If there is one thing that can be said of Google, it is that they play a great chess match. Many moves today are done for the greater chess board position later in the technology game. Some of Google’s pieces in play: Bandwidth at the ready, Google Video & Youtube, Open-Source Android, Voice recognition superiority, semiconductor purchasing. Pieces lost in this round: Cellular bandwidth purchase, and a stuck Google Voice piece. Result: Google repositions it’s mobile strategy with a more encompassed target of appliance-based technologies rather than a just a mobile phone based focus.
There is much speculation being posted recently looking for the reason why Google recently purchased Agnilux. a super secret early stage start-up in the semiconductor sector. Agnilux was founded by members of Palo Alto Semiconductor. These founders left after the purchase of that company by Apple for $278 million in the Spring of 2008. Steve Jobs stated to the New York Times that Apple made the purchase of Palo Alto Semiconductor to promote and manufacture system-on-chips for iPhones and iPods.
Many of the technology pundits say that Google is interested in the Agnilux purchase to assist in possibly creating a new generation of power friendly servers. (A company founded with experts in mobile friendly chips are now focused on making energy efficient server chips? Maybe, but I don’t buy it 100%.) I fail to see Google’s motivation to invest in these types of semiconductor companies to install these (lower powered, energy efficient) chips into their server farms for basically two reasons. Reason one – Efficiency: Isn’t it more efficient to power down a mutli core chip than to ramp up dozens of more energy efficient chips to handle the same job? Reason two – Long term viability: Even if today’s use is not maximizing out the server’s processing power, future data demands will be more, not less. You will have the processing headroom in place to handle the additional demands before you need to do a hardware upgrade. I fail to see the long term ROI being viable on building a server cluster with more underpowered processors in spite of their energy efficient appeal.There’s a reason we use a school bus to transport kids rather than fleets of Yugos.
Google’s recent purchase of Agnilux, along with their purchase of PeakStream (in 2007), starts to position Google with hooks into the real physical world of hardware needed to develop internet appliances. Using the energy efficient Angilux processors coupled with the PeakStream GPU Platform, gives Google a nice position on the technology chessboard. The forward thinking moves of Google will leave them with a nice checkmate position as we continue to play the mobile and set top appliance game, not greener servers.