Is the handheld market the PC revisited?

Posted: 04.18.05

We've all heard that the handheld market is just like the PC market was, and the analogy is based on processors getting more powerful and more applications being added to the platform. It's true that there are some similarities, and the handheld industry has benefited from some of the learning curves of the PC.

Among the similarities between the PC and the HH are color bit-mapped screens. The introduction of color bit-mapped screens gave the PC market growth a kick and it had the same effect on the handheld platform. However, cameras had a tremendous kick in the sales curve for handhelds, and they didn't have much effect at all on PCs. Similarly, PC screen sizes grew in size and resolution, while handheld screens are limited in physical size and are just now approaching VGA. SoCs are critical to handhelds, and hardly used in PCs.

But graphics, and especially 3D graphics, which helped the PC, will unquestionably help the handheld platform. The advances in graphics on workstations and PCs over the last 15 years have to a large extent been applied to the handheld platform in just a few years, albeit with certain reductions to save memory size and/or computational load.

And it's been that challenge, and subsequent successes meeting the challenges, that has attracted so many companies to the handheld market. Indeed, at this point the similarities start to accumulate. Just as we once saw in the PC market, we see an explosion in suppliers who either are building, or threatening to build 3D hardware accelerators. There are three ways to get 3D hardware acceleration to a handheld system: with a co-processor like the type Alphamosaic, ATI, and Nvidia are offering; with an application processor with dedicated graphics acceleration like Qualcomm and Renesas offers; or with an SoC like TI and Renesas are offering and Freescale, Intel, and Qualcomm and are working on, which have embedded IP from some graphics IP company like ATI, BitBoys, DMP, Falanx, or Nvidia.

We now have 33 companies that are in one or more of these categories (but counted only once). Contrast that with the PC market, which reached a peak of 75 companies offering a 3D accelerator, and which has now consolidated down to seven including IGC king, Intel.

FIGURE 1. Graphics market value.

The graphics chip market, which extends to other devices besides PCs such as POS terminals, medical, and industrial systems, and gaming machines (which in total are about 20% of the PC/workstation/server market), is worth something over $5 billion (see Figure 1).

Although the handheld market will exceed the PC and Other market in annual shipments by 2.5:1 or more, the ASP of handheld graphics is much smaller than the PC's graphics ASP, and will sink to an order of magnitude smaller as the market shifts from coprocessors to IP and SoCs. The PC and Other market ASP will not gain much in the coming years, due to pricing pressures and more IP-embedded sales (e.g., all the new game machines).

FIGURE 2. Handheld graphics market vlaue.

Our forecast for graphics accelerators, discrete and embedded, in handheld devices shows dramatic growth this year, and then a decline due to the move to embedded solutions (see Figure 2). Note that this curve, although counter to every market curve you've ever seen, and certainly not in agreement with the overall handheld market curve, represents just the graphics value and not the total multimedia market or the total sales of the companies involved in these markets. For example, companies providing graphics may also be providing the total SoC and/or the baseband processor—so don't panic, the sky's not falling.

The point I'm trying to make is that, although we are predicting incredible growth in numbers for handhelds and for graphics/media accelerators for those handhelds, the platform ASPs and the subsequent percentage of the BOM for the graphics/media accelerators is not as large as the venerable PC and Other market.

Companies not doing well in the PC space, therefore, are not advised to go running into the handheld market with its smaller ASPs and overpopulation of suppliers. Yes, this is still the early days of handheld graphics, and yes, it is during these chaotic periods that great strides and sometimes fortunes are made, but it's going to be a little different this time around. The companies already in this market—the DSP, baseband, and radio suppliers—are already big, really big, companies. They are not going to be muscled out by clever pixel-pushers from PC land. They will assimilate those companies or just squash them. The handheld market is, and will be characterized by partnerships, licensing deals, and a lot of acquisitions and mergers.

The handheld market isn't just like the PC, not at all. The PC market was wide open and wild; the handheld market comes out of a staid and established industry run by giant companies with dramatically different business models, qualification procedures, regulations, and standards, and the most incredible market growth ever seen. If you can hitch onto this train it's going to be one hellofa ride.