Musings About: SpaceX and Swarm Technologies

SpaceX’s acquisition of Swarm Technologies last week will be interesting to observe as it plays out. SpaceX is known for its reusable rockets and its Starlink broadband constellation. The company usually builds its rockets and satellites with little to no outside help. SpaceX’s move to acquire at all is unexpected.

Heart of the Swarm

Swarm Technologies’ focus is different from that of Starlink’s. The company’s mission is straightforward and appears achievable for Swarm: Connect people and devices anywhere, at all times, at the lowest cost. Swarm competes with companies such as Iridium, Globalstar, Hiber, Orbcomm, and others. Similar to SpaceX, Swarm builds its satellites and most of its ground station componentry in-house. This may be simpler for Swarm to accomplish with its satellites than SpaceX with Starlink, because of the size of Swarm’s satellites.

Swarm manufactures and uses some of the tiniest of satellites (called SpaceBEEs), 400 grams, to form a global, if limited, communications constellation. About 650 SpaceBEEs would add up to a mass equivalent to a single 260 kg Starlink satellite. The satellites are small because Swarm wanted to lower the costs of launching them by lowering their mass. According to Sara Spangelo, one of Swarm’s founders:

“If you make the satellite smaller, the overall constellation cost gets much lower, and we’re at a point now where we’re able to put up our entire network of 150 satellites for far less than our Series A, which was $25 million…” (~19:15)

$25 million is less than the cost of four Orbcomm OG2 satellites and less than the cost of a single second-generation Globalstar satellite.

Swarm’s proposed 150-satellite constellation uses two-way low-bandwidth communications equipment (1 kbps) between isolated “Internet of Things” (IoT) devices and ground stations linked to the rest of the world. Swarm’s satellites (currently ~121) serve as “store and forward” nodes, basically communicating with the constellation’s ground stations and customers when they fly overhead, saving the data for transmitting to one or the other at the appropriate time.

For those wondering about the kinds of devices used in IoT and subsequently requiring Swarm’s satellites, one example would be a water pump in the middle of nowhere. That water pump would be critical for communities immediately around it--but those communities don’t generate enough income to justify having a “pump expert” on site to maintain and repair it. However, a small solar panel powering a small computer (Raspberry Pi) or controller (Arduino) and sensors for water pressure, pump resistance, etc., could be attached to the pump to monitor its condition. When the sensor generates an error code, it alerts the closest expert to travel out, inspect, and repair the pump.

Of course, the computer/controller requires a connection to a network, which is where the SpaceBEEs come in. Swarm makes it relatively simple to hook up a communications board (the company calls it a “Tile”) to these devices. When paired with an antenna, the Tile provides communications between the device and the satellites. The error code/status of the device gets pushed from the device to the satellite, which keeps it in memory until it gets to the nearest ground station. At that point, the SpaceBEE pushes the device’s message to the ground station, which then gets shunted to the appropriate recipient.

That’s just one example, but basically, anything operating remotely without terrestrial communications networks available may benefit from Swarm’s offering (and the offerings of other space IoT companies). As in the space industry, many consultants are willing to project the growth of IoT devices. Suffice it to say, the projections are optimistic, going into the tens of billions within a few years from now. And, like the space industry projections, it would be prudent to take those IoT projections with a salty grain. Swarm, however, seems to believe that enough growth in IoT will occur to justify deploying 150 SpaceBEEs.

A Bold Move?

By the time the news broke about SpaceX’s acquisition of Swarm, Swarm was well on its way to obtaining its goals. The company had deployed ~81% of its planned constellation, and Spangelo noted the constellation could be fully operational with about 70 SpaceBEEs in orbit. Spangelo’s voicing of how little money Swarm required to manufacture and deploy the entire SpaceBEE constellation indicates Swarm may be relatively healthy, moneywise.

The inevitable question of “Why?” seems to be in the minds of most industry observers when it comes to SpaceX’s acquisition of Swarm. According to Margaret White’s and Garry Bruton’s “The Management of Technology and Innovation,” about 60% of acquisitions and mergers fail. So, considering that higher failure percentage, why would SpaceX and Swarm choose to go through with the acquisition? For Swarm, the company gets access to more money through SpaceX. It also might be able to tap SpaceX’s satellite manufacturing capability and take advantage of whatever low launch price (if any) by using the company’s Falcon 9. SpaceX acknowledges these incentives in its FCC filing.

But the “Why?” is probably asked more about how SpaceX benefits from its Swarm acquisition. In that same filing, SpaceX highlights intellectual property, expertise, and the company’s team as reasons for acquiring Swarm. However, SpaceX doesn’t mention the IoT business (Swarm seems to be succeeding) nor the SpaceBEE technology itself (which seems to be working) as reasons for the acquisition.

I can’t pretend to know exactly why SpaceX acquired Swarm--but I can contribute a few observations.

First, Swarm’s leadership and team seem to be quite ruthless in its satellite design conventions. The company could have gone with traditional cubesat designs but decided that even those weren’t quite what it needed. Instead, it designed the old-fashioned hard-drive style of its satellite and all corresponding components in-house, where they are manufactured to this day. It didn’t adopt groundbreaking technology or manufacturing methods but uses available and inexpensive electronics to keep price and mass down. The thought processes behind these decisions appear similar to how Musk and his teams approach their engineering challenges. Musk values this type of ruthlessness.

Swarm also moves quickly--another Musk trait. Swarm Technologies was founded in 2017. In less than four years, it has an operational constellation, with redesigned ground stations, manufactured and offers its plug-in Tile hardware, provided a decent smartphone app, and is selling an evaluation kit. The company could move faster, but that desire already drew the Federal Communication Commission’s ire and a fine (but then, the FCC is the frozen molasses to Swarm’s hare). While kow-towing to the FCC’s desires and schedule serves the FCC’s needs well (as well as aiding potential competitors), it doesn’t serve the companies wishing to save money by not wasting time nor cater to their desires to move quickly.

But having more people on the team who understand how to move quickly and who also have implemented a plan to do so successfully is probably a win for SpaceX.

The last observation probably lends itself to more fanciful prognostications. Swarm knows how to work with electronics hardware and has manufactured its boards for its SpaceBEEs and evaluation kit. Perhaps SpaceX isn’t happy with the Starlink terminals it contracted out to STMicroelectronics and wants to develop its version in-house quickly?

Of course, Swarm is aiming at deploying 150 satellites—far from the thousands of ground terminals SpaceX wants to produce, so it’s lacking the experience of scale. But SpaceX is already working with larger scales for its Starlink constellation.

Like SpaceX’s rationale for acquiring Swarm, none of the above observations have anything to do with Swarm’s IoT business. So, hopefully, Swarm will be allowed to continue down its IoT master plan path. But it’s also my hope that both SpaceX and Swarm will surprise us with something more exciting and valuable.

There’s a 40% chance that might happen.