Monday, January 24, 2011

Android in trouble!

Over the weekend, the blogosphere has been buzzing with "evidence" of Google having copied code from Sun/Oracle's Java and even worse, changing the license terms on that code. While some people see this as a nail in the Android coffin, several others see this as irrelevant, as the code in question is only related to testing, and unlikely to have ever found its way into shipping devices.

The point is, it is entirely irrelevant whether the code in question is important, or whether Google really copied the code, or some third party copied the code, or whether this code ever found its way into any shipping devices. The findings over the weekend will hurt Google quite bad in more important ways.

In these sort of cases, it is entirely upto the plaintiff (Oracle) to prove that Google violated their patents/copyrights. And this is a stiff hurdle. Very rarely is there a smoking gun that can be used as evidence. Especially in a project like Android, that has roots in Apache Harmony Project, and has also seen some third party contributions, it is difficult to pin the blame on Google, even if some violations can be proven. So Oracle has a really tough job on had.

That being said, Oracle has one real big advantage in this situation - time is working for Oracle. The more delay there is in resolving this court case, the more uncertainty there will be in the Android community and amongst hardware manufacturers. And this weekend's news only serves to increase that uncertainty. Already we are seeing articles on Android friendly sites - that things are not looking good for Android. This is the sort of PR that Android just cannot afford. This is the sort of situation that can easily go out of control and start impacting Android.

Microsoft Windows Phone 7 could be the unexpected beneficiary of this confusion. I expect manufacturers to start hedging their Android bets with either WP7 or other alternatives like Bada.

From a legal standpoint also, Google is on weak ground.
- The Counter-suit approach is not viable here - Google has very limited patent portfolio, and that is concentrated heavily around search. They can't find anything to beat Oracle with.
- The clean room defense will not fly - Google has hired way too many Java engineers from Sun - including Sun's former CTO - Eric Schmidt.
- "We didn't do it" will not fly - When it came to the original lawsuit, Google tried to blame Apache Harmony for the 6 pages of code that were alleged to have been copied. Apache promptly pointed out that that code did not come from Apache, and is from Google itself.
- Delayed enforcement will not fly - Oracle sued within months of acquiring Sun. Android itself has just been around for 2 years, and over those 2 years, Sun has made enough noises about not being happy with what Google has done with Java.

But all this is not important now - the main thing is, PR and sentiment tide just turned against Google!

Friday, January 21, 2011

What can $3.9B do?

In Apple's latest results conference, Tim Cook mentioned that Apple has invested $3.9B to get an edge on the competition. He also compared this to Apple's move in 2005 to prepay for enormous amounts of Flash Memory, to not only get a very good price, but to also create scarcity for the competition.

I think the comparison to 2005 is very ingenious. This time around, I don't think most of the money has gone into prepayments, but towards Capital expenditure. Like partnering with other companies to setup factories to make components exclusively for Apple requirements.

For instance, take the iPad - there are rumors going around about the resolution of the iPad going up 4 times - to Retina range (even though it will be only 260 pixels per inch, the iPad is generally held further away from the eyes than an iPhone - so even 260 pixels qualifies as Retina). Displays in this league are currently used for Medical monitor purposes, and for professional film and photo editing - and can cost several thousand dollars a unit. So talk of Apple using a Retina display on the iPad seems improbable. However, the point to be noted, is that if production is done at massive volumes, there is no reason an iPad Retina display cannot be produced at a cost that is just marginally higher than the current display cost. This is the sort of thing that Apple would invest in.

My guess is that the 3.9B is spent on following areas:

- Partnering with Samsung on a Chip Fab Plant for A4 and its successors. Considering a new fab plant costs $2B, Apple is likely to have invested $1B in this. This is likely to be an aggressive move to take the A4 to 28 nm process for now, with plans to move to 22 nm eventually. We should see dual core processors with dual/quad graphics cores integrated. For iPhone and iPad, this chip will be clocked at 1GHz, while for Apple TV, this chip will be clocked at 1.6GHz (which will make the AppleTV a viable alternative as a full fledged gaming platform - able to hold its own against the PS3, Wii, and XBox 360).

- Partnering with LG to make Retina displays for iPad to begin with, and later on for Mac Book and iMac as well. Again, Apple is likely to have invested $1-1.2B in this. I expect a move to E-IPS to reduce power consumption via more efficient backlighting, despite higher resolution.

- Partnering with Samsung or Toshiba to build Flash memory. The thing is, everyone knows that the future of storage is Flash, and that Flash is likely to replace HDD for most purposes in about 5 years time. This is the sort of thing that Apple has the power and clout to make happen a lot quicker. All the technology is in place and is mature - the only thing required is to migrate the technology to a more efficient process, and to scale up volumes dramatically. My guess is that this takes about $750M USD.

- With aggressive moves on processors and display, the new iPad will likely see much higher power requirements. Unless Apple dramatically ups Battery efficiency, this will the battery life would drop. I see this happening with two approaches. A big investment to make custom batteries for iPad, with marginal improvement in chemistry. Any increase in weight of the device because of extra battery capacity would be offset with reduced case weight because of moving to LiquidMetal/carbon composite. Should cost about $500B in terms of capital investment.

- The other, and more interesting technique that we are likely to see is the move to integrate Solar Power in the devices. In case of the iPhone, the move to have a Glass back for the case is a clear indication of the intent to harness Solar Power. The only reason we did not see it in the iPhone 4 already, is that the efficiency of the Solar Panels was too low (< 5% efficiency), and therefore made very little sense from a cost benefit perspective. Apple is likely to invest in Nano Technology based Solar Film Panels that will have an efficiency of around 22% or more. In the iPhone, this can be done easily by placing the film behind the glass on the back - but on iPad, I expect a more innovative approach, placing this screen under the backlight - thereby recapturing some of the wasted energy from the backlight. This positioning will also allow the device to charge itself from ambient light when not in use. I expect this to cost about $750B.

With these moves, Apple will distance itself further away from the competition. Already, no one is able to match the prices of the iPad with even remotely close to the specs of the iPad, and with these moves, Apple will make it tougher for everyone else.