Cost of Surveilling the Cloud

There is a new, careful, non-partisan study from the ITIF out today that has focused specifically on the question of how much the computer industry would lose due to privacy concerns. And their estimate turns out to be $35 billion over 3 years. That comes out to about $12B per year on average.

This matches well with my recent estimate. When I outlined a framework for the forces that affect Internet policy, I did a quick back of the envelope calculation for the strength of each force vector. My guesstimate was that the amount the computer industry stood to lose was about the same as what the NSA was spending on broad surveillance, and that was a number in the "very low double digit billions."

So, to the extent that this new estimate is accurate, it corroborates my previous guesstimate. Of course, such numbers are fundamentally unknowable, and the cascade effects of our cloud companies losing their worldwide dominance could well incur long-term costs that far exceed double-digit billions in the long run.

One twist is that the costs are not uniformly distributed over the three year time span. Instead, they are back-loaded, as there is quite a bit of inertia for cloud customers to switch providers. The losses in the first year are only a few billion, while the losses in the third year are on the order of $20B. So, expect the computer industry to take an increasingly pro-privacy stance as its bottom line takes an ever greater hit.

Overall, the good news is that the numbers, no matter how one looks at them, are sufficiently high that the computer industry is motivated to intervene. And they have already held one secret meeting with Obama to voice their concerns. That's the force vector in action, pulling the Internet towards a freer, better place. Let's hope that this evenly matched tug of war yields some tangible results soon.

Share on Linkedin
Share on Reddit
comments powered by Disqus