Over the past few weeks, we've shared our EZchip Semiconductor (EZCH) investment opinion in a pair of articles, outlining our belief that EZCH's $717m market capitalization, 8.0x 2013E revenue multiple, and 32.7x 2013E GAAP P/E (adjusted to include $0.39/share of stock-based compensation) are wholly unjustified. The Bull case depends on EZCH growing well in excess of the overall router market thorough 2016, a challenge we believe is unrealistic given an avalanche of near and long-term business risks. Customer concentration is dangerously high, newly introduced competition from Broadcom (BRCM) could pressure design wins, Huawei's recent order delay proves that in-house design risk is ever present, and weak end-market demand will pressure EZCH's self-imposed 30% revenue growth hurdle (slide 8). To justify EZCH's almost 2x valuation premium to the S&P 500, it must overcome this hurdle each and every year through 2016. Most EZCH investors discount these risks, instead trusting a management team that has over-promised and under-delivered for years, growing the revenue by just 11% annualized between 2009 and 2012. But since some took the time to reply to our article, we feel a response is due.
A Response to Feltl & Co's 'Short-Seller' Report
Following our March 11th report on EZchip, Feltl & Co. issued a one-page response that addressed some of the arguments in our initial report. Feltl also reiterated its "Strong Buy" recommendation that's been in place since it initiated coverage on August 22nd, 2012 at a price of $32.28. Feltl's research note gave EZCH investors the positive reinforcement they craved. After reading through the March 11th report, we believe that Feltl's response includes the following three key points:
- Broadcom's June 2012 Linley commentary proves that BCM 88030 targets the switch market, not EZCH's core end-market of edge routers;
- Even though Broadcom introduced the BCM 88030 almost a year ago, no explicit customer wins have been announced;
- A long sales cycle suggests no immediate competitive threat
In the interest of continuing the conversation, we've taken the time to address Feltl's response below on a point-by-point basis.
1) More Recent Broadcom Comments Unequivocally State BRCM's Intention to Compete in EZCH's Core Market of Edge Routers. While Feltl cites Broadcom commentary from a June Linley conference, we've placed more credence on December 2012 remarks. At its yearly Analyst Day, Broadcom explained that edge routers are its chief addressable NPU end-market, "Another example of a product that expands our SAM is [a] 100 gig network processor [BCM 88030] that we introduced earlier this year…this is a product that [is] very well-positioned for further penetration into the traditional network processor market in EDGE routers primarily [emphasis added], as well as in some core router applications, as well as packet transport applications." To drive this point home, Broadcom included a full-page graphic in the accompanying Investor Presentation that illustrates this remark. Whether EZCH likes it or not, it will be competing with BRCM for new Edge Router NPU business.
Source: Slide 35, BRCM December 6th, 2012 Analyst Day Presentation.
2) Because Sampling and Testing Occur Over a Protracted Period, the Delay Between Broadcom's BCM 88030 Announcement and Customer Wins Should be Expected. Feltl correctly states that the BCM 88030 was first announced in April 2012. But this passage of time does not imply that the competitive threat is removed. As EZCH investors know through experience, the sampling and design process is iterated with the router vendors over a multi-year period before public customer wins are announced. EZCH first disclosed the research-stage details of its comparable NP-5 product in May 2011 and even after nearly two years, it has yet to provide samples to customers:
<Q - Paul Williams>: The NP-5, when do you think you'll be able to provide samples to your customers on that?
<A - Eli Frutcher (CEO)>: In [the] June to July  timeframe, I would say. (EZCH Q4 2012 call).
And in the case of EZCH, timelines can slip. As of November of last year, EZCH expected to sample an NP-5 "in the first quarter of 2013," a deadline that had previously moved from Q4 2012 (Q1 2012 EZCH call). Now, the expectation is for late Q2 or early Q3 2013. As for Broadcom's timeline, it expects to have customers in 2013, "And by the way, you should be seeing all these products [BCM 88030] in production, right, with customers [in] 2013." (BRCM 2012 Analyst Day).
3) While a Lagging Product Cycle Does Provide EZCH with Run-off Value, EZCH's $700m Market Capitalization Necessitates Substantial Future Growth. EZchip is not a cheap stock. An 8x 2013E Revenue and 32.7x 2013E GAAP P/E multiple are indicative of a rapidly growing business. To achieve the 30%+ annual revenue growth targeted in the Q4 2012 Investor Presentation, EZCH must hit each of its growth levers in unison -- double-digit end-market growth, market share expansion, and 40% 4-year ASP growth. As a business whose revenue fell from $63.5m to $54.7m over the past year, EZCH has yet to deliver on this promise.
Even if EZCH is successful in retaining its current customers, it would be hazardous to assume vendors wouldn't take advantage of BRCM's entry to pressure EZCH on pricing and other terms. We are not alone in voicing these risks; EZchip readily acknowledges the competitive shift in the marketplace:
"We believe that competition in this market will become more intense in the future and may cause price reductions, reduce gross margins and may result in loss of market share, any one of which could significantly reduce our future revenue and decrease our net income." ̶EZCH 2012 20F
For EZCH to sustain its almost 2x valuation premium to the S&P 500, it must not only transition each NP-4 customer to the NP-5, it must also increase unit pricing and networking market share, both of which are questioned in the statement above. This task could be made all the more challenging by continued softness in the router and switching markets. Last week FBR securities released a report stating that the market for routers and switches has hit a "dead end." The analyst went on to state, "Looking ahead, we see the potential for additional negative technological trends that could significantly blur the lines between routers, switches and servers." As networks move toward the cloud and data centers, demand for traditional wireline infrastructure will continue to soften. This sentiment is confirmed by Gartner, a technology research house, who predicts industry-wide Edge Router and Switch growth of just 3.6% in 2013 (Dec 13th report). EZchip is not blind to this trend; it has dramatically rebranded its last three investor presentations (Q3, Q4, and Roth), shifting emphasis from Carrier Ethernet to NPS/Smart Networks. Unfortunately, shareholders won't see NPS revenue until 2016+, if ever.
The Question of Stock-based Compensation
Other industry analysts, such as Paul McWilliams of Next Inning Technology Research, challenged our decision to expense stock-based compensation ("SBC") in EZCH's forward EPS projections. Ours is not a radical idea, and Generally Accepted Accounting Principles (GAAP) mandate that SBC be expensed from EPS in all SEC filings. To justify EZCH's add-back and non-GAAP EPS reporting, as used in Wall Street projections, it has been argued that future cash flows will be used to purchase newly-issued, dilutive stock. We completely disagree with this concept. Investors pay 18x 2013E EBITDA for EZCH stock because they expect to have a claim on future cash flows. If that future cash isn't value additive, and is instead used to correct for past SBC excesses, the earnings multiple used to value EZCH should be materially lowered. In light of this, it makes sense that GAAP dictates that stock-option expense is recognized on the income statement at the time of grant.
While the effect of this dilution creeps up slowly, ensuring that investors barely notice, it can have a drastic impact over an extended period. Over the past four years, EZCH's fully diluted share count has climbed from 23.6m to 29.6m today, resulting in 25% dilution for investors who first purchased the stock at the end of 2008.
While this track record of steady dilution presents serious red flags for long-term holders of EZCH shares, it's doubly egregious when investors are also being misled into believing that non-GAAP financials are indicative of financial performance. In 2012, EZCH reported a 49% non-GAAP net income margin (Q4 Presentation, Slide 14), a figure that is frequently cited by bullish investors and the sell-side analysts. After correcting this figure for 2012's $11.2m of stock-based compensation expense, the margin falls precipitously to 29% (15.902 / 54.707). This is no rounding error. It reminds us that investors need to be vigilant for excessive non-GAAP earnings adjustments.
While entrenched bulls argue that EZCH has zero competitive threats, we continue to believe that Broadcom's more complete networking product line (NPUs, switches, KBPs, multi-core processors, etc.) and lower cost base is a competitive advantage versus EZchip. Cisco's success over Juniper is illustrative of the benefits of scale. One could argue that Cisco was able to take share from Juniper because Cisco offered a wider swath of products at discounted prices. EZCH could face a similar competitive dynamic from its newly emerged larger competitors.
Equally as important, even if EZCH is successful in retaining its current customers, we expect that the presence of alternatives such as BRCM will allow Cisco and other customers to pressure EZCH on pricing and other terms. Given its high customer concentration, the ability for customers to produce NPU chips in-house if they desired to allocate the requisite internal resources, and the presence of competitors like Broadcom, we think that EZCH will perpetually have limited pricing power for its products.
At EZCH's current market capitalization of over $700m, we believe the market remains too optimistic. Revenue expectations for the next two years are $69.7m (2013) and $94.2m (2014), requiring a 31% YoY CAGR, well in excess of the 3.6% Edge Router/Switch end-market growth forecasted by Gartner (December 13th, 2012 report). Revenue growth of 30%+ requires EZCH to fend off Broadcom and in-house ASIC competition, have its product mix drastically shift towards higher-priced NP-4, and be able to add new business from Ericsson, Tellabs, and Huawei. If any one of these levers fails to deliver, EZCH will likely see another year of moribund growth and disappointed shareholders.
Additional disclosure: Please read our disclaimer at kerrisdalecap.com/?page_id=2118.