Near Zero Pricing and the Internet of Things

Lots of things are possible if you rip an order or two of magnitude out of the cost of enabling technologies or whole products.

The Internet of Things is possible because we have reduced some sensor costs about 50 percent in a decade. At the same time, computing device energy efficiency has doubled about every 1.5 years.

Also, bandwidth costs have dropped 40 times in 10 years, while processing costs have dropped 60 percent.

In 1981, for example, storage cost was something more than US$100 for 128 kilobytes. But storage prices drop about 33 percent per year.

As much as some people complain about the cost of bandwidth, the products people now buy are not the same as in 1980 or 1990. People buy the equivalent of DS3 circuits (45 Mbps) or dedicated optical connections, not dial-up modem lines.

So cost per bit, cost of a computation and cost of storage have plummeted.

In 1998, the cost of Internet transit across the backbone was about $1200 per Mbps, per month. In 2015 the cost is likely to be 63 cents per Mbps per month.

As with so many other parts of the telecommunications business, one wonders where the trend to marginal cost pricing, which for most bandwidth services is near zero, also means that one more category of product, namely Internet transit, is destined to become a feature of a network or service, rather than a product.

That seems implausible, since networks are extremely capital intensive. But marginal cost, once the networks are built, is quite literally “near zero.” And if all retail pricing trends, over time, towards marginal cost, near zero pricing is inevitable.

That means stress and change for bandwidth providers, but enables the Internet of Things.