How SpaceX Will Launch a Mobile Network

On SpaceX’s earning call, CEO Elon Musk discussed the company becoming a mobile network operator in the US. To date, Starlink’s satellite connectivity has provided fixed broadband services that are mainly used in areas without good fibre broadband, and a messaging service for smartphones as a back-up if terrestrial connectivity is unavailable or not working.

Launching its own mobile service would be a big move. But how viable is it to use satellite as a primary mobile connectivity technology? And does SpaceX have the assets it needs? In this blog, we analyse Starlink’s assets, technology, partnerships and likely future offerings in mobile services.

For Mass-Market Mobile Services, Satellite Players Need Terrestrial Network Partners

It’s unviable to offer a purely satellite-based direct-to-device mobile service in the US that can fully replace a terrestrial mobile network now, and this won’t change in the immediate future. Signal propagation physics makes it extremely hard to offer indoor coverage through satellite alone. Using small cells hasn’t worked as a mainstay of connectivity in the past, whether they shared home Wi-Fi, femtocells or public network small cells. Adding satellite-backed small cells to Starlink broadband customers’ roofs won’t be sufficient for SpaceX to offer a primary mobile service because small cells can deliver — at best — patchy indoor coverage.

Starlink could target dual-eSIM smartphone owners with a secondary back-up service offer. But this would antagonize the mobile network operators and Starlink would need to vastly increase its marketing spend, which makes this approach unlikely in the US. The three large US mobile network operators have already stated that they’re reluctant to support Starlink with a mobile virtual network operator (MVNO) deal and have formed a joint venture pooling spectrum for direct-to-device services, designed to show solidary against any threat.

Instead, we think that Mr Musk is talking up satellite’s potential for mobile services to improve his negotiating position with terrestrial mobile network operators. If Starlink can demonstrate that satellite services — using Starlink’s V2, second-generation mobile satellites — can help mobile operators save costs in rural areas, then operators’ attitudes to Starlink might change. Mr Musk also needs to ramp up the Starlink business as it’s funding SpaceX’s strategic goals in AI and space exploration.

If Mr Musk’s comments go further than this, they reflect his habit of stating lofty goals that exceed what’s credible, and then later scaling back. Examples include statements that SpaceX will establish a Mars colony or deliver full self-driving cars in just a few years.

Starlink currently lacks sufficient spectrum to support a mass-market mobile service with the customer numbers of companies like AT&T, T-Mobile and Verizon. Each of the three national US mobile network operators have over 100 million wireless customers, and there are more than 400 million active wireless subscriptions in the US in total. The 65 MHz of spectrum that Starlink is acquiring from Echostar is only sufficient for a smaller-scale launch, for example in the low tens of millions. However, Mr Musk rarely aims for less than a market leadership position and therefore we consider it unlikely that this size of mobile service is what he has in mind.

What Starlink Will Really Do with a US Direct-to-Device Service

The most-likely scenario is that SpaceX secures an MVNO agreement with one of the big three operators and offers a hybrid satellite-mobile network service in late 2027. This timing aligns with the arrival of Starlink’s V2 mobile satellite constellation and the availability of 65 MHz spectrum that was formerly owned by Echostar. The FCC announced it was approving the transfer of this spectrum to Starlink in May 2026, but Starlink doesn’t expect the deal to close until the second half of 2027. Such an MVNO deal will require a change in the operators’ current stated positions but, given Starlink’s likely 2027 launch timing, Starlink has time to win the carriers over. This timing could also align with the availability of compatible devices as, importantly, no smartphones currently support these MSS frequencies.

Alternatively, using the large sums raised at its recent IPO, SpaceX could look to acquire one of the major US cellular operators. This would provide ample spectrum and a large existing customer base. SpaceX could then highlight the potential for satellite to lower operators’ spending by removing the need to install and operate cellular base stations in rural — and perhaps even some suburban — areas. Starlink could then seek to sell a rural coverage option to mobile network operators worldwide. This would generate much-greater revenue for Starlink than an add-on layer that purely provides back-up, resilience and remote-area coverage, which the company has offered to date.

In both cases, Starlink would be offering a hybrid satellite–terrestrial mobile service, not a pure satellite-based mobile service. To be a primary mobile network service, Starlink needs either an MVNO deal or to own a mobile network operator.

Starlink Is Key to SpaceX Because of Its Proven Cash Generation

Starlink is extremely important for SpaceX. Unlike its Space and AI segments, the company’s Connectivity segment generates significant income. In 2Q26, the AI and Space units suffered losses of $1,257 million and $542 million respectively, but Connectivity made a profit of $1,656 million. As a result, there’s an incentive for the company to expand its Starlink offerings to fund its investments in AI. These made up by far the greatest share of the total addressable market opportunity that the company identified in its IPO filings.

However, Starlink’s current customer numbers are very modest compared with terrestrial broadband and mobile network operators. In 2Q26, there were 12 million customers of Starlink broadband worldwide with an average revenue per user of $66 per month. Enterprise revenue was 42% of total Connectivity revenue in the second quarter.

In mobile, Starlink has over 7.4 million customers of its direct-to-device service according to its IPO filing, through relationships with mobile operators such as NTT Docomo, O2 UK, SoftBank, Spark New Zealand and T-Mobile US. But these mobile customers have a limited messaging-focused service using Starlink, which is very different to Starlink being the primary connectivity technology.

Mr Musks’s Suggested 5G Approach Isn’t Credible Without Partnerships

On the 2Q26 earnings call, Mr Musk talked up the opportunity in direct-to-device mobile services and went much further than the aspirations documented in May’s IPO filing. He identified using 5G small cells with satellite backhaul to improve indoor coverage, suggesting that many of these will be attached to Starlink broadband terminals on rooftops.

There are multiple issues with such an approach. Current Starlink terminals don’t have cellular support, so Starlink would need to deploy new equipment. This takes time and costs. The company’s existing terminals for home and small businesses tend to be in areas with weak cellular coverage and not in the dense urban areas where most mobile users are, where satellite capacity will be most stretched, and where indoor coverage is hardest. To deploy small cells in those key urban areas, Starlink will need to find and acquire sites. This may require planning approval, especially if the output cellular power level is greater than that of a typical home femtocell and is close to traditional mobile base stations. At best, Starlink could get some limited small-cell coverage of US cities, but with many gaps.

With an MVNO deal with a mobile network operator, Starlink could strengthen its position to offer urban service. The small cells would then act similarly to the way US cable broadband providers have been using home Wi-Fi, taking load off cellular networks and complementing a terrestrial cellular network. By handling some of the data traffic, they would reduce the data charges the cable operator pays as an MVNO for the use of their partner’s cellular network.

Starlink’s Mobile Spectrum Position Only Supports a Rural Service

Satellite capacity in urban areas is another problem. With 5 MHz of spectrum, Starlink’s current V1 mobile satellites deliver 256 beams per satellite, with a beam size of approximately 900 km2 and 128 users per beam. Spectrum capacity is shared across the area of the beam. This is a fraction of the capacity per square kilometre of a conventional terrestrial cellular base station, even a rural one with only a single low-frequency band deployed. For example, T-Mobile has 40 MHz of capacity on the 600 MHz band available in many US markets.

Starlink’s V2 satellites will greatly boost this capacity. Using the spectrum acquired from Echostar, Starlink will have 40 MHz of mobile satellite spectrum (MSS) capacity available, which requires phones to support this MSS band. It’ll also have 25 MHz of IMT spectrum, some of which lies in a supplemental band and must be used alongside an anchor band. The new satellites will increase the number of beams per satellite to 1,024, narrow the beam size to 160 km2 — which boosts the capacity per square kilometre — and raise the number of users of each beam to 512.

In rural areas, a Starlink service using this spectrum and the V2 mobile satellites would be at least as good as, if not better than, rural terrestrial cellular networks. This is Starlink’s key advantage. But in urban areas where users are more concentrated, the capacity per square kilometre of Starlink will not prove competitive, even when used outdoors.

A Starlink Direct-to-Device Mobile Service Will Work Best as a Hybrid Play

With the capacity limitations and the problems of delivering indoor coverage, Starlink must partner with a mobile network to maximize its opportunity. Globally, the company needs a proposition that convinces operators that they can lower their costs by using satellite as the primary network service in rural areas. To do this, Starlink must evangelize the capabilities of its upcoming V2 mobile satellites. SpaceX needs Starlink’s revenue from consumers and enterprises to bridge the gap until its AI and Space segments become profitable. A revenue-generating US mobile service would help to increase that revenue. This is why Mr Musk is hyping the future Starlink mobile service.

He has time to build momentum. The V2 satellites and Echostar spectrum point to a launch in the second half of 2027. Only then will Starlink be ready to offer a mobile network service proposition that’s strong enough to compete.

However, if the company wants to launch a mobile service under the Starlink brand, its best approach is to partner and offer a hybrid satellite–terrestrial network, where the existing cellular network is the primary technology in cities and Starlink improves service quality in rural areas. A US mobile launch would act as a proof of concept for international opportunities. Starlink could use the US example to persuade operators in other countries that satellite is a credible technology and that they should pay Starlink to provide satellite connectivity in rural areas and save the cost of deploying and running rural base stations.