CAPITAL THESIS

Independent equity research · CBRS · full edition

Cerebras SystemsOracle expands the opportunity. Equipment economics still decide the return.

The Oracle partnership adds a potential distribution route, while a short-seller report raises unverified durability allegations. The lower share price improves conditional returns, but June remains the latest reported quarter and funding and replacement economics remain unresolved.

Research date: 3 November 2026 · USD common equity · Reference price: USD 96.36 · Observation: 2026-11-02

Speculative watchlist - wait for evidence on cash economics

The investment decision

At the $96.36 regular close, the unchanged October operating model illustrates 18.3% annualised in base over five model years. That improvement comes from a lower entry price, not an earnings upgrade. The conditional 20% ceiling remains $89.69; a 13%/3.5% perpetual-growth DCF gives only $59.31. Higher replacement spending can erode the apparent bargain.

Reference regular close$96.362 November 2026; Nasdaq, USD
Latest reported cash after capex-$596mH1 2026 operating cash less cash PPE
Base five-year annual return18.3%Retained conditional operating model
Conditional 20% entry ceiling$89.6925x terminal cash; central DCF $59.31

Edition update. New edition as of 3 November 2026. Adds the 27 October Oracle partnership, 28 October short-report allegations, subsequent public news and a dated 2 November closing price. June actuals and the 18 October operating model are retained; revised returns reflect price only. Prior edition remains archived.

01. What changed since 18 October

Reference price USD 96.36 on 2026-11-02. This is the regular-session close, not after-hours trading. It is 24.3% below the earlier edition's $127.21 reference and 11.7% below the prior session's $109.18. The observation is Nasdaq-listed Cerebras Systems Inc., ticker CBRS, not another similarly named issuer. [S3]

DateObserved developmentInvestment interpretation
20 OctoberQ3 release scheduled 12 November after the US closeHard date for the next actual financial test
27 OctoberOracle/Cerebras planned OCI availability in 2027Potential distribution; no quantified commitment added to the model
28 OctoberCulper short report; shares closed $112.56Interested-party durability allegations; investigate rather than assume
2 NovemberClose $96.36; public UBS target revision and law-firm noticesLower price, not a proven reduction in business risk

Oracle is a credible route to customers who already purchase cloud infrastructure. It can reduce distribution friction and widen the range of workloads exposed to Cerebras. However, the announcement sets no minimum revenue, deployment capacity, required capital, exclusivity or profit guarantee. Availability is planned for calendar 2027. We leave the earnings path unchanged until commercial detail supports a numerical revision. [S11]

Public reporting on 2 November says UBS reduced its 12-month target from $330 to $250 while retaining Buy. We did not obtain the original broker note and do not import its estimates. The same day's law-firm notices repeat the prior short-report claims; they are not a new operating disclosure or established misconduct. The observed share decline cannot be cleanly assigned to one headline. [S14] [S15]

The conclusion remains watchlist; wait for proof of lifetime cash economics. Lower purchase cost improves the conditional mathematical return. Neither the announcement nor an allegation establishes the realised replacement cycle, utilisation, pricing or funding terms. Our principal scenarios are the 18 October operating cases repriced, not a forecast upgrade.

02. Faster answers are useful; profitable capacity is the investment

Cerebras sells AI computing systems and operates inference capacity that customers use as a service. Its wafer-scale design keeps a very large amount of computation and high-speed memory on one silicon wafer. Shorter data journeys can reduce the delay between tokens when an AI model writes an answer. That is commercially valuable for interactive products and agents waiting for their next step. It is not proof that this architecture wins every workload, every model or every cost comparison. [S1] [S4] [S5]

The system can complement conventional accelerators: other chips can process the initial prompt while Cerebras handles rapid token generation. CS-4 is a system-level development involving multiple wafers, networking, power and cooling. Company speed comparisons remain configuration-specific claims; this report does not turn “up to” performance or extrapolated model sizes into proven customer economics. The competitive test is delivered cost per useful response at the required latency, reliability and security, not a single laboratory speed figure.

There are two different ways to monetise the technology. A hardware sale transfers a system to a customer and can produce earlier cash without Cerebras operating that capacity for years. Cloud service can produce recurring revenue and a larger lifetime opportunity, but Cerebras must procure equipment, secure power, construct capacity and pay operating costs before collecting all the revenue. Selling more cloud capacity can therefore increase both growth and near-term cash needs. Cloud share of Q2 GAAP revenue was 70%, versus 32% a year earlier. [S1]

Our position is a speculative watchlist, with medium evidence confidence and an incomplete financing underwriting case. The product and contracts support a real opportunity. The current price still requires a large, profitable build-out, while full future funding terms, mature utilisation and lifecycle replacement costs remain unproven. A share-price decline or a large contract number alone does not establish a bargain.

03. A $25.4bn contract book is a delivery obligation

The June filing reports $25.4bn of remaining performance obligations, or RPO: contracted revenue not yet recognised. About 22% is expected within the first 24 months, 43% in the following 24 months and 35% thereafter. Those proportions imply approximately $5.59bn, $10.92bn and $8.89bn respectively. RPO includes some pass-through amounts and a significant OpenAI commitment. It is neither present cash nor a gross-profit measure. [S1]

OpenAI has committed to 750MW of capacity delivered in tranches during 2026–28, with an option for a further 1.25GW through 2030. The optional amount must not be described as secured demand. Separately, management described more than 600MW of live and contracted capacity for delivery by the end of 2027. These numbers have different scopes and dates: “live and contracted” does not mean already energised. Failure to meet delivery or service requirements can allow contract termination. [S1] [S2]

The cash loop matters. Infrastructure must be available, customers must accept and use the contracted service, and billings must become cash after service credits. The OpenAI working-capital loan originally provided about $1.00bn; $918.2m remained at June. It can be repaid through service fees, compute, hardware and other permitted credits as well as cash. June loan offsets already included $86.3m of non-cash billings. Such financing helps build the business today while reducing cash available from some future service delivery. [S1]

Demand is concentrated. The three largest customers represented 34%, 32% and 10% of Q2 revenue, together 76%. Two customers represented 76% of receivables. Consequently, a small number of counterparties have substantial influence over delivery schedules, pricing, payment and technology deployment. The report treats customer security restrictions and the possibility of workloads remaining on a customer’s own infrastructure as risks; analyst channel checks about specific frontier-weight decisions are not confirmed cancellations. [S1] [S7]

04. The new durability dispute: a cash question, not a verdict

Culper disclosed a short position and questioned the working life of Cerebras hardware. Its position benefits if the shares fall; some reported support relies on unnamed former staff or customers. We could not retrieve the primary report because its page and PDF returned 403. The narrow claim used here is that such allegations were published, as documented by contemporaneous coverage. We do not independently establish a specific failure rate, useful life, engineering defect or accounting misstatement. [S13]

The economic mechanism is nevertheless straightforward. If a cloud operator must replace equipment sooner, it spends more before a long customer contract has finished paying out. More replacements can reduce usable capacity, raise maintenance spending, interrupt service and increase financing needs. Even accurate depreciation estimates are accounting allocations, not a guarantee of physical lifespan or replacement cost. The filing itself lists useful-life and warranty estimates as judgments subject to change. [S1]

The appropriate diligence is cohort-level installed-system survival, failure and warranty data; service-level credits and downtime; replacement spending separated from growth spending; and customer economics at comparable workload and utilisation. A company response supported by customer or operational evidence could weaken the allegation. We identified no such public rebuttal in the IR sources reviewed as of this edition; that absence is not an admission. [S6] [S13] [S16]

Our capex stress adds five percentage points of annual sales to equipment spending, then recomputes borrowing and interest. It is an independent sensitivity, not a translation of the short seller's asserted lifespan. Base annual return falls from 18.3% to 8.1%, while peak debt rises from $1.11bn to $2.26bn. Greater cash intensity can therefore matter even if demand remains intact.

05. Growth, adjustments and cash tell different stories

USD million unless statedQ2 2025 actualQ2 2026 actual
Hardware GAAP revenue70.29554.119
Cloud and other GAAP revenue33.027125.991
Total GAAP revenue103.322180.110
GAAP gross profit32.09925.559
GAAP gross margin31.1%14.2%
GAAP operating profit / (loss)(57.182)(477.233)

Q2 GAAP revenue rose 74%, driven by cloud growth of 281%; hardware revenue declined 23%. The gross-profit outcome did not keep pace. Customer warrants, pass-through costs and the IPO compensation event make a simple year-on-year margin or earnings comparison misleading. [S1] [S2]

Q2 revenue reconciliationUSD million
GAAP revenue180.110
Add customer-warrant amortisation44.262
Remove pass-through revenue(14.503)
Management core revenue209.869

Management’s core revenue removes specific accounting items; it is not additional cash received. Core gross margin was about 40.6%, versus 14.2% under GAAP, while core operating loss was $33.6m. Customer warrants are economically meaningful because customers can gain equity and existing holders can be diluted. A stronger adjusted margin is useful for understanding operations, but ignoring the equity claim would overstate the shareholder benefit. [S2]

Q2 stock compensation was $377.0m, including $273.6m associated with the IPO liquidity condition. This one-time vesting event should not simply be annualised. Future compensation still has a cost. Our cash model expresses that cost through an expanded share denominator and further annual dilution; it does not also subtract the same non-cash expense from cash. The prior-year net-profit comparison is also distorted by a $363.3m gain from extinguishing a forward-contract liability, so a historical P/E is not informative. [S1]

H1 2026 cash bridge — actualUSD million
GAAP revenue373.516
Cash used in operations(47.488)
Cash purchases of property and equipment(548.873)
Operating cash less cash capex(596.361)

The $596.4m cash gap is calculated from the same six-month period. It excludes purchases of investments and is not an adjusted-earnings figure. Customer deposits, deferred revenue and working-capital movements influence operating cash, while capex payables are not yet cash spending. The correct question is whether subsequent capacity produces enough cash to cover equipment replacement, rent, power and financing—not merely whether revenue keeps rising. [S1]

06. A substantial cash balance, with substantial claims on it

30 June 2026 balance / commitmentUSD bnTreatment
Cash plus investments7.922Unrestricted liquidity starting point
Restricted cash0.685Excluded from model available cash
Customer working-capital loan0.918Fully reserved before forecasting
Operating lease liability, present value0.569Rent remains in operating costs
Commenced future lease payments, undiscounted0.690Not added again to debt after expensing rent
Additional uncommenced leases, including subsequent agreements2.300Future rent must be absorbed by margins
Undrawn revolving facility0.850Availability is not cash

The May IPO raised approximately $6.2bn net, making external capital the main source of today’s liquidity. The $850m revolving facility had no June borrowings and matures on 14 April 2031. The filed post-IPO pricing is SOFR plus 2.25%, potentially falling to 2.00% on the specified leverage test; our future incremental funding rates of 7.5–12% are house assumptions, not the facility’s current contractual rate. [S1]

Our opening model reserve is deliberately conservative: $7.922bn unrestricted liquidity less the whole $0.918bn customer loan and an assumed $2.0bn second-half cash use leaves $5.003bn for the five-year path. The $2.0bn is a Capital Thesis planning assumption, not a reported balance or company guidance. Restricted cash remains excluded even though some restricted resources can support loan-related obligations; this conservatism can understate available economic liquidity. We reserve the customer loan once and do not deduct its future service-credit repayment a second time.

Rent and power are included in forecast cash operating costs. Therefore the model does not also deduct operating-lease liabilities as financial debt. The $2.3bn uncommenced lease total is an important execution test for that margin assumption, not a second cash subtraction. The model has no annual lease-cohort or equipment-retirement schedule; a detailed project-level build could reveal greater cash needs.

Even this opening liquidity does not make the plan self-financing. The base path requires about $1.11bn of peak new debt, above the currently disclosed $850m facility. The bull reaches $1.76bn and the bear $3.02bn. This is a model funding plug, not verified lending capacity. If additional financing is unavailable, or lenders demand more equity, the forecast must be rebuilt with slower capacity growth, higher dilution or distress. That unresolved access to capital is a reason for the watchlist stance.

The retained base and bull repay all modeled incremental debt in FY2030, before the disclosed revolver matures in April 2031. The bear does not repay its debt and already produces zero common-equity value. The model does not demonstrate that lenders will provide its peak funding or waive covenants. Loan fees, project-level borrowing restrictions and refinancing under a delayed build can make outcomes worse. An additional equity-funding stress below makes one alternative explicit. [S1]

07. The share count cannot stop at the IPO

The latest filed common-share count on the cover is 237.564m as of 5 August. At the $96.36 reference price, that gives a dated basic market-capitalisation proxy of $22.89bn. It is not a fully diluted valuation or a live November share count. The quarterly weighted-average EPS denominator is unsuitable for this purpose. [S1] [S3]

Consistent June economic share bridgeMillion shares
June common classes A, B and N228.204
All OpenAI warrant shares, including conditional tranches33.445
All AWS warrant shares2.697
Outstanding options23.797
Unvested RSUs16.532
Performance RSUs9.000
Gross model opening economic claims313.675

The model assumes all these contingent warrants and awards ultimately become shares, ignores exercise proceeds and adds future new-award dilution of 3.5%, 2.0% or 1.5% annually in bear, base and bull. This is a conservative gross-claim convention, not the GAAP treasury-stock method or a forecast that every performance hurdle will be earned. It can overstate dilution in a weak outcome. The annual allowance covers awards beyond the pool already counted; it is not a second charge for the same grants.

The 10.034m OpenAI warrants exercised in July are already inside the full 33.445m warrant pool; they are not added again through the August common count. Earlier G42 warrant exercises are already in June common shares. Early-exercised shares subject to repurchase are also not added again. This consistent June bridge avoids mixing current common shares with warrants that have since changed form. There is no complete November award roll-forward in the reviewed filings. [S1]

Multiple share classes, customer-linked equity and large performance awards make per-share results more uncertain than revenue results. Governance and incentive alignment therefore belong in the underwriting. The company also reported material weaknesses in financial-reporting controls at June. Remediation progress is a monitoring item, not evidence of misconduct. [S1]

08. The five-year model pays for the equipment

Capital Thesis uses a cash-based enterprise valuation because an earnings multiple that ignores the capacity bill is poorly suited to this build-out. FY2027–31 represent five stylised annual periods from the dated purchase price; they are not an exact calendar-date IRR. The October operating assumptions are retained without revision. All following forecast figures are our independent assumptions, not management guidance or broker targets. Forecast sales use an economic, core-like basis: pass-through revenue is excluded and non-cash customer-warrant contra-revenue is reversed. The customer equity cost is captured through dilution; this is not a formal GAAP revenue forecast.

Base house pathFY27FY28FY29FY30FY31
Revenue, $bn3.007.0011.0015.0020.00
Cash operating margin15%32%40%42%44%
Cash capex, $bn3.004.203.854.054.40
After-capex unlevered cash, $bn-2.613-2.265-0.1021.2722.930
Closing new debt, $bn0.0000.9161.1100.0000.000
Economic shares, m319.95326.35332.87339.53346.32

The mechanism is explicit: revenue times cash operating margin, less cash taxes, capex and incremental working capital, equals after-capex unlevered cash. Cash operating margin is before depreciation, non-cash stock compensation, taxes and financing interest, but after cash operating expenses including rent and power. Taxes are 22% of positive cash operating profit less assumed depreciation. No NOL benefit, interest tax shield or interest income is credited. Working capital absorbs 3% of incremental revenue.

The base requires revenue to rise from a FY26 core-guidance midpoint near $0.885bn to $20bn in FY31, with cash operating margin reaching 44%. Annual capex falls from 100% of revenue in FY27 to 22% in FY31; cumulative five-year capex is $19.50bn. Depreciation is assumed to fall from 20% to 14% of revenue. This is a demanding scale-and-efficiency story: contractual demand alone cannot deliver the assumed utilisation, pricing and replacement economics.

A $1bn minimum cash balance is maintained. Deficits are funded with new debt, interest is charged on average annual debt, and excess cash first repays that debt. Shares compound using the case-specific new-dilution allowance. No dividends, buybacks, acquisitions or new external equity issuance are assumed. A project-specific financing or equity raise would change these results; the current model is not a complete three-statement forecast.

For comparison, management’s August FY26 core revenue guidance was $880–890m, core gross margin 41–43% and core operating margin negative 19–17%; it expected revenue to more than triple in 2027. A supplied August analyst model projected $3.172bn for 2027 and $9.271bn for 2028, while another supplied framework used about $6.3bn for 2028. Our $3bn/$7bn base is independent and does not inherit either target. These sources illustrate forecast disagreement rather than an established consensus. [S2] [S7] [S8] [S10]

No incremental Oracle sales are added to this path: the partnership announcement is not a quantified order. No refreshed analyst consensus was obtained. Historical supplied broker forecasts are retained only to show the earlier disagreement, not presented as current consensus. The operating path must be re-underwritten after Q3 and when contract economics or lifecycle costs become observable. [S11] [S12]

09. Conditional outcomes: the business path is unchanged

House caseFY31 economic salesCash op marginCapex / salesExit cash multipleTerminal / share5Y annual return
Bear$8bn30%25%18x$0.00Total loss
Base$20bn44%22%25x$223.1818.3%
Bull$35bn50%20%30x$692.1548.3%

The bear still grows revenue to $8bn but converts too little into cash: a 30% cash operating margin and 25% capex ratio leave only $93.6m terminal unlevered cash after taxes and working capital. At 18 times that cash, enterprise value is below the debt claim after available cash. Common equity is floored at zero. This is a conditional total-loss illustration, not a default probability or the only possible adverse path.

The base FY31 bridge is $20bn revenue × 44% = $8.80bn cash operating profit, less $4.40bn capex, $1.32bn tax and $0.15bn working capital = $2.93bn after-capex unlevered cash. At 25 times, enterprise value is $73.25bn. Add $4.042bn cash and subtract zero remaining new debt, then divide by 346.322m economic shares: $223.18 per share. From $96.36, the five-year annualised return is 18.3%.

The bull reaches $35bn revenue, a 50% cash operating margin and a 20% capex ratio. At 30 times $7.428bn terminal cash, plus accumulated balance-sheet cash and with 337.917m shares, terminal value is $692.15 and the annualised return 48.3%. It requires faster deployment and higher utilisation, strong pricing and operating leverage, lower future dilution and a premium exit valuation simultaneously. It is an upside boundary, not a central expectation.

Five-year annual return = (terminal share value / $96.36)^(1/5) − 1; no distributions are included. Returns are nominal USD before investor tax, fees and currency effects. No scenario probabilities are assigned. A positive current free-cash-flow multiple is unavailable, so a meaningful like-for-like “no rerating” return cannot be calculated. This report intentionally remains outside the website’s comparable-return ranking rather than inventing that input.

The change from the prior 18 October base return of 11.9% to 18.3% is entirely the lower purchase price. Terminal $223.18, entry $89.69, operating assumptions and share paths remain unchanged. The same applies to the higher bull return. These five stylised FY2027-31 periods approximate a five-year purchase-to-exit illustration; they are not an exact calendar-date IRR.

10. The entry ceiling and discounted cash flow answer different questions

At 25x, the terminal cash yield is 4%. That premium valuation still needs growth durability after 2031. The retained base annual return is 18.3%, short of a 20% hurdle at $96.36. Holding the operating path, terminal cash and share count fixed, the current price requires approximately 27.0x terminal cash to achieve 20% annually. This is a valuation backsolve, not an observed market forecast.

The fixed-model price that would yield 20% annually is $223.18 / 1.20^5 = $89.69. It is a conditional entry ceiling, not fair value or an automatic buy price. It assumes the full operating path, 25x cash multiple, funding access and dilution. Lower quotes can worsen capital access or change award economics; this backsolve does not rerun those terms.

A separate perpetual-growth DCF is only $59.31 per share at 13% discount rate and 3.5% perpetual growth. Across 11-15% discount rates and 2.5-4.5% growth, its range is $43.54-90.28. It discounts after-capex unlevered cash and adds the opening reserved $5.003bn once, dividing by terminal diluted shares. It does not add future balance-sheet cash twice. This is a financing-neutral cross-check with future dilution charged upfront, not the same leveraged equity return calculation.

The DCF assumes growth slows sharply after FY31; the 25x exit case rewards a longer growth duration. Their gap is material and should not be averaged away. Neither method proves intrinsic value when lifecycle capex and project funding are unresolved.

One change to baseTerminal / shareAnnual return20% entry ceiling
Retained base$223.1818.3%$89.69
Cash capex +5pp of sales each year$141.968.1%$57.05
Cash operating margin -5pp each year$159.7210.6%$64.19
Annual new award dilution +1pp$212.5517.1%$85.42
New debt interest rate +3pp$222.9818.3%$89.61
20x terminal cash instead of 25x$180.8813.4%$72.69
30x terminal cash instead of 25x$265.4822.5%$106.69

Cash capex +5pp raises cumulative five-year equipment spending by $2.8bn. The resulting 8.1% base return is conditional on funding the larger need. The stress retains a small debt balance in FY2031 and would require appropriate maturity/refinancing arrangements; it does not demonstrate access or price financing fees. In this model, capital intensity is more influential than a modest rate change.

11. What if debt capacity stops at the disclosed ceiling?

The main cases assume new borrowing at house rates to maintain a $1bn liquidity floor. A supplementary base stress caps outstanding debt at $850m and issues common equity for the remaining shortfall, charging 4% issue costs. The entire revolver is assumed usable; actual letters of credit, covenants and lender conditions can reduce availability. The fixed issue prices below are assumptions, not forecasts or arranged offerings.

Assumed issue priceGross new equityFY31 economic sharesTerminal / share5Y return20% ceiling
$96.36$252.2m349.114m$222.1718.2%$89.29
$50.00$252.2m351.702m$220.5418.0%$88.63

The model raises about $252m gross in FY2028-29. Lower borrowing saves interest, while new shares dilute ownership. At a $50 issue price, more shares are required than at $96.36. Existing award dilution still compounds 2% annually after the raises; the same grants are not charged again as cash. This modest isolated stress does not capture a financing shutoff or a simultaneous operating miss. If neither debt nor equity is available, deployment and the revenue path must fall, and the retained base no longer applies.

12. The next test is measurable evidence, not the next headline

Evidence windowSupport for the thesisRe-underwrite or reject if
12 November Q3 releaseGAAP/core reconciliation, deployed capacity and collections support delivery scheduleDeferrals, service credits or acceptance issues consume the apparent backlog
Hardware durability and replacement dataCohort-level reliability and spending validate economic service lifeRepeated replacement capex or downtime exceeds modeled costs
2027 Oracle availabilityPaid usage, commercial terms and acceptable unit economicsLaunch slips or broad availability produces weak monetisation
Funding and dilutionCommitted capital covers construction with intelligible per-share costsBorrowing unavailable, restrictive covenants or larger issuance changes the path
Reporting qualityAward roll-forward, useful-life assumptions and control remediation clarify economicsAdjusted metrics obscure equity cost, spending or obligations

Cerebras has scheduled third-quarter results after the US market closes on 12 November 2026, with a 5pm EST call (9am AEDT on 13 November). The last company guidance reviewed was Q3 core revenue $214-216m, core gross margin 38-40% and core operating margin minus 25-23%; FY26 core revenue $880-890m. These are August company guidance, not current consensus or already achieved outcomes. [S2] [S12]

The current evidence is enough to maintain a speculative watchlist, not enough to establish that the lower price compensates for risk. A credible reliability record, better cash conversion, visible financing and quantified partner economics would improve the case. Persistent cash-capex overshoot, service problems, customer concentration or adverse award economics would weaken it even with revenue growth.

Financial actuals end 30 June; the research cutoff is 3 November 2026 and the price is the 2 November regular close. The four historical notes from the October edition remain dated background. No new consensus, project-level replacement study, complete November share roll-forward, verified unlock quantity or short-interest estimate is asserted. Insider sale forms and law-firm solicitation notices do not by themselves establish customer economics or misconduct. [S1] [S3] [S7] [S8] [S9] [S10] [S14] [S16]

Financial exhibits

Exact observations accompany every chart. Actuals, management guidance and Capital Thesis assumptions are labelled separately; the chart shapes do not imply probabilities.

Chart 01 · USD per share

The price changed faster than the filed evidence

Observed closes are not a projected recovery path. The repriced model retains its October operating assumptions.

Observed regular close
The price changed faster than the filed evidence (USD per share). Exact figures in the table below.025507510012515016 Oct prior referenceObserved regular close: 127.21 USD per share127.2127 Oct Oracle newsObserved regular close: 119.31 USD per share119.3128 Oct short reportObserved regular close: 112.56 USD per share112.5630 Oct closeObserved regular close: 109.18 USD per share109.182 Nov referenceObserved regular close: 96.36 USD per share96.36
Exact values · USD per share
ObservationObserved regular close
16 Oct prior reference127.21
27 Oct Oracle news119.31
28 Oct short report112.56
30 Oct close109.18
2 Nov reference96.36

Selected actual trading dates, not equally spaced time intervals; no single-cause attribution of price moves. Sources: [S3]

Chart 02 · USD million

Cloud growth changes the business and its capital needs

Q2 cloud and other revenue grew rapidly while hardware revenue declined; the two routes have different cash timing.

Q2 2025 actualQ2 2026 actual
Cloud growth changes the business and its capital needs (USD million). Exact figures in the table below.050100150200HardwareQ2 2025 actual: 70.295 USD million70.295Q2 2026 actual: 54.119 USD million54.119Cloud and otherQ2 2025 actual: 33.027 USD million33.027Q2 2026 actual: 125.991 USD million125.991Total GAAP revenueQ2 2025 actual: 103.322 USD million103.322Q2 2026 actual: 180.11 USD million180.11
Exact values · USD million
ObservationQ2 2025 actualQ2 2026 actual
Hardware70.29554.119
Cloud and other33.027125.991
Total GAAP revenue103.322180.11

All figures GAAP and for the same three-month periods. Core revenue is separately reconciled in the reported-financials section. Sources: [S1] [S2]

Chart 03 · USD million

The first-half capacity build consumed $596m after capex

Cash from operations was already negative before equipment purchases. Revenue growth is not current free cash flow.

H1 2026 actual / arithmetic
The first-half capacity build consumed $596m after capex (USD million). Exact figures in the table below.-600-500-400-300-200-1000Cash from operationsH1 2026 actual / arithmetic: -47.488 USD million-47.488Cash capex outflowH1 2026 actual / arithmetic: -548.873 USD million-548.873Cash after capexH1 2026 actual / arithmetic: -596.361 USD million-596.361
Exact values · USD million
ObservationH1 2026 actual / arithmetic
Cash from operations-47.488
Cash capex outflow-548.873
Cash after capex-596.361

Cash after capex is operating cash less cash purchases of property and equipment; not GAAP earnings or core profit. Sources: [S1]

Chart 04 · USD bn

The base only becomes cash-positive after the early build

Capex absorbs the operating gains until deployment scales; rent and power remain in cash operating costs.

Cash operating profitCash capexAfter-capex unlevered cash
The base only becomes cash-positive after the early build (USD bn). Exact figures in the table below.-5-2.502.557.5102027 houseCash operating profit: 0.45 USD bn0.45Cash capex: 3 USD bn3After-capex unlevered cash: -2.6134 USD bn-2.61342028 houseCash operating profit: 2.24 USD bn2.24Cash capex: 4.2 USD bn4.2After-capex unlevered cash: -2.2648 USD bn-2.26482029 houseCash operating profit: 4.4 USD bn4.4Cash capex: 3.85 USD bn3.85After-capex unlevered cash: -0.1024 USD bn-0.10242030 houseCash operating profit: 6.3 USD bn6.3Cash capex: 4.05 USD bn4.05After-capex unlevered cash: 1.272 USD bn1.2722031 houseCash operating profit: 8.8 USD bn8.8Cash capex: 4.4 USD bn4.4After-capex unlevered cash: 2.93 USD bn2.93
Exact values · USD bn
ObservationCash operating profitCash capexAfter-capex unlevered cash
2027 house0.453-2.6134
2028 house2.244.2-2.2648
2029 house4.43.85-0.1024
2030 house6.34.051.272
2031 house8.84.42.93

Capital Thesis assumptions, not filed forecasts. Unlevered cash also deducts model taxes and incremental working capital; excludes financing interest. Underlying starting evidence: S1/S2. Sources: [S1] [S2]

Chart 05 · USD bn

All three cases assume substantial growth

Even the bear reaches $8bn sales: weak cash conversion and financing can still leave no equity value.

Bear house revenueBase house revenueBull house revenue
All three cases assume substantial growth (USD bn). Exact figures in the table below.010203040FY2027Bear house revenue: 1.8 USD bn1.8Base house revenue: 3 USD bn3Bull house revenue: 4 USD bn4FY2028Bear house revenue: 3 USD bn3Base house revenue: 7 USD bn7Bull house revenue: 10 USD bn10FY2029Bear house revenue: 4.5 USD bn4.5Base house revenue: 11 USD bn11Bull house revenue: 17 USD bn17FY2030Bear house revenue: 6 USD bn6Base house revenue: 15 USD bn15Bull house revenue: 25 USD bn25FY2031Bear house revenue: 8 USD bn8Base house revenue: 20 USD bn20Bull house revenue: 35 USD bn35
Exact values · USD bn
ObservationBear house revenueBase house revenueBull house revenue
FY20271.834
FY20283710
FY20294.51117
FY203061525
FY203182035

Independent scenario assumptions; not probabilities or management guidance. Contracted revenue alone does not guarantee the model outcomes. Sources: [S1] [S2]

Chart 06 · Percent per year

Five more cents of equipment per sales dollar changes the return

Borrowing and interest are recomputed. This is a house spending stress, not a measured hardware lifespan.

Conditional five-year annual return
Five more cents of equipment per sales dollar changes the return (Percent per year). Exact figures in the table below.05101520Retained baseConditional five-year annual return: 18.291 Percent per year18.291Capex +5ppConditional five-year annual return: 8.0576 Percent per year8.0576Cash margin -5ppConditional five-year annual return: 10.6351 Percent per year10.635120x terminal multipleConditional five-year annual return: 13.4222 Percent per year13.4222
Exact values · Percent per year
ObservationConditional five-year annual return
Retained base18.291
Capex +5pp8.0576
Cash margin -5pp10.6351
20x terminal multiple13.4222

From $96.36 on 2 November. Same operating vintage as October; no dividends. Additional funding is assumed available. Sources: [S1] [S2]

Chart 07 · USD bn

The base needs more debt than the disclosed facility

Cash generation repays modeled debt in FY2030, but the peak funding has not been committed.

Base modeled debtDisclosed facility ceiling reference
The base needs more debt than the disclosed facility (USD bn). Exact figures in the table below.00.250.50.7511.252027 houseBase modeled debt: 0 USD bn0Disclosed facility ceiling reference: 0.85 USD bn0.852028 houseBase modeled debt: 0.9162 USD bn0.9162Disclosed facility ceiling reference: 0.85 USD bn0.852029 houseBase modeled debt: 1.1097 USD bn1.1097Disclosed facility ceiling reference: 0.85 USD bn0.852030 houseBase modeled debt: 0 USD bn0Disclosed facility ceiling reference: 0.85 USD bn0.852031 houseBase modeled debt: 0 USD bn0Disclosed facility ceiling reference: 0.85 USD bn0.85
Exact values · USD bn
ObservationBase modeled debtDisclosed facility ceiling reference
2027 house00.85
2028 house0.91620.85
2029 house1.10970.85
2030 house00.85
2031 house00.85

Facility ceiling is a comparison, not a forecast of availability. The actual agreement matures 14 April 2031; base debt is repaid before that. Letters of credit/covenants may reduce usability. Sources: [S1]

Source register and research basis

Evidence checked for this edition on 3 November 2026. Dates below identify source vintages. Company guidance and house scenarios are conditional; source links and the downloadable model allow the assumptions to be inspected.

S3 · CBRS historical share prices

2 November 2026, 4pm EST regular close; checked 3 November · USD 96.36, independently corroborated by Alpaca SIP daily bars. The $96.49 after-hours quote is excluded. Provider volume differs slightly; no volume thesis is made.

https://stockanalysis.com/stocks/cbrs/history/
S4 · Introducing Cerebras CS-4

18 August 2026 · Primary company product claims. Architecture and workload-specific performance require independent customer validation. Extrapolated maximum-model claims are excluded from the thesis.

https://www.cerebras.ai/blog/introducing-cerebras-cs-4
S5 · Cerebras CS-4 product overview

Accessed 18 October 2026 · Primary product description. Performance comparisons depend on model, configuration and workload; they are not universal GPU comparisons.

https://www.cerebras.ai/cs4
S6 · Cerebras investor relations

Checked 3 November 2026 · Primary IR/news index: Oracle announcement 27 October; next quarterly results scheduled 12 November. Index search results reviewed; direct snapshot access returned 403.

https://investors.cerebras.ai/
S7 · Morgan Stanley — Quick Thoughts on Ultrafast

1 October 2026; supplied research note, pp.1–3 · Secondary analyst commentary and channel checks about competing inference offerings, delivery and frontier-weight security. The linked URL identifies the publisher; the supplied note is not publicly hosted at that link. Its USD177.65 price and USD279 target are dated, not the present price or a Capital Thesis valuation.

https://www.morganstanley.com/
S8 · Morgan Stanley — Semiconductors Weekly: MU Preview

28 September 2026; supplied research note, CBRS comparison rows · Mostly Micron/sector material. Only dated CBRS forecast/valuation rows inform the disagreement map; unrelated memory commentary and dated short-interest data are excluded. Publisher link is not a public copy of the note.

https://www.morganstanley.com/
S9 · Morgan Stanley — Supernova Recap: Cerebras Unveils CS4

19 August 2026; supplied research note, pp.1–2 · Secondary technology interpretation, checked against company product materials. Stale USD220.01 share price and maximum-model extrapolations are not carried into the model. Publisher link is not a public copy of the note.

https://www.morganstanley.com/
S10 · Barclays — 600MWs Only the Starting Point

13 August 2026; supplied research note, pp.1–4 · Secondary 2026–28 forecasts, core-results reconciliation and target framework. Forecasts are labelled separately from filed actuals and house scenarios. Publisher link is not a public copy of the note.

https://www.home.barclays/