Gilead Sciences GILD

Reference table

> REPORTED TODAY. The 8-K in this archive is dated 2026-08-04 — the same day as this dive — and announces second-quarter results. Exhibit 99.1, which carries the financials, was NOT extracted, so no company-stated figure is available to us. The vendor payload does carry a June-quarter income statement: revenue $7,803M against a $7,399M consensus — a 5.5% beat and +10.2% year on year — and an operating loss of $10,394M from expensed asset acquisitions. earn_cal.epsActual is still null. The verdict is set with that incompleteness as an explicit input.

1. What the business is, and the four transactions

Chief executive Daniel O'Day; 17,000 employees; NASDAQ-listed; CIK 0000882095; reporting in USD.

Product mix, and seg_prod does not sum. The five fiscal-2025 lines total $26,042M against reported revenue of $29,442M — a $3,400M (11.5%) shortfall. The fiscal-2023 entry contains a "Hepatitis B Virus / Hepatitis Delta Virus Product" line of $1,017M that is absent from fiscal 2025, and the liver-disease franchise as a whole (hepatitis C, hepatitis B and primary biliary cholangitis) is not represented in the current-year entry. The block is used for the lines it does contain and the shortfall is stated.

Two things follow. HIV at 70.5% of revenue and growing 5.8% is the company. And Cell Therapy is DOWN 8.1% — which matters more than its size, because it is the franchise the $7.1 billion Arcellx purchase is meant to revive.

seg_geo ties exactly. United States $20,876M + Europe $5,064M + Other International $3,503M = $29,443M against reported revenue of $29,442M. Verified clean. The United States is 70.9% of revenue — among the highest domestic concentrations in this batch.

The 2026 transactions, from the 10-Q filed 2026-05-07 — none of this appears in any vendor field:

Approximate net cash cost of the three 2026 deals: $7.1B + $3.2B + $0.85B = $11.15 billion, essentially all expensed. The 10-Q's wording is explicit in each case — for example on Arcellx: "We expect to account for this transaction as an asset acquisition in the second quarter of 2026 since the acquired intellectual property rights related to anito-cel represent substantially all of the fair value of the gross assets acquired."

This is not new behaviour. The 10-K explains fiscal 2024's operating income of $1,662M — against $7,605M in 2023 and $11,701M in 2025 — as reflecting "a $3.8 billion acquired in-process research and development ('IPR&D') expense related to the acquisition of CymaBay Therapeutics, Inc. ('CymaBay') in 2024, which did not repeat in 2025." Gilead's reported operating income is therefore violently lumpy for a reason that has nothing to do with the operating business, and any multi-year margin series built from it is meaningless without the adjustment.

2. The income statement, and why it swings

All EPS figures are computed by us — the vendor's epsdiluted field is NULL in all six annual and all nine quarterly rows, the fourth instance of this defect in the batch.

Operating margin of 5.8% (2024) and 39.7% (2025) describe the same business. The difference is $3.8 billion of expensed CymaBay IPR&D in one year and none in the next.

The June quarter is the strongest revenue quarter in the file — $7,803M, up 10.2% year on year and 5.5% ahead of the $7,399M consensus — and it carries an operating loss of $10,394M. The two facts are unrelated: the first is the business, the second is the accounting for $11.2 billion of purchased assets.

Three consecutive non-GAAP beats before it (+16.0%, +2.8%, +6.3%).

Two reconciliation problems in the June quarter that we cannot resolve and therefore flag. First, the vendor books the entire charge inside cost of revenue, producing a gross profit of MINUS $5,436M on a company whose normal gross margin is 79%. Gilead reports "Acquired in-process research and development expenses" as a separate operating line — the 10-Q shows $107M and $253M on that line in the comparable quarters — so the vendor's presentation is wrong even though the operating-income total may be right. Second, the $10,394M operating loss becomes a $2,021M net loss — an $8.4 billion gap below the operating line that no document available to us explains. A tax benefit on an $11.2 billion charge would be roughly $2.4 billion at 21%. The remaining gap is unexplained. Separately, the consensus in the same payload was minus $7.26 per share — roughly minus $9.0 billion — against the payload's own minus $2,021M. We report all three figures, use none of them for valuation, and note that the 10-Q covering this quarter is not yet in the archive.

3. Balance sheet — correct at its date, and its date is the problem

Correction 1 — netDebt omits $3,041M of short-term investments. The vendor nets total debt against cash and equivalents only. The vendor reports $17,028M; the correct figure at the same date is $13,987M; an 18% overstatement.

Correction 2 — and this is the larger issue — the balance sheet predates roughly $11 billion of cash outflow, and there is NO June-quarter balance sheet in this archive. The most recent filed balance sheet is 2026-03-31, from the 10-Q filed 2026-05-07. Since then Gilead has paid $7.1 billion for Arcellx (closed April) and expected to close Tubulis (~$3.2 billion) and Ouro (~$1.7 billion, half reimbursed) in the second quarter. Against year-end cash and investments of $10,605M and roughly $2.4 billion a quarter of free cash flow, the arithmetic does not close without new borrowing. We state plainly: current net debt is materially above $13,987M, it is not observable from any document in this archive, and every leverage ratio on every screen for this name is currently wrong in the same direction.

Correction 3 — the vendor's enterprise value and net-debt fields disagree. enterpriseValueTTM of $182.493B less market capitalisation of $167.947B implies net debt of $14.546B, against the netDebt field's $17.028B. Neither matches our corrected $13.987B, and all three predate the acquisitions. We do not publish a rebuilt enterprise value or an EV multiple for this name, because the inputs required to do so honestly do not exist in this archive.

Correction 4 — the ratio block excludes the June quarter entirely. priceToEarningsDilutedRatioTTM of 18.429 at $135.27 implies trailing EPS of $7.34 and trailing net income of roughly $9.17 billion. The correct trailing figure including the June quarter is $5,235M of net income and $4.19 of EPS, giving a P/E of 32.3x. grossProfitMarginTTM of 79.35%, operatingProfitMarginTTM of 38.26% and netProfitMarginTTM of 30.99% are all computed on the same pre-June window. They are correct descriptions of the underlying business and incorrect descriptions of the trailing twelve months. We use them as business descriptors and reject them as trailing metrics, saying which is which at each use.

What is genuinely strong: interestCoverageRatioTTM of 11.3x; returnOnInvestedCapitalTTM of 20.3%; and goodwill plus intangibles of $25,292M — 42.9% of assets — which is high but is mostly the 2020 Immunomedics purchase behind Trodelvy rather than unallocated goodwill. priceToBookRatioTTM of 7.14x and debtToEquityRatioTTM of 0.943x rest on equity of $22,702M and are the reason the vendor's rating scores GILD 1 out of 5 on both price-to-book and debt-to-equity while giving it 5s on return on equity and return on assets — a composite that measures the same denominator twice in opposite directions, and one we reject.

4. Cash flow

A 32.1% free-cash-flow margin on capital expenditure of 1.9% of revenue is exceptional and it is the reason this business can afford what it is doing. freeCashFlowYieldTTM is 6.09% and price-to-free-cash-flow is 17.8x.

The dividend is the larger part of the return and it is well covered: $4,003M against $9,456M of free cash flow — 2.4x coverage — with dividendPayoutRatioTTM of 43.8% on the pre-June earnings base. Total shareholder yield is approximately 3.53% (2.38% dividend plus 1.14% buyback).

Note the tension the acquisitions create. Free cash flow of roughly $9.5 billion a year, less a $4.0 billion dividend and $1.9 billion of buyback, leaves about $3.5 billion a year of discretionary capacity — against $11.2 billion committed in a single year. The gap is the balance sheet, and the balance sheet is what we cannot see.

5. Valuation — priced in or room?

At $135.27 (market capitalisation $167.947B; no enterprise-value multiple is published for the reasons in Section 3):

est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year — arithmetically impossible — and both the EBIT and EBITDA rows are REJECTED. FY2027: ebitAvg $8.415B against netIncomeAvg $12.039B. FY2030: $10.019B against $16.265B. A company cannot earn more net income than earnings before interest and tax while paying both. The ratios are additionally fixed: ebitdaAvg / revenueAvg = 34.84% and ebitAvg / revenueAvg = 26.01% in all eight years, to four significant figures. Per the data contract, all forward valuation runs on epsAvg.

FY2026 EPS is negative and no multiple is computed on it. FY2029 (4 analysts) and FY2030 (4) are excluded. The FY2028 line carries 10 analysts and is used for the bull case with coverage disclosed. The base rests on FY2027 (16 analysts).

Peer context. The vendor peer set is a large-pharmaceutical grab-bag: Boston Scientific, Danaher and Stryker are device and life-science companies, not comparables. The defensible names are Merck ($317.3B), Amgen, Bristol-Myers Squibb, Pfizer, GSK, Sanofi and Novo Nordisk ($196.6B). Gilead at $167.9B is mid-sized within that group. No peer multiples are supplied, so no peer-multiple comparison is drawn. Note that ABT, also covered in this batch, lists GILD in its own peer set — a device and diagnostics group is not a therapeutics comparable and the two dives should not be read against each other on multiples.

5a. What today's price assumes (the inversion)

At $135.27 — 14.0x fiscal-2027 consensus — the price embeds:

• Earnings return to $9.653 in 2027 after a loss year. (Consensus; 16 analysts, range $9.03-$11.11.) That is 19.0% above the $8.113 delivered in 2025, which requires the base business to grow and the expensed assets not to require further spending. The nearest check is the third-quarter print, which is not dated in this file.
• The roughly $12 billion spent in 2026 produces revenue. (Filing figures; our framing.) This is the most fragile assumption in the price. Consensus revenue of $32.356B in 2027 and $34.216B in 2028 is +9.9% and +16.2% on 2025 — from a base that has compounded at 1.9% over five years. The increment can only be anito-cel, TUB-040 and gamgertamig.
• HIV, at 70.5% of revenue, keeps growing near 5.8%. (Filing figure.) There is no exclusivity-runway disclosure in the extracted filings, and this is the single largest unquantified structural risk in the name.
• Leverage returns to normal. (Our derivation.) Roughly $11.2 billion of cash committed against about $3.5 billion a year of discretionary free cash flow after the dividend and buyback. Debt must have risen and we cannot see by how much.
• The market keeps paying 13-15x on the forward year. (Our number.) At 11x FY2027E the stock is $106; at 17x it is $164. Large-pharmaceutical multiples are policy-sensitive and the knowledge base records "tariff pressure [driving] US investment pledges" including "Gilead +$11B/3000 jobs."

5b. The return bridge (why the multiple moves)

Because fiscal 2026 EPS is negative, the conventional bridge does not apply and we state the alternative. Expected return to end-2028 decomposes as: adjusted EPS going from a loss in 2026 to $9.653 (2027) and $10.555 (2028) — a sign change followed by 9.3% growth — + multiple drift (essentially HOLD, from 14.0x to 14.0x) + shareholder yield (3.53% a year, ~7% compounded).

Our base assumes the multiple HOLDS rather than expands or compresses, and the reason is that 14.0x forward is neither cheap nor expensive for a large-capitalisation pharmaceutical with a 70%-concentrated franchise, and we have no basis for predicting a re-rating in either direction. Almost all of the expected return is the earnings recovery, and that recovery is entirely mechanical — it is the absence of the 2026 acquisition charges, not an improvement in the business.

If the multiple simply HELD at 14.0x on FY2028E the price would be $148 — which is exactly our base; if it compressed to 11x on FY2027E it would be $106, below the 52-week low. The bull case at $190 is 16x the FY2028 consensus HIGH of $11.86 — a modest re-rating plus a pipeline beat, above the street's $180 high. The bear case at $108 is 12x the FY2027 consensus LOW of $9.03 and sits 2.1% below the 52-week low of $110.28.

5c. Variant perception (where we differ, what would surprise)

• We differ from every screen on the trailing multiple and it is a large gap. The vendor's priceToEarningsDilutedRatioTTM of 18.43x excludes the June quarter entirely — it implies trailing EPS of $7.34 against a properly computed $4.19, and the correct trailing P/E is 32.3x. The same window contamination affects the gross, operating and net margin fields. Anyone screening GILD today on trailing multiples is looking at a company that has not yet spent $11 billion.
• We think the accounting treatment is the most under-appreciated fact here and it cuts both ways. All three 2026 transactions are asset acquisitions, so the full purchase price is expensed. That makes 2026 look catastrophic and 2027 look like a recovery when nothing has changed operationally. It also means Gilead carries NO new goodwill or amortisable intangibles from $12 billion of spending — the balance sheet stays clean and the income statement takes the whole hit at once. That is a more honest accounting outcome than capitalising and amortising, and it is why the 2027 consensus of $9.653 is a cleaner number than it looks.
• We differ on where the risk sits. It is not the balance sheet in the abstract — interest coverage is 11.3x and free cash flow is $9.5 billion. It is that we cannot see the balance sheet at all: the most recent filed one is 2026-03-31 and roughly $11 billion has left since. Watchable event: the 10-Q for the June quarter, which is not in this archive. It is the single most important missing document on this name.
• We take the knowledge base's Arcellx claim seriously because it predates the acquisition. An independent biotech channel judged on 2025-12-15 that Gilead/Arcellx's anito-cel "has a cleaner safety profile than Carvykti; filing fourth-line in 2026, it becomes preferred where labeled and gains share as it expands earlier." That is a dated, specific, falsifiable claim about the asset Gilead bought four months later for $7.1 billion, and it is the only external validation of the transaction available to us.
• Positive surprise that would force a re-rate: the June 10-Q showing leverage lower than feared; an anito-cel approval or filing milestone; Cell Therapy revenue inflecting from its current −8.1%; or HIV growth above 6% with the exclusivity runway clarified.
• Negative surprise that would break the thesis: a further large expensed acquisition before the current ones produce revenue; an anito-cel clinical or regulatory setback; HIV growth decelerating below 3%; or a debt raise materially above expectations. At 14.0x forward with a 2.38% dividend there is some cushion, but a fifth acquisition would test it.

Synthos fair values

Anchors are multiples of the fiscal-2027 and fiscal-2028 consensus adjusted-EPS distributions (FY2027 mean $9.653, low $9.03, high $11.11 on 16 analysts; FY2028 mean $10.555, low $9.35, high $11.86 on 10 analysts — disclosed). Fiscal 2026 is excluded because consensus EPS is negative.

• Bear ~$108 — 12x the FY2027 consensus LOW of $9.03. Cross-check: 11.5x FY2028E mean; 2.1% BELOW the 52-week low of $110.28; 10.5x free cash flow. The scenario: the purchased pipeline disappoints, HIV growth decelerates, another large expensed deal arrives before the current ones deliver, and Gilead de-rates to a low-teens multiple appropriate to a single-franchise pharmaceutical. −20.2%.
• Base ~$148 — 15.3x the FY2027 consensus MEAN of $9.653. Cross-check: 14.0x FY2028E; 5.9x fiscal-2027 consensus revenue; 19.5x free cash flow. Sensitivity, stated openly: 13x FY2027E gives $126, 7.1% BELOW spot, and 17x gives $164, above the street's median. The whole answer lives in a 13-17x band. The scenario: 2027 earnings arrive roughly on consensus as the acquisition charges annualise out, HIV holds mid-single-digit growth, and the multiple stays where it is. +9.4%, plus about 7% of compounded shareholder yield.
• Bull ~$190 — 16x the FY2028 consensus HIGH of $11.86. Cross-check: 19.7x FY2027E; above the street's $180 high, which is disclosed as an outlier position. The scenario: anito-cel launches and takes share as the knowledge base's independent claim anticipates, Cell Therapy inflects from its −8.1%, TUB-040 and gamgertamig advance, and Gilead re-rates as a company with a pipeline rather than a franchise in runoff. +40.5%.

Base is 9.4% above spot; asymmetry roughly 2.0:1 to the upside (20.2% down, 40.5% up). What holds this at Watch is that the company reported hours before this dive with the release exhibit unavailable, the June balance sheet is not published, the payload's own ratio block excludes the quarter that just happened, and our base sits below the street.

6. Knowledge base — nine usable claims, and the useful ones are recent and specific

Raw hits: 17 (15 entity, 2 text). Used: 9. Discarded: 8. Searches covered Gilead, Biktarvy and Trodelvy.

The attribution problem — the third instance of the same defect in this batch. SEVEN claims dated 2018-2021 are filed under a channel slug named for the chief executive of a different pharmaceutical company and describe remdesivir's trial data, manufacturing scale-up and combination strategy in operator detail: "Remdesivir supply scaled ~40-fold to 2M+ treatment courses by year-end; cycle time cut from 12 to 6 months via 30+ partners"; "Remdesivir... cut recovery time and improved mortality in three randomized trials". These are management-adjacent COVID-era voice, six years stale, and describe a product now at $911M and falling 49.4%. All seven are excluded from the conviction pool. The same channel slug carries IBM's and Abbott's chief executives elsewhere in this batch, which confirms it as a systemic catch-all rather than a single error.

The nine claims used, and six are from one biotech channel across 2025-2026 — a concentration we disclose and weight as one source:

> 2025-12-15 · bullish · conviction 65 · thesis · entities: Gilead Sciences, Arcellx, Johnson & Johnson, Legend Biotech

> "Gilead/Arcellx's anito-cel has a cleaner safety profile than Carvykti; filing fourth-line in 2026, it becomes preferred where labeled and gains share as it expands earlier."

> The most valuable claim in the lane. It is dated FOUR MONTHS BEFORE Gilead paid $7.1 billion for Arcellx, it names the competing product, and it is falsifiable on a filing and a label. It is the only external validation of the transaction available to us.

> 2025-10-10 · bullish · conviction 70 · fact · entities: Arcus Biosciences, Merck, Gilead

> "Arcus's casdatifan (HIF-2α) phase 1 in RCC shows 12.2mo PFS vs Merck belzutifan's 5.6mo and 31% response, looking better than Welireg in a $5B market."

> 2025-10-10 · neutral · conviction 55 · entities: Arcus Biosciences, Gilead, AstraZeneca

> "Arcus/Gilead's FC-silent TIGIT (domvanalimab) persists where FC-active rivals were shelved; FC-silencing touted as key differentiator."

> 2025-08-28 · bullish · conviction 65 · thesis · entities: Gilead, AbbVie, J&J

> "Pharma increasingly favors smaller tuck-ins and licensing deals over picked-over, overvalued public M&A targets; Gilead active on the small side in inflammation."

> Partially superseded: the Tubulis and Ouro deals fit the description, but the $7.1 billion Arcellx purchase of a PUBLIC company does not.

> 2025-08-21 · neutral · conviction 55 · entities: Arcellx, Gilead

> "Firms moved trials to MRD endpoints per FDA guidance, but [a named official] opposes MRD; if CBER reverses it, trials lengthen materially, creating uncertainty."

> A specific regulatory risk to the exact asset Gilead has just bought, from an independent source, five months before the acquisition.

> 2025-08-21 · neutral · conviction 50 · fact · entities: Bristol Myers Squibb, Gilead, Teva

> "Tariff pressure drove US investment pledges (BMS $40B, Gilead +$11B/3000 jobs, Teva $30B), though much may be repackaged spend rather than truly additive."

> An $11 billion domestic-investment commitment that appears in no vendor field and is not addressed in the extracted filings.

> 2026-01-16 · bullish · conviction 68 · thesis · entities include Gilead

> "Generalists are rotating into pharma/biotech because it looks cheap versus an expensive AI-driven S&P; interest is broadening beyond mega-caps to Merck, Bristol, Pfizer, Gilead, Amgen."

> A flow claim rather than a company claim; noted as colour. It has aged reasonably — GILD is +19.5% over twelve months against the index's +24.3%, so it participated without leading.

> 2021-01-07 · bullish · conviction 80 · thesis · entities: Kyverna, Gilead, Vertex, Astellas

> "Cell/gene therapy is expanding beyond oncology into autoimmune, IBD, diabetes and Parkinson's; following these multi-year trends beats chasing single molecules."

> Five years old, from the excluded channel, but the CONTENT is a structural claim rather than management voice and it describes exactly what Gilead bought in 2026 — a CAR-T, an ADC and a T-cell engager for AUTOIMMUNE disease. Used at reduced weight for that reason and the frame disclosed.

> 2025-02-13 · bullish · conviction 50 · principle — "BARDA functions as an ATM for big pharma... Tamiflu orders quadrupled Gilead's stock." Historical colour, low weight.

Discards beyond the seven management-voice claims: a 2020 remdesivir hope from a named independent speaker at conviction 40 (superseded), and a 2025 claim about an AbbVie/Gilgamesh psychedelics transaction where the entity tag reads "Gilgamesh" and Gilead appears only in the thesis text — a near-homograph we caught and excluded.

Net conviction: positive-low. Breadth 9 raw, roughly 3 effective after de-concentrating the single biotech channel. NO claim carries a named speaker. No promoter-class voice appears. The lane's distinguishing feature is that it is unusually asset-specific — it discusses anito-cel, casdatifan and domvanalimab by name with comparative clinical data — and that its central claim predates the $7.1 billion transaction it describes by four months.

7. Data integrity — what we rejected from the vendor file and why

Nine findings. GILD's file is not defective so much as mid-update, and the timing is the problem.

1. The ratio block excludes the June quarter that the income statement includes — REJECTED as trailing metrics. priceToEarningsDilutedRatioTTM of 18.429 at $135.27 implies trailing EPS of $7.34 and net income of roughly $9.17 billion. Summing the four quarters in inc_q to 2026-06-30 gives $5,235M of net income and $4.19 of EPS — a trailing P/E of 32.3x. The same window contamination affects grossProfitMarginTTM (79.35%), operatingProfitMarginTTM (38.26%) and netProfitMarginTTM (30.99%), all of which describe the pre-acquisition business. We use those three as business descriptors and reject all four as trailing metrics, labelling which is which at each use.

2. The June-quarter charge is booked inside cost of revenue, producing a NEGATIVE gross profit — presentation REJECTED. grossProfit for the quarter ended 2026-06-30 is minus $5,436M on revenue of $7,803M, implying cost of revenue of $13,239M for a company whose normal gross margin is 79%. Gilead reports "Acquired in-process research and development expenses" as a separate operating line (the 10-Q shows $107M and $253M there in comparable periods). The operating-income total of minus $10,394M may be right; the line it sits on is not. No gross-margin figure from the June quarter appears in this dive.

3. An $8.4 billion gap between the June quarter's operating loss and its net loss is unexplained — flagged, not resolved. Operating income minus $10,394M; net income minus $2,021M. A tax benefit on an $11.2 billion charge would be roughly $2.4 billion at 21%. The remainder is unexplained by any document in this archive. Separately, the consensus in the same payload was minus $7.26 per share — roughly minus $9.0 billion — against the payload's own minus $2,021M, and earn_cal.epsActual remains null. Three figures, three different answers, and the 10-Q that would settle it is not yet published. We report all three and build on none.

4. netDebt omits $3,041M of short-term investments AND predates roughly $11 billion of cash outflow. The vendor reports $17,028M; the correct figure at 2025-12-31 is $13,987M. But the 10-Q discloses $7.1 billion paid for Arcellx (closed April 2026) and expected second-quarter closings for Tubulis (~$3.2 billion) and Ouro (~$1.7 billion, half reimbursed by Galapagos) against year-end cash and investments of $10,605M. There is NO June-quarter balance sheet in this archive. Current leverage is not observable and is materially higher than $13,987M. We do not estimate a figure.

5. No enterprise value or EV multiple is published for this name. The vendor's enterpriseValueTTM of $182.493B implies net debt of $14.546B, against its own netDebt field of $17.028B and our corrected $13.987M — three inconsistent figures, all predating the acquisitions. We decline to publish a rebuilt EV because the inputs to do so honestly do not exist in this archive, and we say so rather than producing a number with false precision.

6. est.netIncomeAvg EXCEEDS est.ebitAvg in every forward year — arithmetically impossible — and both rows are REJECTED. FY2027: $12.039B against $8.415B. FY2030: $16.265B against $10.019B. The ratios are additionally fixed at 34.84% and 26.01% of revenue in all eight years. All forward valuation runs on epsAvg.

7. epsdiluted is NULL in every income-statement row — annual and quarterly. All six annual and all nine quarterly rows. Every EPS figure here is computed. This is the fourth instance in this batch (with UNP, QCOM and DIS) and it now looks systematic across the vendor's coverage rather than name-specific.

8. seg_prod does not sum — an $3,400M (11.5%) shortfall. The five fiscal-2025 lines total $26,042M against reported revenue of $29,442M. The liver-disease franchise (hepatitis C, hepatitis B, primary biliary cholangitis) is not represented; the fiscal-2023 entry does contain a "Hepatitis B Virus / Hepatitis Delta Virus Product" line of $1,017M that is absent from fiscal 2025. The block is used for the lines it does contain, with the shortfall stated at every use.

9. lastDividend: 3.22 is the trailing-twelve-month total. Fiscal-2025 dividends paid were $4,003M on roughly 1,255M shares — about $3.19 per share.

seg_geo ties exactly — a clean check. United States $20,876M + Europe $5,064M + Other International $3,503M = $29,443M against reported revenue of $29,442M. Verified clean.

Share count — consistent. Market cap of $167,947,156,856 divided by $135.27 gives 1,241,566,000, against a June-quarter diluted weighted average of 1,243M — agreement to 0.1%. No share-class error.

quote.yearHigh/yearLow ($157.29 / $108.46) disagree with tech.hi52/lo52 ($155.80 / $110.28) by 1.0% and 1.7%. We use tech.

Vendor composite rating — mechanism disclosed, partially rejected. rating gives B+ / 3, with 1 out of 5 on both price-to-book and debt-to-equity against 5s on return on equity and return on assets. A rating that simultaneously awards the worst possible balance-sheet score and the best possible equity-return score is measuring the same $22,702M equity base twice in opposite directions. We reject the composite and use returnOnInvestedCapitalTTM of 20.3% as the honest return measure — noting that it too is computed on the pre-June window.

Non-equity tripwire — checked and passed. GILD is common stock, $0.001 par, NASDAQ Global Select Market, per the 8-K cover page. Price $135.27 is not par-like; beta 0.336 — the lowest in this batch, which is a genuine characteristic of a chronic-therapy franchise rather than a data artefact; the dividend is variable with a record of increases; volume 9.50M shares (~$1.28B of turnover); the 52-week band of $110.28 to $155.80 is a 41% range. This is common equity.

8. Technicals

• Price $135.27. −13.2% from the 52-week high of $155.80; +22.7% above the low of $110.28. Position within the annual range: 55th percentile.
• Moving averages, both below and both close: +4.2% above a 50-day of $129.84 and +2.5% above a 200-day of $131.92. The two averages are within 1.6% of each other — a range-bound stock that has just broken slightly higher.
• RSI 56.6 — neutral. MACD +0.71 — marginally positive.
• Maximum drawdown from peak over the trailing year: −13.2%. Shallow.
• Beta 0.336 — by far the lowest in this batch (the next is Abbott at 0.607). This is what a chronic-therapy franchise with 79% gross margins and 70% of revenue in a single indication looks like in a price series.
• Relative performance: 3-month +1.9% vs SPY +7.6%; 6-month −5.3% vs SPY +11.1%; 12-month +19.5% vs SPY +24.3% and QQQ +30.8%. The second-best twelve-month return in this batch, achieved with the lowest beta.
• Sentiment: 0 strong buy, 38 buy, 19 hold, 1 sell across 58 analysts. Consensus $158.67 (+17.3%), median $162.50, high $180 (+33.1%), low $120 (−11.3%).

Today's move and the entry

GILD closed 2026-08-04 at $135.27, up 3.14% or $4.12 from $131.15. It opened at $131.15 — exactly the previous close — traded $129.54 to $135.50, and closed within 0.2% of the day's high on 9.50M shares. The 8-K in this archive is dated 2026-08-04 and announces second-quarter results, so this is a post-print move, and the direction is consistent with the revenue beat visible in the payload ($7,803M against $7,399M).

The honest read: we do not have enough to underwrite this today, and the reason is documentary rather than analytical. The company reported hours ago; the release exhibit was not extracted; the earnings calendar has no actual; the June balance sheet does not exist in any filing available to us; and roughly $11 billion of cash has left the company since the last one that does. Three separate figures in the same payload disagree about what the quarter earned.

What would change the answer:

• The 10-Q for the June quarter, which would settle the balance sheet, the accounting presentation and the reconciliation gap in one document. This is the single most important pending item on this name.
• Or a price near $120 — roughly 12.4x FY2027E, and the street's own low target — which would put our $148 base at +23% with the bear case 10% below.

9. Insiders

The insider file contains no material open-market activity in either direction and is reported as neutral. There are no open-market purchases and no discretionary open-market sales of size. The days around a scheduled earnings release are a standard blackout period, which is the likely operative constraint, so the absence carries less information here than it would elsewhere.

The economically significant capital action is not in the Form 4s. Gilead committed roughly $12 billion to four acquisitions in a single year — against annual free cash flow of $9.456 billion, a $4.003 billion dividend and a $1.922 billion buyback. That is a decisive statement of priorities: the board has chosen pipeline over distribution, and the buyback at 1.14% of market capitalisation reflects it.

10. Verdict, kill-criteria and flip conditions

Watch.

What is genuinely strong: an HIV franchise at $20,752 million growing 5.8% with a 79.4% gross margin; free cash flow of $9,456 million on capital expenditure of $563 million — a 32.1% margin at 1.9% capital intensity; a dividend covered 2.4 times with a 2.38% yield; beta of 0.336, the lowest in this batch; interest coverage of 11.3x and 20.3% return on invested capital on the pre-acquisition base; June-quarter revenue of $7,803 million, up 10.2% and 5.5% ahead of consensus — the best revenue quarter in the file; three consecutive non-GAAP beats before it; a clean geographic reconciliation; and an asset-acquisition accounting treatment that expenses $12 billion immediately rather than parking it on the balance sheet as amortisable goodwill — which is the more honest outcome and leaves the balance sheet uncluttered.

What we are declining to underwrite today: a company that reported hours before this dive with the financial exhibit unextracted; no June-quarter balance sheet in any filing available to us, after roughly $11 billion of cash left the company; a payload whose ratio block excludes the quarter its income statement includes, making the screened trailing P/E of 18.4x wrong by 14 turns; an $8.4 billion unexplained gap between the quarter's operating loss and its net loss, and a consensus estimate in the same file that disagrees with the payload's own net income by roughly $7 billion; 70.5% of revenue in one indication and 70.9% in one country; Cell Therapy revenue down 8.1% in the franchise the $7.1 billion purchase is meant to revive; and a base fair value of $148 that is 9.4% above spot and 6.7% below the street's $158.67.

The distinction that matters. This is not a judgement that Gilead has done something wrong. The strategic logic is coherent, the accounting treatment is conservative, the revenue quarter was the best in the file, and the knowledge base contains an independent, dated, pre-acquisition endorsement of the largest asset bought. What we are saying is that the documentary basis for a position does not yet exist: the quarter that reframes this company was reported today and the filing that explains it has not been published. In the Synthos frame that is a Watch by construction, and the flip condition is a document rather than a price.

Pre-registered KILL criteria — what would take this to Avoid:

• A fifth large expensed acquisition before anito-cel, TUB-040 or gamgertamig produces revenue. Four in one year at roughly $12 billion is already the outer edge of what $3.5 billion a year of discretionary free cash flow supports.
• An anito-cel clinical, regulatory or label setback. The knowledge base independently flags endpoint risk — "if CBER reverses [MRD guidance], trials lengthen materially" — and $7.1 billion rests on this asset.
• HIV revenue growth below 3%, or any disclosure of a nearer-than-expected exclusivity event. It is 70.5% of the company.
• Net debt disclosed above roughly $25 billion in the June 10-Q without a corresponding EBITDA step.
• Cell Therapy revenue declining a further 10%, which would mean the franchise anito-cel joins is contracting faster than the new asset can fill.

Pre-registered FLIP conditions — what would take this to Buy — Tactical:

• The June 10-Q published showing net debt below approximately $22 billion with the reconciliation gap explained. This is the most likely path to a Buy and it is a document, not a price.
• Or a price near $120 — roughly 12.4x FY2027E and the street's own low target — with the picture unchanged. At $120 the base case is +23% and the bear case is 10% below.
• An anito-cel regulatory filing or approval milestone, which would begin converting $7.1 billion of expensed purchase price into a revenue line.
• Cell Therapy revenue inflecting positive from its current −8.1%.
• HIV growth sustained above 6% with an exclusivity runway disclosed.

Where GILD fits in the Synthos Framework Portfolio. The large-capitalisation therapeutics sleeve, at 0% today with a 1.5-2% target on the June 10-Q or a fill near $120. On batch overlap: GILD and ABT are the two healthcare names here and they are not the same exposure — Gilead is a concentrated therapeutics franchise with patent-cliff risk and no leverage problem; Abbott is a diversified device and diagnostics group that has just levered to 2.7x. Holding both is defensible. Gilead's 0.336 beta makes it the single most defensive holding available in this batch. Logged as a tracked Synthos call (Watch) as of 2026-08-04 at $135.27, with the explicit note that the verdict is documentation-driven and is scheduled for review when the June 10-Q is filed.

Single biggest risk: roughly $12 billion of purchased pipeline, expensed in full, on assets that have not yet generated a dollar of revenue. Arcellx cost $7.1 billion in cash plus a contingent value right; Tubulis approximately $3.2 billion; Ouro approximately $1.7 billion. The 10-K already records a $3.8 billion CymaBay charge in 2024 whose payoff is not yet visible in the revenue line, and the 2020 Immunomedics purchase produced Trodelvy at $1.4 billion of annual sales. Meanwhile the franchise these assets join — Cell Therapy — is currently shrinking 8.1%. Consensus revenue of $34.216 billion in 2028 is 16.2% above 2025 from a base that has compounded at 1.9% over five years, and the entire increment is these assets. If they do not deliver, Gilead is a 70%-HIV company at 14x forward earnings with a smaller cash cushion than it started the year with.

Most fragile assumption in the price: that fiscal-2027 EPS of $9.653 arrives. It is 19.0% above the $8.113 delivered in 2025 and it depends on the acquisition charges simply not recurring. The mechanical part of that is sound — asset-acquisition charges are one-off by construction. The fragile part is that consensus also has revenue rising 9.9% by 2027 from a base growing 2.4%, which requires the new assets to contribute within eighteen months of purchase. If 2027 revenue lands at $31.0 billion rather than $32.4 billion and EPS at $8.60 rather than $9.65, then at 15.3x the stock is $132 — three percent below today's price.

Provenance & disclosures

• Traceability: 17 raw knowledge-base hits (15 entity, 2 text); 9 used; 8 discarded; breadth 9 raw and roughly 3 effective after de-concentration. The attribution finding — the THIRD instance of the same defect in this batch: SEVEN claims dated 2018-2021 are filed under a channel slug named for the chief executive of a DIFFERENT pharmaceutical company and describe remdesivir's randomised-trial results, 40-fold manufacturing scale-up and combination strategy in operator detail. They are management-adjacent COVID-era voice, six years stale, describe a product now at $911M and falling 49.4%, and are excluded from the conviction pool. The same slug carries IBM's and Abbott's chief executives elsewhere in this batch, confirming it as a systemic catch-all rather than a single error. A further concentration: SIX of the nine used claims come from ONE biotech channel across 2025-2026 and are weighted as one source. NOT ONE claim in the lane carries a named speaker. The single most valuable claim is dated 2025-12-15 — FOUR MONTHS BEFORE Gilead paid $7.1 billion for Arcellx — and states that anito-cel "has a cleaner safety profile than Carvykti; filing fourth-line in 2026, it becomes preferred where labeled and gains share as it expands earlier." It is the only external validation of the largest transaction in this dive. A companion claim from the same channel flags the specific regulatory risk ("if CBER reverses [MRD guidance], trials lengthen materially"), and a third records an $11 billion Gilead US investment pledge that appears in no vendor field and is not addressed in the extracted filings. A near-homograph was caught and excluded: a 2025-10-10 claim entity-tagged "Gilgamesh" (an AbbVie psychedelics target) mentions Gilead only in its thesis text. No promoter-class voice appears. Searches: Gilead, Biktarvy, Trodelvy. All quotes are verbatim from the stored claim text.
• Data as-of: fundamentals — annual statements to 2025-12-31 from the vendor payload; quarterly income statement to 2026-06-30 (the quarter reported TODAY); most recent filing-verified financials 2026-03-31, from the 10-Q filed 2026-05-07 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873600 = 2026-08-04T20:00:00Z ($135.27, +3.14%; 50-DMA $129.84; 200-DMA $131.92; RSI 56.6; MACD +0.71; beta 0.336) · knowledge-base claims 2026-08-04. Fiscal note: Gilead's fiscal year is the calendar year. All figures come from the Synthos vendor data file for GILD or from the SEC filings in the GILD archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents and its critical gap: 10-K filed 2026-02-24 (fiscal 2025); 10-Q filed 2025-11-07 (September 2025 quarter); 10-Q filed 2026-05-07 (March 2026 quarter); 8-K filed 2026-05-07; 8-K filed 2026-05-20; 8-K filed 2026-08-04 — TODAY — announcing second-quarter 2026 results. All carry preserved financial tables (tables: true). THE GAP: the 2026-08-04 8-K is present as the cover document ONLY. Exhibit 99.1, which carries the financials, was NOT extracted and the file contains no numbers. There is consequently NO filing-verified figure of any kind for the June 2026 quarter, and NO balance sheet later than 2026-03-31 — a period across which roughly $11 billion of cash left the company for acquisitions. Every June-quarter figure in this dive comes from the vendor payload and is labelled as such, and the reconciliation problems in Section 7 are a direct consequence of that gap.
• Where the filings contradicted or corrected the vendor (detailed in Section 7): the entire trailing-ratio block rejected as trailing metrics after priceToEarningsDilutedRatioTTM of 18.429 was shown to imply $7.34 of trailing EPS against a properly computed $4.19 — i.e. the block is calculated on a window ending 2026-03-31 and excludes the June quarter the same file's income statement contains, with grossProfitMarginTTM (79.35%), operatingProfitMarginTTM (38.26%) and netProfitMarginTTM (30.99%) similarly contaminated and retained only as descriptors of the pre-acquisition business; the June quarter's charge shown to be booked inside cost of revenue, producing a negative gross profit of minus $5,436M where the 10-Q reports "Acquired in-process research and development expenses" as a separate operating line; an $8.4 billion unexplained gap identified between the quarter's $10,394M operating loss and its $2,021M net loss, alongside a consensus in the same file (minus $7.26 per share) that disagrees with the payload's own net income by roughly $7 billion; the entire 2026 acquisition programme sourced from the 10-Q because NO vendor field contains it — Arcellx $7.1B cash closed April 2026 plus a $5-per-share contingent value right on $6.0B of cumulative anito-cel sales through 2029, Tubulis ~$3.2B upfront plus $1.9B of milestones, Ouro Medicines ~$1.7B upfront plus $500M with Galapagos funding half, all three stated by the filing to be accounted for as ASSET ACQUISITIONS and therefore expensed; the fiscal-2024 operating-income collapse explained from the 10-K as "a $3.8 billion acquired in-process research and development expense related to the acquisition of CymaBay"; netDebt corrected from $17,028M to $13,987M for the omission of $3,041M of short-term investments, and then flagged as unusable because roughly $11 billion has left since and no June balance sheet exists; NO rebuilt enterprise value published at all, because the vendor's own EV ($182.493B) implies net debt of $14.546B against its own netDebt field of $17.028B and our corrected $13.987B — three inconsistent figures, all pre-acquisition; est.netIncomeAvg rejected as arithmetically impossible (above ebitAvg in every forward year) and on a fixed-ratio signature (34.84% and 26.01% of revenue in all eight years); epsdiluted found NULL in all fifteen rows with every EPS figure computed; seg_prod failing to sum by $3,400M (11.5%) with the liver-disease franchise absent; and lastDividend of $3.22 identified as the TTM total against roughly $3.19 of fiscal-2025 dividends per share. Where vendor and filing AGREED — worth recording: seg_geo summing exactly to reported revenue ($29,443M against $29,442M); the share count (implied 1,241,566,000 against a June diluted count of 1,243M); and the quarterly revenue series against the earnings calendar, including the June quarter's $7,803M against a $7,399M estimate.
• Basis note: est.epsAvg is a non-GAAP EPS measure — the earn_cal actuals ($2.47, $1.86, $2.03) differ from our computed GAAP diluted EPS ($2.43, $1.74, $1.61) in every quarter. The fiscal-2026 consensus of MINUS $0.754 evidently INCLUDES the asset-acquisition charges, since a non-GAAP measure excluding them would be strongly positive — which means the "non-GAAP" label is being applied inconsistently across years in this block, and we flag that rather than resolving it. All forward valuation uses epsAvg for fiscal 2027 and 2028 only; fiscal 2026 is excluded because the figure is negative and its basis is ambiguous. The trailing GAAP multiple of 32.3x is computed by us and is shown once.
• Estimate coverage: 16 analysts on FY2026 EPS and 19 on revenue; 16 and 18 on FY2027 — the anchor year; 10 and 21 on FY2028; 4 and 10 on FY2029; 4 and 10 on FY2030 — FY2029 and FY2030 are excluded entirely. The base case rests on FY2027 and the bull case on FY2028, with the 10-analyst coverage of the latter disclosed at each use.
• Peer note: the vendor peer set mixes therapeutics with devices and life sciences — Boston Scientific, Danaher and Stryker are not comparables. The defensible names are Merck ($317.3B), Amgen, Bristol-Myers Squibb, Pfizer, GSK, Sanofi and Novo Nordisk ($196.6B). ABT, also covered in this batch, lists GILD in its own peer set; a device and diagnostics group is not a therapeutics comparable and the two dives should not be read against each other on multiples. No peer multiples are supplied, so no peer-multiple comparison is drawn.
• Fair-value caveat: the $108 / $148 / $190 anchors are multiples of the fiscal-2027 and fiscal-2028 consensus EPS distributions — 12x the FY2027 low of $9.03, 15.3x the FY2027 mean of $9.653 (16 analysts), and 16x the FY2028 high of $11.86 (10 analysts, disclosed). Fiscal 2026 is excluded entirely because consensus EPS is negative. Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 13x FY2027E gives $126, 7.1% BELOW spot; 17x gives $164, above the street's median. The whole answer lives in a 13-17x band. The bear case sits 2.1% below the 52-week low of $110.28 and 10% below the street's own $120 low; the bull case sits above the street's $180 high and is disclosed as an outlier position. No enterprise-value cross-check is offered, for the reasons in Section 3.
• Timing — the dominant input to this verdict: Gilead reported second-quarter 2026 results TODAY, 2026-08-04. The 8-K in this archive carries that date and announces the release; Exhibit 99.1 was not extracted and no company-stated figure is available. The vendor payload carries a June-quarter income statement showing revenue of $7,803M — a 5.5% beat and +10.2% year on year — and an operating loss of $10,394M; earn_cal.epsActual remains null. The 10-Q covering the quarter, which would settle the balance sheet and the reconciliation gap, is not yet in the archive. The verdict is explicitly documentation-driven and is scheduled for review when that filing appears. The stock rose 3.14% on the day, closing within 0.2% of its high.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.