Every year, the Internet community waits breathlessly for Mary Meeker's Internet Trends presentation, where the celebrated Kleiner Perkins Caufield & Byers tech analyst lays out what's happening in a series of insightful charts and graphs.
I've embedded her 2014 presentation below, and as always (see all Meeker's Internet trends presentations going back to 2001 here) it's chock full of interesting analyses, covering everything from the growing "mobile gap" to the future of big data. (I may delve deeper into those charts in future Network World posts.) But for me -- and I think the majority of Network World readers -- the most compelling part of Meeker's presentation addressed falling prices for many of the basics of modern enterprise technology.
KPCB Internet trends 2014 from Kleiner Perkins Caufield & Byers
You can check the numbers in the presentation for yourself, but the points boil down to this: "Processing costs falling rapidly while the cloud + availability rise."
Specifically, storage costs have fallen 38% annually since 1992, compute costs have declined 33% a year since 1990, and bandwidth costs have dropped 27% a year since 1999. Even smartphone prices are falling, down 5% a year since 2008.
It's easy to get caught up in the year-to-year economics of IT budgets and so on, but it's worth stopping for a moment to think about that. The prices of the key things enterprise buy have fallen by a bout a third every year for the past 15 to 25 years or so.
The compound effects of those trends is so enormous as to have reshaped amost every aspect of enterprise IT. The cost of 1 million transistors was $527 in 1990... and a nickel in 2013! A gigabyte of storage cost $569 in 1992, but just 2 cents in 2013. And 1,000 Mbps cost $1,245 in 1999 but just $16 last year. Smartphones have dropped in price from $430 in 2008 to $335 in 2013, a significant change that's gone largely unnoticed by comparison.
This all kind of puts petty budget fights over 3% salary increases into perspective, doesn't it? More to the point, it forces forward-thinking enterprises to make strategic and tactical technology decisions based on the assumption that computing, storage, and bandwidth are not scarce commodities, but abundant ones. That the real costs of enterprise technology no longer lie there, but reside elsewhere in the technology stack.
The companies that figure out the implications of that fundamental shift will gain an increasing competitive advantage over time. Of course, it's far from obvious where the next set of critical costs and challenges will be, but it's still a very big deal to understand what they're not. Finally, it seems to me that trends like this ought to be spreading a little joy and optimism in our world. At least, I know I felt better to seeing those cost lines scamper for the lower right hand corner of the charts.