Project Hunt...Belated, Incomplete and Off-the-Cuff MiniReport: ARM Holdings Ltd.
Someone recently asked about the status of rickus123 (my brother) and my Project Hunt report on ARM Holdings. My apologies for our not producing the report last year as promised. We had a majority of our research done and were getting ready to start writing the report last spring when a family member had a serious illness, and we put the project on hold. Neither of us have been able to get back to it, and I haven’t followed ARM too closely in the past year, and I’ve continued following the thread in lurk mode. This is nowhere near as detailed as we planned, but in the spirit of contributing to the discussion, here’s what I remember about our not-quite-written report (and this is almost all from memory, so the details may be off). This is not meant to be complete or up-to-date, but hopefully it will whet someone’s appetite to do ARM justice in a more detailed report. Rick didn’t participate in writing this, which he may be thankful for, given it’s sketchiness. Neither of us owned any shares in ARM last year when doing the research, nor do we now.
Overview
ARM is an IP house for embedded RISC microprocessors – high performance, low cost, energy-efficient processors used by almost all digital devices. They are totally fabless (like Rambus) and their revenue primarily comes from licensing the technology and collecting royalties for units sold. The overall RISC market is clearly in hypergrowth -(units shipped more than tripled from 1998 to 1999, and tripled again in 2000.) However, it’s helpful to break the demand down by market - they participate in a variety of markets, some which definitely aren’t in hypergrowth (game machines, printers, fax modems, mini-disc players) and others that are or have been or may soon be in hypergrowth (digital cell phones, digital cameras, mass storage devices, computer networking solutions, pda’s). They seem to be in just about everything. As of a year ago, about half of their licensees were in the mobile phone market. And at the time, I remember seeing the estimate that their processors were in about 90% of all digital cell phones – roughly the same percentage as Intel’s hold on the PC market. The latest stat I’ve seen is that they now control about 75% of the overall RISC market.
Is there a discontinuous innovation?
Rickus’ answer was yes – he argued the RISC processor itself was the discontinuous innovation - and it has become the standard for the embedded industry. He was very convincing at the time but I forgot all his arguments (hey, I said this was incomplete).
Is there a proprietary open architecture?
Our preliminary answer was yes. They controlled the patents, and licensed use and development to business partners, chipmakers, software developers, etc.
High barriers to entry and high switching costs?
Again, our answer was that it seemed to be yes on both counts. Long development cycles, patent protection, a very well developed value chain, software and applications and hardware being designed specifically for ARM chips would make it both difficult for competitors to create compatible chips and costly for partners to switch to the competition.
Here is one question we did have about their ability to accumulate true enabling technology gorilla power (We posted this on the ARM thread last year and never really got a clear answer): There may not be a compelling reason for the industry to coalesce around a single architecture, ARM's or anyone elses. If several cell phone makers become ARM licensees, is there any reason that others can't go on using a competitor's architecture?
Is there a tornado?
We argued whether that should be answered using overall stats from the embedded RISC market, or whether each market they participated in should be looked at separately. For example, they may well be the Gorilla of embedded chips in the digital cell phone market, and participating in potential Gorilla Games in the pda and digital camera markets, and participating in other markets that aren’t in tornado. I don’t have overall embedded chip growth in front of me, but ARM has (through their licensees) tripled unit shipments for two years running.
Value chain?
Definitely, all over the place. Pretty much everyone in the cell phone industry, WIND, Intel, TI, Sony…We pondered how much control they actually exert over the value chain. They only receive about 10-15c royalty per chip, and it’s debatable whether they can leverage much power they way a Microsoft could over their value chain.
Financials:
I don’t remember much of the details, which saves me from exhibiting my lack of knowledge of financials in general. But I was surprised how low their revenues actually are when compared to the Gorillas/Kings commonly talked about here. Again, while they receive royalties on a huge number of chips sold, they royalty itself is low. Their PE was out of sight when we first started looking at them, and while they got hit the past year like everyone else in the tech sector, their PE is still very high. Their gross margins are extremely high (not in front of me), and their operating margins were about 25% in 1999 – I’ve got no figures for 2000.
Caveats:
They have relatively low revenues for a company to be considered for gorillahood. Their valuation (and I know this shouldn’t be considered for gorillahood) has always seemed way out of sight, even with their correction over the past year. It was a question for us just how much leverage they actually have over their value chain.
Also, they are a British company. All the Gorillas discussed on this board are American companies, and the concepts talked about here as well as the companies tracked in ‘The Gorilla Game’ were spawned in a very specific ecosystem (United States) with its laws, regulations, competitive environment, etc. We just didn’t have much of a handle how a company that grew up in a different ecosystem should be considered. It may not matter much, but it’s worth noting.
Conclusions:
As of last winter, we were leaning to stating that ARM should be considered at the very least a Gorilla candidate of embedded chips for the digital mobile phone industry, if not the outright Gorilla. I wish I could say that either of us have kept up with developments related to ARM since then, but we haven’t. Like I said in the beginning, I hope this report will at least spur some discussion or encourage someone to actually do a more extensive and updated Hunt Report. And if you ask me to defend anything I’ve written, at least give me time to read up on ARM again before I answer. :)
Sam Johnson |
