Shaun Maguire, the Sequoia Capital partner who led the firm’s investments in SpaceX, published a company-wide financialmodel on X on 30 October that puts SpaceX’s revenue at $731.1 billion in 2028, up from $51.5 billion in 2026 and $273.8 billion in 2027. He wrote that “Sequoia Capital has a large position in SpaceX so I’m extremely biased.”
By our arithmetic, his 2028 figure is about 12.5 times SpaceX’s third-quarter 2026 sales run rate: the $14,575 million SpaceX reported for the quarter, multiplied by four, is about $58.3 billion. The model is his projection, not company guidance.
The key insight: As we read it, Maguire’s model treats SpaceX mainly as a seller of computing: Terrestrial Compute is about 78% of his $731.1 billion 2028 revenue, by our arithmetic.
Maguire wrote that he built the model after SpaceX’s earnings call and that “I think we are about to see a revenue ramp at SpaceX whose only parallel is NVIDIA’s ramp over the last three years.” He set that view against what he listed as possible exceptions: war, an AI moratorium or revolution.
His table’s third-quarter 2026 revenue, $14,575 million, is the figure SpaceX reported for the quarter, which we covered in SpaceX Q3 revenue jumps 161% to $14.6B on AI and Starlink. From there, his table shows quarterly revenue of $105.3 billion by the fourth quarter of 2027 and $237.4 billion by the fourth quarter of 2028.
Elon Musk replied on X, “I think we can do better.” Maguire answered that “This was my 25th-50th percentile estimate.” In the article itself he wrote, “But this is my 50th percentile forecast.”
Shaun Maguire, a Sequoia Capital partner who led the firm’s SpaceX investments, published an X Article on 30 October titled ‘People don’t understand SpaceX’, with his model of SpaceX revenue rising to $731.1 billion in 2028. He writes that Sequoia has a large position in SpaceX and that he is extremely biased. It matters because the model puts a number, in his view, on how much of SpaceX’s revenue could come from selling compute.
@shaunmmaguire @Indian_Bronson I think we can do better
Elon Musk replied to the model the same evening: ‘I think we can do better.’ Maguire answered that the model was his 25th-50th percentile estimate. Musk’s reply gives no figures of its own; it matters because SpaceX’s chief executive responded publicly to an outside forecast of his company.
Shaun Maguire’s forecast of SpaceX revenue by segment for 2026-2028, from the model table in his 30 October 2026 X Article; his projection, not company guidance. The 12.5x comparison with SpaceX’s third-quarter 2026 revenue of $14,575 million times four is our arithmetic.
Where the Revenue Comes From
The table splits SpaceX into three segments: Space, Connectivity and SI. For 2028 it lists $17.3 billion for Space, $88.5 billion for Connectivity and $625.2 billion for SI.
Inside SI, the line labelled Terrestrial Compute is $569.0 billion in 2028, which is about 78% of his total, by our arithmetic. The other SI lines are “Grok + Cursor + Other” at $76.1 billion, Orbital Compute at $31.0 billion and Terrafab at $5.2 billion. The SI segment is where the AI customer behind 34.5% of SpaceX’s third-quarter revenue sits, as we reported in one AI customer made up 34.5% of SpaceX’s Q3 revenue.
In Connectivity, the table lists Starlink, including Starshield, at $64.7 billion in 2028 and Direct to Cell at $23.9 billion. Maguire wrote, “Personally, I’d wager that Direct to Cell will do more revenue than Starlink in 2029 or 2030 under almost any scenario.”
Business Pill · Capex and Depreciation Explained
Capex and depreciation explained: why the cost and lifetime of AI hardware decide whether compute revenue turns into profit, the variables at the centre of Maguire’s model.
The Cost Side of the Model
The table puts capital spending at $88.0 billion in 2026, $290.0 billion in 2027 and $369.0 billion in 2028, about $747 billion over the three years by our arithmetic. Its change-in-cash line is negative $77.8 billion in 2026 and negative $122.6 billion in 2027, about $200.4 billion combined by our arithmetic, before a positive $114.9 billion in 2028.
The assumptions block lists 21.0 gigawatts of terrestrial compute by the end of 2028 and 250 Starship flights in 2028. It sets annual revenue per gigawatt at $40,000 million from the third quarter of 2027 and capex per gigawatt at $30,000 million from the fourth quarter of 2027.
Maguire wrote that “There are wide error bars in certain areas, especially: revenue per gigawatt, CapEx per gigawatt and Terafab CapEx.” He also wrote, “I also think it’s very unlikely we see anything below $20B/GW within the next ~4 years.” Our earlier coverage of the cost debate around compute in orbit is in Google energy chief says space data centers cost up to 100x.
Orbital Compute and Terafab
The table shows $31.0 billion of Orbital Compute revenue in 2028 on 2.8 gigawatts. In the article Maguire wrote, “We didn’t even talk about Orbital Compute as that really shows up starting in 2029.” Musk’s own orbital claim is in Musk says new orbital compute could top Earth’s by late 2028.
As we read the table, the bet rests on two numbers he set himself: $40,000 million of annual revenue per gigawatt and $30,000 million of capex per gigawatt, which he says carry wide error bars.
As we read the figures, his $731.1 billion is about 12.5 times SpaceX’s third-quarter 2026 sales run rate of about $58.3 billion, by our arithmetic.
Shaun Maguire, Sequoia Capital, X Article, 30 Oct 2026
“SpaceX is the low-cost provider in compute”
Three Implications
FOR SPACEX’S CASH NEEDS Maguire’s table puts capex at $369.0 billion in 2028 and shows cash falling by $77.8 billion in 2026 and $122.6 billion in 2027 before rising in 2028.
FOR COMPUTE PRICING Maguire wrote that he believes rev/GW and CapEx/GW are coupled constants, and that a compute cost below $20B per gigawatt within about four years is very unlikely in his view.
FOR STARLINK Maguire wrote that each Starship flight will be the equivalent of about 20 Falcon 9 launches, about $1B a year of revenue per launch in his math.
The framework states its thesis: “the capex race is no longer a corporate spending cycle.”
As we read it, Maguire’s model sits inside that argument: it has SpaceX spending about $747 billion on capex over three years, by our arithmetic. The framework is the owner’s argument and does not discuss SpaceX.
SpaceX has not published a forecast matching these numbers in the documents we read. Maguire wrote, “I didn’t justify most of my numbers.” He labels his post “not investment advice.”
He describes the same model as his 50th percentile forecast in the article and as his 25th-50th percentile estimate in a reply. The table figures here are as we read them from the image in his post.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
A Sequoia partner with a disclosed SpaceX stake has put a number on SpaceX’s AI bet: $731.1 billion of revenue in 2028, about 12.5 times the company’s third-quarter run rate by our arithmetic, with about 78% of it from terrestrial compute. His model also puts about $747 billion into capex over three years by our arithmetic, and Maguire wrote that he is “extremely biased” and that “I didn’t justify most of my numbers.”
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The forecast figures are from Shaun Maguire’s X Article of 30 October 2026, as we read them from the model table image in his post; SpaceX’s third-quarter revenue is from SpaceX’s filing as we reported it earlier. The 12.5x, 78%, $747 billion and $200.4 billion figures are our arithmetic. We did not contact SpaceX or Sequoia.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
