A wide range of people from academics to Motley Fool writers are trying to predict how the SpaceX Starlink satellite broadband will generate enough revenue to survive and thrive. As a private company, abet one which continues to drop hints it will spin off Starlink for an IPO, SpaceX is not under any obligation to discuss its current financial models. The company and CEO Elon Musk seem to be content to say little and let others fill the vacuum with speculation. 

The best source of information comes from Federal Communications Commission (FCC) filings SpaceX has made over the years.  Starlink currently is licensed to operate up to 1 million user terminals (dishes) in the United States and is seeking to modify that license to go up to 5 million dishes.  Regardless of whether SpaceX can quickly and affordably build 5 million dishes is a different discussion; the company is committing up to 5 million customers in the United States.

The Motley Broadband Model

The Motley Fool’s math starts juggling math using a complex equation aggregating the total amount of bandwidth available from all satellites (23 Gbps/satellite x 12,000 satellites, per FCC filings), assuming all users 1) can access all satellites at once and 2) get 100 Mbps. By this calculation, the Starlink “network” can support up to 2.76 million users on the network at once. Take that number and multiply at the current beta rate of $99/month to come up with $3.3 billion per year.  At an oversubscription of 3x users, you’d get over 7.5 million users and annual revenues of nearly $10 billion.

However, the idea you can aggregate all the accessible bandwidth from all the satellites at once is flawed.  Not all satellites on the network will be accessible equally – there’s a dearth of potential customers at the poles and in the middle of the ocean – therefore you can’t magically tap into “bandwidth” on an unused satellite. If it’s not overhead, it’s not there.

LEO satellites are essentially inverted cell phone towers – users are (typically) in fixed locations with the cell towers (satellites) constantly passing overhead. The limiting issue becomes how many subscribers you can put within a satellite “cell” before service becomes degraded to the point where people switch to a terrestrial alternative or stop using it.

Assuming pricing will remain at $99/month per user is flawed and does not account for any differentials in pricing Starlink may/will introduce in the future for tiers of service based upon speed and type of user (residential/home, mobile business, verticals). It also does not anticipate additional revenue per user for value-added services, such as voice.

Trefis: ARPU closer to the truth

Trefis estimates the LEO satellite broadband constellation could generate annual revenues of over $10 billion by 2025, with a total user base of 14.4 million and a flatline estimate of average revenue per user (ARPU) at $60, around the average broadband ARPU in the U.S. Taking SpaceX’s FCC filing to operate up to 5 million dishes within the U.S., a worldwide userbase of nearly 15 million would make sense, since the company’s beta has already reached over 10,000 customers worldwide, with the majority in the United Sates.

If anything, Trefis is conservative in its estimates because it assumes a decline in price from the $99/month that Starlink is charging in the beta, without allowances for tiered pricing or additional value-added services.  Current positioning for Starlink pricing is based up pent up demand in areas where there are inferior or no broadband options, including existing DSL, cable, and legacy satellite services.

Moving out of beta and into production will enable Starlink to offer differentiated services for both consumers and businesses, adding other value-added features over time to further increase ARPU and hence total revenues.

SpaceX, as noted by Northern Sky Research, is likely to develop a “very loyal customer base” with lower churn levels than other broadband providers – one might call it the Tesla of satellite connectivity options.  Individual/residential style customers are more likely to pay more for baseline connectivity and value-added service.

Exploring ARPU: Average and increasing the average

Instead of building a comprehensive global model of Starlink network capability with simplified assumptions, developing a fuller discussion of Starlink revenues, customers, and value-added services may provide a more informed conversation as to how SpaceX can develop and increase proceeds from the users it captures over the next three to five years.  Satellite and network capabilities will evolve through software upgrades, more satellites, and larger satellites with more power and resources, continuing to increase Starlink capacity.   

There are several ways to frame an ARPU-based discussion, including tiered pricing for residential/consumer customers, value-added services for consumers, consumer vs. business customers, value added services for business customers, and tiered pricing based upon verticals.

Tiered broadband pricing for consumers

Pricing based upon broadband speeds is nothing new for the telecommunications industry. Faster speeds mean more money from DSL, cable, and fiber customers, despite the fact the network equipment is all the same at the point of service delivery. 

SpaceX Starlink beta customers were told they would get downlink speeds of 50Mbps to 150 Mbps, with actual performance varying depending on location, latency of 20ms to 40ms, and a cost of $499 for the equipment and $99/month for the service, billable directly to a credit card. Speeds and services would improve as more satellites were launched and software improved.

In February, SpaceX CEO Elon Musk tweeted that Starlink speeds would double to 300 Mbps later this year and 20ms latency.  This would seem to open the door for Starlink to offer tiered services, charging $99/month for up to 150 Mbps and more for speeds up to 300 Mbps.  

Initial representations of Starlink services by Elon Musk cited gigabit speeds while a February Starlink FCC filing said speeds up to 10 Gbps would be available in the future. SpaceX has provided no public timetable as to when it expects to reach 1 Gbps and 10 Gbps speeds or what improvements are necessary to deliver those speeds to customers.

Following broadband industry practices of “good, better, best,” Starlink could offer tiered services at 150 Mbps, 300 Mbps, and 1 Gbps in the near-term, with 10 Gbps speeds likely reserved for business customers when it becomes available. 

It is unlikely SpaceX will offer speed/pricing package below $99/month in the near-term, due to the expense of producing the Starlink antenna.  While the company offers the dish at a price of $499, estimates on the actual cost to build it range from $2000 to $2500, meaning the company is heavily subsidizing the dish and is amortizing the capital expense across multiple months.

Value-added services for consumers

Based on FCC filings and Starlink engineer comments in Reddit, at least two value-added services will be available for consumers in the future.

Voice services are clearly in the works, based upon SpaceX’s FCC filing to formally register Starlink as a phone company.  SpaceX is still examining its options on how to delivery voice, either by offering multiple third-party services or building and maintaining its own branded service; in either case, the company will have to support connection to 911 emergency call center services as well as providing a “lifeline” dial tone in case of a power outage.

Providing regulated voice services with an associated phone number, 911 public safety access and a battery to keep (VoIP) dial tone operating are additional expenses that won’t come for free. In case of a power outage, SpaceX says it will provide a battery to keep dial tone services available for up to 12 hours.  Basic “dial tone” service could start at $10-15/month.

During the Starlink beta, retail customers are locked into a specific location on the ground and can’t move their dishes very far from that location without having service turned off.  The Starlink dish incorporates a GPS receiver chip to provide dish location, a necessity due to the portability of the dish and making sure Starlink remains in compliance with regulatory requirements around the globe. Licensing of satellite dish operations is done on a country-by-country basis, with Russia among those saying they will not permit Starlink to operate within their boarders.

Digital nomads, RVers, people with second homes, and boaters want the ability to relocate their dishes at a minimum with some desiring operation while physically on the move. On Reddit, Starlink engineers said roaming will be possible in a future software upgrade.  It is possible to likely customers will have to pay for the ability to freely move a dish around the U.S. on a regular basis, more so if involves services delivered to a moving vehicle. 

Consumer vs. business customers

Business customers are typically more demanding than consumers, requiring more bandwidth, additional and symmetrical bandwidth, a responsive help desk/call center in case of service issues, and service level agreements (SLAs) to ensure uptime, latency, and broadband speeds.  Supporting multiple locations and mobility are additional requirements which will result in a premium above the basic $99/month cost.

Value added services for businesses

Hosted PBXes, call centers, unified communications providing unlimited calling plans, voice mail to text, fax services, extensions, CRM integrations and a laundry list of other business features could be used to pump up ARPU.  SpaceX is likely to follow a strategy it described for implementing voice services, first implementing a white-label third party service and then potentially moving services in-house so it can reduce outsourcing costs, improving profitability and increasing ARPU.

Vertical pricing – Better than $99/month

Cellular backhaul is one of the markets SpaceX and other LEO providers wish to pursue. Given the requirements and needs for LTE and 5G, gigabit-class downlink and 100Mbps or higher uplink speeds could be requirements in more developed countries to present a consistent user experience, while evolving economies might “make due” with the expected 2H2021 Starlink 300 Mbps downlink and possibly 20+ Mbps uplink speeds.

Telecommunications carriers will expect SLAs and rapid response technical support to ensure service quality – an email in-box isn’t going to do it. Cellular backhaul services will be priced at what carriers will bear, at some multiple of thousands of dollars per month.

SpaceX is known to be testing Starlink inflight services on up to 5 private jets, per FCC filings.  It is also testing SpaceX onboard its fleet of ocean vessels, including the SpaceX barges where the company lands the Falcon 9 first stages at sea.  Such activities indicate SpaceX plans to sell its service into the private aviation industry and commercial fleets, along with to the various sectors of the maritime industry ranging from private yachts and commercial cruise lines to container ships.

Government customer are going to pay much more than $99/month per dish, due to their requirements and needs.  The U.S. Army and Air Force have and are testing Starlink through a series of exercises and contracts, expecting reliability, customization, and high availability. Revenues in these cases will not be per dish/per month, but multi-year, 6 to 7 figure contracts with dedicated support staff.

At the end of the day, it’s all about the ARPU

Regardless of the customer mix and types of services, SpaceX has already shown it will have an expansive customer base above and beyond a simple $99/month consumer model.   However, it is also clear that SpaceX needs – and most likely has — paths to boost and increase ARPU within its consumer model through value-added services and potentially through tiered broadband speed offerings. 

SpaceX has been much less transparent about its business model, but the company is talking to the oil and gas industry and has its eye on both maritime and aviation sectors at a minimum. Business customers will most certainly bring in more than $99/month/dish.

There are no indications currently SpaceX plans to build a large-scale reseller force, since corporate culture – i.e. Elon Musk – prefers to own the customer directly rather than having a middle man.  However, the company may become more active in establishing key relationships within certain verticals, such as aviation and maritime, in order to get to “better,” higher revenue customers more quickly.