Governments are Scrambling for a Starlink Alternative. Does One Even Exist?

Governments are Scrambling for a Starlink Alternative. Does a Viable Alternative Even Exist?

Over the past five years, Starlink from SpaceX has changed from a niche service that very few people even knew existed to a critical enabler for many national armies - with the war in Ukraine being a prime example of its military utility.

This growth in utility has been accompanied by erratic behavior on the part of its owner, Elon Musk, and the desire of relevant governments for direct control over such critical assets. As such, many nations on Earth now find themselves reliant on a service which can be switched off on the whims of one man (or by an American government currently operating at the whims of another erratic man). This has led to a growth in demand for alternative services, although few seem to realize just how difficult it will be to fully replace Starlink in the short to medium term.

Starlink and its military use

On a modern battlefield, the ability of even the smallest units and individual weapon crews to send and receive vast amounts of data is absolutely critical for effectively waging a war. Modern sensors, from cameras on small reconnaissance drones to air defense radars, are constantly generating reams of the stuff, all of which needs to be analyzed and distributed to all the relevant units by battlefield management systems.

But most battlefields inevitably lack the normal civilian communication infrastructure, and most of the legacy military communication systems are either focused on enabling voice communication, or they are absurdly large, power-hungry, expensive, and available in tiny numbers - all of which often makes them unsuitable or at least insufficient for current battlefield needs.

Without the ability to rely on the local ground infrastructure, satellite communications have for a long time been attractive to military users, but until recently, satellite communications used to suffer from very low performance and being ridiculously expensive. Part of the problem was that due to the distance (GEO is 35,786km above the Earth's surface), connections via satellites in geostationary orbit suffer from high latency, need large terminals to get usable connection bandwidth, and are vulnerable to jamming. Historically, it wasn’t considered economically viable to build a communication system in low Earth orbit, where speed and latency would be much better, but where any such constellations had to comprise hundreds or thousands of satellites.

With Starlink, SpaceX has shown that the recent technological advances in signal processing, phased-array antennas, and rocket reusability make it possible to resolve these problems and provide a high-speed, low-latency communication system that is resistant to jamming and which has so much redundancy that disrupting it with a physical attack would be very difficult.

But this has required an unprecedented engineering effort to reduce the cost of satellite construction and launch. And, despite those achievements, Starlink is definitely not a cheap system to build. Even without accounting for the recent frequency license purchases and development of the hugely ambitious Starship rocket, SpaceX has likely spent around $15B on Starlink so far. Importantly, SpaceX has a reputation of being very cost-efficient - exactly the opposite of what tends to be true for most governments and their service providers. A great example of SpaceX cost-efficiency is that the current production cost of a Starlink satellite, which provides 100Gb/s capacity, is reported to be $600-800K, and the internal cost of the whole dedicated Falcon 9 launch, including overhead, is around $20M - compare that to a Viasat-3 satellites in GEO that each cost $700+M while providing 1Tb/s capacity.

What military users care about is that Starlink provides high-bandwidth, low-latency connectivity almost anywhere in the world, while being resistant to jamming, and since terminals are small and cheap, they can be installed on almost everything - including single-use strike drones.

Given that Starlink has become so obviously useful for military forces, it is reasonable to ask where all the Starlink competitors are. Clearly there is a hunger to diversify away from a service that - however useful - can be switched off by a single man not known for his predictability. Indeed, go wading through European social media, and you will find multiple calls for the continent to ditch any connections to Starlink and find an alternative.

The harsh truth is that there are only a tiny number of viable competitors that may start providing a service in the near future. Many others have already failed, or are now dealing with serious issues. The good news is that Starlink competitors can now take stock of the last decade of developments and see what does and doesn’t work. In other words, they can copy SpaceX’s successful engineering approaches to reduce their development risks and costs.

Requirements and risk factors

Let's talk for a moment about what it really takes to build such a satellite megaconstellation.

First, you need the ability to design satellites and terminals that effectively balance reliability, performance, cost, and ease of manufacturing.

*Second, you need production systems that are able to make at least hundreds of satellites and hundreds of thousands of user terminals per year (and preferably an order of magnitude more of both).

Beyond that, there’s the need for launch capacity to deliver produced satellites into LEO and do so at an affordable cost. This requires at least partial rocket reuse.

And then there’s the ability to deal efficiently with highly complex problems. After all, any engineering project of this scale and nature will inevitably have its share of complications and issues, and if you can't resolve them quickly and effectively, large cost increases and delays are a certainty.

All the above should also be, if possible, done in-house, to avoid delays and cost increases inherent to outsourcing. Which brings us to our final point…

You will need to have lots of money. Don’t bother even showing up to this race if you don't have at the very least $10B available, and preferably several times more than that.

As should be clear from the above, making a Starlink competitor with a team of companies or countries clubbing together will be much harder than if there is a single decision maker. The bloat, duplication, and endless meetings about having future meetings for which Europe in particular is infamous would end with designed-by-committee equipment, and a budget that runs at multiple times its original goal. There is, after all, a reason why the European Space Agency remains a player comparable with Japan’s JAXA, rather than a behemoth like NASA.

But authoritarian countries can struggle just as much due to the mix of corruption, avoiding responsibility, and decisions often being made based on who proposed the solution, instead of which solution is the best. Those wanting evidence need look no further than Russia’s space program, which if nothing else performs the valuable service of making Europeans feel less upset about their own effort.

But even the lean American companies that should be snapping at SpaceX’s heels suffer from their own disadvantages. While some have recently started landing their orbital rocket first stages and, in doing so, have taken baby steps towards regular booster reuse, everyone appears to still be very far behind SpaceX, as can be seen by even the quickest overview of the competition.

Eutelsat OneWeb

The first potential competitor for Starlink was the European OneWeb project. Despite being announced and launching its first satellites in the same years as Starlink, and now having completed its 1st-generation constellation of 600+ satellites, OneWeb has suffered huge numbers of technical and organizational failures that have turned this initially promising European project into another space industry cautionary tale, and at one point forced the company into bankruptcy.

Some of its problems were down to just bad luck, like Russia stealing dozens of its satellites and downpayments for future launches after the start of the 2022 invasion of Ukraine. But the company has also stuck with an “old space” mentality by outsourcing as much as possible, and not being able to manage those suppliers. This resulted in an inferior satellite design that was not only technically worse than early Starlink satellite design, but was also ten times more expensive per satellite, with the same rockets able to carry (and thus launch) fewer of them than they could Starlink satellites.

Today, OneWeb service is operational, but with very limited capacity - it has 600+ satellites in orbit, each with 10Gb/s capacity, but the whole network is only advertised to have “over 1.1Tb/s capacity” due to the lack of inter-satellite laser links that would allow utilization of satellites that are not over areas with ground station coverage. For comparison, each Falcon 9 launch with current Starlink satellites delivers into orbit about 2.4-2.9Tb/s in network capacity.

As a result, OneWeb is a commercial disaster. But the importance of Europe having access to such a service has meant that this project is now being propped up by UK and EU funding, with plans to integrate it into the IRIS² initiative. Since 2024, OneWeb has contracted construction of 440 new satellites for its replacement satellite constellation, which is expected to start launching into orbit in 2028. Given its planned number of satellites, it will be one of the smaller such systems, along with the Russian Rassvet.

Amazon LEO

Amazon LEO, earlier known as Project Kuiper, was announced in 2019 and is so far the most promising commercial attempt at providing an alternative to Starlink - but it has suffered from more than its share of schedule delays. Amazon LEO is planned to be smaller than Starlink, but still very capable. Its main problem has been that while its factory is mass-producing satellites, it doesn't have rockets to launch them, so they are sitting in a warehouse, because while Amazon has contracted over 100 rocket launches, it focused on contracting rockets from all Western heavy-lift launch companies except SpaceX, and those other companies have all suffered severe delays in their new rocket programs.

Fortunately, by July 2026, it has already launched enough satellites to start the initial limited beta service in some countries (possibly including Ukraine) in late 2026 or early 2027, after which the coverage will gradually be expanded as more satellites are launched. In the longer term, the complication for the future of this project is the lack of internally available cheap launch capacity, from which Starlink benefits - without that, its business case is weaker, and that’s important for a project that isn't funded by any government.

China’s Sovereign Mega-Constellations (Guowang & Qianfan)

While the Chinese government didn't earlier pay much attention to this area, after the success of Starlink during the initial phases of the Russian 2022 invasion of Ukraine, it has started investing a lot in both replicating this capability and in trying to develop counters for it.

Two major projects for LEO satellite internet mega-constellations have been launched, each with announced plans to operate over 10,000 satellites. While the Guowang project is mostly relying on conservative designs, Qianfan appears to be more risky, but also more dynamic - making sure that at least one is likely to succeed. Chinese national security reasons apparently justify spending large sums of money on two constellations.

As of mid-2026, both have launched a few hundred satellites, and are expecting to start the initial service at some point in 2027. While the launch rate has been lagging behind schedule, despite China's launch rate being only behind that of the USA, the recent success in the area of rocket reusability appears to be a good sign for the future prospects of an increased launch rate. But perfecting this technology will take time.

During the Chinese launch campaign for these satellite constellations, there have been several orbital-debris-generating events involving rocket upper stages, which need to be addressed - otherwise, this debris will over time become a serious risk for both Chinese and foreign satellites in orbit.

Rassvet

While most of the Russian space projects announced over the past 20 years have stayed firmly in the PowerPoint stage of project execution, one of the lessons of the Russian 2022 invasion of Ukraine has been that modern LEO satellite communication systems are critical for effectively waging a war, and so the Russian government has provided sufficient resources to rush the development and start the deployment of its Rassvet satellite internet constellation.

In Russia’s circumstances, it makes sense for the system size, as announced, to be comparable to OneWeb, but it is interesting that the orbital inclination of the Rassvet satellites isn't optimized to provide coverage over the battlefield in/around Ukraine as quickly as possible. The result of the decision is that with just 32 satellites launched so far (of which one has died and three more appear to be struggling), it is very likely that Rassvet won't be able to start providing continuous service until at least 2028 - by which time the war in Ukraine is likely to have ended.

And for military purposes, the smaller constellations like OneWeb and Rassvet are likely to be more vulnerable to jamming, due to a much smaller number of satellites being overhead at any time.

IRIS²

On paper, the EU's IRIS² project should integrate OneWeb and other similar national and multi-national services into a coherent and resilient multi-orbit satellite constellation, but the constant disputes (especially the petty debates about what components will be built in what country) and changes have so far only generated a lot of useless paperwork and wasted time and money. As a result, besides keeping OneWeb alive, the European efforts in this area have been fragmented and chaotic, with no operational system even being on the horizon, unless the 2nd generation of OneWeb constellation turns out to perform better than expected. But even in that unlikely case, it isn't likely to be fully operational before 2030.

There is also the minor detail that Europe doesn't have the capacity to launch a large number of satellites. This is largely because it has spent the last decade funneling €5B into the Ariane 6 rocket, which provides only a slight improvement over its predecessor, is produced at a high cost in only small numbers, and which has no element of rocket reuse. Being European in the 21st Century is to accept that collective action has become code for self-sustaining bureaucracy, but it is still at least somewhat impressive to witness again and again how self-defeating and unambitious our continent can really be.

What Comes Next

It is clear that over the next several years, several services will start providing real alternatives to Starlink, which will pose a problem for armies that until now were able to suppress the opponent's technological capabilities by denying them the ability to use services provided by Western companies. On the other hand, for those able to access more of these services, it will mean that SpaceX and Starlink will no longer hold life-and-death power over them by being the only option.

But it needs to be recognized that none of this will be cheap - and that some of these services, like OneWeb and Rassvet, will exist on a much more limited scale than Starlink. As a result, they will likely only be relevant to government and military users, due to the lack of capacity to provide commercial services at a significant scale. In those cases, the relevant governments will need to cover all the ongoing costs of such projects and treat them just like any other military assets that require constant investment. The success of Starlink should be as replicable as the successes of the Ford Motor Company or IBM, or any other groundbreaking company of old. Judging by the current competition, this may not be the case for some time to come.