Gilead Sciences, Inc. (NASDAQ: GILD) — A Better HIV Engine, a Much Larger Capital-Allocation Bill

Prepared by: Independent equity research Report date: 2026-08-13 Analyst: Independent analyst Price reference: $135.87 (2026-08-12 close); market cap approximately $168.47B; current enterprise value approximately $191.54B; 1,239,955,155 shares outstanding at 2026-07-31

⚡ Claude’s Take

This block is the author’s subjective opinion and is provided for general information only. It is not investment advice. The analytical body that follows takes no position and discusses valuation only through embedded expectations and scenarios.

Verdict — HOLD / quality-franchise-at-a-fair-price. The HIV evidence improved, but the capital-allocation evidence deteriorated by almost the same amount. I would build aggressively only below approximately $115–120 and would not chase above approximately $150.

Tag: “The best HIV franchise in the cohort is now carrying an oncology shopping bill big enough to matter.”

Gilead remains the rare large-pharma name whose core asset is unambiguously excellent: a dominant US HIV franchise, non-GAAP product gross margin of 86.9% in Q2, very low capital intensity, and a long record of retaining virally suppressed patients. The August print strengthened that case. HIV sales rose 12%, the company lifted full-year HIV growth guidance to 9–10%, quarterly PrEP sales exceeded $1B, and Yeztugo reached $232M with more than 70% of eligible patients returning for their six-month injection. Yet the same quarter made the other half of the thesis materially worse. Gilead spent $11.3B of cash on acquisitions in the first half, recorded $11.2B of acquired IPR&D expense and a further $1.75B Trodelvy-related impairment, and ended June with approximately $23.1B of net debt versus approximately $14.3B at the prior report date. The moat strengthened; the evidence that management can reinvest outside it weakened.

At $135.87 the shares trade at approximately 15.8x the midpoint of clean 2026 earnings power after reversing the $9.08-per-share transaction charge. Conventional trailing FCF yield is 7.68%, but it falls to approximately 6.53% after charging stock compensation and an explicit $1B annual external-science reserve. A reverse DCF asks for approximately 3–4% five-year owner-FCF growth with a 14x terminal cash multiple. That apparently modest hurdle assumes the company can bridge Biktarvy’s 2028 price reset and expected 2036 adult generic-entry boundary while avoiding another cycle of oncology impairments. The current enterprise value of approximately $191.5B also stands well above central no-growth earnings-power value. Investors are paying a defensible premium for customer captivity and the patent ladder—not getting the pipeline for free.

What keeps me from being more constructive is still the same unanswered question: is lenacapavir economically additive, or is it mainly a more convenient way to migrate Gilead’s own patients? Management disclosed that Yeztugo switches are sourced roughly one-third from Apretude, one-third from Descovy, and one-third from generics, while treatment-naive starts are improving but remain unquantified. That is better than the prior fear of two-thirds direct self-cannibalization, but it is not proof of category expansion. BIC/LEN is itself positioned primarily as a switch regimen, and the Merck-partnered islatravir/lenacapavir program will carry weaker economics than wholly owned Biktarvy. Framing: quality compounder at a fair price, not contrarian value. Conviction: medium. Flips bullish if total HIV compounds through the 2028 Medicare reset, PrEP-treated lives expand with a rising naive share, and the new oncology assets begin earning attractive incremental returns. Flips bearish if the negotiated Biktarvy price cut is severe, long-acting products mostly recycle existing Gilead revenue, and one more large oncology program fails or is impaired.

Five-Year Price-Action Event Map

The share-price record is best read as a sequence of thesis revisions rather than a smooth compounding story. Daily prices are split-adjusted closes from Alpha Vantage/AZI through 2026-08-12; event interpretations are tied to company filings and releases.

Changes Since the 2026-06-12 Report

The update is not a new story; it is a more polarized version of the old one. The table below marks what changed the underwriting rather than repeating durable background.

Bottom line: the prior balanced underwriting survives. Better HIV execution raises the base-case operating path, but added leverage, transaction cost, and impairment evidence offset much of the improvement.

The next update should be evidence-led: regulatory decisions change probability, while commercial mix, patient additions, cash returns, and debt reduction change intrinsic economics.

1. Executive Summary

Gilead Sciences is an approximately $30B-revenue, asset-light, very-high-margin biopharmaceutical company whose economic center of gravity is one dominant therapeutic franchise: HIV. In Q2 2026, HIV generated $5.69B, or nearly three-quarters of product revenue, and Biktarvy alone produced $3.77B. Around that fortress sit four secondary businesses in varying states of health: a declining HCV book partly re-based by fast-growing Livdelzi; an almost exhausted COVID tail; a shrinking cell-therapy unit; and a growing but fiercely contested oncology ADC in Trodelvy. The Q2 10-Q and earnings exhibit are the principal current financial sources for this report (SEC 10-Q, filed 2026-08-06; Q2 earnings exhibit, filed 2026-08-04).

The investment is a study in contrasts. Q2 base-product sales excluding Veklury grew roughly 10%, HIV grew 12%, and management raised full-year base-product guidance to $29.8–30.1B and HIV growth guidance to 9–10%. The core therefore looks healthier than the five-year consolidated topline suggests. Yet Gilead is fighting two known, dated threats: Biktarvy’s 2028 Medicare negotiated price and its early-2030s patent exposure. Its answer is a family of lenacapavir regimens. Twice-yearly Yeztugo is already commercial; once-daily BIC/LEN has an Aug. 27 PDUFA; a once-weekly oral PrEP application has a Feb. 2, 2027 PDUFA; and Merck-partnered islatravir/lenacapavir is a potential weekly treatment option. The strategy is coherent, but the economics differ: wholly owned prevention can expand the market, while switch regimens and partnered therapies may preserve patients at lower incremental value.

Capital allocation is competent on shareholder distributions but negative on recent M&A evidence. First-half operating cash flow of $6.12B and free cash flow of approximately $5.86B comfortably funded $2.07B of dividends, but acquisitions consumed $11.32B. Gilead also repurchased $774M of stock while issuing debt and drawing down cash. At June 30 it held $3.18B of cash and marketable securities against $26.25B of debt. The result is not a solvency problem; it is a lower margin of safety and a much higher hurdle for Arcellx, Tubulis, Ouro, and the remaining Immunomedics asset base.

On valuation, GILD is fair versus large-cap pharma at approximately 15.8x clean 2026 earnings, 12.5x clean EBITDA, and a normalized owner-FCF yield near 6.5%. A live own-history percentile was unavailable because the authenticated fundamentals endpoint could not be refreshed; the prior June reading—95th percentile on price-to-sales—should therefore be treated as dated context, not a current fact. The thesis still swings on one question: does lenacapavir expand treated populations and preserve economics, or mainly migrate existing patients?

2. Business Overview

What Gilead does. Founded 1987, IPO 1992, headquartered in Foster City, California, with ~17,000 employees, Gilead discovers, develops and commercializes prescription medicines for HIV, viral hepatitis, liver disease, oncology and inflammation. It is a research-and-marketing organization wrapped around a small number of blockbuster molecules: it manufactures little physical plant (capex <2% of sales) and earns its returns on intellectual property, clinical-development capability and a global commercial infrastructure. Revenue is overwhelmingly US-weighted and concentrated in a handful of branded products.

Revenue segmentation (FY2025 product sales $28.9B; total revenue $29.4B incl. $0.5B royalty/contract). (FACT — FY2025 10-K MD&A product table.)

Current operating snapshot (Q2 2026). The quarterly mix shows why the update is simultaneously stronger and more concentrated.

Revenue quality improved because base products—not Veklury—drove the quarter. However, Biktarvy plus Descovy and Yeztugo accounted for most incremental dollars. There is little evidence that diversification has reduced economic dependence on HIV; in fact, the mix has become more dependent on it as cell therapy and Veklury contracted.

End markets & customers. Gilead sells to specialty distributors, pharmacies, hospitals, government programs (Medicaid, Medicare Part D, the VA, 340B covered entities) and public-health buyers (e.g., Ryan White). Demand is disease-driven and largely non-cyclical; pricing is an agency relationship — physicians prescribe, payers reimburse, patients bear the clinical risk of switching — which historically conferred durable pricing power now being partially capped by the IRA (Section 3).

How it makes money, and how durable each dollar is. (INTERPRETATION.) The quality of Gilead’s revenue varies enormously by line:

• HIV treatment (~$17B+) is the crown jewel: chronic, lifelong daily therapy where virally-suppressed patients almost never switch (rebound/resistance risk). The most durable, recurring, moat-protected revenue in the company.
• HIV prevention / PrEP (Descovy + Yeztugo, ~$3.5B and growing) is recurring but more elastic (adherence-driven) and is the genuine net-new growth category.
• HCV (Epclusa, Harvoni) is the curative-drug paradox in action: the medicine cures in 8–12 weeks, so every sale removes a future customer. This is the self-liquidating annuity that sank the 2015–2020 Gilead story and is now a small residual.
• Veklury is an episodic pandemic tail, nearly exhausted.
• Cell therapy (Kite) is one-and-done and, critically, declining.
• Trodelvy is recurring-until-progression but in a hyper-competitive class.

Verdict: A high-margin, cash-generative franchise whose durable core (HIV treatment plus the PrEP build, ~$20B+) is excellent, but whose diversification lines are mostly declining or unproven. The single most important business fact is concentration: one product is nearly half of sales and one franchise is roughly 72%.

3. Industry Dynamics

HIV therapeutics — a structurally attractive oligopoly. The US market is chronic, reimbursed, guideline-driven, and concentrated among Gilead, ViiV Healthcare, and Merck. Entry requires large, long clinical programs against an unusually high efficacy, safety, resistance, and tolerability bar. The moat is not that Biktarvy is uniquely efficacious forever: ViiV’s Dovato has demonstrated non-inferiority in relevant switch settings, and ViiV’s long-acting portfolio continues to grow. The moat is the combination of customer captivity, prescriber familiarity, guideline placement, contracting scale, and a deep product ladder. Merck is both competitor and partner through islatravir/lenacapavir. On Greenwald’s test, stable share and scale reinforce each other; patent rights alone would be more transient.

The IRA is the central structural overhang—concrete in timing, unknown in magnitude. In January 2026, HHS selected Biktarvy for Medicare negotiation, with the Maximum Fair Price effective Jan. 1, 2028 and scheduled for publication by Nov. 30, 2026. Management expects the price to be substantially lower and warns of Medicaid and 340B knock-ons, but no defensible net-revenue impact is yet available. The gross CMS spend figure is not company revenue. The correct underwriting is a sensitivity, not false precision. The policy taxes one pillar of the moat—agency pricing—without eliminating customer captivity or clinical scale.

Oncology — the opposite of HIV. The ADC and cell-therapy spaces Gilead is pivoting into are crowded, fast-moving, capital-intensive and low-barrier on durable advantage. In ADCs, AstraZeneca/Daiichi Sankyo (Enhertu; Datroway, which carries a near-term TNBC PDUFA directly threatening Trodelvy), Merck and Pfizer/Seagen are all racing. In myeloma CAR-T, J&J/Legend (Carvykti) and BMS (Abecma) are entrenched incumbents that Gilead’s anito-cel must displace. A PD-1/VEGF bispecific wave (Summit/Akeso) is reshaping solid-tumor standard-of-care beneath Trodelvy. In Marathon capital-cycle terms, this is a “capital-pouring-in, returns-competed-away” zone — heavy R&D and M&A, no supply discipline, rapid obsolescence.

The capital-cycle read. (INTERPRETATION — Marathon lens.) HIV has a favorable supply side (consolidated, disciplined, high-barrier, no new capacity), which justifies premium economics and suspends mean-reversion. Oncology has an unfavorable supply side (capital flooding in through biotech M&A — Gilead itself put ~$11.5B of upfronts into the space in a single quarter). Gilead is redeploying cash from a favorable-supply business into an unfavorable-supply one — the classic asset-growth-anomaly warning that predicts low forward returns on the incremental oncology dollars.

Verdict: HIV and specialty virology form a structurally attractive industry, but with a material regulatory tax. Oncology is hyper-competitive and offers Gilead no demonstrated structural edge. Most current revenue sits in attractive markets while incremental capital is flowing into much harder ones.

4. Competitive Position

The HIV moat is a genuine Greenwald “triple” reinforced by agency pricing — the strongest configuration, and it is financially evidenced. (INTERPRETATION, grounded in the FACTs below.)

1. Intangibles / proprietary technology (primary). Composition-of-matter and formulation patents on the TAF backbone, bictegravir, and lenacapavir. Patents are Greenwald’s most transient advantage — but Gilead runs a patent-ladder / evergreening machine: each new best-in-class regimen (TAF → Biktarvy → the lenacapavir era) re-starts the exclusivity clock before the prior one cliffs, and the bar is now so high that competition has shifted from efficacy to dosing-interval convenience, where Gilead also leads.
2. Switching costs / customer captivity (demand). Virally suppressed patients and prescribers avoid unnecessary regimen changes, while guidelines and formulary access reinforce familiarity. This captivity is meaningful but not absolute: the entire BIC/LEN and islatravir/LEN strategy itself assumes clinically stable patients will switch for convenience or durability.
3. Economies of scale + R&D scale (supply). Gilead runs the deepest HIV R&D engine in the world (60 abstracts at CROI 2026; a seven-launch-by-2033 HIV pipeline). Fixed R&D and commercial infrastructure amortized over >50% share is a per-unit cost/innovation advantage rivals cannot match — the Greenwald economies-of-scale-plus-captivity combination, the most durable advantage there is.

The tests confirm it. (FACT.) On the market-share-stability test, Biktarvy’s US treatment share has risen year-over-year every quarter since its 2018 launch, now >52%; total HIV grew $18.2B → $19.6B → $20.8B across 2023–25. Share that is stable-to-rising over 5–8 years is the textbook signature of a real, durable moat. On the ROIC test, the franchise earns ~87% gross margin, ~47% operating margin and well above the 15–25% bar (capex <2% of sales). The moat is real precisely where the profitability is real: HIV.

Where the moat actually resides. It is tempting to label a blockbuster’s patent as the moat, but that understates the durable sources and overstates the temporary one. Patents exclude identical molecules for a defined period; they do not prevent a rival from developing a different non-inferior regimen. Gilead’s more durable advantage is the installed base of suppressed patients, prescriber comfort with its safety and interaction profile, formulary contracting, and the ability to fund multiple overlapping lifecycle programs. These mechanisms make a rival’s commercial entry harder even after it clears the scientific bar. They also let Gilead cannibalize itself before a generic does. The cost is that self-migration can preserve the franchise without generating much incremental revenue.

ViiV is the best falsification test. Dovato is a credible two-drug oral competitor and its long-acting injectables prove that convenience can change behavior. Apretude was the source for roughly one-third of disclosed Yeztugo switchers, showing that Gilead can win share from a capable rival. But ViiV’s continued HIV and long-acting growth also shows that Gilead does not possess monopoly economics. The market structure is a disciplined oligopoly where multiple participants can earn returns. Gilead’s advantage is breadth and installed scale, not permanent efficacy superiority.

Merck creates a second test. Islatravir/lenacapavir can help Gilead defend weekly treatment leadership, but the partnership means value is shared. A patient who moves from wholly owned Biktarvy to a partnered regimen may count as a clinical and retention success while contributing less operating profit. Investors should therefore monitor product-level net sales and margin, not simply the number of approvals or patients on a lenacapavir backbone.

Pressure-test: durable to 2040, or a patent cliff in disguise? (INTERPRETATION — the central debate.) Orange Book listings show Biktarvy patents through November 2036 and adult generic settlements permitting entry as early as Apr. 1, 2036; Yeztugo listings extend through November 2040. That makes a single 2033 “cliff” the wrong mental model. The real sequence is a 2028 government-price reset, continued Biktarvy cash generation, and migration into lenacapavir regimens before settled generic entry. The transition is feasible because Gilead has executed TAF-to-Biktarvy switching before, but it may preserve customers without preserving the same economics. Patent runway supports durability; it does not prove additive growth.

Oncology — little or no durable moat. (INTERPRETATION.) Kite/cell therapy has no moat: revenue is declining under in-class and out-of-class (bispecific) competition, and its only edge — manufacturing/vein-to-vein logistics — is operational effectiveness, not a Greenwald barrier. Trodelvy is a real, differentiated product (leading 2L mTNBC ADC, NCCN Category 1) but faces a direct PDUFA threat from AZ/Daiichi’s Datroway and the broader ADC arms race; it is patent-protected and growing, but Gilead is sub-scale against the AZ/Daiichi scale leaders. Anito-cel (just acquired via Arcellx) is a potential best-in-disease BCMA CAR-T with a differentiated clean-neurotoxicity profile, but it must displace entrenched Carvykti/Abecma in a market where Gilead’s own Kite franchise is shrinking. Per Greenwald, differentiation without barriers earns average returns; per Marathon, asset growth into a competed-away space predicts low forward returns.

Verdict: Gilead has a durable, multi-source competitive advantage in HIV, with an expected adult Biktarvy generic-entry boundary in 2036 conditional on settlement terms and lifecycle execution. The IRA taxes but does not break the moat. Oncology has no demonstrated durable moat.

5. Growth History and Forward Opportunities

History: a flat topline hiding a violent mix shift. (FACT — FY2025 10-K.) Total revenue ran $27.3B (2021) → $27.3B (2022) → $27.1B (2023) → $28.8B (2024) → $29.4B (2025): ~1.9% CAGR, essentially flat. That flat line is the algebraic sum of three offsetting tides:

1. HIV grew steadily — ~$16–17B (2021–22) → $18.2B (2023) → $19.6B (2024, +8%) → $20.8B (2025, +6%), with Biktarvy compounding $11.85B → $13.42B → $14.33B. This is the durable engine.
2. HCV/Veklury collapsed — Veklury fell from a ~$5.6B pandemic peak (2021) to $0.91B (2025), a ~$5B tailwind that fully reversed and masked the HIV growth; HCV self-liquidated to a small residual.
3. Oncology stagnated — cell therapy −7% in FY2025 (and declining), Trodelvy +6% (accelerating later). Net oncology −2%.

The honest read: stripping the Veklury runoff, the base business compounded mid-single-digits. But the quality of historical growth is narrow — almost entirely one franchise (HIV), increasingly one product (Biktarvy). A decade of diversification spend (Kite, Immunomedics, CymaBay) has not yet produced a second growth pillar; only Trodelvy and Livdelzi are genuinely additive.

Forward drivers. (FACT on guidance/pipeline; INTERPRETATION on sizing.)

• Yeztugo / lenacapavir PrEP—the new-category bet. Q2 sales reached $232M, 40% above Q1, and management retained approximately $1B full-year guidance. More than 70% of eligible patients returned for their second six-month injection. Among switchers, roughly one-third came from Apretude, one-third from Descovy, and one-third from generics. Only the Descovy bucket is direct Gilead cannibalization; Apretude and generic switches add revenue to Gilead without necessarily expanding the PrEP category. Treatment-naive starts are improving slowly but remain undisclosed. The analysis must therefore keep “share gain,” “category growth,” and “new-to-PrEP” separate.
• A regimen ladder through the late 2030s. BIC/LEN has an Aug. 27, 2026 PDUFA and is supported by switch trials, making retention the primary near-term use case. Once-weekly oral Yeztugo for prevention has a Feb. 2, 2027 PDUFA. Islatravir/lenacapavir is Merck-partnered and thus carries less attractive economics per migrated patient than a wholly owned regimen. GS-3242 has demonstrated potential for a four-month injectable interval; six-month dosing remains an aspiration awaiting further clinical evidence.
• Oncology / other. Trodelvy grew 26% in Q2 and its first-line mTNBC addressable population is roughly twice the second-line population, with longer treatment duration. That is a genuine post-June improvement. It is offset by EVOKE-03 failure, a $1.75B impairment, Datroway overlap, and cell-therapy sales down 14%. Anito-cel has a Dec. 23, 2026 PDUFA, but approval is only the first gate; commercial success requires differentiation from Carvykti, whose sales were growing rapidly in the latest peer cross-read. Livdelzi at $167M quarterly is the clearest non-HIV launch success.

Can lenacapavir offset the 2028 price reset and later generic entry? (INTERPRETATION.) The company must preserve a meaningful slice of a roughly $15B annualized Biktarvy product while growing prevention. The designated replacements are lenacapavir-based PrEP, BIC/LEN, partnered weekly oral treatment, and future injectables. The TAF-to-Biktarvy precedent shows Gilead can execute self-cannibalizing renewal. What remains unproven is whether the next migration also expands patients and maintains profit per patient after government pricing and partnership economics.

Milestone bridge and what each milestone proves. Regulatory approval is necessary but insufficient. BIC/LEN approval would validate the fixed-dose dossier and create another switch option; it would not show net-new demand. Weekly oral Yeztugo approval would validate a differentiated dosing interval, but its launch mix could be heavily cannibalistic to Descovy. Anito-cel approval would establish that the acquired asset cleared FDA review; it would not establish manufacturing reliability at commercial scale or share capture from Carvykti. The first high-quality commercial evidence will be quarterly patient additions, duration, gross-to-net, and franchise-level sales after the launches overlap.

The growth bridge is therefore a portfolio of sequential proofs, not one binary clinical event. The 2026–27 approvals can reduce technical risk; the 2027–29 commercial mix will decide whether the moat is merely defended or economically extended.

What the quarter did not prove. First, the 12% HIV growth rate is not a pure volume measure. Management attributed Biktarvy’s growth to realized price and channel mix, inventory, and demand, while noting that the expiry of enhanced Affordable Care Act subsidies slowed the treatment market more than expected. A return to 2–3% treatment-market growth is guidance, not observed evidence. Second, more than 70% Yeztugo reinjection is encouraging persistence but does not reveal how many eligible patients discontinued, why they did so, or how the cohort compares with oral alternatives after adjustment for patient selection. Third, simultaneous Descovy and Yeztugo growth is useful evidence against immediate one-for-one cannibalization, but it does not identify price versus volume or the number of new people entering PrEP.

Fourth, Trodelvy’s first-line opportunity was approved only days before quarter-end, so the 26% growth rate does not contain a meaningful read on the expanded label. Sponsor-reported addressable-population and duration claims are reasonable hypotheses that need several quarters of share, duration, and gross-to-net evidence. Cross-trial hazard ratios against Datroway cannot establish product superiority because populations, controls, follow-up, and safety differ. Fifth, anito-cel’s reported 98% first-pass manufacturing success and 18-day median turnaround come from trials. Commercial centers, scale, and real-world vein-to-vein execution can differ materially.

Finally, a dense regulatory calendar is not the same thing as diversified cash flow. The near-term HIV candidates mostly defend or reformat an existing franchise; the oncology candidates enter crowded classes; and several economics are shared with partners. Counting approvals without subtracting cannibalization, royalties, collaboration shares, launch expense, and acquired capital would overstate growth quality. The report therefore treats Q2 as strong evidence of near-term franchise health, not proof of the full 2028–36 transition.

Verdict: Mixed-quality growth. The HIV/PrEP engine is high-quality and is the real reason to own GILD. Oncology growth is lower-quality and more capital-intensive. Consolidated forward growth can plausibly reach the mid-single digits through the late 2020s, but it depends on the lenacapavir transition absorbing the 2028 price reset and migrating patients well before expected adult generic entry in 2036. This is a fortress annuity attempting one more franchise-renewal cycle, not yet a proven multi-pillar secular grower.

6. Financial Quality

Margins and the structure of profit. (FACT — FY2025 10-K.) GAAP product gross margin is ~78% (COGS includes ~half manufacturing, ~half acquired-intangible amortization and royalties); non-GAAP product gross margin is ~87% and rose ~2 points on the expiry of a long-standing TAF royalty. R&D ex-acquired-IPR&D is ~$5.8B (~20% of revenue, roughly flat), because Gilead supplements internal R&D by buying pipeline. SG&A is ~$5.8B. Non-GAAP operating margin is ~47%, top-quartile. The base HIV business converts revenue to cash about as efficiently as any drug franchise in the world.

The central quality-of-earnings flag: IPR&D makes GAAP earnings non-representative. (FACT.) GAAP operating income swings violently — $7.3B (2022) → $7.6B (2023) → $1.66B (2024) → $10.0B (2025) — and GAAP net income with it: $6.2B (2021), $4.6B (2022), $5.7B (2023), $0.48B (2024), $8.5B (2025). The driver is two M&A P&L lines under asset-acquisition accounting, which expenses deal upfronts immediately:

The 2024 collapse was $3.8B of CymaBay IPR&D plus $4.2B of Immunomedics/Trodelvy impairment. The current episode is larger. Q2 alone included $11.18B of acquired IPR&D, driven principally by approximately $7.0B for Arcellx, $3.1B for Tubulis, and $1.0B for Ouro, plus a $1.75B IPR&D impairment associated with the Immunomedics/Trodelvy asset base. Q2 GAAP operating loss was $10.39B and net loss $10.50B; H1 net loss was $8.48B. Reported diluted EPS of negative $8.45 and company non-GAAP EPS of negative $6.75 both fail as measures of ongoing earnings because the transaction and tax impact was approximately $9.08 per share. Reversing that item implies clean full-year 2026 earnings power of approximately $8.43–8.78 per share, essentially unchanged from the prior $8.45–8.85 framework.

But non-GAAP is not clean either. (INTERPRETATION — the honest QoE point.) Acquired IPR&D is episodic for any one molecule but recurring for Gilead’s business model. Calling $11B of actual cash investment “non-GAAP” does not make it free. H1 operating cash flow of $6.12B less $257M of capital spending produced approximately $5.86B of reported free cash flow, helped by working-capital timing and therefore not safely annualized. A sustainable owner-earnings anchor remains approximately $9.5–10.5B of annual free cash flow, before discretionary acquisitions but after ordinary operating reinvestment. The lower end preserves the prior conservative estimate; the upper end recognizes the stronger base-product trajectory. Investors should separately charge acquisition cash against capital allocation rather than contaminate quarterly operating margins with it and then pretend it never happened.

Balance sheet — still manageable, no longer conservative. (FACT.) At June 30, cash and marketable securities were $3.18B, down from $10.61B at year-end. Current and long-term debt totaled $26.25B, producing net debt of approximately $23.07B, up roughly $8.8B from the prior report’s year-end anchor. First-half financing included $4.09B of debt proceeds and $2.78B of repayments. The company has ample cash generation and no immediate solvency issue, but debt is now roughly two-plus years of normalized free cash flow rather than approximately one and a half, and the cash buffer is much thinner. Goodwill remains $8.31B; intangibles fell to $14.03B from $16.98B at year-end after impairment and amortization. Equity fell to $11.74B from $22.62B because the expensed acquisitions and impairment ran through earnings. Book value and ROE have become even less economically meaningful.

Cash conversion, ROIC, and the aggregator trap. (FACT + computed.) The Q2 balance sheet reports approximately 1.241B shares outstanding, modestly below the 2025 diluted base; H1 buybacks of $774M continued to offset dilution. A commercial data feed displayed a nonsensical trailing EV/EBITDA above 600x because transaction accounting reduced its denominator to only $263M. That number must not be used. Similarly, consolidated ROIC mechanically improves after an acquired asset is impaired because the denominator disappears. A normalized franchise calculation still supports very high returns in HIV, but a through-cycle enterprise calculation must retain the cash cost of failed assets. On that economic basis, Gilead’s legacy HIV returns are excellent and incremental acquisition returns are poor or not yet observable.

Acquisition-inclusive cash tells the sharper story. Conventional H1 free cash flow was $5.859B: $6.117B of operating cash less $257M of capital spending. Subtracting $11.318B of acquisition cash produces negative $5.459B of acquisition-inclusive owner cash before dividends and repurchases. This is not the same as saying acquisitions are an operating expense; it is a capital-allocation measure that prevents the analysis from describing a serial acquirer as a perpetual 6% free-cash-flow-yield stock while ignoring the cash repeatedly required to sustain its growth narrative.

The distinction matters across time. For one isolated acquisition, normalized free cash flow is the better measure of existing franchise earning power. For a model that repeatedly supplements internal R&D with purchased pipeline, acquisition-inclusive cash is a necessary companion measure. Gilead can fund the current cycle because HIV generates exceptional cash; shareholders benefit only if the purchased projects later generate cash above their cost of capital. Until then, the balance-sheet increase is observable and the return is a hypothesis.

A reconstructed 2025 ROIC illustrates denominator sensitivity. Using reported capital produces a normalized return around 26%. Restoring 2021–25 acquired IPR&D and impairment charges to the invested-capital base lowers the estimate toward 18%. Both remain above a reasonable capital cost, because the HIV franchise is so profitable, but the roughly eight-point gap is economically meaningful. It is the historical cost of looking through write-offs. Neither figure should be presented as precision; the spread is the insight.

Verdict: Economics improve with scale in the HIV core but not demonstrably in the consolidated enterprise. The right earnings anchor is approximately $8.43–8.78 of clean 2026 EPS and $9.5–10.5B of normalized annual free cash flow, with the $11.3B of acquisition cash treated as invested capital that must earn a future return—not as an operating expense to capitalize, nor as a harmless adjustment to ignore.

7. Capital Allocation

Philosophy: return a meaningful share of FCF, recycle the rest plus balance-sheet capacity into pipeline M&A. (FACT — 10-Q; 2026 proxy.) In H1 2026, Gilead paid $2.07B of dividends and repurchased $774M of shares while spending $11.32B on acquisitions. Shareholder returns remained covered by operating cash; acquisitions required the cash balance and debt market.

• Dividend — exemplary. Grew from $3.71B (2022) to ~$4.0B (2025); ~$3.16/share, ~2.5% yield, raised every year since initiation in 2015. At ~$4.0B against ~$9.5B FCF (~42% payout) it is safe and well-covered even through GAAP-loss years, because the losses are non-cash IPR&D while OCF holds ~$10B. This is the genuine shareholder value transfer.
• Buyback — honest but unremarkable. $1.4B (2022), $1.0B (2023), $1.15B (2024), $1.9B (2025), ~$0.4B in Q1-26 — enough to sterilize ~$0.9B/yr of SBC and hold the share count flat, but not a valuation-sensitive lever.
• R&D intensity is modest (~$5.8B, <20% of sales), because Gilead outsources a chunk of R&D to M&A, run through the IPR&D line.

M&A track record — scored honestly. (FACT on price/date; INTERPRETATION on grade.)

The 2026 spree has now become cash fact rather than guidance: $11.32B left the company in H1, approximately one full year of normalized owner earnings. In Marathon’s capital-cycle framework, the warning is not simply that Gilead bought assets; it is that cash from a concentrated, supply-disciplined HIV oligopoly is being redeployed into ADC, CAR-T, bispecific, and inflammation markets attracting enormous industry capital. High scientific promise can coexist with low shareholder return when many well-funded competitors chase the same endpoints and standards of care evolve before an asset launches. The burden of proof should therefore be deal-level incremental ROIC after failures—not management’s aggregate non-GAAP EPS.

Deal-level hurdle rates. Arcellx must do more than obtain anito-cel approval: it must take meaningful share from Carvykti and emerging bispecifics, scale manufacturing, and generate enough after-tax cash to justify approximately $7B of incremental cash cost on top of the earlier collaboration investment. Tubulis and Ouro have earlier-stage, longer-duration risk and should be valued probability-weighted, not at cost. Trodelvy’s 26% Q2 growth and expanded first-line TNBC population are commercially encouraging, but the $1.75B impairment—after the prior $4.2B charge—shows that indication breadth and terminal value remain far below the original Immunomedics underwriting. Livdelzi is the cleanest recent positive counterexample: a focused specialty launch with $167M of Q2 sales, more than double the prior-year period.

The counterfactual matters. Every acquisition dollar could instead retire debt, repurchase stock when expected returns are attractive, or remain available for later opportunities. At current normalized free cash flow, the $11.3B H1 acquisition outlay equals roughly one year of distributable owner earnings and more than four years of the current dividend. That comparison does not make dividends inherently superior to R&D; it defines the opportunity cost. A high-quality acquisition should plausibly compound value well above the cost of equity, not merely replace revenue that an expiring product would otherwise lose.

Recent peer experience reinforces the warning. ABBV demonstrates that a large incumbent can execute a successful molecule baton pass, but also that successor concentration can lead to another wave of expensive deals. REGN shows that a genuine discovery engine and strong branded share do not prevent severe total-market erosion after loss of exclusivity. MRK and PFE illustrate the wide difference between reported earnings and economic reinvestment when large acquisitions and acquired IPR&D dominate a period. Gilead’s HIV transition is more credible than its oncology diversification because the customer base and commercial scale already exist; the acquired oncology returns still have to be created.

Incentive alignment — a real governance gap. (FACT — 2026 DEF 14A.) The annual cash incentive is 60% financial (Adjusted Net Product Revenue + Adjusted Operating Income; the financial weight was raised from 50% in 2025) + 40% strategic; long-term incentives are 50% performance shares (3-yr relative TSR + a newly added multi-year adjusted-EPS-growth metric), 25% options, 25% RSUs. There is no return-on-capital metric anywhere in the proxy — zero mentions of ROIC/ROTC/ROE. For a company whose central question is whether ~$11.5B/yr of M&A earns its cost of capital, executives are paid on revenue, operating income, EPS growth and TSR — all of which a serial acquirer can hit by buying revenue regardless of value created. CEO Daniel O’Day (Chairman & CEO since 2019) earned ~$28.4M in 2025 SCT total (base <7% of total — good alignment on form, wrong metrics). Say-on-pay support is ~91% (adequate, not strong), and a stockholder proposal for an Independent Board Chair recurs (O’Day holds the combined role) — investor concern aimed precisely at oversight of the capital allocation flagged here.

Insider signal—mildly negative, low weight. A refreshed five-year issuer-filtered sweep covered 316 Gilead Form 4 filings. It found no code-P open-market insider purchase. Since the prior report, O’Day sold 30,000 shares, Dickinson 6,000, and Mercier 6,000, for 42,000 shares and approximately $5.38M in aggregate; every sale was disclosed under a Rule 10b5-1 plan. Planned selling without discretionary buying is directionally negative, but retained holdings and pre-scheduled plans make it weak evidence rather than an event-driven warning.

Verdict: Mixed-to-negative on M&A, positive on shareholder returns. The dividend is exemplary; the buyback is honest. But the M&A machine — outside the wasting-asset Pharmasset win — is a record of overpayment and write-offs, and the 2024–26 spree reads as buying growth under cliff pressure. Management returns the cash-cow’s output competently but has not shown it can convert M&A into durable franchise value beyond HIV; the missing ROIC metric and combined Chair/CEO compound the concern. Capital allocation is a thesis risk, not a strength.

8. Changes and Headwinds — Last Two Years

(FACT timeline — 8-Ks, transcripts, AZI news feed.)

• Feb 2024: CymaBay (~$4.3B, Livdelzi/PBC); $3.8B IPR&D + $4.2B Trodelvy impairment crush FY24 GAAP NI to $0.48B.
• 2025 H2: Yeztugo (lenacapavir PrEP) launch — FY26 guide later raised to ~$1B, a potential first-year blockbuster; total-HIV growth guide raised 6%→8%. (POSITIVE — the optionality underwriting the bull case.)
• Jan 2026: Biktarvy selected for Medicare price negotiation, negotiated price effective 2028—a structural pricing headwind on the approximately 49%-of-sales flagship, years before expected adult generic entry under settlements beginning in 2036. (NEGATIVE, slow.)
• Dec 2025: drug-pricing agreement with the US government (lower Medicaid pricing) + proposed ACA changes → ~2% growth headwind in 2026. (NEGATIVE, quantified, manageable.)
• 2026-03-31: Galapagos collaboration restructured around gamgertamig (BCMAxCD3 T-cell engager); ~$1.675B upfront split 50/50 + up to $500M milestones.
• 2026-04-28: Arcellx closed (~$7.8B, anito-cel CAR-T myeloma).
• Q2-2026: Arcellx, Ouro and Tubulis transactions produced $11.18B of Q2 acquired IPR&D expense; H1 acquisition cash use reached $11.32B. Net debt rose to $23.07B. (NEGATIVE for financial flexibility; outcome still open.)
• 2026-06-08 (POSITIVE): islatravir/lenacapavir once-weekly oral HIV (with Merck) Phase 3 positive — extends lenacapavir into weekly oral dosing, reinforcing the post-Biktarvy bridge.
• 2026-06-08/09 (NEGATIVE): EVOKE-03 (Trodelvy+Keytruda, 1L NSCLC) discontinued for futility; the stock traded lower. A meaningful Trodelvy label-expansion indication lost — directly undercutting the $21B Immunomedics thesis.
• 2026-06-15 (POSITIVE): FDA accepted once-weekly oral Yeztugo for PrEP, with a Feb. 2, 2027 PDUFA. The regimen can expand convenience but will explicitly compete with Descovy and generics (company release).
• 2026-07-24 (POSITIVE): European CHMP issued a positive opinion for Trodelvy plus Keytruda in first-line PD-L1-positive metastatic triple-negative breast cancer, reinforcing the near-term breast franchise.
• 2026-08-03 (LEGAL POSITIVE / LONG-TAIL RISK REMAINS): the California Supreme Court ruled for Gilead in the TDF/TAF “duty to innovate” litigation. The decision reduces an unusual liability theory but does not eliminate ordinary product-liability risk.
• 2026-08-04 (MIXED): Q2 base products +10%, HIV +12%, Trodelvy +26%, and guidance raised; cell therapy −14%, $11.18B acquired IPR&D, and $1.75B impairment.
• Ahead (thesis-critical): BIC/LEN Aug. 27, 2026; anito-cel Dec. 23, 2026; weekly oral Yeztugo Feb. 2, 2027. These are regulatory catalysts, not evidence of commercial adoption.

Verdict: Net mixed. The HIV/lenacapavir engine strengthened enough to raise near-term earnings confidence, while oncology and balance-sheet evidence weakened enough to offset much of that gain. The five-year price chart shows a market that already recognizes the HIV transition: despite the IRA selection, EVOKE-03 failure, and transaction losses, the stock remained above its long-term trend and closed 8.65% above the prior-report reference. That resilience is supportive momentum, but it also means the operating improvement is not undiscovered.

9. Risk Analysis

The dominant risks are structural and dated rather than acute. The Biktarvy maximum fair price takes effect Jan. 1, 2028; CMS is expected to make the negotiated price public by Nov. 30, 2026. The reported $3.904B of gross Medicare spend is not Gilead revenue and should not be multiplied mechanically by an assumed discount. Product mix, rebates, Medicaid and 340B knock-ons, and commercial contracting all determine the ultimate earnings impact. A catastrophic loss remains remote; the realistic bear outcome is prolonged stagnation, lower returns on incremental capital, and a de-rating.

Patent and exclusivity nuance. Orange Book listings now support Biktarvy patents extending from 2032 through November 2036, with adult generic settlements allowing launch as early as Apr. 1, 2036, and Yeztugo patents extending from 2034 through November 2040. Patent listings do not guarantee litigation outcomes or commercial durability, but they make the prior shorthand “2033 cliff” too crude. The more accurate risk is a 2028 price reset followed by a managed regimen migration and a multi-year patent ladder. This reduces binary cliff risk while increasing the importance of switch economics.

10. Valuation Discussion (Embedded Expectations)

Starting capitalization (2026-08-12 close). (FACT + calculation.) At $135.87 and 1,239,955,155 shares, equity value is approximately $168.47B. Adding $23.07B of net debt produces current enterprise value of approximately $191.54B. The reported 2026 earnings guidance remains distorted by transaction accounting. Adding back the disclosed $9.08-per-share transaction impact to the negative $0.65 to negative $0.30 non-GAAP guide produces clean earnings power of approximately $8.43–8.78 per share, or 15.5–16.1x at spot (15.8x midpoint). The annual dividend yield is approximately 2.41%.

TTM revenue is $30.457B, producing price-to-sales of 5.53x and EV-to-sales of 6.29x. Reconstructed clean TTM operating income is $14.947B, a 49.1% margin, and clean EBITDA is approximately $15.318B, producing 12.50x EV/clean EBITDA. Price-to-book is economically unhelpful because expensed transactions and impairments reduced equity to $11.74B. Reported trailing EV/EBITDA is worse: one vendor showed more than 600x because acquired-IPR&D charges collapsed its denominator to $263M. Any comparison using that number is invalid.

TTM cash from operations was $13.552B and capex $609M, producing $12.943B of conventional FCF and a 7.68% equity yield. Charging approximately $909M of stock compensation lowers the yield to 7.14%. Because external science is economically recurring for this company, the valuation work also subtracts an explicit $1.0B annual pipeline-replenishment reserve, leaving a rounded $11.0B normalized owner-FCF anchor and 6.53% yield. The reserve is an analytical smoothing assumption, not an accounting fact or management forecast.

Embedded expectations / reverse DCF. (INTERPRETATION.) A smooth Gordon inversion on $11.0B of owner FCF implies perpetual growth of 1.5%, 2.5%, or 3.5% at 8%, 9%, or 10% costs of equity. A five-year model is more informative because it can expose terminal dependence: at a 9% discount rate, current capitalization requires a 17.6x terminal FCF multiple if owner FCF is flat, 14.7x at 3% CAGR, 12.3x at 6%, or 10.3x at 9%. The most balanced reading is approximately 3–4% five-year owner-FCF growth and a 14x terminal cash multiple. That is not heroic, but it requires continued HIV durability and disciplined reinvestment.

Earnings-power value. Normalize clean operating income at $14.947B, tax it at 20%, capitalize the resulting $11.958B of after-tax operating earnings, then subtract $23.067B of net debt. At 8.0%, 8.5%, 9.0%, and 10.0% costs of capital, equity EPV is approximately $126.4B, $117.6B, $109.8B, and $96.5B—75%, 70%, 65%, and 57% of current capitalization. At the central 9% rate, approximately 35% of equity value depends on growth, excess returns, franchise longevity, or pipeline value. It is therefore incorrect to say oncology is “free” merely because current earnings ignore its future contribution.

This EPV gap is not automatically overvaluation. Greenwald’s framework permits value above no-growth EPV when a company has defensible barriers and can reinvest at attractive returns. Gilead clearly meets that condition inside HIV; it has not demonstrated it in acquired oncology. The central valuation question is therefore whether future incremental capital is deployed inside the moat or outside it. A dollar used to migrate captive HIV patients can preserve high-return earnings; a dollar used for a crowded early-stage ADC must create an advantage that did not previously exist.

Scenarios (five years, to approximately 2030; clean EPS basis; assumptions, not forecasts).

The scenarios are deliberately wide because the 2028 Biktarvy MFP is not public and the new acquisitions have no observable commercial return. “Coverage” is present value divided by current equity capitalization, not a price target. Current capitalization maps to the base assumption set: a successful HIV bridge, near-current margins, flat shares, and acceptable—not exceptional—returns from the rest of the portfolio. The bear combines lower cash flow with a lower terminal multiple; the bull requires both operating delivery and evidence that the moat extends beyond defensive product migration.

Peer and history cross-check. Current local reports place GILD above cliff-discounted PFE and BMY, broadly around the middle of profitable large pharma on clean earnings, and below companies rewarded for clearer growth or diversified innovation. Exact peer multiples are sensitive to each company’s own acquired-IPR&D adjustments, so false precision is not useful. The prior June authenticated history snapshot placed GILD at the 95th percentile of its own price-to-sales history, but the endpoint could not be refreshed because its token was unavailable. That observation remains historical context only; it is not represented as a current percentile.

Verdict: At $135.87, GILD’s approximately 15.8x clean earnings multiple, 12.5x clean EBITDA, and 6.53% normalized owner-FCF yield are consistent with a high-quality, low-to-mid-single-digit compounder. Reverse valuation requires roughly 3–4% five-year owner-FCF growth with a 14x terminal cash multiple and assigns a meaningful premium above no-growth EPV. The shares are neither priced like a broken cliff asset nor like a proven multi-franchise growth company. No price target.

11. Variant Perception

Positioning. Short interest is approximately 1.97% of float and fell 4.3% month over month, so there is no crowded short or squeeze thesis. From the prior report through Aug. 12, the shares returned 8.65%; three-, six-, and twelve-month total returns were +2.79%, −11.72%, and +16.16%. The $135.87 close was 3.46% above its 50-day EMA and 6.00% above its 200-day EMA, with the 50-day above the 200-day. Momentum is constructive but not euphoric.

FactorsToday attribution supports the same read. The dominant modeled exposure was Biotechnology (+0.998), followed by Health Care (+0.442) and Dividend Yield (+0.209), with a negative beta exposure (−0.577). Momentum and Value were effectively zero and Quality was only +0.038; model R-squared was 0.410. In plain English, GILD is held as a defensive biotech/dividend exposure rather than a crowded quality-momentum trade. The only 252-day factor extreme was Dividend Yield at a +2.01 z-score. This positioning can cushion drawdowns, but it offers no independent evidence that the stock is mispriced.

Strongest bull case. Lenacapavir is a second HIV product cycle, not merely a Biktarvy patch. Yeztugo is already approaching a $1B annualized objective, quarterly PrEP sales have exceeded $1B, the treated prevention population can grow, and listed patents extend to 2040. BIC/LEN, weekly oral PrEP, and islatravir/lenacapavir offer a dosage ladder across treatment and prevention. If total HIV keeps growing through the 2028 price reset, Gilead proves that customer captivity and R&D scale can migrate the franchise before patent expiry. Trodelvy in first-line breast cancer, Livdelzi, and anito-cel would then be additive rather than necessary.

Strongest bear case. Biktarvy remains nearly half of product sales and faces a 2028 price reset before the new portfolio proves its economics. Long-acting and weekly regimens mostly switch patients—including Gilead’s own Descovy and Biktarvy users—while partnered products dilute profit per patient. Oncology remains a serial capital consumer: Kite is shrinking, Forty Seven was written off, Immunomedics has now absorbed at least $5.95B of disclosed impairments across 2024 and Q2 2026, and the new deals consumed $11.3B of cash. At approximately 15.8x clean earnings and a normalized free-cash-flow yield below 6%, the market offers limited protection if growth normalizes after the launch cycle.

The assumptions that matter most, and what falsifies each side.

1. Lenacapavir additive vs. cannibalizing (the dominant swing factor). Separate three tests: additive to Gilead, additive to the PrEP category, and new-to-PrEP patients. Switching from Apretude or generics is additive to Gilead but not necessarily to the category; switching from Descovy is internal cannibalization; treatment-naive initiation is the cleanest category-expansion evidence. Falsifies bull: total HIV flattens or declines through 2028 and naive PrEP starts fail to accelerate. Falsifies bear: HIV compounds through 2030 with a rising naive share and stable economics.
2. Depth of the 2028 Biktarvy Medicare cut + breadth of future IRA selection. Falsifies bull: negotiated price + Medicaid/340B knock-on cuts Biktarvy realized price >25–30%, with more HIV products selected. Falsifies bear: the cut is modest/limited-mix and no further core-HIV products are selected near-term.
3. PrEP TAM is real net-new demand (0.5M→1M+) vs. a re-shuffle. Falsifies bull: PrEP market growth decelerates from ~14%; Yeztugo persistency/naive share disappoint. Falsifies bear: the PrEP-treated population visibly steps up and Yeztugo sustains a multi-billion trajectory.
4. Oncology ever earns its cost of capital (or is correctly valued at ~zero). Falsifies bull-as-upside: more EVOKE-style failures / Trodelvy share loss to Datroway / anito-cel disappoints. Falsifies bear-as-black-hole: Trodelvy 1L + anito-cel turn oncology into a genuine second pillar with positive incremental ROIC.
5. The multiple: durable-annuity re-rate vs. cliff-discount de-rate.

Verdict: The position is un-crowded and defensively exposed, while the price trend already recognizes improving HIV evidence. The variant perception cuts both ways: the bull sees a patent-laddered annuity that can compound through 2040; the bear sees a concentrated franchise financing acquisitions whose aggregate economics remain poor. The dominant swing factor remains the three-part additivity test above.

12. Fact vs. Interpretation Table

13. Open Questions

1. What is the magnitude of the 2028 Biktarvy Medicare negotiated-price cut, and what share of Biktarvy revenue is Medicare-exposed? Management gives no number; this is central to the 2028→2033 bridge.
2. What is the net-new vs. switch split of Yeztugo/PrEP demand? The single most thesis-critical unknown — the difference between additive growth and a lower-price reshuffle.
3. How defensible are the Orange Book patent ladders in practice? Listed Biktarvy patents extend through November 2036 and Yeztugo through November 2040, but formulation litigation, settlements, and actual regimen migration determine economic exclusivity.
4. Will anito-cel (Arcellx) or gamgertamig (Ouro) break the pattern of value-dilutive oncology M&A? No clinical/commercial readout yet; decides whether the 2026 spree was renewal or destruction.
5. How much of the $21B Immunomedics/Trodelvy thesis survives EVOKE-03? The mTNBC franchise plus remaining indications versus the lost lung-cancer leg.

14. What Must Be True

For the bull case to be right:

• Lenacapavir-based regimens must ramp fast enough to hold total HIV revenue growing through the 2028 price reset and migrate patients well before expected 2036 generic entry, with visible net-new PrEP patients—not just intra-franchise switching.
• Falsification test: if, by FY2028, total HIV revenue is flat-to-down, naive PrEP initiation fails to accelerate, and new regimens mainly migrate existing Gilead patients at weaker economics, the engineered transition is failing.

For the bear case to be right:

• The 2028 Biktarvy Medicare cut must be deep (>~30% realized-price erosion on the Medicare book), lenacapavir must prove net-cannibalizing, and oncology must keep destroying capital — collapsing the franchise toward a cliff-pharma multiple.
• Falsification test: if HIV total revenue compounds mid-single-digits through 2028–2030, PrEP grows double-digits with rising naive share, and Trodelvy 1L / anito-cel launch into genuine second-pillar revenue with positive incremental ROIC, the bear “stagnating taxed annuity” thesis is broken and a durable-annuity re-rate is warranted.

The analysis above takes no investment position and contains no price target; valuation is discussed solely as embedded expectations and scenarios. The single, clearly labeled exception is the Claude's Take block, which is the author’s subjective view.

15. Source Appendix

See the separate Source Appendix (GILD_source_appendix.md) for the full enumerated source list with URLs, dates and filing references.

APPENDIX A — Standard Diligence Questionnaire

Gilead Sciences, Inc. (NASDAQ: GILD) — as of 2026-08-13

Supplemental to the research memo. Fact, interpretation, and assumption labels are applied where they matter.

General

What thoughtful questions have other investors asked about this company? The dominant debate is whether lenacapavir is additive to Gilead, additive to the PrEP category, and new-to-PrEP; those are three different tests. Other questions: How deep is the 2028 Biktarvy Medicare cut? Can BIC/LEN preserve economics, not merely patients? Will any oncology acquisition earn its cost of capital? Did the $11.3B H1 acquisition spend buy a second pillar or only raise leverage? Is the dividend safe through an accounting-loss year? Yes on current cash flow, though the balance-sheet cushion has narrowed.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? (Interpretation.) Neither—pharma earnings are disease-driven, not macro-cyclical. GILD is at a mix inflection: Veklury has nearly disappeared, HCV has largely bottomed, and lenacapavir is ramping before the 2028 Biktarvy price reset. Clean 2026 EPS of approximately $8.43–8.78 and normalized FCF of $9.5–10.5B are reasonable earning-power anchors.

Driven by the external environment or internal actions? Predominantly internal (product launches, pricing, R&D/M&A), with a growing external overlay from IRA/Medicaid/ACA policy (~2% headwind in 2026, rising).

How stable are revenues? Very stable in aggregate, but concentrated: Biktarvy remains nearly half of product sales. The dated risks are the 2028 negotiated-price reset and later generic entry under settlements beginning in 2036, not a single 2033 cliff.

Outlook for products/services? HIV treatment durable; PrEP a genuine growth category; HCV/Veklury declining; cell therapy declining; Trodelvy growing but contested (EVOKE-03 NSCLC failed); Livdelzi small/fast-growing.

How big will this market be? Global HIV ~$30B+ and growing modestly; US PrEP-eligible pool ~2.2M (only ~0.5M treated) — the principal expansion vector. Oncology TAMs are large but hyper-competitive.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? HIV: stable oligopoly, but ViiV (long-acting injectables) and Merck (capsid/islatravir) are pushing on dosing-interval convenience; the IRA adds a regulatory competitor (government as price-setter). Oncology: intensely and increasingly competitive.

How profitable is the business (ROIC, ROE)? (Fact + computed.) Q2 non-GAAP product gross margin was 86.9%; the core remains exceptionally profitable. A 2025 normalized ROIC is approximately 26% on reported capital but closer to 18% after restoring 2021–25 IPR&D and impairments. ROE and book multiples are distorted by expensed acquisitions and write-offs.

How profitable is the industry — competitors, barriers? HIV is highly profitable with formidable barriers (~$1B+/decade per regimen). Oncology earns average returns under low durable barriers.

Can the business be easily understood? Mostly — it is “HIV cash cow + diversification bets,” but the IPR&D-driven GAAP volatility (2024 NI $0.48B; 2026E GAAP loss) requires normalization to interpret.

Can it be undermined by foreign low-cost labor? No—IP and regulation, not labor cost, drive the economics. Generic entry after exclusivity and settlements beginning in 2036 is the relevant low-cost threat.

Do brands matter? Physician/guideline entrenchment and prescriber relationships function as the “brand” — real switching costs, but ultimately patent- and data-backed rather than consumer-brand.

Nature of competition? Efficacy/safety/resistance and increasingly dosing convenience; payer access; guideline positioning.

Customers’ switching costs? High for virally-suppressed treatment patients (clinical rebound/resistance risk); lower for PrEP (adherence-elastic).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The HIV franchise’s intangible value (patents, lenacapavir pipeline) is internally generated and not capitalized — economic value far exceeds the ~$18/sh book.

Off-balance-sheet liabilities? No hidden liability changes the thesis. Contingent milestones, royalties, collaboration obligations, and CVRs are disclosed. The larger current issue is visible on the balance sheet: $26.25B of debt against $3.18B of cash and securities.

How conservative is the accounting? Reasonably conservative on its face — M&A upfronts expensed immediately as IPR&D, impairments taken promptly. The flip side: this documents ~$10B+ of capital destruction this cycle rather than hiding it.

How CapEx-hungry? Very light — capex <2% of revenue (~$0.5B), asset-light pharma.

Capital Allocation & Management

How much FCF, and how is it used? H1 conventional FCF was $5.859B, versus $2.069B of dividends and $774M of buybacks. Acquisition cash was $11.318B, making acquisition-inclusive owner cash negative $5.459B before shareholder distributions. Normalized annual FCF is approximately $9.5–10.5B.

Significant acquisitions recently? Yes — CymaBay (2024, ~$4.3B); Arcellx (closed Apr-2026, ~$7.8B); Ouro + Tubulis (2026, closing Q2); Galapagos restructuring (Mar-2026). Track record outside Pharmasset is overpayment + write-offs.

Buying back shares? Modestly—H1 repurchases were $774M and the share base is approximately 1.241B outstanding. The program mainly offsets dilution rather than materially shrinking the company.

Issuing large amounts of stock to insiders? SBC modest (~$0.9B, ~3% of sales — low for biopharma).

Compensation policy? (Fact — 2026 proxy.) CEO O’Day ~$28.4M (2025); low base, heavy PSU/option. Metrics: revenue, operating income, EPS growth, relative TSR — no return-on-capital metric. Say-on-pay ~91%. Combined Chair/CEO with a recurring independent-chair proposal.

Motivations of management? Form-aligned (equity-heavy) but metric-misaligned for a serial acquirer (no ROIC hurdle). Strategy is explicitly “maximize long-acting HIV + accelerate the oncology/inflammation pipeline build.”

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a US C-corporation common stock (NASDAQ: GILD); standard 1099 dividend treatment.

Dividend policy? ~$3.16/share annualized, ~2.5% yield, raised every year since 2015 initiation; ~42% FCF payout; safe.

How profitable is the business? See above — franchise extraordinarily profitable; consolidated returns diluted by M&A.

Is net income diverging from cash from operations? (Fact — yes, materially.) GAAP NI swings on non-cash IPR&D (2024 NI $0.48B vs OCF $10.8B; 2026E GAAP loss vs ~$10B OCF). Use FCF/normalized earnings, not GAAP NI.

Risks & Downside

What would cause the stock to decline? A deep 2028 Biktarvy price cut; evidence that new regimens migrate existing Gilead users at weaker economics; slower naive PrEP adoption; further oncology failures or impairments; acquisition leverage without returns; or a BIC/LEN/anito-cel regulatory setback.

Risk of catastrophic loss? Low—net debt has risen to $23.07B, but recurring HIV cash flow and manageable maturities make an acute financing event remote.

Chance of a total loss? Negligible — no solvency path is visible; the realistic bear is multi-year stagnation + de-rating, not impairment.

Recent News & Events

Has the business environment changed recently? Yes—positively on HIV (Q2 +12%, guide raised to 9–10%, Yeztugo $232M and >70% reinjection) and negatively on capital allocation (H1 acquisition cash $11.3B, net debt $23.1B, $1.75B impairment). Trodelvy’s first-line TNBC approval is a genuine offset, while cell therapy fell 14%.

Significant acquisitions? Arcellx, Ouro, and Tubulis closed/entered the Q2 accounts and drove most of the $11.18B acquired-IPR&D charge; Galapagos was also restructured.

Change in accounting policies? None material; IPR&D expensing is longstanding.

Recent changes — new markets, facilities, management? Leadership stable (O’Day Chairman & CEO since 2019; Dickinson CFO; Mercier CCO; Berger CMO; Perettie EVP Kite). Strategic pivot deepening into oncology/inflammation.

APPENDIX B — Source Appendix

Gilead Sciences, Inc. (NASDAQ: GILD) — Research updated through 2026-08-13

Sources are primary-first. All financial figures reconciled to SEC filings / EDGAR XBRL; management commentary treated as hypothesis and validated against filings and external data.

Primary — SEC filings (EDGAR, CIK 0000882095)

1. Gilead Sciences FY2025 Form 10-K, filed 2026-02-24 (gild-20251231.htm). Statements of Operations, Balance Sheet, Cash Flows; MD&A product-sales-by-geography table; Item 1 Business; Item 1A Risk Factors; debt, goodwill, intangibles, IPR&D and acquisition notes.
2. Gilead Form 10-Q filings for FY2024 through Q2 2026, accessed through SEC EDGAR.
3. Gilead Form 8-K filings covering earnings, acquisitions, executives, litigation and regulatory events, accessed through SEC EDGAR.
4. Gilead DEF 14A Proxy Statement 2026, filed 2026-03-20 (gild014721-def14a.htm): compensation structure (annual incentive 60% financial / 40% strategic; LTI 50% PSU on relative TSR + EPS / 25% options / 25% RSU); CEO O’Day SCT ($28.4M, 2025); say-on-pay ~91%; independent-chair stockholder proposal. Prior proxies 2022–2025 for trend.
5. SEC Forms 3/4/5 insider corpus: 330 Form 4 documents, of which 316 were Gilead-issuer filings after excluding 14 filings for other issuers. The complete five-year issuer sweep found no code-P open-market purchase. Post-baseline sales totaled 42,000 shares and were all under disclosed Rule 10b5-1 plans.
6. EDGAR XBRL company facts and company concepts (refreshed 2026-08-13): revenue, operating income, net income, ordinary R&D, SG&A, goodwill, intangibles, debt, equity, stock compensation, operating cash flow, capex, repurchases, dividends, depreciation and shares. Primary filing statements control where tagging differs.

Primary — earnings calls & investor events

7. Q1 2026 Earnings Call, 2026-05-07 — FY26 guidance (base business $29.4–29.8B, +5–6%; total product $30–30.4B; ex-transaction non-GAAP EPS $8.45–$8.85; GAAP loss $(1.05)–$(0.65); $11.5B IPR&D; tax 140–190%; ~2% policy headwind; non-GAAP GM ~87%, op margin ~47%, tax 18.3%); HIV +10% YoY; Biktarvy $3.4B, >52% US share; Descovy $807M (+38%); Yeztugo $166M (+72% q/q, raised to ~$1B FY26); Trodelvy +37%; cell therapy −12%; Livdelzi +200%; capital return ~60% of FCF.
8. Q4 2025 Earnings Call, 2026-02-10 — FY25 results; initial FY26 guidance.
9. Q3 2025 / Q2 2025 Earnings Calls (2025-10-30; 2025-08-07).
10. Bernstein 42nd Strategic Decisions Conference, 2026-05-28 — “no major LOE until 2036”; lenacapavir ~17-yr development; PrEP TAM 1.2M→2.2M, ~0.5M treated → >1M by mid-2030s; once-yearly IM lenacapavir target 2028; FY25 absorbed ~$1.2B Part D redesign; M&A cadence/discipline.
11. Galapagos NV / Gilead M&A Call, 2026-03-31 — collaboration restructuring around gamgertamig; ~$1.675B upfront split 50/50 + up to $500M milestones.
12. Special Call, 2026-04-07 — Arcellx / Ouro / Tubulis transaction overview.

Secondary — data feeds & news (treated as signal, validated to primary)

13. AZI fundamentals feed (accessed 2026-06-12) — snapshot (sector, employees, description, analyst target ~$158, rating 4.13, short interest 1.9%, institutions ~92%); valuation_index own-history percentiles (P/E 53rd, P/B 88th, P/S 95th, composite 79th).
14. AZI news feed (accessed 2026-06-12) — EVOKE-03/KEYNOTE-D46 discontinuation (2026-06-08/09, negative); islatravir/lenacapavir once-weekly oral Phase 3 (2026-06-08, positive); Cencora/Kite distribution (2026-06-02). Underlying company releases validated.
15. Yahoo Finance historical snapshot (2026-06-12) — price, capitalization, balance-sheet and peer-market-data context; superseded by the current filing-rebuilt figures in the update.

Analytical frameworks

16. Analytical frameworks — Greenwald & Kahn, Competition Demystified (moat taxonomy; market-share-stability and ROIC tests; barriers to entry); Marathon Asset Management, Capital Returns (supply-side capital cycle; asset-growth anomaly).

2026-08-13 update — primary operating and regulatory evidence

17. Gilead Sciences Q2 2026 Form 10-Q, Gilead Sciences, Inc./SEC, filed 2026-08-06, accessed 2026-08-13, regulatory filing. Q2 financial statements; $1.75B NSCLC IPR&D impairment; Biktarvy MFP risk disclosure; Descovy/Cipla litigation. https://www.sec.gov/Archives/edgar/data/882095/000088209526000031/gild-20260630.htm
18. Gilead Sciences Announces Second Quarter 2026 Financial Results, Gilead Sciences, Inc., Form 8-K Exhibit 99.1, published/filed 2026-08-04, accessed 2026-08-13, earnings release. Product sales, margins, acquired IPR&D, cash flow and revised guidance. https://www.sec.gov/Archives/edgar/data/882095/000088209526000028/exhibit991earningspressrel.htm
19. GILD Q2 2026 Prepared Remarks, Gilead Sciences, Inc., published 2026-08-04, accessed 2026-08-13, investor presentation. Management-attributed PrEP market, product-demand, reinjection, share and launch-readiness commentary. https://s29.q4cdn.com/585078350/files/doc_financials/2026/q2/GILD-Q226-Prepared-Remarks-4-August-2026.pdf
20. Gilead Sciences Q2 2026 archived webcast, Gilead Sciences, Inc., webcast dated 2026-08-04, accessed 2026-08-13, management call. Official audio is the primary source for Q&A; exact text recovery used ROIC.ai/earnings-call transcript mirrors where no issuer transcript was published. https://investors.gilead.com/financials/quarterly-results/default.aspx
21. Orange Book Data Files, US Food and Drug Administration, monthly patent/exclusivity files dated 2026-07-10, accessed 2026-08-13, regulatory database. NDAs 210251 (Biktarvy), 208215 (Descovy), and 220018/220020 (Yeztugo). https://www.fda.gov/drugs/drug-approvals-and-databases/orange-book-data-files
22. Medicare Drug Price Negotiation Program: Selected Drugs for Initial Price Applicability Year 2028, Centers for Medicare & Medicaid Services, published 2026-01-27, accessed 2026-08-13, government fact sheet. Biktarvy selection; $3.904B gross Medicare expenditures and 101,000 users in the measurement period; statutory timeline. https://www.cms.gov/files/document/factsheet-medicare-negotiation-selected-drug-list-ipay-2028.pdf
23. CMS Announces Manufacturer Participation in Third Cycle of Medicare Drug Price Negotiation, Centers for Medicare & Medicaid Services, published 2026-03-13, accessed 2026-08-13, government fact sheet. Gilead participation and offer/counteroffer/negotiation timeline. https://www.cms.gov/newsroom/fact-sheets/cms-announces-manufacturer-participation-third-cycle-medicare-drug-price-negotiation
24. FDA Approves Sacituzumab Govitecan-hziy as Monotherapy and in Combination with Pembrolizumab for First-Line TNBC, US Food and Drug Administration, published 2026-06-24, accessed 2026-08-13, regulatory approval notice. US labels and ASCENT-03/04 efficacy. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-sacituzumab-govitecan-hziy-monotherapy-and-combination-pembrolizumab-first-line
25. European Commission Approves Trodelvy as a First-Line Treatment for mTNBC Patients Not Candidates for PD-(L)1 Inhibitors, Gilead Sciences, Inc., published 2026-06-23, accessed 2026-08-13, issuer regulatory release. EU monotherapy label only. https://investors.gilead.com/news/news-details/2026/European-Commission-Approves-Trodelvy-as-a-First-Line-Treatment-for-Metastatic-Triple-Negative-Breast-Cancer-Patients-Not-Candidates-for-PD-l1-Inhibitors/default.aspx
26. CHMP Recommends Trodelvy Plus Keytruda in PD-(L)1-Positive First-Line mTNBC, Gilead Sciences, Inc., published 2026-07-24, accessed 2026-08-13, issuer regulatory release. Positive opinion; European Commission decision remained pending at cutoff. https://investors.gilead.com/news/news-details/2026/CHMP-Recommends-Gileads-Trodelvy-Plus-Keytruda-in-PD-L1-Positive-First-Line-Metastatic-Triple-Negative-Breast-Cancer/default.aspx
27. FDA Approves Datopotamab Deruxtecan-dlnk for Unresectable or Metastatic TNBC, US Food and Drug Administration, published 2026-05-22, accessed 2026-08-13, regulatory approval notice. Pre-cutoff competitive context for first-line PD-(L)1-ineligible mTNBC. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-datopotamab-deruxtecan-dlnk-unresectable-or-metastatic-triple-negative-breast-cancer
28. FDA Approves First Treatment for Chronic Hepatitis Delta Virus Infection, US Food and Drug Administration, published 2026-05-22, accessed 2026-08-13, regulatory approval notice. Hepcludex accelerated approval and indication. https://www.fda.gov/news-events/press-announcements/fda-approves-first-treatment-chronic-hepatitis-delta-virus-hdv-infection
29. CDC Recommends New Injectable HIV PrEP, US Centers for Disease Control and Prevention, published 2025-09-18, accessed 2026-08-13, government clinical recommendation. Twice-yearly lenacapavir recommendation and 2.2M US PrEP-eligible estimate. https://www.cdc.gov/nchhstp/director-letters/new-injectable-hiv-prep .html
30. About AtlasPlus, US Centers for Disease Control and Prevention, current page accessed 2026-08-13, surveillance/methodology page. Explains suspension of current PrEP-coverage reporting. https://www.cdc.gov/nchhstp/about/atlasplus.html
31. HIV Diagnoses, Deaths, and Prevalence: 2026 Update, US Centers for Disease Control and Prevention, published 2026-04-30, accessed 2026-08-13, surveillance release. 2024 US diagnosed-prevalence and diagnosis data. https://www.cdc.gov/hiv-data/nhss/hiv-diagnoses-deaths-and-prevalence-2026.html
32. Global HIV & AIDS Statistics — Fact Sheet, UNAIDS, 2025 edition using 2024 data, accessed 2026-08-13, international surveillance fact sheet. https://www.unaids.org/en/resources/fact-sheet

2026-08-13 update — competitor-primary and market-data evidence

33. Q2 2026 Results Announcement, GSK plc, published 2026-07-28, accessed 2026-08-13, competitor earnings release. HIV, Dovato, Cabenuva and Apretude sales. https://www.gsk.com/media/jayh41pz/q2-2026-results-announcement.pdf
34. Dovato Is as Effective as Biktarvy in First Head-to-Head Study in Treatment-Naive Adults, ViiV Healthcare/GSK, published 2026-07-27, accessed 2026-08-13, competitor clinical release. VOGUE Phase IIIb topline; sponsor-reported pending full independent review. https://www.gsk.com/en-gb/media/press-releases/viiv-healthcare-s-2-drug-regimen-dovato-is-as-effective-as-3-drug-regimen-biktarvy-in-first-of-its-kind-head-to-head-study-in-treatment-naive-adults-living-with-hiv/
35. Merck to Present New Data on Daily, Weekly and Monthly Options Across Its HIV Pipeline at AIDS 2026, Merck & Co., Inc., published 2026-07-15, accessed 2026-08-13, competitor pipeline release. https://www.merck.com/news/merck-to-present-new-data-on-daily-weekly-and-monthly-options-across-its-hiv-treatment-and-prevention-pipeline-at-aids-2026/
36. Johnson & Johnson Reports Q2 2026 Results, Johnson & Johnson, published 2026-07-15, accessed 2026-08-13, competitor earnings release/supplement. Carvykti sales. https://www.investor.jnj.com/investor-news/news-details/2026/Johnson--Johnson-reports-Q2-2026-results-raises-2026-outlook/default.aspx
37. SEC Forms 4 — post-baseline senior-executive transactions, SEC, filed 2026-06-16 through 2026-08-04, accessed 2026-08-13, insider filings: Andrew Dickinson https://www.sec.gov/Archives/edgar/data/1310204/000131020426000020/wk-form4_1781646803.xml and https://www.sec.gov/Archives/edgar/data/1310204/000131020426000022/wk-form4_1784239680.xml ; Johanna Mercier https://www.sec.gov/Archives/edgar/data/1782135/000178213526000024/wk-form4_1781647325.xml and https://www.sec.gov/Archives/edgar/data/1782135/000178213526000026/wk-form4_1784235279.xml ; Daniel O’Day https://www.sec.gov/Archives/edgar/data/1638643/000163864326000026/wk-form4_1783023256.xml and https://www.sec.gov/Archives/edgar/data/1638643/000163864326000028/wk-form4_1785879633.xml
38. AZI GILD historical-price download, AZI Trading, accessed 2026-08-13; secondary market data. https://azitrading.com/controls/download-data.php?t=GILD
39. FactorsToday methodology and GILD endpoints, FactorsToday, accessed 2026-08-13; statistical exposure, not causal attribution. https://www.factorstoday.com/about
40. GILD Key Statistics, Yahoo Finance, snapshot accessed 2026-08-13 for period 2026-07-31, secondary short-interest vendor page. https://finance.yahoo.com/quote/GILD/key-statistics/