contributor in the PC industry, desktops are seen as the rust belt of the business. Notebooks, with
their skyrocketing growth rates and plunging prices, have been getting all the attention lately. But
desktops are set for a metamorphosis. In the next phase of development, they will again become
an exciting sector where technological ferment will hold attention and drive sales.
Definition
In an era of converging technologies and form factors, it is perhaps first necessary to differentiate
a desktop from a notebook. Although there are anywhere from 14 to 20 distinct client (desktop and
notebook) form factors, depending on definitions, the great divide is really between desktops and
notebooks. The difference between the two all boils down to this: the notebook is mobile and the
desktop stays more or less in one place. One could quibble over the all-in-one desktop, which is
luggable or semi-portable. Another definition snags on the same issue: notebooks have displays
attached, but so does an all-in-one. However, an all-in-one is not sufficiently portable to be
considered a notebook. It has an external keyboard and mouse, aside from being in the 20-pound
range. You would never take this thing on an airplane. So, it is classified with desktops.
On the other side, the desktop-replacement (or transportable) notebook is pretty big and heavy,
clearly not something to take on the plane with you, and often ends up plugged into many things —
the network, speakers, an external drive, power — as it sits gathering dust in the only place that it
will ever reside. However, big as it is, this seven-to-nine pound monster with the 17-inch display
is clearly a notebook. It is self contained, and, in a pinch, could be closed up and taken
somewhere. Think of the writer in Manhattan who pecks at it day after day all winter, and when
summer comes, throws it in the car and takes it to East Hampton, where the system sits in the airy
third-floor room facing the ocean for the rest of the season as he tries to complete the Great
American Novel. The essential difference, then, is that notebooks move and desktops don't.
Notebooks' Rise to the Limelight
The worldwide notebook ratio (proportion of notebooks to total clients) stands at about 31% today.
About one in three clients shipped is a notebook. This penetration is up enormously from 1999,
when the ratio was about 20% or one in five. The shift in the consumer segment is even more
dramatic, with the 1999 ratio rising from 10% or one in ten up to nearly the same level as the
commercial segment today.
The rush to mobility is driven by the fact that mobility is a good alloyed by only three shortcomings,
all of which have been mitigated to some degree recently. These three limitations are price,
performance, and comfort.
Price
Notebooks in the late 1990s were twice the cost of desktops. The flat panel displays that they
required were often half the bill of materials cost and the hard and optical drives as well as other
components all carried miniaturization premiums. A number of developments have converged to
mitigate the price issue. One is the rise of flat panel use with desktops. Related to that is the
dramatic drop in panel prices. Although the price is CRTs, the old desktop display of choice, is
still below that of flat panels, desktop buyers embraced flat panels with passion once their prices
began to drop. For example, a 17" LDC display that cost $900-1,000 in 2001 goes for around
$250 today. Thus, the oranges-to-oranges comparison between desktops and notebooks is
easier now because the relative cost of a desktop display has actually risen, making the two
packages more equivalent. Meanwhile, as notebook volumes have risen, the cost of notebook
components (e.g., hard and optical drives) has fallen, both in absolute terms and relative to the
cost of desktop components. The single greatest contributor to these price declines has been
volume increases.
This holiday season, desktops and notebooks will be nearly at parity. Promotions for the $399
desktop have become common, and now we're seeing the $499 notebook creep into the product
portfolio of major branded vendors. A flat panel display easily makes up the difference between
the two. Expect to see a lot more of this sort of pricing.
Performance
The story of performance is circuitous. Desktops have always outperformed notebooks because
they are less constrained in terms of thermal envelope. New technologies are often large and
hot. It takes a year or two to put the same functionality into a smaller, cooler package. Notebooks
are heavily constrained by both physical space and available thermal envelope. For this reason,
top gaming parts, with their blistering wattage, are reserved for desktops. But a funny thing has
happened to performance. Except for a few constituencies, notably gamers, the available
performance in mainstream systems is now well above what normal application loads require,
beyond what most people need. This abundance of performance holds true for both desktops
and notebook. While desktops continue to outperform notebooks at the high end, both types of
systems are well beyond what most people need. Thus, the performance penalty to choosing a
notebook is less significant than it used to be. Sure, a notebook is not quite as fast as a desktop,
but most people don't need the extra juice.
Another funny thing has happened to performance. When Intel hit the 4GHz barrier, it declared that
it would henceforth take a new direction. No longer would it seek pure clock speed increases — a
metric it had touted for years — but would aim at total system performance instead. The inefficient
desktop architecture had come up against a major roadblock. Further increases in clock rate
produced unacceptable increases in power consumption and heat. At that point, the desktop
team took a page out of the notebook team's playbook and began to look toward "balanced
performance." The analogy is: a head with a big brain (the processor) needs to be supported by a
strong neck and body (the buses, core logic, and other subsystems). This philosophy makes
eminent sense. In addition, Intel was trying to groom the desktop for a role in the living room,
where being large and loud (from multiple fans, spinning at high speed, desperately trying to
dissipate the heat) would be considered uncivilized. Thus, the company began to steer toward
using mobile technologies in desktop parts, technologies that emphasized power savings.
Today, the gap between mainstream desktop and notebook performance is as small as it has
ever been, and in the future it will be even smaller. Meanwhile, the performance of both form
factors exceeds most requirements.
Comfort
The desktop has always had an advantage over the notebook in terms of comfort. After all,
portability must come at some sacrifice, and poor ergonomics, a cramped keyboard, and a
suboptimal pointing device seem like the immutable price to be paid for the ability to "clap it and
go." Desktops have monitors that can be arbitrarily large, big enough to see several windows
simultaneously at 100% size, reducing eyestrain, and desktop monitors can also be angled just
so and raised to a height that won't cause curvature of the spine. An external mouse can be large,
extra sensitive, and fit perfectly to the hand. After all, the real estate constraint is all but totally
relaxed. Audio can be run through high quality external speakers, giving a far better experience,
particularly for home users, who have PCs as often for entertainment as for any other application.
In the end, comfort, of the three comparative dimensions, may offer the strongest argument in
favor of the desktop.
Core Desktop Constituencies and The Future Notebook Ratio
So, with the notebook ratio climbing every year at a rate of 2-4% for the past five years, one might
wonder where it will all end. An analysis of this issue leads to a conclusion that 8-10 years from
now, the worldwide notebook ratio will approach 60%, with the developed regions (the United
States, Western Europe, and, of course, Japan) being as much higher, approaching 70-75%
(Figure 1).
Figure 1
long-term theoretical asymptote of, say, 65%, which stabilizes 20 years from now. This
asymptote represents the limit at which it is not possible to convert hardcore desktop buyers to
notebooks, those people for whom mobility is not a value worth paying or sacrificing for, indeed,
for whom mobility may in some cases actually be an anti-value, a negative.
This group includes gamers, who want performance at any price, workstation users, whose
needs are similar to those of gamers, white box assemblers, who like the modularity of desktops
for flexibility, many home entertainment buyers, who just want a system to play digital content in
the living room, and corporate buyers who outfit task workers. Each deserves some further detail.
Gamers
Although the truly serious gamers represent a relatively small group (less than 1% of all PC
buyers), this crowd needs the biggest heat envelope it can get. Cases for gaming desktops
routinely have capacities of between 40 and 60 litres, and these cases can have multiple
graphics subsystems, multiple processors, lots of add-in cards, and even specialized cooling
systems to keep it all from blowing up. The needs of this segment will always outstrip the
capabilities available in mobile form factors.
Workstation Users
The heavy duty tasks undertaken by these users — software engineers, CAD/CAM operators,
architects, designers — require maximum performance and capability. Their requirements mirror
those of the gamers, but in the commercial sector.
White box
There will always be some push from white box assemblers to maintain the desktop form factor.
Because it is more modular than a notebook, the desktop allows the assembler to change out
parts opportunistically or to meet some individual requirement of an end user. Although this
market will gradually become less important as notebooks take over the mainstream, white box
assemblers will continue to provide desktops to end users for many years to come.
Home Entertainment
This category is potentially quite large. Today, not many PCs reside in the living room, but as they
get quieter and better looking, and as media center functionality becomes more mainstream,
consumers will begin to accept PCs in the place where they relax and entertain. A desktop in the
living room will be like a stereo component, and, once installed, will not likely move. It will be
plugged into other elements of the entertainment stack and behave just like a civilized consumer
electronics device should.
Task Workers
Other than home entertainment buyers, a subset of task workers is likely to be the biggest
category of desktop users. These are workers whose employers do not want their systems going
anywhere. Managers of task workers would prefer it if the systems were nailed to the floor. Shift
workers come in, they boot up their applications and sign in, they work, they sign out, and leave
the systems to the next shift. These PCs are anything but personal, and their owners consider
mobility an anti-value.
The Future Desktop
In the future, the desktop will bifurcate into two distinct forms: the big honkin' case for the power
users, and the small desktop for everyone else. The box power users buy will be a direct
descendent of today's workstation or gamer PC. It's the desktop for the rest of us that will evolve
into something new and interesting.
Disaggregation
Particularly for home users, but also for task workers, the desktop will disappear into its
peripherals. The desktop, which was once thought of as a large box with things — a monitor,
mouse, and keyboard — attached to it, will disaggregate. Its parts will each be optimized for their
own sake. For example, the monitor for the home entertainment center will grow large and
potentially hang on the wall. It may get its data streamed to it over broadband wireless. The
keyboard and mouse will be wireless, esthetically pleasing, and may be supplemented by a
remote control. The box itself will be quiet and small and may be attached to the back of the
monitor or hidden away inside a piece of furniture. Various pieces may be across the room from
each other or even in other rooms.
Many technologies will help with this development. For example, flash memory is now reaching
densities that make PCs with no moving parts possible. Eventually, flash will be cheap enough
to use for non-volatile storage in PCs. Desktop processors will have increasing power-savings
capabilities built into them. These capabilities, which will allow the desktop to run more coolly,
will allow smaller chassis, which will be able to approach the size of a notebook. The economics
of notebook components will ameliorate as well. With the rise of notebook shipment volumes,
desktops will be able to incorporate notebooks' smaller hard and optical drives economically.
Conclusions and Recommendations
Although notebooks have long since stolen the limelight from desktops in the public eye and in
some cases are already more important in terms of revenue and profit contribution, it is too early
to count the desktop out entirely. First of all, desktops maintain some long term advantages over
notebooks, particularly in the area of comfort, and, second, certain audiences are predisposed to
buy desktops for other reasons (i.e., anti-mobility, modularity, heat capacity). Notebooks will
continue to rise, taking share from desktops, but a core of buyers will still want desktops.
The desktop itself will morph as technologies allow new versions to be smaller and quieter.
These new desktops will all but disappear into the optimized peripherals to which they connect.
Many OEMs have greatly deemphasized or abandoned desktops entirely. Some major vendors
— notably Dell, HP, and Gateway — are still slugging it out in what has become a thankless
ground war. Still, those vendors able to ride the wave of technological development and changing
economics to create a newly relevant desktop form factor stand to inherit what will still be a major
market, even a decade from now.
