As important as network links are in today's hyperconnected world, the fact that some three or so billion dollars (a sum which could buy a lot of other things, from providing millions of people with clean water to patching up bridges and power plants) was easily found for a 60-millisecond speed increase is mostly to do with being massively useful for high-frequency algorithmic trading. Objectively, it makes no difference whether a transaction between London and Tokyo takes 170 or 230 milliseconds to take place—though whether the transaction gets in before or after the rest of the market is the difference between profit and loss. Already, a significant part of the global financial system resembles a game of Core Wars played with real money; large amounts of wealth are conjured into being in finance houses by wartrading bots created from GPUs and FPGAs by extremely well-renumerated geeks, and many of the brightest minds of our age are eschewing the vows of poverty which go with the academic life or the modest salaries promised by pure science and medical research and instead going into creating the bots that will outcompete the current generation of bots. As such, there's all the money in the world for faster network links between global financial centres, and the Arctic link should tide the traders of London and Tokyo over until someone opens a finance house on Novaya Zemlya or in the Canadian arctic and beats both sides to the punch. After all, 299,792,458 metres per second is not just a good idea; it is about as iron-clad a law as there is.
The article suggests that, while algorithmic trading will benefit from the link, it will also be open to general traffic. Though, since the reduced latency is a competitive advantage worth countless millions, I wonder whether civilian access to the cable will be specially configured to slow packets down by a few milliseconds.