Definitions:
- Falcon 9 is its rocket system that launches satellites, people & cargo into orbit.
- Falcon Heavy is Falcon 9 w/2 extra rocket boosters for much larger launch capacity.
- Dragon is its spacecraft that carries people & goods.
- Starship is its next-gen rocket currently moving from testing to commercial use.
A lot of Launch Segment Firsts:
- 2008 -- First private company to successfully launch a liquid-fuel rocket into low Earth orbit.
- 2012 -- First private company to guide & dock a spacecraft at the International Space Station.
- 2015 -- First to use a rocket's own downward thrust to land (propulsive landing).
- 2017 -- First rocket re-flight.
- 2020 -- First private company to dock astronauts at the International Space Station.
Dominant Launch Segment Positioning & Roadmap:
- Falcon 9 has a 99%+ mission success rate. This is 85% cheaper per kilogram to launch than average (per NASA). The lead has only grown from there. Falcon Heavy is 92% cheaper per kilogram than the average substitute.
- Cost edges come from things like the top piece of their rocket (the upper stage) costing SpaceX $5M to build vs. $50M for legacy competition. The power of vertical integration.
- Falcon 9 + Falcon Heavy have an 80% share of global mass to orbit since 2023.
- Starship is next. This will come with a fully reusable design that can immediately refuel and launch again. Starship can also carry 3x-4x the weight vs. Falcon Heavy in low Earth orbit... making it perfect to deploy giant satellites & data centers. They hope this will boost their cost per kilogram lead from 92% with Falcon Heavy to 99% for Starship use cases.
- They aim to charge $90M/launch on Starship, which is far cheaper than other options.
- A lot of launch capacity right now is dedicated to Starlink rather than 3P customers. This gives them a lot more flexibility to cater to demand while servicing their own needs. It also means they're not stuck with excess capacity if a partner cancels or delays a launch. They can use it for Starlink.
- While margins for the segment are pressured due to investments right now, they make more than $67M per launch & it costs them $15M to run it. The unit economics are good & should support healthy long-term margins.
Starlink & Connectivity:
- 2022 -- First company to build "consumer-grade phased-array terminals." These are flat, stationary satellite dishes that use software to steer radio signals so they can track moving satellites without actually moving themselves. That simplifies design, lowers cost & improves durability.
- 2025 -- First commercially scaled direct-to-mobile satellite cluster... bypassing the need to connect to cell tower service that is often unavailable.
- Starlink has grown from just over 2M users in 2023 to 10M+ in Q1 2026.
- Extending connectivity to deeply remote areas & using it for mission-critical things like emergency response around the world are resonating (not to mention commercial opportunities with airlines).
- One thing to keep an eye on is pricing power. They cut pricing from $99 in 2023 to $81 for 2025. It fell from $86 in Q1 2025 to $66 in Q1 2026. While that's not amazing, it was to take advantage of excess capacity already in place & to steal share faster. It's also a byproduct of mix-shift to lower ARPU countries.
- Starlink has a 63% EBITDA margin.
- The larger, more distributed low orbit satellite base creates a 95% latency edge vs. the competition due to closer proximity to the point of consumer connection. Setup is way easier, there are no cancellation fees & customers can easily pause service whenever they want to. Pay when you need to use it.
- While overall cost for Starlink is a bit higher than some more antiquated solutions, the vastly better performance means cost per megabit is a lot lower.
xAI:
- Part of SpaceX as of January 31st.
- Grok has 550M monthly active accounts.
- These investments are why GAAP EBIT swung from a 2024 gain to a 2025 loss. Connectivity is deeply profitable. Launch Systems aren't losing nearly as much.
- Floated a $28.5 trillion TAM for this segment. That’s not a typo.. but take with a very large grain of salt.
- Anthropic will pay SpaceX $1.25B per month for access to compute within its AI segment. That equates to $15B per year for a segment that did just $3.2B in 2025 revenue.
Quick Take:
I’m not interested in owning this company at the current market cap. I find the business deeply impressive and I think there are a lot of extremely attractive assets here. I just won't invest in them unless I can do so at what I find to be the right price. I think I'm going to keep following this company and covering major news so I’m prepared to hopefully own it in the future if it's not quite so expensive. There’s a decent chance that never happens, but I need to stay disciplined and accept that as a risk.