GLOBALFOUNDRIES GFS

Reference table

1. What they actually sell — and a serious disclosure gap

The segment data for FY2025 is incomplete, and the gap is large.

The FY25 seg_prod block contains only the $769M pre-fabrication services line. The wafer fabrication line — which by subtraction must be approximately $6,022M, or 88.7% of revenue — is simply absent from the file. We can infer it, and we state that inference explicitly rather than presenting it as disclosure.

Geography is complete and it tells a real story:

This is the most informative table in the file and it is worth dwelling on. United States revenue has fallen from $4,262M to $3,395M — down 20.3% in two years — while EMEA rose 11.5% over the same period and accelerated to +20.2% in FY25. The entire FY25 revenue stabilisation ($6,750M to $6,791M) came from Europe offsetting a continued United States decline.

Two readings are possible and the data cannot distinguish them: either European automotive and industrial demand is recovering ahead of the United States, or GlobalFoundries is losing share with a large domestic customer. Given that the stock has fallen 31% below its 50-day average, the second possibility deserves more weight than it would otherwise — and it is exactly the kind of thing tomorrow's print might reveal.

Note also the description of the business from the company profile: microprocessors, mobile application processors, baseband and network processors, radio-frequency modems, microcontrollers, power management units, and microelectromechanical systems. Not one of those categories is an AI accelerator. That is the structural fact underpinning the entire growth assessment.

2. The cycle — four years of no growth, and a real margin repair

Read the revenue row honestly: $8,108M in FY22, $6,791M in FY25. This company is 16.3% smaller than it was four years ago. That is the single most important fact about GlobalFoundries and it is often lost in semiconductor enthusiasm. It rides an automotive, industrial and handset cycle that has been in a downturn, not the AI cycle.

The FY24 loss deserves explanation. The full-year operating loss of -$214M was produced almost entirely by a single quarter: Q4'FY24 operating income of -$701M against roughly +$150-185M in each of the other three quarters. That is a large impairment or restructuring charge; the file does not identify it. Excluding it, FY24 was a normal, profitable, low-growth year. But it demonstrates that this business takes large periodic write-downs, which is characteristic of capital-intensive manufacturing.

The quarterly recovery is genuine but modest:

Gross margin is the real improvement: 22.4% to 27.6%, up 520 basis points in five quarters. That is the operating leverage of a capital-intensive business filling capacity it already built.

But revenue growth is +3.1%, and consensus needs +6.7% this year and +11.0% next. No reported quarter has delivered anything close to the acceleration the estimates require. That gap between the reported trend and the modelled recovery is the central fundamental tension in this name.

Note the GAAP-versus-adjusted divergence in Q1'26: GAAP diluted EPS $0.18 against adjusted $0.40 — the widest gap in the recent series and worth watching. Across the five quarters shown, adjusted has averaged roughly 1.2x GAAP; in Q1'26 it was 2.2x.

3. The balance sheet and cash flow — the genuinely good news

FY25 year-end: cash $1,809M plus short-term investments $1,241M = $3,050M; inventory $1,577M; total current assets $6,205M; property, plant and equipment $7,792M; total assets $17,141M; short-term debt $86M, long-term debt $1,065M, total debt $1,707M; total liabilities $5,158M; equity $11,928M.

• Our computed net cash: $3,050M − $1,707M = $1,343M, giving an enterprise value of $26,098M. The vendor reports $27,316M, implying net cash of only $125M — it nets the $1,809M cash line against debt and ignores $1,241M of short-term investments. All EV figures here use our corrected number, which takes EV/Sales from the vendor's 3.99x to 3.82x.
• Debt has been reduced 38% in two years: $2,754M (FY23) → $2,320M (FY24) → $1,707M (FY25), with $840M of net debt repayment in FY25 alone. Equity of $11,928M against $1,707M of debt is a genuinely conservative structure, and debt-to-equity of 0.147x is the second-lowest in this batch after the net-cash software names.
• Property, plant and equipment of $7,792M is 45% of total assets. This is a heavy manufacturer, and it has been reducing net PP&E ($10,164M in FY23 to $7,792M in FY25) as depreciation exceeds capital expenditure — the definition of a build phase ending.

This is the strongest part of the story. Capital expenditure has collapsed from $3,059M (FY22) to $722M (FY25) — down 76% — while operating cash flow held at $1.7B, converting a -$435M free cash flow drain into $1,009M of genuine free cash flow. At our enterprise value of $26.10B that is a 3.9% free-cash-flow yield; against the market capitalisation, 3.7%.

Two cautions. First, capital expenditure of $722M is well below depreciation of $1,314M — the company is spending materially less than its assets are wearing out. That is normal at the end of a build cycle and unsustainable indefinitely; a genuine demand recovery would require capital expenditure to rise, and the FY22 figure shows how far it can go. Second, $682M went to acquisitions in FY25 — the largest use of cash after debt repayment — with no detail in the file. Goodwill rose from $75M to $407M and intangibles from $585M to $961M, so roughly $708M of the purchase price was allocated to intangible assets. What was bought, and why, is not disclosed here.

Note the dividend anomaly. The company profile reports a last dividend of $0.12 and dividendPerShareTTM of $0.12 (a 0.24% yield), but the FY25 cash flow statement shows $0 of common dividends paid. A dividend was therefore initiated during 2026 and is not yet reflected in any annual cash-flow data available to us. Flagged rather than reconciled.

4. Valuation — priced in or room?

At $50.01 (market cap $27.44B, our corrected EV $26.10B), trailing revenue is $6,840M, trailing EBITDA is $2,187M, trailing free cash flow is approximately $1,009M, and trailing GAAP diluted EPS is $1.39.

FY29E is incoherent — revenue growth of +0.1% following +13.1%, and EPS of $3.120 sitting BELOW FY28E's $3.314 — driven by a collapsing analyst panel (11 to 5 revenue and 7 to 2 EPS contributors). **Excluded from all analysis. FY26E, FY27E and FY28E carry 7-13 EPS contributors and are the reliable base.

6a. What today's price assumes (the inversion)

At $50.01 — 20.1x FY27E consensus adjusted EPS, 3.82x trailing EV/sales and 11.9x trailing EV/EBITDA — the price embeds roughly:

• Revenue reaccelerates from the reported +3.1% to +6.7% this year and +11.0% next, adding $1.20B of annual revenue by FY27. (Source: consensus, 15-16 analysts.) THIS IS THE MOST FRAGILE ASSUMPTION IN THE PRICE. No reported quarter in the last five has grown faster than +3.4%, and revenue is still 16.3% below its FY22 peak. Falsifier: two more quarters below +5% growth.
• Gross margin holds at or above the current 27.6% and continues expanding toward 30%, delivering the FY27E adjusted EPS of $2.491 on $8.042B of revenue. (Our inference from the reported margin trend of 22.4% → 27.6%; the consensus margin series is rejected as corrupt.) Falsifier: gross margin below 25% in any quarter.
• Capital expenditure stays near $722M rather than reverting toward depreciation of $1,314M or the FY22 level of $3,059M. The 3.9% free-cash-flow yield depends entirely on this. (Cash flow statement.) Falsifier: quarterly capital expenditure implying an annual run rate above $1.2B.
• The market keeps paying ~20x forward. At an unchanged price the multiple rolls to 15.1x FY28E. At today's price the market is arguably NOT paying 20x — it is pricing the bear case, given that $50.01 is precisely the lowest analyst target and 36% below the $78.25 consensus.
• No further large impairment. FY24 included a -$701M single-quarter operating charge in a capital-intensive business with $7.79B of property, plant and equipment. Falsifier: any new write-down.

6b. The return bridge (why the multiple moves)

Expected return ≈ EPS growth + multiple drift + shareholder yield.

Our base case of $62 is built as 25.0x FY27E consensus adjusted EPS of $2.491 = $62.3.

Note what this assumes and, more importantly, what it does not. Today the stock trades at 26.4x FY26E. Applying 25.0x to FY27E means holding roughly today's forward multiple constant and rolling it forward one year. There is no re-rating in the base case at all — the entire 24% return comes from consensus adjusted earnings growing from $1.896 to $2.491, a 31.4% increase, partially offset by a marginal multiple give-back.

That is a deliberately conservative construction and it is the right one here. Unlike LPLA or ZM, where the base case needed a re-rating, or CW and SYM, where it needed compression, GFS's base case works if the earnings simply arrive. The fragile leg is therefore not the multiple — it is the FY27 revenue reacceleration to $8.042B, which no reported quarter yet supports.

Shareholder yield is negligible: a 0.24% dividend and $5M of buybacks in FY25 (down from $200M in FY24, as cash went to debt repayment and acquisitions instead). Do not attribute any return to capital return. The $840M of debt repayment is genuine value accrual to equity holders, but it is now largely complete.

6c. Variant perception (where we differ, what would surprise)

• We are more constructive than the price and materially less constructive than the street. Consensus targets average $78.25 (+56%) with a range of $50.00 to $125.00. The stock trades at $50.01 — the lowest target, to the cent. That configuration is unusual and informative: the market is currently pricing the single most bearish covering analyst's view, while the average sees 56% upside. Our $62 sits deliberately between them.
• Our specific variant perception is about the revenue reacceleration, and it is a cautious one. Consensus needs +6.7% and then +11.0% revenue growth; the last five reported quarters delivered +2.3%, +3.4%, -2.9%, 0.0% and +3.1%. We think the FY27 revenue estimate is optimistic relative to the observed trend, which is why our base of $62 is 21% below the street's $78.25 despite our being constructive on the entry point. We are buying the oversold technical and the margin repair, not the modelled recovery.
• Where we are MORE positive than the market: the balance sheet and the cash flow. Net cash of $1.34B, debt cut 38% in two years, capital expenditure down 76%, and $1.01B of free cash flow are not the characteristics of a business that should trade at a 44% drawdown. This is a genuinely improved company at a genuinely distressed price.
• Positive surprise that would force a repricing: tomorrow's print showing revenue growth above +6% with gross margin above 28% and constructive second-half commentary. That would validate the consensus reacceleration and make the $78.25 target credible. Note the company has beaten the consensus EPS line in six consecutive quarters by an average of 13.9%. Watchable at: 2026-08-05, tomorrow.
• Negative surprise, and the one the price is already sniffing: a demand warning, a lost customer, or continued United States revenue decline. The 20.3% two-year fall in United States revenue against +20.2% EMEA growth in FY25 is the one genuinely worrying pattern in this file, and a 31% collapse below the 50-day average is consistent with the market knowing something about it that we cannot see. Watchable at: 2026-08-05, tomorrow.

Synthos fair values

• Bear ~$34 (-32%). The mature-node cycle rolls over again as it did in FY23-24 (a 16.8% peak-to-trough revenue decline), the consensus reacceleration does not arrive, revenue stays near $6.8B, and a capital-intensive foundry with a 3.9% free-cash-flow yield re-rates to 13.5x depressed FY27E adjusted EPS of $2.491 = $33.6. This is 8% above the $31.51 52-week low, so the market visited a comparable level within the last twelve months.
• Base ~$62 (+24%). 25.0x FY27E adjusted EPS of $2.491 — approximately today's forward multiple, rolled forward one year. No re-rating assumed; the return is entirely earnings growth.
• Bull ~$100 (+100%). The essential-chip and Western-manufacturing thesis re-rates the whole company as the automotive and industrial cycle turns: 30.0x FY28E adjusted EPS of $3.314 = $99.4. Note this is 20% below the street's high target of $125 and 11% above the 52-week high of $89.96 — so even the bull case is a return to recent highs plus a modest premium, not a new regime.

5. Knowledge base — zero substantive claims

A systematic search on GFS, GlobalFoundries and Global Foundries returns exactly one hit, and it carries no investment view.

lex_fridman (2021-09-26, conviction 65, bullish), entity-tagged Intel and Global Foundries, categorised under semiconductor manufacturing and photolithography:

> "Silicon's moat is manufacturability — photolithography, ion implantation, ever-bigger wafers — enabling ~1000 processors per wafer at low unit cost; a scalability physics itself permits."

Global Foundries is named as an example within a general observation about semiconductor manufacturing physics. There is no view expressed on the company as a business or as a security. It is five years old and predates the FY22 revenue peak, the FY23-24 downturn, the FY24 impairment and the entire margin repair.

We therefore record kb_claim_count: 0 and kb_breadth: 0. This dive is fundamentals and quant-driven only — an accepted, precedented outcome, and never a reason to manufacture conviction.

One idea is worth extracting even so, because it is directly relevant. lex_fridman's framing — that silicon's moat is manufacturability rather than design — is precisely the argument for owning a mature-node foundry rather than a fabless designer. The value in GlobalFoundries is that it physically makes things at scale in the United States and Europe, which is exactly what the geographic disclosure shows ($3.40B United States, $1.71B EMEA). That is an observation, not a claim, and it is weighted as such.

6. Technicals — the most extreme setup in the batch

This section carries more weight than usual, because the technical picture is the reason this dive reaches Stage-In rather than Watch.

• Price $50.01 (+0.04% today), -44.4% from the $89.96 52-week high (the quote series shows a $92.55 high), +58.2% above the $31.62 low.
• 30.8% BELOW the 50-DMA ($72.27). This is the defining fact. A 31% gap between price and the fifty-day average means a collapse of extraordinary speed — the average has not had time to follow the price down, which implies the fall happened within roughly the last two months.
• Fractionally BELOW the 200-DMA ($51.55) — 3.0% below. The primary trend line has just been broken. This is the single most important level in the chart: reclaiming ~$51.55 would signal the break was a flush rather than a regime change.
• RSI 22.6 — the most oversold reading in this batch (the next lowest is Reddit at 27.3). Readings below 30 are conventionally oversold; below 25 is rare and typically resolves in a bounce. MACD -6.49, deeply negative.
• Relative performance shows the whole shape: 12-month +33.8% against SPY +19.9%, 6-month +18.5%, but 3-month -23.0% against SPY +5.1%. The stock ran from roughly $37 twelve months ago to $90, and has given back half of that in a quarter.
• $50.01 is within one cent of the street's LOWEST price target of $50.00, against a consensus of $78.25 and a high of $125.00. The market is currently trading exactly at the most bearish published view.
• Beta 1.796 — the third-highest in this batch. Average volume 4.7M shares.

The technical read is genuinely constructive on a contrarian basis and genuinely dangerous on a trend basis, and both are true simultaneously. An RSI of 22.6 at the 200-DMA at the lowest analyst target is the kind of confluence that produces sharp reversals. A 31% gap below the 50-DMA is the kind of collapse that usually has a cause. With earnings twenty-four hours away, the correct response is a partial position, not a full one.

7. Moat & competitive position

GlobalFoundries' position is qualified mature-node capacity with a Western manufacturing footprint — and it is a moderate moat, not a wide one.

The genuine advantages. Semiconductor process qualification is expensive and slow: once a customer designs a part for a specific fab and process, moving it requires requalification and re-certification, which for automotive and industrial parts can take years. That produces real stickiness and explains how revenue held roughly flat through a demand downturn. The geographic position adds a second layer — $3.40B of United States and $1.71B of EMEA revenue means GlobalFoundries operates capacity outside Asia at a time when supply-chain geography carries strategic value. lex_fridman's observation about manufacturability being the moat applies directly.

The genuine limitations, and they are what cap the score at 5.

1. Returns are structurally low. Return on invested capital is 4.9% and return on equity 6.7% on the trailing data. A business earning 5% on capital is not compounding value, regardless of how sticky its customers are. Compare the 19.6% return on equity at Curtiss-Wright.

2. Capital intensity is punishing. $7,792M of property, plant and equipment and $1,314M of annual depreciation against $6,840M of revenue. The FY22 capital expenditure of $3,059M — 38% of that year's revenue — shows what a genuine expansion cycle costs.

3. Mature-node foundry is price-competitive. Named peers include United Microelectronics and STMicroelectronics; capacity is available globally, much of it heavily subsidised. Gross margin of 25-28% against a leading-edge foundry's 50%+ is the market's verdict on pricing power.

4. No AI exposure. This is the honest structural fact. GlobalFoundries does not participate in the demand event driving semiconductor valuations, and it never will at mature nodes.

Verdict on the moat: real but moderate — qualification stickiness and Western capacity in a capital-heavy, low-return, price-competitive industry. It justifies a fair multiple on a recovered earnings base. It does not justify a premium.

8. Data integrity — what we rejected and why

Vendor data corruption checking is standard for these dives. For GFS:

• REJECTED: the consensus EBIT and EBITDA series. The est block reports FY22E EBIT of $138.1M against reported FY22 operating income of $1,167M — off by a factor of 8.5. FY25E shows EBIT of $1,132M against reported $797M, a 42% overstatement. The series is not internally consistent in any year and is excluded. Only revenue, net income and EPS from est are used, which is why section 6a contains no consensus margin bullet and the gross-margin assumption is derived from reported results instead.
• REJECTED: FY29E entirely. Revenue growth of +0.1% following +13.1%, and EPS of $3.120 falling below FY28E's $3.314 — driven by a collapsing analyst panel (11 to 5 revenue contributors, 7 to 2 EPS contributors). Not a coherent forecast. Excluded.
• CORRECTED: enterprise value. The vendor reports $27,316M, netting only the $1,809M cash line and ignoring $1,241M of short-term investments. Correct net cash is $1,343M and enterprise value $26,098M, taking EV/Sales from 3.99x to 3.82x. All EV figures use the corrected number.
• FLAGGED: the interest-expense field is zeroed. interestExpense reads $0 in every quarter of FY25 and FY26 despite $1,707M of total debt, and interestCoverageRatioTTM is 0. The FY24 annual figure of $145M and FY23's $137M confirm the company does pay interest. The recent interest-expense series is unusable, and any margin analysis that depends on it is avoided.
• MAJOR GAP: the FY2025 product segment disclosure is incomplete. seg_prod for FY25 contains only the $769M "Engineering and Other Pre-Fabrication Services" line; the wafer fabrication line — approximately $6,022M, or 88.7% of revenue — is absent. Every prior year discloses it. We infer the figure by subtraction and label it as an inference throughout.
• FLAGGED: dividend inconsistency. The profile reports lastDividend of $0.12 and dividendPerShareTTM of $0.12 (0.24% yield), while the FY25 cash flow statement shows $0 of common dividends paid. A dividend was initiated during 2026 and is not captured in any annual cash-flow data available to us. Noted, not reconciled.
• FLAGGED: a share-count irregularity. Q3'FY25 reports basic shares of 563.6M and diluted shares of 559.0M — diluted BELOW basic, which is arithmetically impossible. Adjacent quarters show basic 555M / diluted 557-561M. The Q3'FY25 share figures are treated as unreliable; all per-share work uses the Q1'FY26 figures (555M basic, 561M diluted).
• FLAGGED: the FY24 impairment is unexplained. Q4'FY24 operating income of -$701M against roughly +$150-185M in the surrounding quarters. The file gives no description. Its existence is used as evidence of write-down risk; its cause is not asserted.
• THE CENTRAL GAP: no news, guidance or company commentary of any kind. This data file contains financial statements, prices and estimates. It contains nothing that explains why the stock is 31% below its 50-day average. For a name whose entire near-term setup is a violent unexplained decline, this is the binding constraint on the analysis, and the staged entry is constructed around it rather than despite it.
• GAP: no customer concentration disclosure and no capacity-utilisation data. For a foundry, utilisation is the primary margin driver and it is absent.
• NOTE: insider activity is routine. Several directors had in-kind share withholdings of ~1,700 shares each at $49.02 on 2026-07-29 (tax withholding on vesting, not sales), and Chief Legal Officer Samak Azar sold 335 shares at $48.93 on 2026-07-30. Three director entries show zero securities and zero price — likely new-appointment filings. No open-market purchases and no meaningful sales. Notably, no insider bought into the collapse, which is a missing bullish tell.

9. Verdict, kill-criteria & flip conditions

Stage-In. GlobalFoundries is a mature-node essential-chip foundry with $1.34B of net cash, debt cut 38% in two years, capital expenditure down 76% from its peak, $1.01B of free cash flow, gross margin up 520 basis points in five quarters, and a FY24 loss turned into $885M of FY25 profit. Consensus models adjusted earnings roughly doubling from $1.90 to $3.31 by FY28, putting the stock at 26.4x, 20.1x and 15.1x. Base fair value $62 (+24%), built by holding today's forward multiple and rolling it one year — no re-rating required.

And it is trading at a genuine oversold extreme: RSI 22.6, the lowest in this batch; 31% below its own 50-day average; sitting on a just-broken 200-DMA; -44.4% from the 52-week high; and at $50.01, precisely the lowest price target any covering analyst carries, against a consensus of $78.25 and a high of $125.

The reason this is Stage-In rather than Buy is stated without hedging: our data does not explain the collapse, and earnings report tomorrow. A 31% fall below the fifty-day average is not ordinary volatility — it is the market pricing information. This file contains financial statements and prices; it contains no news, no guidance and no company commentary. The one worrying pattern we can see is that United States revenue fell 20.3% over two years while EMEA grew 20.2% in FY25 — a mix shift consistent with either a European recovery or a domestic customer problem, and we cannot tell which.

Staged entry (build a 1.5-3% position in 3 tranches):

1. Tranche 1 — now (~1/3 of target), at ~$50. RSI 22.6, at the 200-DMA, at the street's low target, 20.1x FY27E, net cash, free cash flow positive. Underwrite the balance sheet and the margin repair, not the print.

2. Tranche 2 — the reaction to tomorrow's 2026-08-05 print. Add on a beat with constructive second-half commentary and gross margin above 28% (buying confirmation), or on a capitulation flush toward $40-42 that holds. Do not add in the middle.

3. Tranche 3 — only on evidence of the revenue reacceleration: a quarter above +6% growth, or United States revenue returning to growth. This tranche is gated on the fundamental question, not on price.

Pre-registered KILL / avoid-adding criteria:

• Two more quarters below +5% revenue growth, which would break the consensus path to $8.04B in FY27 — the assumption the entire base case rests on.
• Gross margin below 25%, reversing the 520-basis-point repair that is the best fact in this file.
• Capital expenditure rising toward the $1.3B depreciation level or beyond, which would eliminate the 3.9% free-cash-flow yield.
• Continued United States revenue decline below the FY25 $3,395M — the pattern most consistent with a customer or share problem rather than a cycle problem.
• Any new large impairment in the mould of the FY24 -$701M charge.
• A decisive loss of the 200-DMA (~$51.55) on company-specific bad news, opening the $31.51 52-week low as the reference.

Pre-registered FLIP TO HIGH-CONVICTION (upsize toward 3%):

• Revenue growth above +6% for two consecutive quarters with gross margin above 28% — this would validate the consensus reacceleration and make the street's $78.25 credible, lifting our base toward $75.
• The 200-DMA reclaimed and held, confirming the collapse was a flush rather than a regime change.
• United States revenue returning to growth, removing the single worrying pattern in the geographic data.
• Open-market insider buying into the drawdown — conspicuously absent so far.
• Any disclosure of the FY25 $682M acquisition's contribution, or of capacity utilisation.

Where GFS fits in the Synthos Framework Portfolio. The semiconductor / essential-silicon sleeve, at a 1.5-3% core weight. Its portfolio function is specific and deliberately contrarian: it is the non-AI semiconductor exposure, levered to automotive, industrial and handset demand and to Western manufacturing capacity rather than to accelerator capital expenditure. That makes it genuinely diversifying against an AI-heavy technology allocation — but note the beta of 1.796 and a 52-week range of $31.51 to $92.55, a 2.9x swing. Size for volatility, not for conviction. Logged as a tracked Synthos call (Stage-In) as of 2026-08-04 at $50.01.

Single biggest risk: a 31% collapse below the 50-day average whose cause is invisible in our data, with earnings twenty-four hours away. Markets do not move a $27B semiconductor company that far that fast without information, and this file — financial statements, prices, estimates, no news — cannot tell us what that information was. The most fragile assumption in today's price is the consensus revenue reacceleration from the reported +3.1% to +6.7% this year and +11.0% next, which no quarter in the last five supports and which the entire base case depends on. If the collapse turns out to reflect a demand warning or a lost United States customer, the reacceleration does not happen, FY27 adjusted earnings of $2.491 do not arrive, and $50 is not the bottom — the $34 bear case is. Buy a third; that is what the uncertainty is worth.

Provenance & disclosures

• Traceability: ZERO substantive knowledge-base claims. A systematic search on GFS, GlobalFoundries and Global Foundries returned one hit — lex_fridman (2021-09-26, conviction 65), in which Global Foundries is named as an example within a general observation about semiconductor manufacturing physics, with no view expressed on the company or the security. Five years old and predating the FY22 peak, the FY23-24 downturn, the FY24 impairment and the entire margin repair. Recorded as a mention, not counted as a claim. kb_claim_count: 0, kb_breadth: 0. This dive is fundamentals and quant-driven only — an accepted, precedented outcome.
• Data as-of: fundamentals 2026-03-31 (Q1'26, filed 2026-05-05) · estimates 2026-08-04 · prices 2026-08-04 ($50.01; 50-DMA $72.27, 200-DMA $51.55, RSI 22.6, MACD -6.49, beta 1.796) · KB search 2026-08-04. Sole data source: scripts/deepdive/vti_data/GFS_data.json.
• Rejected data: the consensus EBIT and EBITDA series (FY22E EBIT of $138.1M against reported operating income of $1,167M, off 8.5x; FY25E off 42%); FY29E entirely (revenue growth of +0.1% following +13.1%, EPS falling below FY28E, on a collapsing 5-and-2 analyst panel); the Q3'FY25 share counts (diluted 559.0M reported below basic 563.6M — arithmetically impossible).
• Corrected data: vendor enterprise value of $27,316M ignores $1,241M of short-term investments; corrected to $26,098M on net cash of $1,343M, taking EV/Sales from 3.99x to 3.82x.
• Flagged data: interestExpense reads $0 in every FY25 and FY26 quarter despite $1,707M of debt (FY23-24 annual figures of $137M and $145M confirm interest is paid) — the recent series is unusable; a dividend of $0.12 reported in the profile against $0 of dividends in the FY25 cash flow statement, indicating 2026 initiation not yet captured; a FY24 fourth-quarter operating charge of -$701M whose cause the file does not identify.
• Disclosed gaps: the FY2025 product segment disclosure is incomplete — only the $769M pre-fabrication services line appears, with the ~$6,022M (88.7%) wafer fabrication line absent, so that figure is INFERRED by subtraction and labelled as an inference throughout; no news, guidance or company commentary of any kind, which is the binding constraint given the unexplained 31% collapse; no customer concentration disclosure; no capacity-utilisation data, which for a foundry is the primary margin driver; no detail on the $682M of FY25 acquisitions.
• Non-GAAP note: all forward P/E multiples use the consensus adjusted EPS series, which runs above GAAP (Q1'26: $0.40 adjusted against $0.18 GAAP diluted). The trailing GAAP multiple of 36.0x is shown alongside. Labelled at each use.
• Fair-value caveat: the $34 / $62 / $100 anchors are scenario multiples applied to consensus adjusted EPS, labelled at each step. The base case assumes NO re-rating — approximately today's forward multiple rolled forward one year — so the return derives entirely from consensus earnings growth, and the fragile leg is named as the FY27 revenue reacceleration rather than the multiple. This is scenario arithmetic, not a discounted-cash-flow model.
• Not investment advice. Independent research, educational and informational only, never personalized. Nothing here is a recommendation to buy or sell any security.
• Version: 2026-08-04-full.