Ambarella Looks Like A Short

| About: Ambarella (AMBA)
This article is now exclusive for PRO subscribers.

When shares of Ambarella (NASDAQ:AMBA), a vendor of multimedia system-on-chip products, traded at about $13 earlier this year, I was bullish on the name, believing the shares were worth about $18 (~38% upside). Interestingly enough, the shares went on a pretty massive run and as of writing trade just shy of $31 per share. It is my belief that this stock, which was once a compelling long opportunity, now looks as though it could be a very compelling short opportunity from a risk/reward standpoint particularly as the massive run in the shares has had much more to do with hype and news-flow than with the underlying fundamentals.

Explaining This (Insane) Run

On Dec. 20, Ambarella announced that it was working with Google (NASDAQ:GOOG) on a "new class of wearable cameras for the Helpouts application and service." For those of you unfamiliar with Google Helpouts, it is an online collaboration service launched back in November that allows users to share their expertise through live videos and provide real-time help from either their PCs or mobile devices.

It's a neat idea, and the service could very well take off (with or without dedicated wearables for this space). However, the "insanity" here is that the stock gained 26% or $190 million in market capitalization on this news - more than a year's worth of sales. Now, it is completely understandable that anything and everything attached to the buzzwords, "wearables" and "internet of things" tends to generate an obscene amount of hype, but it's important to actually step back and try to quantify this opportunity.

The barrier to entry just isn't all that high

The SoC that Google and Ambarella will be using for this reference design is the A7LW. Here is the block diagram:

While the above may look a bit foreign, the chip is not at all dissimilar to a modern smartphone or tablet system-on-chip (it's actually less sophisticated in many ways). It's a great chip for the specialized applications that it's designed for (i.e. cameras), but there is nothing "revolutionary" about what Ambarella is doing here and, if the market were to become interesting enough, there is absolutely no reason why the likes of Nvidia (NASDAQ:NVDA), Qualcomm (NASDAQ:QCOM), Intel (NASDAQ:INTC), or Samsung (OTC:SSNLF) would not whip up an SoC to sell into the wearable camera market. It would probably - given the resources that these three companies have and how adjacent mobile chips are to camera SoCs - be trivial for any one of these players (so we can rule "acquisition" out).

Indeed, even Ambarella recognizes this in its Form 10-K,

Currently, our competitors range from large, international companies offering a wide range of semiconductor products to smaller companies specializing in narrow markets. Our primary competitors in the camera market include CSR Plc (who acquired Zoran Corporation in August 2011), Fujitsu Limited, HiSilicon Technologies Co., Ltd. and Texas Instruments Incorporated, as well as vertically integrated divisions of consumer device OEMs, including Canon Inc., Panasonic Corporation and Sony Corporation. In the automotive aftermarket camera market, we compete against Core Logic, Inc., Novatek Microelectronics Corp. and Sunplus Technology Co. Ltd. Our primary competitors in the infrastructure market include Intel Corporation, Magnum Semiconductor, Inc. and Texas Instruments Incorporated. Certain of our customers and suppliers also have divisions that produce products competitive with ours. We expect competition in our current markets to increase in the future as existing competitors improve or expand their product offerings and as potential new competitors, such as Broadcom Corporation, NVIDIA Corporation, Qualcomm Incorporated and Samsung, enter these markets.

Valuation Looks Absurd

As of the time of writing, Ambarella trades at 4.6x EV/Sales, 25.3x EV/EBITDA, and 30x EV/FCF. Now, what's interesting is that looking at this from a discounted cash flow standpoint, it would take a 20% 10-year FCF CAGR followed by 7% of long-term growth in order to justify a fair value of roughly $30 (and in this analysis I am generously not discounting the cash on the balance sheet). Note that there is further lenience here in assuming that we have this kind of visibility over 10 years and that Ambarella will continue this meteoric growth over this time. Looking for a 20% 5-year FCF CAGR with long-term growth of 7-9% yields a fair value (including cash) range of between $19 - $20.

So, to further illustrate the absurdity here, it is worth stepping into the realm of the slightly speculative and running some potential numbers. The stock is up significantly on the "news" that the company will be partnering with Google on a wearable camera reference design. What could the potential incremental revenue look like from such an opportunity?

Well, it's worth stepping back and making a few assumptions:

  1. ASP for an Ambarella SoC for this space (in particular the A7LW shown above and its successor) could be between $10-$20 (assume $15 at midpoint) at gross margin level of about 60% (in the MRQ GM saw 200bps improvement on higher infrastructure mix, implying camera/wearable SoC GMs are lower than corporate average)
  2. Ambarella has a monopoly in the Google-sponsored wearables market (this is one that is very likely to break, but we are trying to establish an upper bound for the opportunity)
Google "Hangouts Camera" Units sold Incremental Revenue @ $15 per SoC Incremental Gross Profit @ 60% GM EPS Impact (at ~11% tax rate, assuming minimal incremental R&D/SG&A)
20,000 $300,000 $180,000 $0.0064
40,000 $600,000 $360,000 $0.0128
60,000 $900,000 $540,000 $0.0192
100,000 $1,500,000 $900,000 $0.0321
1,000,000 $15,000,000 $9,000,000 $0.321

What's interesting is, again, Ambarella gained $190 million in market capitalization on this opportunity. But, as we see here, that even if this exploratory reference design actually goes into production for sale, it would likely need to sell about a million units to have just a 32-cent impact on EPS. Assuming that from that 32-cent baseline (again, this requires 1 million units to be sold) that this market grows like gangbusters and that there is little - if any competition - leading to a 25% CAGR in that incremental EPS growth rate over 5 years before leveling off to a long-term growth rate of about 10%, then this opportunity may be worth ~$6-7 per share today.

But in that model, we are assuming no competitive pressures, no incremental opex, and a market that actually ends up materializing. Investors are quite literally buying a close-to "best case" scenario following that PR. And, as we've seen in many other fallen tech stock angels, the "best case" scenario - particularly with these small SoC companies with valuations that bake in hyper growth - is very rarely what pans out.

A Real World Example: We've Seen This One Before

Take a look at Ambarella's chart:

AMBA Chart

AMBA data by YCharts

Now, here's a chart of a stock that exhibited largely the same "pattern" that I expect will ultimately befall Ambarella, although perhaps not quite to the extreme degree shown here:

SIGM Chart

SIGM data by YCharts

Sigma Designs (NASDAQ:SIGM) is a system-on-chip company that built video processing chips for set-top boxes, TVs, Blu-Ray players, and so on. After spending years waddling around in a $5-$10 trading range, it too saw a pretty extreme "breakout." Why? Well, back then there was some pretty hard-core euphoria about IPTV. Sigma, like Ambarella today, essentially had a monopoly on chips in that market. Eventually, it turned out that the opportunity wasn't really as large as had been hoped for and what had once been a monopoly position now has become a competitive bloodbath with heavyweights like Broadcom, ST-Micro, Marvell, Qualcomm, and others muscling in on various portions of its business.

A Few More Words

Barring a miracle, Ambarella's shares look meaningfully overvalued today and at the first sign of a slip-up (i.e. growth not coming in as expected, poor media/consumer reaction to Ambarella/Google joint reference design, etc.), the shares are likely to see a very sharp correction. Following its decline, Sigma did this and was buying back shares in the $50s; Cirrus Logic did the same when its stock fell from its $40+ grace). This is a classic pattern that has been seen many times with these small cap semis that strike it rich (or strike it momo on promises of striking it rich) time and again. Look at the charts of Sigma Designs circa 2006-2007, Nvidia circa 2007, and Mellanox in 2012:

CRUS Chart

Overhyped and Overbought, Ambarella Looks Like A Short

From a risk/reward standpoint, it seems likely that following the "reveal" of this Ambarella/Google joint project reference design at CES, there could be a real "sell the news" event and shares could give up most - if not all - of the gains on the Google hype, which would imply 21% downside. Downgrades could follow and the stock could be further re-priced. Indeed, even assuming a more modest 15% FCF growth rate over the next 5 years and a fairly aggressive 10-12% LT growth rate only supports a valuation range - cash included (and not discounted) - of about $17-$19, implying 43-49% downside over the next 6-12 months as the hype fades.

Disclaimer: The opinions expressed here are solely my opinion and should not be construed in any way, shape, or form as a formal investment recommendation. Investors are reminded that before making any securities and/or derivatives transaction, you should perform your own due diligence. Investors should also consider consulting with their broker and/or a financial adviser before making any investment decisions. Any material in this article should only be considered general information.

Disclosure: I have no positions in any stocks mentioned, but may initiate a short position in AMBA over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: I am long INTC, BRCM, NVDA.