New ways to trade data

IN 2016, according to Cisco, an American technology group, the volume of data flowing through the internet each month passed a zettabyte, enough to fill some 16bn 64GB iPhones. By 2025 it will be many times greater. Immeasurably more data sit outside the public internet on company servers. Most of these data are valuable information, which means that people are keen to trade it.

Typically, data deals are at present worked out between someone holding the information and those who want to extract insights from it. For instance, Uber has deals allowing many cities to access data generated by its fleet of drivers. This helps city planners understand traffic flows.

Such deals can be clunky to set up, however. They tend to concentrate on datasets that hold obvious value. They may also involve data physically moving between one computer and another, which makes it vulnerable to abuse, as in the recent scandal surrounding Cambridge Analytica’s use of Facebook data. New schemes, created as part of the crypto-currency boom, aim to change all that.

One of these, called Fetch, was announced on March 28th. It was founded by Humayun Sheikh and Toby Simpson, respectively an investor in and early employee of DeepMind, a British artificial-intelligence company that is part of Alphabet. Instead of sending blobs of data around the internet, Fetch allows an organisation to ask questions about datasets residing on another organisation’s servers. The network will keep track of which datasets are used to answer these questions, allowing future queries to be directed to the right place automatically. A local weather-forecasting group, say, might have its algorithm tap into performance data from a power grid to improve its predictions (the frequency at which electricity moves in cables is related to the air temperature).

Fetch, which plans to launch itself in 2019, is a non-profit organisation and sees itself as a custodian of this question-and-answer network. Payments to ask questions will be made in the form of digital tokens. Unlike some make-a-buck crypto schemes, Fetch says that its tokens will not be available for public purchase until it is up and running, and has demonstrated its value. Fetch’s financial backer, Outlier Ventures, has bought future rights to these tokens rather than shares in the company. The idea is that as more organisations make their data searchable, and more people pay to ask questions with tokens, the value of the tokens will go up.

Another project, called IOTA, operates a similar scheme. Bosch, a German engineering giant, thinks that it could use IOTA to earn money from the data its domestic appliances generate. It has bought IOTA tokens through its venture-capital arm.

These new data markets face stiff challenges. Maintaining individual privacy and monitoring questions to prevent corporate leaks will be difficult. The cryptography securing the network needs to be airtight. Perhaps the biggest challenge will be convincing people to use them. The take-up of similar efforts, such as Solid, developed by the Massachusetts Institute of Technology, and Maidsafe, a Scottish data-sharing network, has been lacklustre. Nevertheless, Fetch says several large European firms are lined up to give it a go. And, like other digital currencies, IOTA’s token value has soared and fallen.

This article appeared in the Science and technology section of the print edition under the headline "Exchange value"/economist