I’m on a long plane ride to Phoenix. Outlook corrupted its .ost file so I can’t do email. I don’t want to see another Keanu Reeves and Sandra Bullock movie. My books are in the checked luggage. Time to listen to music and think about trends.
Andy Grove famously said that only the paranoid survive. It’s not clear to what extent Intel’s past success has been driven by paranoia and continued innovation vs. the x86 compatibility requirement for PCs stemming from the relationship with Microsoft (and, let’s not forget, IBM) from 25+ years ago. Otellini certainly has big challenges to deal with. AMD is winning the server business so Intel has certainly had some innovation challenges. And now, there are macro-level movements that challenge the x86 compatibility requirements for PCs. This doesn’t just affect Intel. It will affect the entire PC industry.
The x86-to-PC binding has been a tight one since the beginning of the industry. Apple tried to get in the game with Motorola and then PowerPC chips with little success. The reason was software–most of it and the best of it was only available on x86-compatible machines. Throughout the history of computing, loose & late binding has trumped tight & early binding. The latter is simply more efficient from an economic standpoint being a super-set of the former. Hence, there are market forces trying to break the x86-to-PC binding and, in my opinion, the next 5-10 years will bring a major shift in the PC industry. The main driving force is Asia. The main enablers are a host of new technologies and approaches to delivering software & services that are rapidly maturing.
I’m blogging this on a plane so my research capability is limited. If I recall correctly, economists predict that about half a billion (!) new consumers from India and China alone will enter the world market in the next few years. I use “consumer” to denote a person whose disposable income is high-enough to enable her to consumer goods and services on the world market. The first computing device she’ll likely own (probably already does) is a phone. The phone won’t solve all of her computing needs because of its I/O limitations and, to a secondary extent, storage requirements. (IMO, compute power is not going to be a big issue for much longer.) So, at some point, a good portion of these 500M consumers will want a PC (a laptop, to be precise).
My expectation is that demand will be very elastic. The reason is that income inequality will remain very high in both India and China–most of the net new consumers won’t have very high disposable incomes but a strong desire to own technology that has dual use (entertainment and education/work). PC manufacturers will therefore be incented to offer very low cost machines. What’s the magic price? MIT has been going after the $100 laptop, albeit w/o a disk drive. A team in China is aiming for sub $200 on a MIPS-like chip. HCL in India makes a $200 PC already. Another Indian startup whose name escapes me is doing an even cheaper PC based on a mobile chipset. Walmart and Dell are already selling well-loaded laptops with drives for under $400. I don’t know what the magic price is but I’m sure it’s well within reach in the next few years, especially considering some of the options on the table.
So, how do you get the price of a PC down? The highest-cost (and producer margin) items are the IP-heavy parts, e.g., the microprocessor and the software. Hence, to lower the price of the PC without significantly reducing its capabilities you’d have to throw out the proprietary parts while preserving the key features of the end-user experience. Starting from the top of the stack:
- Increasingly, the software that consumers use these days is Web-based. AJAX and Flash create desktop-like experiences using SaaS delivery. In addition, major file formats (from Office to PDF) are becoming more open. There goes the need for Microsoft Office and many other downloadable applications. Go Web 2.0 with a few desktop-based tools such as backup, desktop search, etc. May be get OpenOffice, though I’m wondering whether it’s too complicated for the target audience. (Videogames will be hardest one tackle, both because of their compute intensity and because of their tight binding to the hardware and expensive video cards. Virtualization may offer a temporary solution but ultimately new types of videogame experiences may spring up. I expect Asian videogame companies will aggressively target this opportunity. Note how Cyworld doesn’t need the fanciest gamer box.)
- Why pay for Windows if you don’t need the MS desktop application software? Go with Linux, hacked to your government’s specs. Government specs? Yup.
- If you’ve gotten this far, then why do you need x86 compatibility? A Web browser with AJAX support can be ported to any HW/OS combo w/o much trouble. Adobe will automatically do the Acrobat Reader and Flash ports for any platform that has significant volume in order to maintain market leadership. The target customer is unlikely to want to buy as many HW add-ons as in the developed world, which will also help.
- Perhaps you even skip the disk drive, as the MIT approach suggests. This assumes a lot about connectivity but also removes a component with, relatively speaking, high failure rates.
The governments of both India and China have a very strong incentive to push for broad deployment of PC-level computing to build up an educated workforce and a high-tech industry around the consumers of PC content/applications. I’d expect them to use significant leverage–from protectionism to government-level bargaining–to achieve this goal. Given the numbers involved, why shouldn’t China back local companies to build everything from the chip up to the customized OS? At the volumes we’re talking about, all software vendors that have downloadable software will be pushing to port/certify their apps on the new platform. Certainly, I’d expect both countries to push new PC specs using some type of industry consortium approach because specs enable network effects. These fast growing economiest don’t have the time to wait around for markets to figure things out, at least in the beginning. They’d want to focus innovation first and worry about making the wrong choice later. It’s hard to really screw up, BTW. It’s not like PCs are rocket science.
Intel and Microsoft should be worried. Microsoft ultimately has a way out–turn Windows Live into the best set of network services for consumers and make them dirt cheap in the third world. Go Ray! What’s Intel’s way out? I’m not sure… The breakup of the tight binding between the software stack and the hardware in the PC business may be the end of Intel as we know it.
Another possibility is the merging of game console, set-top box and PC capabilities. The trouble I have with this version of the future is that (a) consumers loose portability (we’ve seen a clear trend towards laptops and away from desktops), (b) costs go up for the initial device and (c) there are usage conflicts–watch TV or watch YouTube or play a game?
The icing on this cake will be the education market in the developed world. Ever anxious to maintain leadership in new technologies, developed countries are starting to experiment with introducing technology earlier in the education cycle, all the way down to one-to-one computing starting around 4th grade. (In the US, Maine was the first state to do this. They went with iBooks. NH and MI are following.) Cost pressures and the blank slate that kids are (they don’t expect to use Powerpoint so why do they need it?) may push state’s and governments towards third-world-style PCs. Once kids in the developed world grow up with something different than the PC/Mac as we know it, everything will change. So, take the half billion people in India and China and add a few hundred million kids. Still don’t think this is a huge opportunity?
Who makes money? Well, anyone who’s in the media and advertising business will have a ton more eyeballs. On the software side, Adobe is a good bet, provided it sustains Flash momentum online and positions Flash Lite as the best presentation layer on mobile (remember, the Asian consumers will have phones first). Microsoft’s and Google’s success in software will depend on how they evolve their online apps. Then, whoever gets “blessed” as the chip provider for mass-market PCs in India and China may become the new Intel. I really do expect the governments there to play a hand in this… The same applies to chip-level integrators. Finally, there is a stunning services opportunity–think GeekSquad on steroids.
The weakest link in my argument is the implicit assumption about broadband deployment in Asia. I’m comfortable that it is not too outrageous because most of the new consumers will initially be in the cities and the cities are getting good broadband. Last but not least, 3G is big there.