Tesla
Completion
- TSLA
- 376.37
- Cap
- $1.41T
- Tape
- +0.0%
- Trenches
- 32
- Avg slip
- +25 mo
- $B to finish
- $248B
Peers, not press
Compare the two empires
Tesla vs SpaceX on the board, then each against the company that actually ships the peer product. Scores are forensic estimates from public artifacts — not KPIs, not a shrine.
SpaceX
Completion
- SPCX
- 185.00
- Cap
- $2.44T
- Tape
- +0.0%
- Trenches
- 39
- Avg slip
- +9 mo
- $B to finish
- $902B
Vehicles all-time
10.0M
Deliveries this Q
480k
FSD subs
1.48M
Energy this Q
13.5 GWh
Energy this Y
22.3 GWh
Starlink working
11k
Starlink launched
13k
Starlink users
10M
Falcon flights
680
Starship flights
13
Shared
21
Trenches
92
Head-to-head
Average completion and merge relevance across scored trenches. To-finish is $B remaining, midpoint.
Trench status
Count of cards, not dollars. Shared sits in gold.
Product lanes
Who actually has an industrial product. Score is forensic presence, not a TAM slide. Peer fights below are vs the company that ships.
| Product lanes | Tesla | SpaceX |
|---|---|---|
| Volume vehicles | 92 | 4 |
| Grid electrons | 68 | 18 |
| Constellation | 12 | 84 |
| Launch cadence | 0 | 86 |
| Unsupervised miles | 22 | 2 |
| Humanoid labor | 14 | 6 |
| AI compute | 44 | 58 |
| Silicon independence | 16 | 14 |
| National security | 8 | 88 |
| Cash generation | 82 | 36 |
Giga fleet 
No auto product Volume vehicles
Giga fleet
92
10.0M cumulative. The P&L that still funds every other slide.
No auto product
4
Adjacent only via in-cab Starlink — a related-party SKU.
Tesla already shipped. SpaceX is not a car company.
Megapack + Powerwall 
Pad load Grid electrons
Megapack + Powerwall
68
Multi-GWh quarters. Megapack 3 in production Aug 2026. Quiet compounder.
Pad load
18
Starbase and Colossus offtake. Buyer, not a grid OEM.
Storage is Tesla’s second act. SpaceX is the customer on the invoice.
In-cab dish 
Starlink Constellation
In-cab dish
12
A SKU on a roof rack. Not a network operator.
Starlink
84
~11.1k working birds. ~9–10M users. The cash engine.
Starlink is the network. Tesla is a reseller of a related-party dish.
None 
Falcon + Starship Launch cadence
None
0
No launch product. Semi and Cybercab do not go to orbit.
Falcon + Starship
86
Falcon is industrial. Starship is not yet operational orbit + payload.
Medium-lift is won. Cheap heavy lift is still vs physics.
Robotaxi pilot 
None Unsupervised miles
Robotaxi pilot
22
Geofenced paid rides. Low-million cumulative. Dozens of Cybercabs.
None
2
Starlink in the cab is connectivity, not autonomy.
The honest peer is Waymo, not SpaceX. See Peer fights.
Optimus Academy 
Mars slide Humanoid labor
Optimus Academy
14
Gen 2 public. Fremont line. Useful factory labor still ~0.
Mars slide
6
Optimus-on-Mars is a deck. High merge relevance, near-zero execution.
A two-demo race against Figure, not a shipment against each other.
Cortex (captive) Colossus hosting AI compute
Cortex (captive)
44
Training for FSD/Optimus. Not rented. Still in the Nvidia queue.
Colossus hosting
58
Named CSAs. ~$1.11B/mo cited from Dec. 90-day outs. Related-party.
SpaceX is already renting the cluster. Tesla is still building picks-and-shovels.
Dojo / AI5 tenant 
Starlink silicon Silicon independence
Dojo / AI5 tenant
16
Designs exist. Wafers still come from TSMC-class foundries.
Starlink silicon
14
Custom comms silicon, still a foundry customer. Terafab is the shared LOI.
Terafab is the merge argument. TSMC is the factory. 8% is not a process node.
Shanghai volume 
Starshield + NSSL National security
Shanghai volume
8
Largest factory in China. The FOCI problem the merge slide underwrites.
Starshield + NSSL
88
Classified bus + monopoly cadence this month. A FOCI magnet if Tesla is the parent.
They do not share a regulator. See China vs Starshield below.
Public auto + energy 
Starlink + NSSL + CSA Cash generation
Public auto + energy
82
Cash pile ~$43–45B class. Auto still pays for the robot and the fab slide.
Starlink + NSSL + CSA
36
Capex furnace. Starlink and national-security cash keep the pad lit. No Tesla-class pile.
Tesla is the balance sheet. SpaceX is the growth story funded by birds and pads.
Shape of the two companies
Labeled opinion on six axes. China-safe is inverted exposure: Shanghai volume vs a US national-security parent. Not a score you should average.
Scale
70 · 78 /100
Cars+energy vs Starlink+launch. Both are real products.
Cash
82 · 28 /100
Tesla pile ~$43–45B. SpaceX is a capex furnace funded by Starlink and NSSL.
Cadence
64 · 86 /100
Giga output vs Falcon cadence. Starship cadence is not Falcon cadence.
AI
55 · 62 /100
Cortex captive vs Colossus for rent. Related-party either way.
China-safe
18 · 86 /100
Shanghai is Tesla’s volume factory. SpaceX is a US national-security co.
Robot
14 · 6 /100
Optimus Academy vs Optimus-on-Mars slides. Neither is a labor product.
Completion vs merge relevance
Each dot is a scored trench. X is % done. Y is merge relevance. Size is $ remaining to finish. Click a dot to open the card.
Attention share
Forensic proxy of Elon-hours and capital follow — not a calendar feed.
2023
FSD v12 + Cybertruck + Starship IFT-1.
70 · 55
2024
Robotaxi unveil; Starship catch; political year.
62 · 72
2025
xAI combination talk; Starship cadence; pay package.
58 · 80
2026 H1
SpaceX IPO dominates. Tesla is the other ticker.
64 · 90
2026 Q3
Cybercab paid rides + Flight 14 orbital filing. Capital is following both.
78 · 88
Valuation vs execution vs remaining bill
Market-cap share vs weighted completion vs trench count vs remaining dollars. A private mark is not a finished Starship.
Cash already in vs still to spend
Midpoints of cost-to-date and cost-to-finish. Shared is the gold-lane bill.
China vs Starshield
Tesla’s volume factory and a US national-security launch company do not share a regulator. This is the silent killer of the merge slide.
Shanghai volume
- Shanghai is the volume plant — not a rounding error.
- A US defense parent is a Party-class problem in Beijing.
- Export-control and data-localization land on FSD weights and Optimus.
US national security
- Starshield and NSSL are the identity, not a side hustle.
- FOCI mitigation can force a carve-out that kills the merge math.
- Winning the pad is not the same as surviving CFIUS.
Tesla’s volume factory sits in Shanghai. SpaceX’s identity is US national security — Starshield, NSSL, classified customers. They do not share a regulator. CFIUS-class review in Washington and Party-class review in Beijing can each say no, or yes-with-conditions that wreck the synergy slide.
Related-party invoices
The gold lane is not synergy. It is Elon sitting on both sides of the invoice.
Gold-lane overlap
The merge case lives here. Shared nodes, related-party invoices, slides that need both P&Ls.
Slip board
Average months late on dated trenches. Context-only cards are excluded.
Tesla
+25 mo
SpaceX
+9 mo
Peer fights
The honest comparison is not Tesla vs SpaceX on every axis. It is Tesla vs Waymo on miles, SpaceX vs Kuiper on birds, Terafab vs TSMC on wafers.
Unsupervised paid miles
Tesla robotaxi
22
Low millions cumulative. Q2 ~2.5M paid robotaxi miles cited. Dozens of Cybercabs.
Waymo
78
Multi-million per week class. Years of geofenced scale in several cities.
Waymo is the scaled unsupervised service. Tesla is a paid pilot with a bigger narrative.
Mass to orbit
SpaceX
88
Dominant. Falcon 9 cadence is industrial. Starship is not yet operational orbit + payload.
ULA / Blue Origin
28
ULA: national-security cadence, no reuse factory. Blue: New Glenn early, not Falcon 9.
SpaceX already won medium-lift. Heavy cheap lift is the open fight, and it is vs physics more than vs ULA.
Constellation size
Starlink
82
~12.9k launched / ~11.1k working (Sep 2026). ~9–10M users.
Kuiper
18
Early hundreds to low thousands depending on the week — not a peer network yet.
Starlink is the network. Kuiper is a funding story with rockets on order.
Inference silicon independence
Terafab (Tesla+SpaceX)
8
Cited ~$55B-class. Framework agreement. 8% complete. Not a fab.
TSMC / Nvidia tenant
92
The actual queue both companies sit in today. Capacity, export controls, packaging.
Terafab is the merge argument. TSMC is the factory. Do not confuse a LOI with a wafer start.
AI compute as a business
SpaceX Colossus hosting
58
Named CSAs (Anthropic, Google, plus ~$1.11B/mo from Dec). 90-day outs. $100B ARR is a run-rate claim.
Tesla Cortex
42
Captive training for FSD/Optimus. Not rented. Still buying Nvidia-class supply.
SpaceX is already renting the cluster. Tesla is still building picks-and-shovels. Related-party either way.
Humanoid labor
Tesla Optimus
14
Gen 2 public. Gen 3 unshown. Fremont Academy. Useful factory labor still ~0.
Figure / Apptronik class
16
Paid pilots and demos at similar maturity. Nobody has a labor product at auto volume.
A two-demo race. Do not score a pay-package number as a shipment.
Grid storage
Tesla Megapack
62
Megapack 3 in production Aug 2026. Multi-GWh quarters. Quiet compounder.
Fluence / CATL packs
48
Real competitors on interconnect-limited sites. Tesla leads on factory cadence, not on market design.
Storage is Tesla’s second act. Interconnect queues still sit outside every OEM.
Rooftop solar
Tesla Solar Roof / panels
22
Panels exist. Solar Roof never industrialised. Energy GWh is storage-led.
Sunrun / residential installers
58
The actual US rooftop channel. Ugly, permitted, attached.
Do not let Megapack launder Master Plan ‘every roof.’ Distinct trenches.
Cell manufacturing
Tesla 4680
48
Giga Texas run-rate >40 GWh claimed. Still not the whole auto pack mix.
CATL / Panasonic
80
The pack Tesla still buys. Scale, chemistry, and price Tesla has not displaced.
Vertical materials is 30–50% of a factory stack, not of energy independence.
National-security launch
SpaceX NSSL
90
Only active flyer in public Phase 3 trackers as of Sep 2026.
ULA / others
22
Paper providers until they fly. A monopoly cadence is cash and a political target.
Winning NSSL is not the same as surviving a Tesla-China parent. Starshield is the sister file.
Next gates
Cited dates, not a schedule you can take to a bank. Ceremonial opens and NET slips live here.
Live tape, split
Headlines tagged Tesla vs SpaceX. Shared sits in gold. Stale when RSS is down.
Tesla
- Tesla Semi Sparks plant inauguration set for 24 September — ceremonial, not nameplate
Tesla / Nevada · Sep 11 2026
- Tesla tape: live price is not robotaxi unit economics and not a China FSD approval
Yahoo Finance · Sep 11 2026
- Optimus Academy units at Fremont — useful factory labor still scored ~0
TRENCHES desk · Sep 11 2026
- Megapack 3 in production — energy GWh is storage-led, not Solar Roof
Energy notes · Aug 1 2026
SpaceX
- Starship Flight 14 remains NET 15 September — orbital attempt, not a ship catch
FAA / pad trackers · Sep 11 2026
- SpaceX AI hosting: third cloud deal cited ~$1.11B/month from 1 December, 90-day outs
CFO commentary 10 Sep · Sep 10 2026
- T-Mobile calls satellite-from-space 5G threat 'laughable' as EchoStar spectrum file sits at FCC
T-Mobile / FCC · Sep 10 2026
- Colossus 2 described as gigawatt-class; Memphis power and emissions remain the constraint
S-1 / local load · Sep 3 2026
Shared
- Terafab remains a cited ~$55B framework — no site, no node, no tool order
TRENCHES desk · Sep 11 2026