same excuse in page 3/24 of the press release regarding increased expenses due to R&D, making investors believe the increase is non-recurring when it is recurring and worsening.
All major expenses increased in Q1 12, including R&D, Sales & Marketing, and General & Administrative. The share-based payments are hidden in plain site, where the company discloses where the costs are allocated to in the financials, but still has a confusing process. Again, since this is a cost, it is reporting it in british pounds, because it LOOKS lower. Total share-based payment costs and related payroll tax charges of £9.8 million in Q1 2012 were included within cost of revenues (£0.5 million), research and development (£5.9 million), sales and marketing (£1.8 million) and general and administrative (£1.6 million).
• Number of licenses signed in 2012 were the lowest in years. A total of 39 in Q1 2011, 29 in Q2 2011, and 35 in Q4 2011. This also means lower revenues in the future, since licenses translate into future royalties.
Other Important Items
• European exposure is not a permanent negative catalyst, but it is a catalyst today, in case the European financial crisis deteriorates further. Although it may not affect ARMH's business as much as it would a European based manufacturer, a crisis can easily trigger an attack on the stock's momentum, as it would be deemed guilty by association. This was clearly shown during the summer of 2011, when the stock dropped 25% in a matter of days, before partially recovering.
• Insiders have been selling this stock in the same fashion (or even more so in some cases) as our friends at Green Mountain (GMCR). Selling part of your shares is okay, but getting rid of a stock as if it were poisonous is giving us more confidence that management at least understands the overvaluation of their stock. With one of the company's founders and president, Tudor Brown, resigning (effective 5/12), we are safely assuming that the majority of his shares will be disposed of shortly. Major insider sales always remind me of a game of musical chairs -- only now the competition has an unfair advantage and I don't want to play.
• In all conferences, management always relies on outside analysts' projections of the industry rather than having their own. We find it strange that management believes the analysts have a better sense of their own business than they do.
• Recent Q1 earnings miss was blamed in part on industry hard drive issues. We believe this is false, as this company sells 2 billion chips to customers that, for the most part, don't use the ARMH products in technology that has hard drives.
• Earnings seem to be less reliable than one would think from a Royalty & Licensing business, as unexpected large "one-off" deals occurred a few times to make the quarter.
• Nearly half of previous hot selling IP is currently obsolete as shown by no additional sales in recent quarters. ARM7™, ARM9™, ARM11™. Earnings reports claim they are adjusted for licenses that are no longer expected to generate royalties, yet license totals are not growing in count and are bound to expire in large size at one point when phones are obsolete and replaced. Specific risk for ARMH is with its old chips (ARM7, ARM9, ARM11), since all are now replaced by Cortex A's three versions. It's strange that it reports licenses it sold many years ago. It's like MSFT wanting credit, or at least recognition, for selling Windows 95.
Under the current business model, we break down valuation into two parts: Current/Future business + Patents. With the limited growth in smartphones (available to ARMH) and a product price that has declined every year, we see the value of the business and its earnings power at $1.5 to $2 billion (3.25-4.50 per share).
The main asset belonging to ARMH is its patents. Although this asset represents the only reason this company even exists today, it is difficult to determine how valuable this asset will be in the future, as these patents inevitably become obsolete and are no longer hailed as "the best idea," well before the patents expire. We believe a fair valuation of those patents today is anywhere between $2 to $4 billion (4.50-9.00 per share). Please note that this value continues to decrease each day ARMH or its competitors get closer to having superior technology.
Since the first business would not exist without the patents, we believe that a fair intrinsic value for ARMH is at some middle point of approximately $10/share. The big (short) catalyst that can become a reality with any technology company is its reaction once it loses its "best idea" status. If it does not react fast enough for the market, the value of the business that will remain will be its "old" patents. Just think of how valuable your old VHS VCR is. A good example of the swing from Great to Obsolete in the current market would be Research In Motion (RIMM).
Yes, we wish we shorted RIMM, too. The good news is that RIMM is not the first or the last falling star in the technology world. As in the GMCR short position, we are consistently implementing derivative strategies to mitigate some of the short-term risk of the stock going against us. Incidentally, ARMH was also our most profitable short-term trade of 2011, when we took advantage of its temporary fall from $30 to $23 during the summer.
The most significant difference is the fact that we are alone with our bearish view of the company. I actually think that my thesis is being relatively generous with the middle ground valuation of $10/share. In addition to Deutsche's recent negative ratings change, there was only one Sell Side firm that we found that had a similar view, but they're a small shop in Europe without the same influence as the traditional sell-side analysts most are familiar with. Otherwise, everyone is too busy being bullish.
Quite frankly, we see the ARMH business model as completely flawed and think the company will either be out of business or bought out for a fraction of its current value within the next five years. The only reason it has had no competition from its own customers is because it offered its service first, and at a cheaper price than it would cost its customers to develop it.
The problem that the sell side doesn't realize is that the major customers recognized that ARMH has no brand value whatsoever, and therefore, forced the company to continuously cut prices as it sells more and more chips. Further, the industry giants realized that they need to develop chips that have the same power efficiency, and have had a couple of years to do so. This is a perfect commoditized plus competition formula that could lead to ARMH's extinction.
Its model and asset-less business will not allow it to stay independent. Its technology has a limited longevity value, and therefore, it is merely a matter of time before the end comes. The insiders' actions imply that they agree with us. Lastly, there is an increased likelihood that it will have to take at least a $250M impairment charge, as its goodwill increases for the overpriced purchases it has made.
If you had $11.5B to spend, would you buy a company that is generating $850M in sales, $50M in net free cash flow, has a price/unit model that goes down 10%/year, has limited physical (easy to value) assets, and has internal projected growth that's shrinking by as much as 50%? If the answer is yes, it means that, unless something drastic changes, it would take you over 100 years to get your investment back.
Risks To Our Thesis
Analysts are already uber-bullish on this stock, so an upgrade is not really in the cards. Bullish catalysts that can go against our short could be:
1. Earnings/revenues grow higher than expected. ARMH projected at least $860M in revenues, while we project a numbers closer to $850M. Neither one would justify the current valuation. Risk would be if revenues grow faster than that.
2. MSFT/ARMH products come to market sooner and in more models than expected, and change the world by becoming the new standard overnight.
3. MIPS gets bought out at some astronomical valuation, like 15 times its revenues.
4. ARMH signs deals with INTC and/or AMD to design the next generation of chips together that would work with the X86 architecture. This would by far be the most detrimental to our short, since it would change quite a few things. I don't believe that either company will sign with ARMH for a variety of reasons.
They say "the fish rots from the head down," and ARMH is starting to smell.
Disclosure: I am short ARMH.