Who Gets High-Speed Internet Satellites First: Rural America or High-Frequency Traders?

Infrastructure isn't just roads and bridges anymore

Today President Biden unveiled his $2.25T (with a ‘T’) American Jobs Plan, aka the infrastructure proposal. There’s been plenty of debate on the plan’s merit, but this blog isn’t about that1. This blog is about high-speed satellites. The plan allocates $100B “to build out the nation’s high-speed broadband infrastructure to 100 percent coverage, including in remote and rural areas.”

Today the Wall Street Journal published a video about SpaceX’s new satellite internet service: Starlink. Starlink is currently in early beta testing, and the results have been… sub par thus far. In some places, it works reliably, but at lower speeds than the U.S. average:

In others, the download speeds drop significantly if there’s any weather interference like clouds, rain, and even wind. That’s a problem.

What’s the solution? More satellites. Tens of thousands of satellites. Literally:

To the point where a little inclement weather won’t stop anyone from having access to high-speed (sort of) in every corner of the world ever again!

Naturally, building and launching satellites ain’t cheap. That’s where the $100B from Biden’s plan presumably comes into play. With government subsidies being added to rapidly improving technology, high-speed internet for rural areas seems like a foregone conclusion at this point to me.

And Musk isn’t the only “World’s Richest Man” title-holder with this goal internet-for-all. Current “World’s Richest Man” Jeff Bezos has his own low-orbit satellite project, OneWeb Global, that aims to do the same thing.

This is all to say that high-speed satellite internet for rural areas is not an if anymore, it’s a how. As in, how fast can we get make it?

This brings me to an article published today by the same Wall Street Journal. This one was also about the quest for high-speed satellite internet, but the intended beneficiary was not rural areas. Not even close. It was high-frequency traders.

High-frequency trading represents ~50% of trades made in the stock market, and the name of the game for HFT firms is speed. So much so that they’ll spend a great deal of money to gain just milliseconds on the competition2. Below are the ways in which they transmit data to execute trades these days. Most firms use the subsea fiber because it’s more reliable than the speedier shortwave radio signals that travel through air and space:

These guys would rather lay a goddamn fiber optic cable across a motherfucking ocean than lose a fraction of a millisecond.

Enter high-speed satellite internet:

Of course, this is nothing new to HFT firms, they’ve been eyeballing satellite technology for years now. Anything to get an edge, right?

In 2016 a startup called LeoStat secured their first HFT customer after claiming they could transmit data from New York to Tokyo in 130 milliseconds using their network of satellites. Obviously, that’s a long way to go from the 29.6 millisecond time offered by shortwave radios, but hey, it’s a start right? Not quite. Investors didn’t like the progress they were seeing so they pulled their money and the project was scrapped.

But other firms that run global networks for high-frequency traders like BSO Network Solutions and Telesat remain interested in high-speed satellite networks and continue to devote resources to R&D.

This is clearly something that will happen in the long term. If we had $1 billion, we probably would have launched it already.

Michael Ourabah, CEO BSO Network Solutions

Right now, a London-based firm called Azuries Space Mission Studios claims to have designed a network of satellites that would be 20% faster than the current subsea fiber technology that HFT firms are using. Azuries’ network, Angel, would consist of just 111 satellites positioned strategically to transmit data from New York-London and Chicago-Tokyo.

So instead of using thousands of satellites zooming around the world at thousands of miles an hour to cover everywhere, like Starlink intends to do, Angel would rely on a concentrated constellation of satellites that were designed specifically for HFT.

Comparing Starlink and Angel is like comparing a Tesla and a Formula One car. A Tesla is designed for the consumer market whereas a Formula One car is designed to win on a small number of very specific circuits.

J Cooke, founder Azuries

My knee-jerk reaction is: pump the breaks, dude. You designed a constellation of satellites out of your space boutique. That’s like drawing a picture of food and saying you’ve got a plan to cure world hunger.

He also says that with funding of $155M he could have the network up and running within 3 years. He has launched zero satellites to date, by the way.

I don’t know who Cooke’s life coach is but he deserves a raise.

But hey, maybe it’s the real deal. Who am I to say? I’m not a satellite expert.

The point is that as farfetched as some of this stuff sounds, the inevitability of a constellation of satellites being used to transmit data for HFT is the same as it is for rural areas: it will happen.

Now, I know it shouldn’t matter, but I’m curious to see who will achieve their goal first? The government-subsidized giants like Musk and Bezos feeding the technologically impoverished in every corner of America? Or high-frequency traders ushering in the next wave of technology on Wall Street?

What do you think?


Personally, I think it’s far too excessive. That, combined with my lack of confidence in the government’s ability to competently appropriate funds has me leaning against it. It’s certainly well-intentioned, but it needs serious trimming.


Flash Boys by Michael Lewis is a great read on HFT