Phase Space AI

Financial Model Notes

Seagate Technology [STX]

STX — Financial Model Notes

Sources, transformations, and every place a vendor figure was rejected in favour of a primary filing.

Data provenance

Item Source used Why not the alternative
FY2026 revenue, margins, EPS, balance sheet, cash flow 8-K 0001137789-26-000153, EX-99.1 (filed 2026-07-29) Most recent primary source; postdates every API's latest period
Quarterly revenue / operating income / R&D, FQ1'24–FQ3'26 EDGAR companyfacts XBRL (RevenueFromContractWithCustomerExcludingAssessedTax, OperatingIncomeLoss, ResearchAndDevelopmentExpense) AV agrees to within 1.9%; EDGAR is primary
Quarterly gross profit / gross margin AV INCOME_STATEMENT (grossProfit) cross-checked against the 8-K GAAP reconciliation STX does not tag GrossProfit in XBRL. AV's FQ4 figure ties exactly to the 8-K ($1,898m)
Shares outstanding 8-K: 227m (2026-07-03); 10-Q XBRL 224.4m + dei cover 224.2m (2026-04-27) AV BALANCE_SHEET.commonStockSharesOutstanding = 229.0m is the diluted WASO, not outstanding — rejected
Diluted shares for valuation 232m (FQ4'26 GAAP diluted, from the 8-K) Includes exchangeable-note dilution that management explicitly says is in FQ1'27 guidance
Debt and cash 8-K balance sheet: $185m current + $3,380m long-term = $3,565m; cash $1,704m AV longTermDebt and shortTermDebt were checked and are consistent one quarter earlier, but the newest period is only in the 8-K. The coordinator's longTermDebt-returns-zero defect was specifically checked: STX's XBRL LongTermDebtNoncurrent = $3,465m at 2026-04-03, non-zero, so that defect did not fire here.
Exabytes shipped Earnings-call transcripts, CFO prepared remarks Not in XBRL and not in the press release; only in the call and the (non-EDGAR) supplemental
Opex lines for the bridge 8-K income statement: product development $188m, marketing & administrative $147m (FQ4) Per the coordinator's warning, sellingGeneralAndAdministrative from AV was not used for the bridge
Prices, realised vol Alpaca snapshots (last trade) and bars with adjustment=all
Splits AV SPLITS: empty — no splits. Verified explicitly per the KLA precedent

Vendor figures REJECTED, with magnitudes

Field Vendor value Correct value Error
AV BALANCE_SHEET.commonStockSharesOutstanding (2026-03-31) 229.0m 224.4m +4.6m / +2.1%
AV OVERVIEW.MarketCapitalization $190,974m (implies a $851.69 price) $171,356m at 224.2m × $764.20 +$19.6bn / +11.4%; built on a stale price
AV OVERVIEW.EPS $9.64 $13.90 FY26 GAAP diluted −30.6%
AV OVERVIEW.TrailingPE 77.52 55.0 on FY26 GAAP EPS +41%
AV OVERVIEW.EBITDA $3,514m $4,220m LTM non-GAAP EBITDA (8-K) −$706m / −16.7%
AV OVERVIEW.OperatingMarginTTM 35.7% 28.3% (TTM GAAP op income $3,119m / revenue $11,010m to 2026-03-31) +7.4pp — consistent with the coordinator's finding that AV's ebit is not operating income (AMAT: 9.7pp)
AV OVERVIEW.RevenueTTM $11,010m correct as of 2026-03-31, but one quarter stale vs FY26 $12,195m −9.7%
AV INCOME_STATEMENT.operatingIncome (2025-10-03) $707m $694m (EDGAR) +$13m / +1.9%
AV EARNINGS_ESTIMATES estimates: [] endpoint returns nothing on this name
AV EARNINGS_CALL_TRANSCRIPT quarter=2026Q2 December-2025 call fiscal, not calendar labelling off by 2 quarters

Not rejected: AV's ebitda field was never used (brief §"Two known AV defects"), and EBITDA was taken from the 8-K's own reconciliation, which correctly adds positive D&A of $276m to EBIT.

Model structure

Revenue build: exabytes × revenue-per-exabyte, by segment (data centre vs edge/IoT), reconciled to reported revenue. This is the only defensible build for this business because both terms are disclosed and both are moving.

FY2026A FY2027E FY2028E
Revenue $12,195m ~$17.6bn ~$21.1bn
growth +34.1% +44% +20%
Non-GAAP gross margin 46.1% ~52% ~53%
Non-GAAP opex $1,170m (9.6%) ~$1,240m (7.0%) ~$1,315m (6.2%)
Non-GAAP operating margin 36.5% ~46% ~47%
Non-GAAP diluted EPS $15.58 ~$31.60 ~$35.80

FQ1 FY2027 is guided ($4.1bn, $7.30). FQ2–FQ4 FY2027 are a sequential progression consistent with (a) nearline capacity almost fully allocated through CY2027, (b) mid-20s% data-centre exabyte growth, and (c) mid-single-digit to high-single-digit revenue-per-TB increases. FY2028 applies management's stated "minimum 20%" revenue growth target.

Not modelled, deliberately: capacity expansion capex beyond the 4–6% of revenue target (management has ruled out unit-capacity investment); any acquisition; any change to the exchangeable-note capital structure beyond scheduled maturity.

Terminal-period assumptions and where they came from

Parameter Value Derivation
m_gross,T 44.0% 1.6pp below FY2026 actual 45.6%; 8.3pp below FQ4 exit 52.3%
R&D 4.5% of revenue FY26 product development $755m = 6.2%; near-fixed, so falls with scale; held at 4.5% not 3.5% for post-HAMR roadmap
S&M + G&A 4.0% FY26 marketing & admin $577m = 4.7%
Other 0.5% restructuring, amortisation
m_EBIT,T 35.0% bridge residual; above FY26 GAAP actual 33.6%
t 16% FY26 GAAP effective 13.7%; guided ~16%
g 4.0% terminal
ROIC 40% FY26 actual 66% (NOPAT $3,533m / invested capital $5,326m), decayed
WACC 10.0%
Exit EV/EBIT 12.6x from (1−t)(1−g/ROIC)/(WACC−g), not asserted
FCF margin (interim) 26.0% FY26 actual 25.5%; FQ4 30.8%

Known limitations

  1. The shipped reverse_dcf.py is terminal-only and ignores five years of interim FCF, which on a 26%-FCF-margin business overstates the required CAGR by ~5.1pp (39.9% vs 34.8%). Both figures are reported; the FCF-inclusive one is used.
  2. No options data was priced, so no IV−RV measurement exists and no derivative vehicle is admissible.
  3. The FQ4 FY2026 earnings-call transcript was not available at the time of writing (call was scheduled for 2:00 PM PT today; AV returned an empty transcript array). Contract-term and allocation language is therefore as of the FQ3 call (2026-04-28), three months old.
  4. Exabyte series has a disclosure break in Dec-2024 and two segment renames. Only the last three quarters are strictly comparable.
  5. entity_public_float (223.559m) ≤ SharesOutstanding (224.229m) ≤ diluted (232m); float × spot $170.9bn ≤ market cap $171.4bn. Invariant holds.