Business Description: Nvidia is a leading producer of graphics chips for professional workstations, desktop and notebook PCs, video game systems, and has an emerging presence in mobile devices. Nvidia invented the graphics processor unit (GPU) in 1999. A GPU is a separate processor that has historically executed only graphics instructions, freeing the central processing unit (CPU) to run other instructions it is better equipped to handle. The company was founded in 1993 by Jen-Hsun (pronounced Jensen) Huang who remains CEO and owns 3.5% of the company and is very engaged in the business. Nvidia has a low capital intensity business model as it outsources production of its chips to TSMC. The result of this is a business that generates excellent cash flow and high returns on invested capital over the course of the cycle.
Investment Thesis: Nvidia and competitor ATI (acquired by AMD in 2006) essentially act as a duopoly in the market for "discreet" GPUs. While Intel and AMD offer integrated solutions and are contemplating embedded solutions (i.e. graphics that are part of the CPU) for the low end market, everyone we have talked to believes there will always be a significant market for "discrete" GPUs because high end graphics require a specialized chip. We believe that Nvidia's discreet chip business (which represents the bulk of its profits) will continue to grow and generate excellent cash flow as consumers and businesses upgrade their PCs. We think Windows 7 will drive a substantial PC upgrade cycle over the next several years and Nvidia will participate in that cycle. We believe normalized earnings power (including stock options expense) for the core graphics chip business is at least $0.58 per share assuming a 10.8% normalized operating margin which the average operating margin of the company from 1998-2009 on a $3.6 billion top-line (TTM revenue). The company has $2.98/share of cash on its balance sheet. So we are essentially buying the business ~11x normalized EPS. We believe there will be times when the business will earn much more than this (see the Fiscal year ended Jan 28, 2008 when it generated $1 billion in FCF on GAAP EPS of $1.31), but we think this is a conservative cut at the earnings power of the business.
Additionally, Nvidia has a multi-purpose processor called Tegra for smartphones and other handheld devices (games, MP3 players, etc). Given the growing popularity of graphics intensive smartphones (e.g. iPhone) we think that Nvidia is well positioned to secure design wins for Tegra and has the potential to grow this into a multi-hundred million dollar product over the next several years. Nvidia has claimed their will be Android smartphones based on Tegra by the holidays. Finally, Nvidia is also a leader in supercomputing where multiple Nvidia GPUs are being used in parallel for general purpose computations as a low-cost way to enhance the performance of supercomputers. Should this architecture be adopted more broadly in lower end computers someday, it would be a boon to Nvidia's business.
Finally, Nvidia would be an accretive acquisition for lots of players in the industry. Intel is the perpetually rumored buyer though I am not sure an acquisition stands up to regulatory scrutiny. Other buyers like Broadcom or Qualcomm could also make some sense though.
Bottom-line, we think the market is valuing Nvidia's core discreet graphics business too cheaply. You also get a free call option on Tegra chip for smartphones and Nvidia's Tesla super-computing efforts which are promising but in their early stages at this point.
Why has the stock been a dog?
First, Nvidia messed up the current product cycle. Their chip based on their new architecture called Fermi was late to market. When it did finally arrive, the high end chip (F100) ran hot, used a lot of power, and was expensive. Basically, ATI had a better product offering. Nvidia lost a TON of market share (about 8%) in the most recent quarter as a result of not really having a competitive product offering. The company recently made a huge cut to guidance and I suspect this was a combination of cleaning out inventory and losing share (more details to come out on their August 12 earnings call). However, in July a streamlined version of the new Fermi chip (F104) came out for the lower end of the market ($199 price point for finished cards) and the reviews have been very positive. While it is hard to predict quarter to quarter shifts in market share over the history of the industry Nvidia has always responded when they missed a product cycle (see 2004 when they really missed the July quarter). We would expect this time to be no different.
Second, there has been some bad luck. Rambus won a ruling at the ITC recently. This could results in a a couple hundred million payment by Nvidia to Rambus ultimately, but in the mean time Nvidia will continue to appeal. Additionally, Nvidia lost a socket in Apple's latest version of the iMac to ATI.
All of this has combined to make Nvidia one of the worst performers in the S&P 500 this year (-49% YTD). What we see is really a series of short-term problems that probably don't impair the long-term value of the business.
Intel and AMD plan to capture more and more of the market with their integrated and embedded solutions. Historically, demand for high end graphics has always outstripped the processor companies' ability to execute on integrated/embedded solutions to serve anything but the lowest end of the PC market, but this could change. As it stands today Apple still refuses to use integrated graphics which I think is a testament to how bad the performance of integrated solutions are. If for some reason embedded integrated graphics end up being able to serve the whole market including the high-end, that is a problem for Nvidia.
The GPU business is cyclical. If the cycle turns down earnings may disappoint. We believe Nvidia's strong cash position should enable it to weather any downturns, but the stock may temporarily suffer until the cycle turns back up.
Nvidia has a lawsuit with Intel and will not be able to develop integrated chips to go on motherboards with Intel processors if Intel wins. For now, Nvidia has shifted its resources away from integrated chips for Intel. Total integrated chips represented (including AMD and Intel) 26% of revenues in F'2009, but a substantially lesser percent of profits given that these are low margin products. The loss of this product stream will be a headwind. However, the recent FTC settlement with Intel may keep this market alive for Nvidia.
The ITC recently ruled in Rambus's favor in an ongoing dispute between Nvidia and Rambus. Nvidia claims they can just take the European license form Rambus at a cost of less than 1% gross margin is a worst case scenario. However, I would expect potentially as much as a couple hundred million in back payments if this is the case. No big deal given Nvidia's cash balance, but never something you like to see.
The Tegra mobile chip platform faces stiff competition from Broadcom, Intel, Marvell, Mediatek, NEC , Qualcomm, Renesas, Samsung (powers iPhone), Seiko-Epson, Texas Instruments, and Toshiba. While the mobile market is a potentially huge market, Nvidia will have to work hard to get its piece of the pie.
This company is a stock options (ab)user. Beware of dilution. I have fully expensed options in my EPS estimates though so I think I have been fair.
Nvidia regains market share as new discreet graphics chips gets traction. This may start to happen as soon as next quarter, but certainly with 12-18 mos.
Tegra gets some traction in 2H 2010 or 1H of 2011 as it powers tablets and Android smartphones.
Intel, Broadcom, Qualcomm, etc. bids for Nvidia (possible but not counting on it).