Despite having execution problems and ferocious competition (detailed in my last article), I still think NVIDIA Corp. (NASDAQ:NVDA) will still be at least moderately profitable in the long-term, and the 30% drop in its stock price following one bad quarter is an over-reaction. NVDA has several factors in its favor.
NVDA is unlikely to lose market share, and might even gain some share. Both technological and market factors account for this. Technologically, GPU chips are second only to CPUs in complexity, meaning that competition from Asian fabs are unlikely, and the only competition is likely to remain ATI/AMD and Intel (NASDAQ:INTC).
ATI has a decent reputation with its Radeon cards, but its parent AMD is bleeding cash at an unsustainable clip, and may soon find itself unable to support R&D in CPUs and GPUs simultaneously. While NVDA has stumbled recently, I consider it very likely that in the long-term, ATI/AMD will begin to fall behind technologically (if it doesn’t implode completely first).
Intel, of course, poses a much more formidable threat. It is certainly capable of throwing enough money in R&D to at least equal NVIDIA technologically. However, Intel already holds a large market share in GPUs, and PC makers will not want to give additional market power to Intel if they can help it. The only way NVIDIA can lose market share is if it’s chips prove to be consistently unreliable. Considering that NVIDIA backs its chips with strong warranties, management at NVIDIA is unlikely to tolerate a repeat of this quarter’s chip failure fiasco.
The demand for graphics chips is set grow rapidly. Future operating systems are likely to be increasingly graphics-intensive. Already, Windows Vista’s Aero GUI has graphic card demands previously seen only in games, and the Mac OS X’s Aqua GUI relies on the NVIDIA GPU in every Mac for its sleek look. TIn addition, now that the Blu-Ray/HD-DVD battle has concluded, sales of Blu-ray movies will place additional demands on graphics cards during DVD playback.
While recent GPU sales has stagnated (along with CPU sales and the tech sector in general), this is due largely to US domestic sales slowing down as the market for desktops nears saturation, and consumers generally feel that that current software and hardware are adequate and there is no need to upgrade. However, upgrading cannot be indefinitely postponed, and at some point in time, advances in hardware and software will make it worth the while to upgrade. In the meantime, the rare bright spot in the US economy is the export sector, and export of high-end chips is likely to go up as increasing numbers of middle-class consumers in developing countries buy their first PCs.
NVDA has competent management. CEO Jen-Hsun Huang is a co-founder of the company, and was a chip designer. He owns 5.1% of the NVIDIA (including options), and is known to be a demanding boss to work for. Before the latest stumble, NVIDIA has historically taken great pride in never missing a production deadline. As the entire chip industry hits the thermal and power consumption bottleneck, it is likely that NVDA will have to retool its chip designs, either going the Intel/AMD route and putting multiple cores on a single die, or the ATI route and putting multiple GPUs on a single card. Having a CEO who understands the technical complexities of this tricky transition is a big plus for NVIDIA.
NVDA has $1.6 billion in cash and no debt, and has a market cap of approximately $6.4 billion (assuming $11.50 stock price). In 2007, it had $4 billion in revenues, $836 million in operating income (20.9% operating margin), and $64 million in interest income. After $103 million in taxes (a 11% tax rate), net income is $798 million in 2007. In 2008, it had $1.1 billion in revenue in the first quarter, and has just reported a likely revenue of $875-$950 million in the second quarter, with a $150-200 million charge for bad chips.
Assuming that total revenue for 2008 come in at $4 billion, and margin drops to 17.5%, operating income will be $700 million, reduced to $500 million after the $200 million bad chip charge, and assuming a 10% tax rate, net income will be $450 million. Since NVDA has excess cash of $1.6 billion, the enterprise value of the company is only $4.8 billion, which works out to an earnings multiple of 10.7. Bear in mind that this low multiple was derived off an artificially low earnings hit by a one-time $200 million charge.
Furthermore, NVDA’s revenue has grown by 30-40% annually before stagnating in 2007-2008. Even if margins were to remain permanently depressed at 17.5%, resumption of even modest revenue growth as the rest of the world demands more graphics chips should bring the earnings multiple up. I believe that, at a minimum, NVDA should be worth 15 times its long-term earnings of $700 million (sans charges), making it worth about $21.80. The current stock price essentially assumes no revenue growth plus enormous margin loss, which is highly unlikely.