ARM Holdings (ARMH) is the world's largest vendor of microprocessor-related IP. Products containing its IP span the smallest and lowest-power micro-controllers to the most powerful smartphone and tablet chips. Of course, as the smartphone boom substantially drove the need for increasingly more powerful processors, ARM has not only seen robust volume growth, but it has consumed more of the system-on-chip bill of materials by expanding into graphics and physical IP licensing. Further, ARM has taken its success in mobile and is trying to run with it in brand new areas such as network processors and even micro-servers. The business is well-run and sound.
However, despite a truly excellent underlying business, the shares remain expensive, trading at 36 times non-IFRS (that's similar to non-GAAP as it excludes stock based compensation) fiscal 2014 estimates,which already bake in 30%-plus growth over the current year. However, expensive isn't itself a red flag, but expensive with slowing growth is.
ARM's quarter was a beat, but it missed on royalties
ARM actually beat consensus top-line estimates -- $286.7 million versus consensus of $274 million -- but posted an in-line bottom line at $0.24 per share. The top-line beat was largely driven by a surge in licensing revenue. That is, the initial one-time, up-front payment to have access to either the architecture (so that the licensee can design its own chip) or off-the-shelf processor IP was strong. In theory, that bodes well for future royalty streams. However, the growth in royalties -- the "cut" ARM takes from the actual sales of products that contain its IP -- was up only 14% year over year.
ARM has quite a lot in the long-term pipeline to help juice royalty rates, including the broad shift toward Cortex A50/ARMv8-based chips across many of its end markets and increasing adoption of its graphics IP in the mobile space. However, it's tough to ignore the fact that the high-end smartphone market is showing signs of having saturated and that the fundamental shift from feature phone to low-end smartphone is almost complete.
Only 18% of ARM's processor shipments in the quarter were Cortex-A (i.e. something that would be higher-end like in a phone or tablet) with the simpler ARM7, ARM9, and Cortex-M dominating the processor shipments. So while ARM has been moving up the value chain in a number of end markets, it has also seen substantial unit growth in the lower-end markets, which aren't as juicy per unit on the royalty side of things. So, what does this royalty miss imply?
Growth expectations may be overdone
In the Oct. 10 piece, "ARM Holdings Is Best in Class, But It's Already Priced In," I noted that even a very optimistic DCF model -- 25% FCF growth over the first 10 years before leveling off to 11% long term growth that it appears this company may be overvalued. While this seemed plausible, it seems that such expectations have now pushed deeper into the realm of unrealistic. It's bad enough that Intel (INTC 1.28%) is making a very credible push in tablets (with smartphones coming next year), which could serve as a headwind that manifests itself over the next two years, but even unopposed the royalty growth rate has stalled.
Now, to be fair, the steeper ramp of the ARMv8 products from the likes of Broadcom and Cavium Networks in the enterprise networking space could drive meaningful growth (although a realistic upper bound here would be an incremental $51 million – the entirety of MIPS' revenue base). In addition, ARM has a real opportunity for its physical IP in the low/mid-range portions of the smartphone market. But once the mix shift up from feature phones to smartphones is finished, will growth truly continue to be so robust?
Share-based compensation and new valuation
On top of this, while ARM's non-IFRS operating results look particularly appealing, it's tough to ignore that the IFRS results (i.e. those that don't factor out stock based compensation) are significantly less impressive with the company earning just $0.17 per share in the most recent quarter. While share based compensation is an important part of most tech compensation schemes, it is not something that should be ignored when attempting to value the stock.
Indeed, it looks as though on a non-IFRS basis, the company is on track to earn $0.55/share this year. Lowering my 10 year IFRS growth CAGR to 20% and then lowering the long-term growth expectations to 10%, and adding back in tangible book value, it's hard to go beyond a fair value of $25. Letting the share based compensation scheme slide and using non-IFRS results (but same growth estimates and a forgiving discount rate), it's still tough to go beyond a fair value of $38 today.
Best in class, but overpriced
ARM is best in class, but its shares do look significantly overvalued at the $50 level today. Investors wanting to get in on this story should wait on a pullback, particularly as concerns around royalty rate growth and high share-based compensation cannot be permanently masked by even a great growth story.
