Starlink, and the battle for the edge of the network and heart of the customer (Telco Diary)

From the newsletter (October 2; sign-up or follow if you want it sooner): Satellite is not about to replace terrestrial networks – but neither is the Starlink threat imaginary. As orbital capacity scales, the real contest is over the edges of connectivity, spectrum, infrastructure and, ultimately, who controls the customer. Video killed the radio star – catchy, but wrong. There’s other stuff happening outside of our little terrestrial RAN world – which RCR has obsessed over for the last week. There is lots of news (and talk) about satellite comms, too. So let’s consider it, as part of a Friday afternoon ramble, and agree at the outset that a big problem – with politics, tech, society, journalism (!) – is how binary everything seems to be; like there is no middle ground (and we know what happens when the middle collapses). So as a journalistic exercise (and an anti-fascist rant) we should correct some of the populist rhetoric peddled by both sides in the Starlink debate – and fold-in some news, along the way. The debate is reductive. In the red corner, you have noisy ‘influencers’ on LinkedIn talking like Elon Musk is about to replace terrestrial operators with a busy constellation of flying routers. (Really, they should be required to declare their stock investments.) In the blue, you have an army of dyed-in-the-wool radio engineers, plus some displaced software programmers, insisting that physics, spectrum, and basic network economics make the whole scheme laughable. Both positions are too neat. Musk, of course, has been winding the crank arm on his hype machine. “It is increasingly probable that Starlink will carry a majority of Earth’s IP traffic long-term,” he said a couple of days ago. It was a line repeated from his firm’s IPO brochure (or its last quarterly results?). “Starlink would de-facto become the internet, and everything else would connect to [it],” he said. Which is a big statement, but it is hardly a proper forecast. But think on it: the answer is not that satellites beat fiber. They don’t; fiber, in the red corner, is good at carrying high volumes of predictable traffic where lots of people live. Satellite, in the blue, puts capacity in the sky – and somewhere on the ground. Starlink is upping the ante, too: its new satellites (V3) are geared for 1Tbps of downlink capacity; its latest launch vehicle (Starship) is intended to deploy as many as 60 at a time. The first operational batch went up last week, delivering 26 satellites. The rate at which orbital capacity can be manufactured and deployed is skyrocketing. But 26Tbps of satellite capacity is not 26Tbps of useful capacity – over New York at 8pm. Orbital geometry, spectrum availability, beam usage, radio interference – all matter. So do terminals, gateways… and demand. As a news aside, there is a useful counter-example to the Starlink binary from Spanish NB-IoT group Sateliot, which has just put five satellites into orbit on a SpaceX transporter rocket (Falcon 9) – to expand its constellation to 11, about-double its daily ‘passes’, and halve its ‘revisit’ times. SpaceX’s vertical integration, essential to its Starlink ambitions, works for the rest of the market too. Sateliot has a roadmap: 500 satellites at some point, moving from IoT to D2D with its Tritó launches from 2027. Alén Space, which manufactured its new spacecraft, did so in half the time – versus its last Sateliot mission. So SpaceX, most of all, is supporting the industrialisation of orbital telecoms. Which doesn’t settle the capacity argument. Some of the criticism is right. Orbital capacity isn’t a magic global pool. A Starlink satellite over the Pacific cannot lend unused capacity to the Starlink customers on the Atlantic seaboard. Fiber is where the workloads are, already; so is terrestrial cellular. But not always. The opposite is also wrong. Rural environments, maritime sectors, remote industries, disaster zones are not imaginary markets. These fringe markets are where terrestrial economics always breaks down. Satellite does not have to replace fiber or 5G to be relevant – and increasingly strategic for telecoms. It just has to be good enough, cheap enough, ubiquitous enough to cream off the edge. And the satellite industry looks like it has plenty of road: the FCC has just opened more than 1,000MHz in the 12.7GHz and 42GHz bands for satellite comms, and is proposing more spectrum besides. Infrastructure is political. Spectrum is political. Tech is political. Space is political. But the game is still open, to an extent. Yesterday, AT&T, T-Mobile, and Verizon formally launched their US satellite joint venture to pool spectrum resources and develop a common D2D platform – as a complement to their terrestrial mobile networks. They are talking about common technical specifications, standards-based compatibility, and a tech-neutral approach. It is a clever defensive move. Their challenge, in the end, is to make satellite connectivity another layer in their networks before they become another layer in satellite networks. That is your existential crisis, right there. The trick is to keep their logo on the handset, and at the top of the bill. Europe is doing similar, for a different reason: IRIS² is being built as a sovereign European connectivity system, with the constellation expanded to 348 satellites, and launches planned in 2029. The EU framing is security, resilience, and digital independence – for critical infrastructure and agencies. It is about who owns the network – which is political, and not just about telcos. Nokia knows this; it announced a deal, also yesterday, about the politics of ownership with Finnish space group ICEYE. The pair will develop sovereign LEO systems where nations own and control the satellites, ground segments, and terminals. (It will go neatly with Nokia’s macro-private 5G pitch.) So far as Starlink goes, Musk doesn’t need to carry most of the world’s IP traffic. He just needs to build enough of the space stack to capture them in a pincer movement – so terrestrial players have to connect to it. (AT&T et al might do well to bring along the tier-twos, as well.) The US telcos are trying to retain the customer relationship. Europe is trying to ensure independence and control. Other satellite operators are just looking to wholesale airtime in the middle of it all. The threat is that telcos will be marginalized, to a degree, and not that they will disappear. Satellite might make them richer, too – by delegating their expensive edges. If they can just retain control of the heartlands.