← Space Exploration Technologies Corp.
SPCX · Forward model · Elon case
The Elon case, 20 quarters out
Each segment is projected from its own operating driver, rolled up into consolidated cash flow, and discounted back to a fair value per share. The assumptions below are editable — change them and every number on this page moves with them.
SpaceX reports three segments; this model carries six. Connectivity is split into Starlink and Starlink Mobile, AI into Terrestrial AI and Space AI, and Developer Platform is added outright, because each pair mixes two businesses with different economics and very different risk. Starlink is the dish-and-antenna business - broadband, enterprise, maritime, aviation, rail, Starshield and today's light direct-to-cell messaging - modelled bottom-up on subscribers x ARPU. Starlink Mobile is the enhanced direct-to-handset tier that the V2 Mobile satellites and the EchoStar spectrum unlock, and it earns nothing until 2028 Q1. Terrestrial AI is modelled on energised compute capacity x utilisation x price, because revenue there is gated by what is switched on rather than by demand; Space AI on the same shape but from zero, with six quarters of no revenue before the first demonstrator earns anything; Space carries only EXTERNAL launch and Dragon revenue, modelled on flights x price (internal Starlink launches are not segment revenue). The two new verticals have NO reported history and no carved-out base. SpaceX has never published a sub-split of Connectivity or of AI, so rather than invent one, the actuals stay cut the way they were reported and both new lines start from zero in the projection. Only 2026 Q2 is a reported segment split ($4.29B / $0.96B / $2.56B, $7.81B total). The 2025 quarters are apportioned from the disclosed full-year segment totals ($11.4B / $4.1B / $3.2B) and 2026 Q1 is apportioned from the reported $4.694B consolidated quarter - every one of those is marked ESTIMATED on this page. Developer Platform is Cursor and Origin, acquired for $60B in all stock and closed August 14, 2026 — after the end of every quarter shown here, which is why it is zero across the whole reported history and starts earning in 2026 Q3.
Musk's own stated targets, taken at face value and translated into per-vertical growth: Starlink 10x on V3 bandwidth, Terrestrial AI at 10 GW and $30-50 a watt, Starlink Mobile at hundreds of millions of devices, Starship at a flight a day, and orbital compute scaling toward a million satellites. Two legs have no Musk post behind them and are ours. Developer Platform is our extension of the announced Cursor purchase and the Origin launch - Origin displaces GitHub outright and agent-driven seats keep compounding, which gets the line to roughly $31B a year against Cursor's $4B run rate today. Model layer is pushed to a 60% terminal share of the pool: Grok does not just compete with Anthropic and OpenAI, it becomes the default, worth about $36B a quarter by 2031 Q2 against $0.55B today. Note what the case does NOT reach: his $100B ARR by end-2026 needs a step-change no driver here produces off a $7.81B quarter, and $1T by 2030 sits above even this case. And note what the 60% share costs to believe - it is a bigger business than every megawatt SpaceX leases to everyone else, which is the arithmetic the case has to answer.
Latest: $229.16B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $4.30B |
| 2025Q2 | $4.58B |
| 2025Q3 | $4.80B |
| 2025Q4 | $5.02B |
| 2026Q1 | $4.69B |
| 2026Q2 | $7.81B |
| 2026Q3E | $10.59B |
| 2026Q4E | $12.63B |
| 2027Q1E | $14.97B |
| 2027Q2E | $17.65B |
| 2027Q3E | $20.70B |
| 2027Q4E | $24.17B |
| 2028Q1E | $28.43B |
| 2028Q2E | $33.44B |
| 2028Q3E | $39.34B |
| 2028Q4E | $46.29B |
| 2029Q1E | $54.44B |
| 2029Q2E | $64.00B |
| 2029Q3E | $75.20B |
| 2029Q4E | $88.30B |
| 2030Q1E | $103.59B |
| 2030Q2E | $121.68B |
| 2030Q3E | $142.80B |
| 2030Q4E | $167.38B |
| 2031Q1E | $195.96B |
| 2031Q2E | $229.16B |
What drives each segment
Starlink
Subscribers × ARPUThe whole dish-and-antenna business: consumer broadband, the better-priced enterprise, maritime, aviation and rail book, Starshield government work, and the light direct-to-cell messaging service partners resell today. Consumer ARPU falls as emerging markets mix in; the non-consumer book is what carries the segment.
Latest: $44.82B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $2.40B |
| 2025Q2 | $2.75B |
| 2025Q3 | $3.00B |
| 2025Q4 | $3.25B |
| 2026Q1 | $3.50B |
| 2026Q2 | $4.29B |
| 2026Q3E | $4.92B |
| 2026Q4E | $5.62B |
| 2027Q1E | $6.39B |
| 2027Q2E | $7.25B |
| 2027Q3E | $8.19B |
| 2027Q4E | $9.23B |
| 2028Q1E | $10.37B |
| 2028Q2E | $11.63B |
| 2028Q3E | $13.01B |
| 2028Q4E | $14.54B |
| 2029Q1E | $16.23B |
| 2029Q2E | $18.09B |
| 2029Q3E | $20.16B |
| 2029Q4E | $22.45B |
| 2030Q1E | $25.00B |
| 2030Q2E | $28.09B |
| 2030Q3E | $31.58B |
| 2030Q4E | $35.50B |
| 2031Q1E | $39.89B |
| 2031Q2E | $44.82B |
Assumptions & reasoning
- Modelled bottom-up on subscribers x ARPU rather than a growth rate, because the two halves move in opposite directions: the base keeps compounding while price per user falls on mix. A single growth number would hide exactly the tension that decides this segment.
- The non-subscriber book is not derived - it is backed out as segment revenue less subscribers x ARPU, which puts it at $1.91B a quarter, about 45% of the segment against the disclosed ~42% enterprise/government share. The residual absorbs any error in the ARPU split.
- Margin glides 61% to 66% because constellation cost is largely fixed once launched, so each added subscriber and each enterprise contract drops through at high incremental margin. Capex intensity falls 34% to 22% as V3 satellites raise capacity per launch.
- The load-bearing risk is ARPU, not subscribers. Held flat at $66 instead of drifting down, this segment alone is worth materially more; at the $45 floor it is worth much less. Kuiper is the reason the floor might bind sooner than modelled.
- Today's direct-to-cell service belongs here, not in Starlink Mobile. It runs on ~650 first-generation satellites at roughly 2-4 Mbps - texting, location and light apps in dead zones - and it is sold as a bolt-on by carriers, so it is already inside this segment's non-subscriber book. About 22M monthly actives as of the August 2026 all-hands, up from ~10M earlier in the year.
Starlink Mobile
Subscribers × ARPUThe enhanced direct-to-handset service, not the messaging bolt-on that exists today. It needs the V2 Mobile satellites - ~50 per Starship flight, roughly 100x the data density - and the EchoStar spectrum before there is enough capacity to sell rather than to bundle. That makes it a 2028 business with a 2026 announcement, and the gap is the whole risk.
Latest: $4.93B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $0.00 |
| 2025Q2 | $0.00 |
| 2025Q3 | $0.00 |
| 2025Q4 | $0.00 |
| 2026Q1 | $0.00 |
| 2026Q2 | $0.00 |
| 2026Q3E | $0.00 |
| 2026Q4E | $0.00 |
| 2027Q1E | $0.00 |
| 2027Q2E | $0.00 |
| 2027Q3E | $0.00 |
| 2027Q4E | $0.00 |
| 2028Q1E | $56M |
| 2028Q2E | $142M |
| 2028Q3E | $263M |
| 2028Q4E | $421M |
| 2029Q1E | $620M |
| 2029Q2E | $863M |
| 2029Q3E | $1.15B |
| 2029Q4E | $1.50B |
| 2030Q1E | $1.90B |
| 2030Q2E | $2.36B |
| 2030Q3E | $2.88B |
| 2030Q4E | $3.48B |
| 2031Q1E | $4.16B |
| 2031Q2E | $4.93B |
Assumptions & reasoning
- Modelled on monetised users x wholesale ARPU because the economics are a revenue share, not a price SpaceX sets. Growth comes from carriers signing, not from users churning in and out.
- Zero revenue for six quarters is the point of this vertical. The light service already earns money and is counted in Starlink; what is modelled here is only the step-change the V2 satellites and the EchoStar spectrum unlock, and none of it can be sold before end-2027.
- Margin glides 40% to 68% - higher than the dish business, because the incremental cost of a direct-to-cell user is close to zero once the satellites are up. Capex intensity starts at 140% while the dedicated V2 fleet is built and falls to 25%.
- Reaching roughly 70M monetised users by 2031 sounds aggressive against ~22M light actives today, but the constraint is carrier packaging rather than coverage, and coverage is already near-global. The load-bearing assumption is ARPU: at $3 rather than $6 this vertical is worth half, and the carriers set that number, not SpaceX.
Space
Units × priceExternal commercial and government launch plus Dragon crew/cargo and Starship development milestones. A cost centre with strategic optionality: it subsidises the other two segments at internal cost and only approaches break-even as Starship reusability lands.
Latest: $4.15B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $1.20B |
| 2025Q2 | $1.05B |
| 2025Q3 | $950M |
| 2025Q4 | $900M |
| 2026Q1 | $450M |
| 2026Q2 | $960M |
| 2026Q3E | $1.03B |
| 2026Q4E | $1.11B |
| 2027Q1E | $1.20B |
| 2027Q2E | $1.29B |
| 2027Q3E | $1.38B |
| 2027Q4E | $1.49B |
| 2028Q1E | $1.60B |
| 2028Q2E | $1.72B |
| 2028Q3E | $1.85B |
| 2028Q4E | $2.00B |
| 2029Q1E | $2.15B |
| 2029Q2E | $2.31B |
| 2029Q3E | $2.49B |
| 2029Q4E | $2.67B |
| 2030Q1E | $2.88B |
| 2030Q2E | $3.10B |
| 2030Q3E | $3.33B |
| 2030Q4E | $3.58B |
| 2031Q1E | $3.86B |
| 2031Q2E | $4.15B |
Assumptions & reasoning
- Only external customers are counted. Internal Starlink launches are by far the larger share of flights but are an internal cost transfer, not revenue — treating them as revenue would double-count the constellation the Connectivity segment already pays for.
- Modelled on flights x price because both move independently and in opposite directions: Starship adds cadence while reusability lowers what the market will pay per kilogram. A revenue growth rate would collapse two opposing forces into one number.
- Margin starts deeply negative at -35% and glides only to 8% by the end of the horizon. Starship development is expensed here while its benefit accrues to Connectivity, so this segment carries cost for value that shows up in another line.
- This is the segment where the model is least sensitive and least confident. It is roughly 5% of revenue by the final quarter, so even a large error barely moves the fair value — but a Starship schedule slip would also slow the constellation, and that does move it.
Terrestrial AI
Capacity × utilisation × priceGround-based compute leasing is the explosive line, and it is capacity-gated: revenue tracks energised megawatts times what is contracted times price. xAI/Grok and X advertising ride on top. The highest-variance vertical in the model - the contracts are real, the pricing power past the first tranche is not yet proven.
Latest: $79.12B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $700M |
| 2025Q2 | $780M |
| 2025Q3 | $850M |
| 2025Q4 | $870M |
| 2026Q1 | $744M |
| 2026Q2 | $2.56B |
| 2026Q3E | $2.65B |
| 2026Q4E | $3.41B |
| 2027Q1E | $4.30B |
| 2027Q2E | $5.36B |
| 2027Q3E | $6.59B |
| 2027Q4E | $8.03B |
| 2028Q1E | $9.70B |
| 2028Q2E | $11.65B |
| 2028Q3E | $13.90B |
| 2028Q4E | $16.52B |
| 2029Q1E | $19.54B |
| 2029Q2E | $23.03B |
| 2029Q3E | $27.07B |
| 2029Q4E | $31.73B |
| 2030Q1E | $37.11B |
| 2030Q2E | $43.31B |
| 2030Q3E | $50.46B |
| 2030Q4E | $58.70B |
| 2031Q1E | $68.19B |
| 2031Q2E | $79.12B |
Assumptions & reasoning
- Modelled on energised capacity x utilisation x price because revenue here is supply-gated, not demand-gated. The contracts are signed well ahead of the power being available, so what is switched on sets the ceiling — a demand-side growth rate would be modelling the wrong constraint.
- The build rate is the engine: 400 MW a quarter compounding at 10%. That is derived from roughly $14B of quarterly capex at current build costs, which is why capex intensity starts at an extreme 520% of segment revenue and decays as revenue catches up with the asset base.
- Price is assumed to fall 1.5% a quarter. Every hyperscaler is building at once, so the pricing power visible in the first tranche of contracts is unlikely to survive multi-gigawatt supply. Holding price flat is the single most flattering change available in this model.
- This is the highest-variance segment and the one that decides the answer: it goes from a third of revenue to over half, and its present value is deeply negative inside the horizon because the capex lands years before the cash. If the utilisation target is wrong, the valuation is wrong.
- Orbital compute was previously carried here as unmodelled optionality. It is now its own vertical, so every megawatt in this line is on the ground.
- Fifteen points of utilisation were moved to the Model layer vertical. 1,800 MW x 15% x $2.03M is $548M a quarter - the capacity Grok runs on. Internal inference is not a third-party lease, so billing it here and again as model revenue would count the same megawatt twice.
Space AI
Capacity × utilisation × priceOrbital data centres: compute in vacuum, cooled by radiation and powered by uninterrupted sunlight, launched on Starship at a marginal cost only SpaceX has. Pre-revenue today and the only vertical here that is pure option value - it is worth nothing at all if Starship cadence slips, and it is worth more than the rest of the model combined if it works.
Latest: $52.23B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $0.00 |
| 2025Q2 | $0.00 |
| 2025Q3 | $0.00 |
| 2025Q4 | $0.00 |
| 2026Q1 | $0.00 |
| 2026Q2 | $0.00 |
| 2026Q3E | $0.00 |
| 2026Q4E | $0.00 |
| 2027Q1E | $0.00 |
| 2027Q2E | $0.00 |
| 2027Q3E | $0.00 |
| 2027Q4E | $0.00 |
| 2028Q1E | $241M |
| 2028Q2E | $668M |
| 2028Q3E | $1.33B |
| 2028Q4E | $2.30B |
| 2029Q1E | $3.64B |
| 2029Q2E | $5.44B |
| 2029Q3E | $7.80B |
| 2029Q4E | $10.82B |
| 2030Q1E | $14.64B |
| 2030Q2E | $19.43B |
| 2030Q3E | $25.35B |
| 2030Q4E | $32.61B |
| 2031Q1E | $41.48B |
| 2031Q2E | $52.23B |
Assumptions & reasoning
- Deployment is derived from launch cadence, not assumed directly: flights per quarter x the share of the manifest given to compute x 40 satellites x 160 kW. At one flight a day with half the manifest, that is about 290 MW a quarter — 156 flights to the gigawatt.
- Cadence ramps toward one flight a day and stops there. One a day with half the manifest is about 291 MW a quarter, so the base case is a claim that orbital compute settles at roughly a gigawatt every three and a half quarters — not that cadence keeps compounding. Faster or higher belongs in the bull and Elon cases.
- This replaced a fixed 3 MW a quarter compounding at 15%, which reached only ~120 MW by 2031 — roughly a tenth of what the flight arithmetic supports. The old number was not a view about launch capability; it was a placeholder that had never been reconciled against one.
- Utilisation stays well below the ground business. Latency rules out inference and anything interactive, so this competes for training and batch work only, and it prices at a premium to ground only while power and permitting are the binding constraint there.
- The swing factor is cadence, and it is unproven. Every megawatt needs a launch slot, so if Starship holds at one flight a week rather than one a day this line stays a rounding error. Move the ceiling slider to see it: the assumption is doing more work here than anywhere else in the model.
- The manifest is not a single pool. This share is of Starship flights, while the Space vertical counts external customer missions across Falcon and Starship both. Do not read the two flight numbers as competing for the same slots — but do treat a high compute share as an implicit claim that Starlink deployment has moved off Starship or become far more efficient per flight.
Model layer
Capacity × utilisation × priceGrok sold as a model rather than as a megawatt: tokens, seats and API calls, competing with Anthropic and OpenAI for a pool that booked a combined $18.3B in the June quarter. Modelled as market x share, because share is the only quantity the argument for this line is ever made in - and the finding is that the pool's growth rate matters more than the share does.
Latest: $36.13B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $0.00 |
| 2025Q2 | $0.00 |
| 2025Q3 | $0.00 |
| 2025Q4 | $0.00 |
| 2026Q1 | $0.00 |
| 2026Q2 | $0.00 |
| 2026Q3E | $894M |
| 2026Q4E | $1.30B |
| 2027Q1E | $1.77B |
| 2027Q2E | $2.31B |
| 2027Q3E | $2.94B |
| 2027Q4E | $3.67B |
| 2028Q1E | $4.50B |
| 2028Q2E | $5.46B |
| 2028Q3E | $6.56B |
| 2028Q4E | $7.83B |
| 2029Q1E | $9.28B |
| 2029Q2E | $10.93B |
| 2029Q3E | $12.83B |
| 2029Q4E | $15.00B |
| 2030Q1E | $17.47B |
| 2030Q2E | $20.29B |
| 2030Q3E | $23.51B |
| 2030Q4E | $27.18B |
| 2031Q1E | $31.37B |
| 2031Q2E | $36.13B |
Assumptions & reasoning
- SpaceX does not break out Grok. There is no line in the segment disclosure that corresponds to this vertical, so it carries zero across every reported quarter and starts from nothing in the projection, like Space AI and Developer Platform.
- The $0.55B starting run rate is not new revenue. It is the capacity Grok already consumes, moved here out of Terrestrial AI's utilisation - fifteen points of 1,800 MW at $2.03M. Reported June-quarter revenue is unchanged; only its character is.
- Market x share is a landlord's model turned inside out. Terrestrial AI prices the building; this prices the tenant. Both are guesses about the same megawatts, which is why one of them had to give up fifteen points of utilisation.
- The pool growth rate is worth more than the share. Ten percent of a pool compounding at 10% a quarter beats a quarter of one compounding at 5%, and both beat half of a pool that stops - so a reader who wants to argue with this line should move the build rate, not the share.
- Capex at 8% terminal is software economics, and it is the assumption most likely to be wrong: if serving Grok needs megawatts SpaceX would otherwise lease, the cost of this line is Terrestrial AI's revenue rather than its own capex.
Developer Platform
Subscribers × ARPUCursor and Origin: the AI coding editor SpaceX bought for $60B in all stock on August 14, 2026, and the code-hosting layer it shipped three days later. Cursor was annualising above $4B on its own before the close; Origin is the attempt to take the repository itself, which is the part of GitHub that carries the switching costs. The line is worth what Origin's attach rate turns out to be, not what the editor already earns.
Latest: $7.79B (2031Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $0.00 |
| 2025Q2 | $0.00 |
| 2025Q3 | $0.00 |
| 2025Q4 | $0.00 |
| 2026Q1 | $0.00 |
| 2026Q2 | $0.00 |
| 2026Q3E | $1.09B |
| 2026Q4E | $1.19B |
| 2027Q1E | $1.31B |
| 2027Q2E | $1.44B |
| 2027Q3E | $1.59B |
| 2027Q4E | $1.76B |
| 2028Q1E | $1.95B |
| 2028Q2E | $2.17B |
| 2028Q3E | $2.41B |
| 2028Q4E | $2.68B |
| 2029Q1E | $2.99B |
| 2029Q2E | $3.33B |
| 2029Q3E | $3.70B |
| 2029Q4E | $4.12B |
| 2030Q1E | $4.59B |
| 2030Q2E | $5.11B |
| 2030Q3E | $5.68B |
| 2030Q4E | $6.31B |
| 2031Q1E | $7.01B |
| 2031Q2E | $7.79B |
Assumptions & reasoning
- This vertical has no history and never will have one before 2026 Q3. SpaceX closed the acquisition on August 14, 2026, six weeks into the September quarter, so not one dollar of Cursor revenue sits in any reported SpaceX quarter — the zeros across 2025 and the first half of 2026 are fact, not padding.
- The first projected quarter is nonetheless overstated. The model books 2026 Q3 as a full quarter of ownership when SpaceX owned Cursor for roughly six of its thirteen weeks, which flatters that quarter by somewhere near $0.5B and washes out entirely by 2026 Q4.
- The 2026 Q3 base is Cursor's last circulating pre-acquisition run rate expressed as a quarter, not a growth assumption: $4B annualised is $333M a month, and three of those months is $1.00B — which is what 4.17M seats at $80 comes to. Deutsche Bank's ~$12B Cursor contribution to its $100B year-end target is that same $4B tripled, and it is an extrapolation rather than a reported figure. Cursor's own year-end guide, given in April, was above $6B. This line carries the disclosed run rate; the tripling sits in Bull and Elon, where a claimed number belongs.
- Seats are backed out, price is assumed — the reverse of the usual order, and worth knowing which half to argue with. Cursor discloses a run rate above $4B and more than a million paying customers, but an enterprise customer is not a seat. The $80 blended monthly price is ours; 4.17M seats is what that price implies against the disclosed run rate. Halve the price and the seat count doubles for the same revenue, so the level here is safer than either input.
- The base case takes only a modest share of GitHub, and takes it late. Adds start at 60K seats a quarter — Cursor's own organic pace, no repositories moving — and the ramp is tuned against that organic path so that the extra seats Origin wins are worth about $0.1B a quarter by 2027 Q4, roughly 8% of the line, and compound from there. Penetration of the professional developer base sits near 15% through 2027 H1 and reaches about 25% by 2031, for roughly $2.4B a quarter. Displacing GitHub outright is worth about $4B a year at our estimate of its current run rate, and that sits in the Bull and Elon cases, not here.
- Margin starts at 5% because inference is the cost of goods here and Cursor has been buying it. Inside SpaceX it increasingly will not be, which is the acquisition's clearest operating logic — but note that internal inference consumption is NOT Terrestrial AI revenue, exactly as internal Starlink launches are not Space revenue. Do not read a rising developer margin as a bigger AI segment.
- Capex intensity is 12% falling to 6%, the lowest in the model, because the GPUs this business runs on are built and charged inside Terrestrial AI. If SpaceX ever discloses compute dedicated to Cursor, that spend belongs here and this line is understated until it does.
- The ramp is a glide, not a step, because the subscription driver has no delayed-inflection field — only startsAfterQuarters, which zeroes revenue outright and would be wrong for a business already earning $4B a year. The 2027 H2 inflection is therefore a shape we tuned, not a mechanism the engine enforces. Drag the ramp slider and it moves from the first projected quarter.
- The four cases disagree about this line more than about any other, which is the right outcome for an asset bought eight weeks ago. Bear has it at $0.97B a quarter in 2031 — below where Cursor is running today, a $60B purchase that never grew. Base reaches $2.4B, Bull $4.6B and Elon $7.8B. Nothing in the reported history constrains any of them, so read the spread rather than the point estimate.
Where each case comes from
Elon case — primary sources
The primary sources this case is built from — filings, calls and posts. Where they are claims by an interested party rather than disclosures, every number in the Elon column is what happens if they are taken at face value.
Non-human traffic will dominate
Only Starlink can carry it
Market share and revenue potential
Capacity leap: V3 and beyond
- Aug 8, 2026 V3 satellites are an order of magnitude more capable and an order of magnitude more will be launched — >100X the bandwidth. Even at 10X lower revenue per gigabit, communications revenue >$200B a year.
- Aug 9, 2026 Total constellation including direct-to-cell to exceed 100k satellites. Future versions may exceed 50X the throughput of a V2.
Orbital compute
The $1T company target
- Jun 14, 2026 I think SpaceX might be able to reach approximately $1T revenue in 2030
- Jun 14, 2026 And I would be surprised if revenue is not greater than $1T in 2031
- Aug 8, 2026 Clarifies the $1T target is actual revenue in the year 2030 itself, not cumulative, and that these are internal projections.
From cash flow to fair value
| Present value of free cash flow, 20 quarters | -$58.26B |
| Terminal-year revenue | $735.30B |
| Terminal-year EBITDA | $368.23B |
| Exit multiple, on revenue | 25.0x |
| Terminal value | $18.38T |
| Discounted at 9.5% a year, terminal value becomes | $11.68T |
| Enterprise value | $11.62T |
| Net cash | $5.00B |
| Equity value | $11.62T |
| Diluted shares | 13.18B |
| Fair value per share | $882.19 |
| Against the current price of $134.00 | +558% |
20x is above what the current share price implies, which is about 14x. It is held there because the terminal year is still growing near 40% with a rising mix of connectivity and compute revenue, which is a different business from the one the trailing multiple is priced against. It is still the largest single input in the model, and the least derivable from the operating assumptions above — treat the fair value as a function of this number first and everything else second.
Read the other way round: at $134.00 the market is paying 3.9x terminal-year revenue, holding every other assumption on this page fixed. That is the number to argue about.
Capex outside the verticals
Each vertical's capex is a share of its own revenue, which is the right shape for capacity that scales with sales. These are not that: fixed-size programmes with their own schedule that spend whether or not any segment grows. They are added on top, and they are why free cash flow dips in the years below.
Terafab phase 1 (SpaceX share)
2028 Q1 → 2030 Q4The joint Tesla/SpaceX advanced fab, $16.8B announced for phase 1. SpaceX leads the large-scale manufacturing portion, so 60% is assumed here. County filings point to a far larger build-out — $55B for a first phase and up to $119B complete — which this line deliberately does NOT carry.
The projected path
| Quarter | Starlink | Starlink Mobile | Space | Terrestrial AI | Space AI | Model layer | Developer Platform | Revenue | YoY | EBITDA | Capex | FCF | R40 | PV of FCF |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 Q3E | $4.92B | $0 | $1.03B | $2.65B | $0 | $894M | $1.09B | $10.59B | +121% | $3.74B | $13.91B | -$10.17B | +24 | -$9.95B |
| 2026 Q4E | $5.62B | $0 | $1.11B | $3.41B | $0 | $1.30B | $1.19B | $12.63B | +152% | $4.67B | $15.07B | -$10.39B | +69 | -$9.93B |
| 2027 Q1E | $6.39B | $0 | $1.20B | $4.30B | $0 | $1.77B | $1.31B | $14.97B | +219% | $5.79B | $16.12B | -$10.33B | +150 | -$9.65B |
| 2027 Q2E | $7.25B | $0 | $1.29B | $5.36B | $0 | $2.31B | $1.44B | $17.65B | +126% | $7.11B | $17.10B | -$9.99B | +69 | -$9.12B |
| 2027 Q3E | $8.19B | $0 | $1.38B | $6.59B | $0 | $2.94B | $1.59B | $20.70B | +96% | $8.67B | $18.06B | -$9.39B | +50 | -$8.38B |
| 2027 Q4E | $9.23B | $0 | $1.49B | $8.03B | $0 | $3.67B | $1.76B | $24.17B | +91% | $10.48B | $19.02B | -$8.54B | +56 | -$7.45B |
| 2028 Q1E | $10.37B | $56M | $1.60B | $9.70B | $241M | $4.50B | $1.95B | $28.43B | +90% | $12.60B | $21.31B | -$8.71B | +59 | -$7.43B |
| 2028 Q2E | $11.63B | $142M | $1.72B | $11.65B | $668M | $5.46B | $2.17B | $33.44B | +89% | $15.10B | $23.06B | -$7.95B | +66 | -$6.63B |
| 2028 Q3E | $13.01B | $263M | $1.85B | $13.90B | $1.33B | $6.56B | $2.41B | $39.34B | +90% | $18.04B | $25.15B | -$7.11B | +72 | -$5.80B |
| 2028 Q4E | $14.54B | $421M | $2.00B | $16.52B | $2.30B | $7.83B | $2.68B | $46.29B | +91% | $21.50B | $27.65B | -$6.15B | +78 | -$4.90B |
| 2029 Q1E | $16.23B | $620M | $2.15B | $19.54B | $3.64B | $9.28B | $2.99B | $54.44B | +92% | $25.58B | $30.63B | -$5.05B | +82 | -$3.94B |
| 2029 Q2E | $18.09B | $863M | $2.31B | $23.03B | $5.44B | $10.93B | $3.33B | $64.00B | +91% | $30.38B | $34.16B | -$3.78B | +86 | -$2.88B |
| 2029 Q3E | $20.16B | $1.15B | $2.49B | $27.07B | $7.80B | $12.83B | $3.70B | $75.20B | +91% | $36.03B | $38.34B | -$2.31B | +88 | -$1.72B |
| 2029 Q4E | $22.45B | $1.50B | $2.67B | $31.73B | $10.82B | $15.00B | $4.12B | $88.30B | +91% | $42.67B | $43.28B | -$609M | +90 | -$443M |
| 2030 Q1E | $25.00B | $1.90B | $2.88B | $37.11B | $14.64B | $17.47B | $4.59B | $103.59B | +90% | $50.47B | $49.11B | $1.16B | +91 | $823M |
| 2030 Q2E | $28.09B | $2.36B | $3.10B | $43.31B | $19.43B | $20.29B | $5.11B | $121.68B | +90% | $59.77B | $56.03B | $3.18B | +93 | $2.21B |
| 2030 Q3E | $31.58B | $2.88B | $3.33B | $50.46B | $25.35B | $23.51B | $5.68B | $142.80B | +90% | $70.69B | $64.19B | $5.53B | +94 | $3.76B |
| 2030 Q4E | $35.50B | $3.48B | $3.58B | $58.70B | $32.61B | $27.18B | $6.31B | $167.38B | +90% | $83.46B | $73.78B | $8.23B | +94 | $5.47B |
| 2031 Q1E | $39.89B | $4.16B | $3.86B | $68.19B | $41.48B | $31.37B | $7.01B | $195.96B | +89% | $98.36B | $84.20B | $12.03B | +95 | $7.82B |
| 2031 Q2E | $44.82B | $4.93B | $4.15B | $79.12B | $52.23B | $36.13B | $7.79B | $229.16B | +88% | $115.73B | $97.41B | $15.57B | +95 | $9.89B |
Every row is projected. A year-over-year change is shown only where the quarter it compares against exists — an em dash means there is no comparable quarter, not a flat year.
Model revisions
Assumptions are marked to reality as each quarter prints. Every change is appended here, with the fair value the model produced at the time, so the model's own history stays visible.
| Date | Changed | Fair value then | Note |
|---|---|---|---|
| 2026-08-23 | verticals.developer.driver.rationale, verticals.developer.notes | — | Re-grounded the Developer Platform base on the disclosed run rate rather than on the backed-out seat count: $4B annualised is $333M a month and three months is $1.00B, which is where 4.17M seats at $80 already sat, so no level moves and no fair value changes. Recorded that Deutsche Bank's ~$12B Cursor line is that same $4B tripled to a December exit, against Cursor's own April guide of above $6B, and that the tripling belongs in Bull and Elon rather than in Base. |
| 2026-08-19 | verticals.modelLayer, verticals.terrestrialAi.driver, scenarios | $187.12 | Added the Model layer vertical - Grok sold as a model rather than as a megawatt, driven by the $18.85B quarterly pool x share. Fifteen points of Terrestrial AI utilisation (70% to 55%, target 82% to 67%) move here: 1,800 MW x 15% x $2.03M is the $548M a quarter Grok already runs on, so the June quarter is unchanged and only its character is. Base takes the share to 18% long run, about 12% of a $33.8B pool by 2028 Q4. Elon is pushed to a 60% terminal share of the pool, roughly $36B a quarter. |
| 2026-08-16 | all | $145.29 | Initial model, built off the Q2 2026 print: $7.81B consolidated revenue, 12.0M Starlink subscribers, $66 ARPU, AI segment annualising at ~$10.2B. Horizon set to 20 quarters because free cash flow does not turn positive inside a three-year window. |
| 2026-08-16 | verticals | — | Split Connectivity into Starlink and Starlink Mobile, and AI into Terrestrial AI and Space AI. Starlink keeps everything that earns money today, including the ~22M-active light direct-to-cell messaging service; Starlink Mobile carries only the enhanced tier the V2 satellites and EchoStar spectrum unlock and earns nothing before 2028 Q1. Space AI likewise starts at zero with first revenue in 2028 Q1. Neither new vertical re-cuts reported history: both are held at zero across the six reported quarters. |
| 2026-08-18 | verticals, scenarios, price | $172.13 | Added a sixth vertical, Developer Platform, for Cursor and Origin. The $60B all-stock acquisition closed August 14, 2026, so the line is zero across every reported quarter and starts in 2026 Q3; 4.17M paid seats are backed out of the disclosed $4B run rate at an assumed $80 blended monthly price. Seat adds start at Cursor's own organic pace and the ramp is set against that path so the seats Origin wins from GitHub are worth about $0.1B a quarter by 2027 Q4. All four cases carry the line and disagree sharply about it: Bear ends below today's run rate, Elon at roughly $31B a year. Price snapshot refreshed to the August 17 close. |