The Moat in DoD’s Eye: Starlink, DIU, and Innovation

The Moat in DoD’s Eye: Starlink, DIU, and Innovation

To be very clear, Musk’s attempted shakedown of the DoD at the expense of the Ukrainians in the middle of a conflict is downright stupid. The DoD will remember just how fickle Starlink's service is because one man can’t take criticism. Based on that experience alone, the DoD will NEVER primarily rely on Starlink for military satellite communications. Instead, it will actively encourage other, more reliable commercial partners (it’s already doing so). The U.S. military has paid excessive amounts of cash for reliable, slow communications systems. Musk just made that decision more straightforward.

Ukrainians may also remember his words, in conjunction with other less inspired words about appeasement. Both had the potential to kill Ukrainians. No one will blame them for dumping Starlink when they don’t need it anymore. They’re already fighting off a regional dictator and are probably very willing to keep an overambitious tech bro/monopolist from taking root in their country.

Musk changed his tune, and Ukraine will still get service. But attempted backstabs will likely be long remembered.

It’s hard to believe that SpaceX deployed its first batch of 60 Starlink satellites a little over three and a half years ago in May 2019. The low Earth orbit (LEO) broadband constellation has a total of ~3,177 in orbit, with 3,132 working. The company is the only satellite operator with thousands of satellites in orbit (and it intends to deploy more). OneWeb nor SES are even close. Amazon has yet to deploy any satellite for its Project Kuiper constellation, instead moving the initial launch of its two pathfinders from one fictional rocket to another (which might be less fictional).

Since the launch of the first Starlink batch, SpaceX has also: introduced phased array antenna user terminals for linking with Starlink satellites; gained nearly 800K subscribers; contracted with a few airlines and cruise lines; deployed satellites with laser-interconnects; won U.S. military contracts; updated its phased array antenna; deployed in a theater of war (and worked–until lately); and gained condemnation from governments with dubious aims.

Those accomplishments and activities result from an aggressive company like SpaceX being out on the market. It’s conducting its business and gaining customers, and it HASN’T DEPLOYED ITS FULL CONSTELLATION YET. The company also plans on deploying a larger, more capable generation of Starlink satellites. At least, until its CEO decides not to.

Those facts and plans result from SpaceX pursuing a business plan that theoretically should help keep the company healthy and growing. Moreover, they present capabilities and opportunities for those willing to put in the work to exploit them. That, after all (and according to Peter Drucker), is a characteristic of innovation:

“The overwhelming majority of successful innovations exploit change.”

Competitive Innovation

With Drucker’s characteristic in mind, it’s puzzling to see the following complaint from an organization that has the word, innovation, as part of its name:

Without mentioning Starlink by name, Shimmin said “there is a particular commercial satellite provider in low Earth orbit, for example, that is several years ahead of the rest of the field. They do tend to try to establish interfaces, proprietary interfaces, that don’t necessarily want to play well with others.”

That comment throwing shade at Starlink comes from the Defense Innovation Unit. Yes, the same group produced the “State of the Space Industrial Base 2022” report I critiqued about a month ago. The latest DIU complaint stems from its efforts to develop a hybrid space network. It contracted with a few companies to start the development of the network in June 2022. Only one LEO commercial broadband operator is several years ahead of its competitors–SpaceX–and it’s unclear why Shimmin didn’t just outright say that name (it’s not a Harry Potter novel, for goodness sake). Why even mention the challenge publicly instead of quietly engaging with SpaceX?

A few words about the desire for companies and their products to “play well with others.” That’s not how it works. Amazon’s bouts with Microsoft for government contracts come to mind. There’s also Apple, a company notorious for not playing well with others, fighting against Google, Epic Games, the EU, Spotify, etc. Each of these companies implements strategies and counterstrategies against competitors. SpaceX does this, too, and has succeeded mainly because it’s in an industry where most space companies don’t operate as aggressively or competently. A good argument could be made that innovation results from not playing well with others.

There are other arguments to be made in favor of the “walled gardens” or “moats” that companies like Apple create to lock in customers and others, whether using proprietary hardware (Lightning) or, in the case of developers, Apple’s App Store. The arguments for locking in these customers aren’t great. They certainly aren’t focused on customer and developer needs and experiences, even if the company justifies them as facilitating precisely that. The moats and walled gardens allow a company not to focus on competing (and probably not innovate)–because they lock competitors out. DIU, through Shimmin, believes SpaceX is doing that.

Presumed Importance=Unwarranted Influence?

DIU’s complaint and focus on a proprietary Starlink interface seems at odds with that organization’s mission, which, if it involves innovation, exploits existing market products/services (proprietary warts and all). Innovators tend to see the weakness of moats presented in proprietary systems and implement ways to make them irrelevant. Innovators aren’t seeking to play well. Instead, they’d like to dominate the competition in ways that demonstrate almost an economy of motion through surprising exploitation.

Worse, though, is that while the complaint sounds as if it could be legitimate, it’s a bit hazy about what the problem truly is. DIU’s use of the word “proprietary” creates negative associations. Those associations cause a nearly knee-jerk, unquestioning response since proprietary stuff is terrible, after all. Using “interface” as the descriptor doesn’t tell readers what the problem truly is. Is it the mobile app for setting up the terminal? Is it the phased array antenna? Is it the user terminal? Is it the ubiquitous protocol used for communications? Is it ineffectively fusing felt together? It’s just not clear how Starlink technology is obstructing DIU’s goals.

Whatever the problem is, it also presents an opportunity–something DIU could exploit as an innovation unit. I would hope that someone from DIU has already talked with SpaceX about the challenge that the company’s interface seems to represent to DIU. And if SpaceX does nothing, well, that’s the company’s prerogative. For one thing, DIU and the U.S. military are a much smaller customer subset of Starlink’s business. So why should DIU even think it can influence SpaceX’s business decisions? Does DIU believe it’s just that important? Maybe if the contracts it parses out were more significant…?

Other Perspectives

SpaceX also has existing Starlink customers and contracts based on its current system, and changing the interface could negatively impact both. Other U.S. military organizations seem pretty satisfied with Starlink and aren’t bothered by whatever proprietary technology the company is using. So the moat is a non-issue.

The U.S. Army has been testing Starlink. The descriptions of the constellation’s performance are interesting but should perhaps have been expected:

“The benefit of this system is the amount of time it takes for the signal to go up into space and come back down; it saves us a lot of time on latency,” said Chief Warrant Officer 3 Kyle Neese, the senior battalion network technician for the 50th ESB-E. “The old military satellite communication system uses what's called geosynchronous [satellites], which orbit around the equator at a steady pace, but it takes a little over half a second for the signal to travel up and back down. With Starlink, it comes back at more than twice the speed.”


“So far, we’ve tested version one which comes with a dish, a power injector, and a router. The simplest way to hook the Starlink up to our kit is to take an ethernet cable and plug it into our router, which connects to our cradle point router and the cradle point router connects to our other systems as normal,” said Warrant Officer Corey McClure, a network technician for the 50th ESB-E. “This is the first of its kind to come on the market, and so far, it’s been great. It’s going to allow us to be a lot more flexible in terms of where we can set up and what we can do.”


“Starlink data rates exceed some of our current capabilities,” said Lt. Col. Mallory Wampler, the commander of the 50th ESB-E. “I know they're still doing some engineering and design modifications to make the equipment more ruggedized like our 1.2 meter Hawkeye terminal, T2C2 and our Phoenix E systems.”

Note that those comments were made before Starlink’s successful use in Ukraine. Nevertheless, they are very positive in their assessment of a commercial, partially-implemented constellation that costs less than some single military-spec satellites, is faster, requires little effort to set up, exceeds current U.S. military capabilities, etc. Moreover, these comments come from uniformed members of a signal brigade.

Those in that career work with military satellite communications and intuitively understand how to exploit Starlink to overcome the limitations imposed by those other systems. Unlike DIU, they aren’t focusing on building a future hybrid space network but on immediate mission challenges that Starlink is helping to solve. Some of those challenges result from decisions made in the past by organizations like DIU pushing complicated systems those signal brigade soldiers are forced to use. It’s almost as if the DoD would like to lock its soldiers in by using complexity and training them to deal with that.

DIU should use the brigade’s soldiers as inspiration for being innovative and understand how to use Starlink for its capability. Instead, it lamely attempts to shame SpaceX into changing what might be a fundamental business characteristic (I don’t know–but it appears DIU doesn’t know, either). The companies it has already contracted with for its hybrid architecture seem to have intelligent, capable people working for them. Some specialize in “mesh” networks. It seems they should be able to get Starlink to work somehow with their other networks without changing much.

But maybe those efforts are not necessary. The whole point is to have backup communications in the field, and Starlink has proven that it can provide that–if its founder can keep out of the way. Soldiers in Ukraine and the U.S. seem to believe it’s not just easy to use but provides critical and unprecedented capability in a small package. It’s important enough that the Army is “ruggedizing” some Starlink terminals for the battlefield.

Sometimes innovation happens without government guidance and supervision.