The NVDA Short: Falling Off the Precipice

On April 7th, 2010, as a part of my Fundamental Investing course, I recommended initiating a short on NVIDIA Corp. (NVDA), then trading at $17.16 with a price target of $14 based on a 15x P/E on '10 EPS of $0.79 (normalized), accounting for the ~$2 in cash per share. Since then the trade has been taken off yielding a 70% return via put options.  For full details about the trade, see the attached one-pager and presentation (.pptx). The trade thesis was as follows:

1. Business model has substantial challenges: 

• Very capital intensive
• Effectively a 1-2 product business model (GPU business and everything else)
• Intellectual property is transportable: face cannibalization from other manufacturers integrating NVDA intellectual property into their chipset
• Substantial customer concentration (2 customers comprise 19% of Revenue; see 10-K, page 51)

    2.  Substantial problems with the financials:

• Sustained trend of deteriorating top-line growth 
• Margins under continued pressure
• Deteriorating FCF and net working capital
• Earning’s power likely to remain challenged with operating expenses already returning to peak levels, even at significantly lower revenue rates

The catalysts for the trade were the May-10 Q1 and Fall 2010 Q2 earnings releases, Intel's future release of Larrabee later this year, and a recent wave of insider selling. On May 13th, NVDA reported disappointing results with quarterly-profit of $138mm, or  $0.23/share (which was above analyst mean estimate of $0.22/share), but with a weaker-than-expected sales outlook for Q2. The stock traded down 12% on the day, closing below $13 for the first time since Nov-09.

Key takeaways on NVDA's Earnings:

1. Weaker-than-expected forecast for sales outlook, expecting revenue to be seasonally down 3% to 5% sequentially (Implied $950-970mm range)
2. Top-line: revenue of $1.0bn for Q1 FY2011, up 2% from Q4 FY2010 and up 51% from $664.2 million in Q1 FY2010
3. Bottom-Line: TQ! FY2001 GAAP profit of $137.6 million, or $0.23/share (above analyst mean of $0.22), compared with a loss of $201.3 million, or $0.37/share, for the year-earlier period.
4. Gross margin increased to 45.6% from 44.7% in the previous quarter and 28.6% a year earlier (largely due to Fermi products, targeting the high-end gamer market; this may not be sustainable).

Implications:

1. I am concerned about Management's guidance for a seasonal revenue decline, particularly given new product cycles ramping up (i.e., Fermi, Tegra for phones, etc.), capacity is easing, inventory is increasing and GPU demand has remained robust
2. While GPU revenue has remained relatively constant (+0.2% sequentially), I am concerned that their restatement of including MCP segment revenue within GPU revenue may be masking further top-line deterioration in their core GPU segment; this was a central part of my thesis.
3. They are operating non-core / non-market leading segments (i.e., Consumer Products business: game consoles, etc.) represents only 3% of revenue and is down 31% sequentially this quarter. I think their focus is diluted.

On an un-levered basis, the trade yielded a 16% compound rate of return in just over one month (193% annualized). I implemented the trade through Sep-10 $17-strike put options purchased for $1.08, which yielded a 70% return for the trade (519% annualized).  Accounting for volatility, the Sharpe of the trade was 0.5.