Surprise! Analysts Got This One Wrong
09/20/2012 7:45 am EST
In typical fear-based analyst style, here is one company that dominates a key, growing tech sector, expands its margins and revenue—and gets downgraded, even though there's at least 25% upside in the stock over the short term, writes Paul McWilliams of Next Inning Technology Research.
Mellanox (MLNX) is the undisputed king of the chip market for a high-speed communications technology known as "InfiniBand."
In the beginning, much of MLNX's InifiniBand sales were in module form versus at the chip level. This began to transition during the summer of 2010, and Wall Street misjudged the drop in revenue to imply failure.
What Wall Street missed here is the fact that as MLNX went through this transition, and aggregate demand for InfiniBand grew, MLNX increased its gross profit margin (chips carried a higher margin than modules), and exited the transition with both higher sales and profits.
Recognizing this would likely be the case (higher revenue and higher margins), I suggested buying MLNX in July 2010 when the price dipped into the teens.
QLogic (QLGC) tried to compete with MLNX in the InfiniBand market, but after falling hopelessly behind, QLGC decided last year it couldn't keep pace with the necessary R&D investment to continue the pursuit. As a result, it sold its InfiniBand product line to Intel (INTC).
The short story here is no one was able to match MLNX's 56Gbps InfiniBand speed, and while INTC will likely emerge as a valid competitor in a year or two, in the near term it needed MLNX's speed as well, and signed on this year as a major MLNX customer.
Last June, I forecasted MLNX would report non-GAAP earnings in 2012 of $2.50, and in 2013 non-GAAP earnings of $3. I saw both as bold predictions given the fact that the consensus estimates were then $2.25 and $2.57 respectively, and for most other semiconductor companies I was forecasting 2012 earnings below the then-current consensus estimates.
I reiterated those forecasts in our July 15 earnings preview, and included there my estimated balance sheet value for MLNX was just shy of $6. With these data, I noted in that report that at its then current price of $66.07, MLNX was trading just below the $66.50 midpoint of my projected three- to six-month full-value range.
Based on this view, I suggested hedging MLNX exposure by selling either December $65 or $67.50 covered calls. At the time, those calls were yielding premiums of $12.20 and $10.40 respectively.
This meant that including the call premium, the total value at exit (assumes MLNX was trading above the respective call strike prices at the close on December 21, 2012) would be $77.20 or $77.90 respectively. Given my view that the high side of a full value price range was $72, I saw this as a good way to capture that and reduce the risk exposure.
The short story here is I was wrong—like every analyst covering the stock, I was blindsided by MLNX's amazingly strong second-quarter results and its equally strong guidance for Q3.
After listening to MLNX's conference call on July 18, I posted a review that evening stating I was wrong and that I thought we would see the price of MLNX move into triple digits. However, since MLNX declined to answer numerous questions about its longer-term growth prospects, I also stated there was not enough fundamental data to forecast how far into triple digits, and that I couldn't build a well-grounded model for a full valuation. I still don't have that data for the latter, and neither does anyone else.
The price of MLNX gapped up at the opening on July 19 into the $90s, but fell back the next Monday into the $80s. After what the chart readers would term as some consolidation, MLNX rebounded to finally move above $100 three days later, on July 26, and in spite of its recent sell-off remains above that mark.
A couple weeks ago, the analyst from Stifel Nicolaus lowered his rating of MLNX from buy to hold. While he also increased his earnings estimate for 2013, the basis for the lowered rating was the then-current price of MLNX, which was trending above $120, reflected what he thought was a full and fair valuation.
At the bottom line, he stated that he thought MLNX's revenue would grow in line with INTC's Romley server and storage deployments. I think this is a reasonably logical view, but would add there are other drivers outside Romley for MLNX. The most obvious of these is the fact Oracle (ORCL) uses MLNX InfiniBand parts in its server designs that are based on the SPARC chip originally developed by Sun Microsystems.
That said, even if we were to model MLNX's 2013 non-GAAP earnings at $5—which is above the highest estimate presented by the 13 analysts covering the company—at last Thursday's closing price of roughly $120, MLNX was trading at more than 22 times the $5 estimate plus its balance sheet value.
While 22 times forward earnings plus balance sheet value isn't too much for a growth story that is in its infancy, with a market capitalization of nearly $5 billion at last Thursday's closing price, it is high on both a relative and absolute basis for a company with that market valuation that will also likely face valid competition from INTC within a couple of years.
Investor fears were stoked Monday as short sellers jumped into the fray, with one writing a searing review that was published by SeekingAlpha. This report came on the heels of a press release from MLNX announcing the Chief Financial Officer (CFO) Michael Gray, who had been with the company since before its initial public offering, would retire to spend more time with his family.
Gray has an active 10b-5 selling plan that he has used to manage his stock options, but at last report still owns nearly 28,000 shares and some number of unexercised options. Personally, I have no knowledge of Gray's personal situation, so I cannot comment on his retirement.
In the near term, I see MLNX's most significant risk being demand volatility. In other words, I think there is a fair chance that once the growth rate for the Romley upgrade cycle peaks late this year (or in early 2013), we'll see a more volatile demand pattern for InifiniBand products.
To insure I'm being clear here, I believe we'll see demand rise solidly in 2013 and potentially deliver non-GAAP earnings that approach $5, but with a risk of quarter-to-quarter volatility.
However, even with a projection of $5, I think the $120 price we saw earlier this month represents that view reasonably well, and if I'm right, that leaves only modest room for further upside. By implication, this suggests the upside from the current price range is something in the range of 20% to 35%.