The evidence should remain inspectable after the conclusion is published.

Relevant finding

Describes the consolidated company, AI platform, application businesses, customer concentration, risk factors and pending transaction context used to define the SOTP segments.

Review notes

The website date reflects the model research cutoff where the exact filing date was not separately carried into the publication draft. Filing accession and SEC metadata control. Workbook ID S34 uses this source.

Relevant pages: Business; AI platform; Grok, X and applications; risk factors; capitalization; offering and acquisition disclosures.

Relevant finding

Reports approximate size, hardware, and construction time for the Colossus and Colossus II clusters used as historical build-speed benchmarks rather than proof of future repeatability.

Review notes

The website date reflects the model research cutoff; SEC filing metadata controls. Workbook IDs S16-S18 map to this filing.

Relevant pages: AI Infrastructure and Colossus development discussion.

Relevant finding

Provides the filed basis for the pending all-stock Cursor transaction and its approximately $60 billion implied value; cash flows and shares are included by scenario factor.

Review notes

Workbook ID S19. The website date reflects the model research cutoff; SEC filing metadata controls.

Relevant pages: Transaction description, expected consideration and closing conditions.

Relevant finding

Argues that rapid construction can support a 10 GW 2027 build and scarcity pricing, while also presenting unconfirmed Microsoft offtake and NVIDIA vendor-financing claims excluded from Base.

Review notes

Workbook IDs S23-S25. Retained as an aggressive technical and commercial thesis, not the primary source of record. The URL retains an older headline figure while the live headline refers to $300 billion of ARR.

Relevant pages: Public article and paywalled site/buildout discussion; accessed by the model research cutoff.

Relevant finding

Provides the $133.11 per-share market reference used for the valuation hurdle and reverse-DCF cross-check.

Review notes

Workbook ID S26. The $133.11 reference price is compared with a June 30 balance-sheet valuation date, creating a known date mismatch.

Relevant pages: Historical and current price display for August 7, 2026.

Relevant finding

States a current design target of 75 kW per ton, implying 75 GW for one million tons; thermal, radiation, bandwidth, repair and cost remain milestone-gated.

Review notes

Workbook ID S32. Technical target is not evidence of commercial cost competitiveness.

Relevant pages: Starmind mass, power-density and deployment description; accessed August 7, 2026.

Relevant finding

Presents a Bloomberg-consensus-style 2030 Connectivity cross-check of roughly $107.2 billion revenue, $56 billion operating income and $33.4 billion capex.

Review notes

Workbook ID S51. Research-cutoff access date is used; the figures are a consensus cross-check rather than issuer guidance.

Relevant pages: 2030 Connectivity revenue, operating income and capex estimates.

Relevant finding

Secondary reporting attributes a $300 target, a greater-than-$3.3 trillion 2040 revenue path and free cash flow only in 2035 to Morgan Stanley, reinforcing funding and terminal-state risk.

Review notes

Workbook ID S52. Research-cutoff access date is used; underlying assumptions are paywalled and non-reproducible.

Relevant pages: Reported target, 2040 revenue case and free-cash-flow timing.

Relevant finding

Reports an initial $16.8 billion Tesla/SpaceX investment and a much larger aspirational output goal; the Elon Case uses possible capex-per-watt improvement but no chip-sales revenue.

Review notes

Workbook ID S33. The funded first phase and aspirational long-run scale are kept separate.

Relevant pages: Article discussion of initial investment, ownership and long-run production aspiration.

Relevant finding

Reports the xAI common-control acquisition presentation, Space, Connectivity and AI segment results, H1 operating cash flow and capex, cash, securities, debt, lease obligations, backlog, stock compensation, awards, and reported shares.

Review notes

Primary source of record for filed historical financials. Workbook IDs S01-S11, S21, S47-S48 map principally to this filing and its cited ownership disclosure.

Relevant pages: Cover; financial statements; cash-flow statement; segment note; debt and leases; commitments; backlog and concentration; EPS and awards; subsequent events.

Relevant finding

Contains management's year-end compute targets, power-and-cooling pipeline, current payback claim, additional cloud-contract disclosure, December ARR statement, 2030 revenue aspiration, and longer-shot SpaceX claims.

Review notes

Official management statements are classified as guidance, claims, or speculation rather than audited facts. Workbook IDs S12-S13, S20, S22, S27-S29, S31 and S35 use this source.

Relevant pages: Management discussion and Q&A on compute, AI pricing, revenue aspirations, Starlink V3, Starship, Starmind and financing.

Relevant finding

Secondary reports attribute an Overweight rating and approximately $225–$240 targets to JPMorgan, but full assumptions are unavailable and not used as model inputs.

Review notes

Workbook ID S43. The underlying sell-side report is paywalled; target, timing and rating are non-reproducible cross-checks only.

Relevant pages: Reported JPMorgan rating and target discussion.

Relevant finding

Shows approximately 1.302 billion restricted Class B shares already issued and therefore already included in the reported cover count rather than added again as incremental dilution.

Review notes

Workbook ID S47. Month-start date is a publication placeholder where the exact filing day was not carried into the draft; SEC metadata controls.

Relevant pages: Beneficial-ownership table and restricted Class B share footnotes.

Relevant finding

Secondary reporting attributes a Buy rating and $235 target to Bank of America and says its DCF extends to 2045, but the WACC, site timing and funding waterfall are not public.

Review notes

Workbook ID S44. The underlying report is paywalled; month-start date is a placeholder and the linked page metadata controls.

Relevant pages: Reported rating, target and long-horizon DCF description.

Relevant finding

Secondary reporting attributes a $200 base target and a $900 illustrative long-shot outcome to Citi; the long shot is not reproducible and is used only to frame the low-probability Elon Case.

Review notes

Workbook ID S45. The underlying report is paywalled; month-start date is a placeholder and the linked page metadata controls.

Relevant pages: Reported base and long-shot valuation discussion.

Relevant finding

Reports a Morningstar fair-value estimate near $780 billion and a more cautious view of xAI and orbital-compute economics, retained as an independent downside cross-check.

Review notes

Workbook ID S46. Detailed private valuation assumptions were not fully public.

Relevant pages: Reported valuation and skepticism toward orbital data centers and xAI.

Relevant finding

Discloses 325,000 NVIDIA GPUs plus supporting infrastructure and approximately $1.25 billion of steady-state monthly fees, with a 90-day termination mechanism after the initial period.

Review notes

Workbook ID S14. The exact filing metadata controls if the placeholder month-start date differs from the SEC filing day.

Relevant pages: Capacity, fees, service period and termination provisions.

Relevant finding

Discloses 110,000 NVIDIA GPUs plus related components and approximately $920 million of steady-state monthly fees after ramp, with termination rights after December 2026.

Review notes

Workbook ID S15. The exact filing metadata controls if the placeholder month-start date differs from the SEC filing day.

Relevant pages: Ramp, capacity, fees, delivery remedies and termination provisions.