Qualcomm’s deal to supply radio technology for Starlink Mobile’s terrestrial network strengthens SpaceX’s ambitions to challenge the US mobile incumbents. Analysts see a potentially disruptive, vertically integrated operator taking shape – but warn that building a nationwide network will take years, billions of dollars and some difficult decisions on coverage.
In sum – what to know:
Fourth network – SpaceX’s ambitions are taking shape: Qualcomm’s radio tech will support a terrestrial 5G buildout alongside Starlink’s satellite infrastructure, with deployments from 2027.
Major hurdles – Coverage and spectrum are issues, still; SpaceX has a way to go to match the capacity and reach of US incumbents, and will need wholesale deals or spectrum acquisitions.
Dragonwing win – Qualcomm gets a major RAN opportunity; the deal offers a high-profile greenfield deployment for its Dragonwing platform, potentially extending to 5G Advanced and 6G.
SpaceX is moving closer to becoming a fourth US mobile network operator, it seems, with a multi-year deal with Qualcomm to supply radio technology for a terrestrial 5G annex for its Starlink Mobile network. The deal, which loosely targets deployments from 2027, adds a ground-based RAN to SpaceX’s growing satellite and spectrum portfolio – and brings Qualcomm back into the terrestrial infrastructure game. The partnership could help SpaceX build “its own flavor of RAN”, as Leonard Lee at neXt Currve suggested in response.
RCR covered the news yesterday, including a couple of quick-fire analyst responses; RCR has also published more on the rumors about SpaceX’s purchase of Boost Mobile, not made mutually-exclusive by the Qualcomm tie-up. But given the hyper excitement about the deal, there is sense to round-up further responses from the analyst community, as below. In sum, they broadly agree: SpaceX is serious about terrestrial 5G, if it was in doubt, and its combination of satellite and terrestrial infrastructure could reshape the US mobile market.
What’s not said, and what might be more significant in the mid-term, versus the perceived jeopardy for the ‘big three’ US carriers, is the impact of US-made Qualcomm’s entry into the RAN space on incumbent network vendors – notably Nokia and Ericsson. We will revisit this. It might be considered as well that Verizon chief Dan Schulman said on an earnings call last week, effectively, that it is entrenched enough to know everything that is going on in the market – so there are no surprises, and, presumably, that it had an inkling of the Qualcomm deal ahead of time, even as it was dismissing the challenge out-of-hand last week as a gimme. RCR will also pick up again on this.
But SpaceX’s route to being a terrestrial contender is also far from straightforward, reckons analysts. Spectrum holdings remain well below par, and ubiquitous coverage will likely require wholesale access or further acquisitions. The deal is a “big shot in the arm” for Qualcomm, says RCR regular Prakash Sangam; for SpaceX, the challenge is turning vertical integration into a viable network business. “It won’t be easy or fast”, says Michael Thelander at SRG Research.
Whatever, the Qualcomm deal is a way to strengthen SpaceX’s negotiating position with incumbent operators. “My intuition is that SpaceX’s answer is to make building without AT&T credible enough that refusing a deal becomes less attractive. My bet is a wholesale agreement before SpaceX signs tower leases at national scale.” So says Vish Nandlall, another familiar in these pages. Anyway, here is the view of the analyst community – and, arriving (to RCR) late, maybe the angle from Sam McHugh at BNP Paribas, inserted at the top, is the most interesting.
Sam McHugh, head of telecom, cable and satellite equity research at BNP Paribas:
“As we have written, we do think investors should take SpaceX’s plan to build a network seriously, and at a headline level this is another step towards the company putting the puzzle pieces together, and as such is likely to be taken as a small negative for AT&T, Verizon and T-Mobile, as well as the cable companies. That said, the degree of press headlines similar to this (including SpaceX reportedly using 80,000 towers per Bloomberg, and SpaceX potentially buying Boost per The Washington Post) is unprecedented. This suggests to us there is a clear strategy to try to ramp pressure on one of the big three to fold on an MVNO agreement – which we think is very unlikely – or at least depress multiples to make a potential acquisition of a carrier more accretive for SpaceX.
“What Qualcomm would supply is only a minor part of the puzzle for building out a network, and notably there doesn’t appear to be any firm commitment by SpaceX to actually go ahead with purchasing any equipment from Qualcomm.
“Qualcomm sold its networking chip business to Ericsson in 1999, but has been trying to get back into the business by supporting open RAN deployments; it sells chips and reference designs, and supplies only a small part of what SpaceX would need to build a network. The real constraints on building a network are finding sites, backhaul, spectrum, and time. While Qualcomm… talks about thousands of sites there doesn’t appear to be any firm commitment. A freshly built 5G network using exclusively 64T64R massive-MIMO radios would be state-of-the-art, and would help squeeze as much capacity as possible out of SpaceX’s relative lack of spectrum. The timing is of course interesting, coming off the back of the US mid-terms, with a stronger than expected Democratic result perceived by the market as a negative for Musk’s empire.”
Neil Shah, co-founder and vice president of research, Counterpoint Research:
“Qualcomm and SpaceX’s multi-year deal is a first-of-its-kind move. SpaceX is building a true hybrid terrestrial 5G and satellite network, the epitome of vertical integration, and it has the potential to reshape the telecom industry. The ambition is clearly to build a nationwide competitor, not a niche player. But with roughly 80 MHz of spectrum assembled so far, compared with 300-400+ MHz for each of the big three, matching Verizon, AT&T and T-Mobile on density and capacity will take years and serious capital.
“In the near term, SpaceX can become a credible fourth player with a differentiated hybrid offering, especially for coverage in rural and suburban areas. The implications for incumbents grow significantly if SpaceX acquires more spectrum, builds a greenfield 6G network or buys a player with a large spectrum position. For Qualcomm, this is a strong re-entry into terrestrial RAN and could open the door to a much broader partnership across Elon Musk’s companies, from Tesla’s vehicles and robotics to custom compute for SpaceX.”
Subhadip Roy, senior analyst, Counterpoint Research:
“The key question in this deal is who owns the spectrum. Today, satellite-to-phone services in the US largely depend on carrier spectrum and partnerships. Starlink’s T-Satellite runs on T-Mobile spectrum, and AST SpaceMobile works with AT&T and Verizon. SpaceX is taking a different route by building its own spectrum position alongside its satellite and terrestrial infrastructure, and that could change how D2D works in the US.
“There’s an interesting irony here: Starlink Mobile has reportedly passed 22 million monthly users, largely through carrier partners such as T-Mobile and Rogers, the same operators it could increasingly compete with. The timelines are also converging. V2 direct-to-cell satellites on Starship, Qualcomm-powered terrestrial deployments and the EchoStar spectrum transfer are all expected in 2027, ahead of SpaceX’s own-spectrum mobile service planned for 2028. Next to watch is the FCC’s decision on SpaceX’s 800 MHz deal and its broader satellite-to-phone spectrum rules.”
Vish Nandlall, founder and lead analyst, Vish Nandlall Consulting:
“My SpaceX research is one corkboard and a roll of red string away from a family intervention. The cumulative announcements over the past year seem to point toward a vertically integrated US telecom company, controlling how its satellites and ground network work together. Qualcomm is supplying radio silicon and Layer 1 software. SpaceX says the deal will accelerate development of its base-station architecture. The plan calls for tens of thousands of sites from 2027, although financial terms and binding minimum volumes remain undisclosed.
“Towers add capacity by reusing spectrum across smaller areas. On 20 MHz, the FCC’s final satellite minimum for 2036 works out to 12 Mbps per beam. A three-sector tower at an assumed 4 bits per second per hertz yields 240 Mbps. But those towers create a coordination problem. The FCC found that satellites and terrestrial stations cannot effectively serve the same place simultaneously on the same channels. Someone must decide where and when each system transmits. The 15 MHz AWS-3 uplink needs the same coordination wherever both layers use it.
“This is where I see the value of vertical integration. Control of the scheduling software lets SpaceX develop space and ground as one system. Qualcomm supplies the foundation, leaving SpaceX responsible for the software above it. A key question is wether SpaceX builds the radio. I suspect yes. If you work in the industry, you know Washington’s interest in a US telecom stack. An American operator developing its own equipment architecture carries geopolitical weight, even with global component manufacturing. It would also increase Washington’s dependence on a single company.
“Coverage remains the expensive part… Dish built a network and still needed incumbent coverage to fill the gaps. That explains the reported Boost talks. EchoStar had 7.38 million wireless customers at June-end, mostly through Boost and Gen Mobile. Boost offers a potential route to AT&T coverage while SpaceX builds. BNP says the T-Mobile arrangement is winding down and believes AT&T’s contract contains restrictions that prevent an automatic transfer of wholesale access.
“That means AT&T has leverage. Enabling SpaceX means helping a future competitor. My intuition is that SpaceX’s answer is to make building without AT&T credible enough that refusing a deal becomes less attractive. My bet is a wholesale agreement before SpaceX signs tower leases at national scale. For now, I think the Qualcomm agreement’s greatest strategic value is strengthening that negotiation. If SpaceX signs leases for tens of thousands of sites first, well that will change things. AT&T started off building their own equipment, maybe this is how we complete the circle.”
Prakash Sangam, founder and principal at Tantra Analyst:
“Wow… a big shot in the arm for Qualcomm’s cellular infra business. A large new greenfield operator in the western hemisphere, coming after many, many years, and that too, a high-profile player like SpaceX, deciding on the Qualcomm Dragonwing platform… says a lot about its technological prowess.”
Michael Thelander, president at SRG Research and Consulting Services:
“I haven’t written much about the SpaceX terrestrial ambitions, in part, because the topic has been beaten to death by esteemed analysts and wanna-be technology bloggers. For what it is worth, my two cents are that it can be done if you have the will and the $$$, it won’t be easy or fast, and it needs to differentiate from the three existing mobile networks in a way that I can’t imagine.
“Dish Networks aspired to build out a greenfield 5G network and it was doing so with relatively modest lower frequency FDD spectrum assets compared with the big three. Prior to turning out the lights, Dish had done a respectable job of building out a network in the cities where I had the opportunity to check it out – namely Las Vegas and Minneapolis. However, the coverage wasn’t ubiquitous but at least the smartphone could revert to another operator’s network (AT&T or T-Mobile) for continued coverage.
“Without some sort of MVNO agreement, SpaceX may not have the luxury of not having ubiquitous coverage, while switching back and forth between TN and NTN doesn’t seem practical or compelling from an end user experience perspective. This statement is especially true when the coverage hole is within a building. I showed T-Satellite can work for basic text messaging from the basement of my house but I seriously doubt it would provide a decent broadband data experience. Then again, I’m thinking about coverage in terms of SpaceX competing directly with a big-three operator for a typical eMBB consumer experience.
“Who’s to say that SpaceX doesn’t take its 5G and 6G vision in a different direction and target use cases that don’t necessarily need ubiquitous TN coverage and/or which can leverage NTN more extensively than the typical consumer use cases…
“It is difficult to extrapolate the 5G NTN user experience based on the original, or even current, satellite constellation, which is based entirely on a somewhat modified implementation of LTE that was designed to accommodate the high doppler and latency that is typical for a cell site positioned 200-300 miles above the earth and traveling at speeds reaching 17,000 mph. I’m not sure what to expect myself, but I’m pretty sure that whatever capabilities exist five years from now will be nothing like I’ve experienced to date with the service.”
Read Thelander’s full blog here.
Mike Dano, lead industry analyst, Ookla:
“The new agreement between SpaceX and Qualcomm reinforces the likelihood that Starlink Mobile could become a credible fourth nationwide mobile network provider in the US Of course, there are many more steps along this path, including building radios for a Starlink Mobile terrestrial network, putting them on top of towers, and then activating services for paying customers. And even then, it’s not clear whether Starlink Mobile will be able to successfully compete against the market’s established mobile providers.
“However, it’s worth noting that our recent findings show Starlink Mobile is currently providing needed connectivity outside of terrestrial coverage areas in the US For example, Starlink Mobile’s D2D offering has already helped to reduce cellular coverage gaps in the US and globally. That said, the performance of Starlink Mobile’s network trails the performance of terrestrial networks – for example, roughly one in seven Starlink Mobile D2D texts failed to go through, in RootMetrics’ testing, compared with around one in fifty on the terrestrial network. This is the kind of performance gap that Starlink Mobile will need to address with its terrestrial network buildout.”
Leonard Lee, executive analyst and founder at neXt Curve:
“Qualcomm suggested that SpaceX will begin deployments in 2027 for a nationwide mobile terrestrial network. Instead of the femtocell idea that Elon floated back in August, the Starlink Mobile RAN deployment will feature massive MIMO 64T64R radios for sub-2.5 GHz bands. Qualcomm will introduce innovations in beamforming to support the lower frequency bands that SpaceX. Recall that SpaceX acquired 1.7/2.1 GHz AWS / MSS midband spectrum from EchoStar with rumors that the company will acquire more spectrum in future auctions.
“Directionally, it seems SpaceX is moving toward some form of open RAN architecture. Qualcomm has been a champion of open RAN and may have an opportunity to do with Starlink Mobile what Dish was not able to do with, ironically, the same spectrum assets. This would be a positive for Qualcomm, which has invested tremendously in its O-RAN-compliant RAN accelerators but has struggled to scale markets with the mellowing of the open RAN movement.
“Qualcomm disclosed that it has entered into a multi-year agreement in supporting the ambitions of Starlink Mobile, which will undoubtedly be bold given that the company sits at the nexus of terrestrial and non-terrestrial telecommunications in the AI era. [It] indicated that the partnership will not only support Starlink Mobile through initial deployment and scaling of their terrestrial mobile network, they will also work with them in evolving their networks through 5G Advanced and 6G evolutionary stages.
“One thing is certain, it looks like SpaceX is going to build its own flavor of RAN for the Starlink Mobile MNO.”
Read Lee’s full blog here.
