What would Jen-Hsun Do? February 17, 2006 Alan /.effect Dang, Pfucata.com

Introduction

Page:: ( 1 / 3 )

In the Q4 NVIDIA Earnings Conference Call held Thursday, February 16, 2006, NVIDIA CFO Marvin Burkett said:

“We will continue to make acquisitions. We will continue to acquire, hopefully for cash. We do still have a stock buyback in place, so we will continue to do that.”

This is an interesting comment to explore. Which companies would be ideal acquisition targets for NVIDIA? We can wait for analysts to speculate on the answer using publicly available information, but I think I’d make a good analyst as well. Figuring out which company is in NVIDIA’s sights isn’t too hard. We only need to make two assumptions. Any acquisition deal must:

1. Be beneficial to NVIDIA.
   
2. Be beneficial to the company being acquired.
   

What are NVIDIA’s Weaknesses?

The ideal company for NVIDIA to acquire isn’t a company that will add new products or services or “fill in the gaps” of the existing line-up. Instead, the ideal target for NVIDIA would be a company that augments something that NVIDIA is already actively involved with, but is an area of increasing growth requiring additional resources.

Core Competencies

So let’s look at NVIDIA’s core competencies. When it comes to desktop GPUs, NVIDIA has it set with the GeForce 6 and 7. While NVIDIA would probably jump at the opportunity to buy out ATI’s design team, there’s no reason for NVIDIA to buy out any of the “second-tier” 3D graphics chip manufacturers such as BitBoys or Imagination Technologies (PowerVR). They don’t need extra help. How about workstations? Quadro FX has set the standard for 3D workstation applications and Quadro NVS has gained substantial popularity in the financial world that those products are intended for. There’s no reason for NVIDIA to try to acquire a company like 3DLabs from Creative. NVIDIA already has sufficient talent and success in this area also.

What about motherboard core logic? NVIDIA does benefit from an acquisition here, but we already know the answer to this one: ULi. With ULi on-board, NVIDIA will have all the resources they need for continued growth of the nForce line. They already have the widest range of products for the AMD platform, and they’re already moving to Intel and the mobile markets. This leaves NVIDIA with just three markets that they are actively involved in but still have room for substantial growth.

Page 2

Page:: ( 2 / 3 )

Smart phone / Cell phone / Portable Media Devices

Professional Video and Broadcast

The professional video and broadcast market is a high-margin growth opportunity. NVIDIA knows this and the success of the Quadro FX SDI boards have set the pace. NVIDIA’s SDI solutions power NASCAR Racef/x technology, the “1st & Ten” virtual first down line in NFL games, and the K-Zone virtual strike zone in MLB games. They’ve already got a lot of success in this market. However, in the conference call, Jen-Hsun mentioned that key growth of 2006 would occur in the professional markets of “HD content creation, HD video editing, and Broadcast.”

At the moment, NVIDIA does not have any products in the area of HD content creation or video editing; Quadro FX SDI is an output-only solution. In order to truly capture the market, NVIDIA needs a product with full HD/SD SDI input/output. Additionally, most boards rely on PCI-X. Perhaps NVIDIA's expertise with PCI-Express can play a role here. What gets more interesting is that a lot of “real-time HD editing” and transition effects can be improved with the use of the GPU shader pipelines. NVIDIA clearly has strengths to bring to this market, but they also have areas where they can gain valuable expertise through a company acquisition.

So which companies are at the forefront of HD/SD SDI capture? The list is long: Avid, Digital Voodoo/Bluefish444, Matrox, Canopus, Blackmagic design, etc. It could be any of these companies.

But what would Jen-Hsun do? He’d look to buy a company with expertise in HD/SD SDI video production. Unfortunately, there are so many to choose from. Let’s see what other markets are available to NVIDIA and if there’s a more efficient way.

Medical Imaging

The other area of high-margin growth opportunity is the medical imaging market. As more and more hospitals and clinics are moving to digital imaging, there is clearly room for growth in this market. Moreover, this becomes a self-sustaining market with regular upgrades because most of the systems are leased rather than purchased.

The technical challenges involved for developing medical imaging graphics aren’t difficult. Although a handful of devices are built around old SGI Irix designs, most of today’s medical imaging technology is based on a Windows operating system (which is surprisingly stable when you don’t have any of the normal stuff that gets installed). Likewise, the GPU itself only needs to drive a 5 megapixel display at 10-bit grayscale and provide a custom LUT that can auto-calibrate to a DICOM-standard gamma curve. Today’s Quadro’s can easily handle the rigors needed for medical imaging solutions.

While ATI has a presence in medical imaging market through a deal with Barco, NVIDIA has a fairly small market share in this arena. This is because most hospitals and clinics don’t buy their Medical PACS workstations on their own. They don’t go to Dell or BOXX to buy their workstations. Instead, the PCs come as a big bundle with the actual CT scanner or the entire PACS system (the storage server, network server, etc). So when a hospital decides to buy an AGFA, GE Centricity, KODAK, Siemens, Philips, or Sectra PACS system, they get whatever video card that those companies happen to bundle.

The Quadro would be perfect for these markets, but it’s really tough to break into the market. The medical imaging industry tends to be fairly cautious and prefers not to switch manufacturers for no reason.

What would Jen-Hsun do? If he wanted to break into the medical imaging market, he’d buy a company that has an established relationship with the PACS manufacturers.

So which company supplies most of the GPUs in the medical imaging market? It’s not ATI. It’s not NVIDIA. It’s Matrox.

What would Jen-Hsun do? He’d try to acquire Matrox.

Page 3

Page:: ( 3 / 3 )

NVIDIA has significant room for growth in the professional broadcast market as well as in the medical imaging market. While NVIDIA could try to buy out separate companies to address these two markets, Matrox is the one-stop shop for this sort of problem. Since Matrox is a privately held company, it is unclear how much Matrox is actually worth as a company. That said, Matrox has failed to produce any new graphics core logic since the Parhelia. In fact, the G550 core is still found in Matrox products today; the G550 is nothing more than an upgraded G400 with dual-texture-units per pipeline and small modifications. The Matrox G400 is an NVIDIA Riva TNT2-era product. We’ve seen seven iterations of NVIDIA GeForce hardware since then! It’s not unreasonable to believe that Matrox’s market value is low enough where NVIDIA can afford the acquisition.

A smaller, but still tangible benefit would be that the acquisition of Matrox would open up the door for a Canadian R&D center for NVIDIA. While it’s still 7 hours away from ATI’s headquarters, it’s possible that NVIDIA has failed to recruit key talent from Canada who do not want to move to the United States. As long as Matrox is “cheap enough” to acquire, the benefit to NVIDIA is clear. They gain instant market share in two high-margin growth opportunities where their existing technology (Quadro) can be augmented.

What’s in it for Matrox?

Matrox is therefore faced with two options. We already know that NVIDIA plans to enter the market of HD content-creation. If NVIDIA takes them on as a challenger, it’ll be bad news. Matrox lacks the resources to develop a flagship 3D GPU and they would be at a price disadvantage if they bought GPUs from ATi or NVIDIA as 3rd party. This means that while their strengths in software will help them combat NVIDIA in the short term, NVIDIA will quickly overtake Matrox as more sophisticated filters and transitions can be achieved with something like an SLI-Quadro. On the other hand, if Matrox joins NVIDIA, then they now have the advantage of access to flagship 3D graphics technology.

Summary

It’s evident that NVIDIA has room for significant improvement in the professional broadcast and medical imaging markets, and that a company acquisition could help them. Matrox is a leader in both of those industries, and unless there are two smaller companies that are valued at a lower cost than Matrox with equivalent technical and marketing value, or NVIDIA has recently hired new talent in these areas, it seems like Matrox is a perfect target for NVIDIA acquisition. NVIDIA has the means and the motive for acquiring Matrox. It's just a question of opportunity.

Even though this article is based upon reasonable assumptions, this article is still pure speculation. In fact, market analysts are often wrong. Many professional analysts believed that the original Xbox would sell 100 million units. Analysts just have to make the best guesses with the publicly available information. They put their reputation on the line each time they are asked to make a prediction. But if NVIDIA does acquire Matrox before the 2007, remember, you heard it here first.

I’ve heard that lawyers never forget the first case they lose. Do analysts ever forget the first time they’re wrong? What would Jen-Hsun do? I think he’d participate in the “Donate an Aston Martin to Alan Dang Project.” Now that I’ve gotten my first wrong prediction out of the way, hopefully it won’t sting if I’m wrong with this Matrox prediction. But if NVIDIA does acquire Matrox before 2007, remember, you heard it here first.