Sprott Physical Gold Trust (NYSE Arca: PHYS) — Impeccable Gold, a Leaky Wrapper

Independent equity research. Published 2026-07-25. Initiation of coverage.

⚡ Claude’s Take

This block is the author’s own subjective opinion. It is general information, not investment advice, and not a recommendation to buy or sell any security. The analysis that follows in the main body carries no recommendation and no price target — it discusses value only as embedded expectations and scenarios.

Verdict: HOLD if you already own it with a QEF election in place. AVOID-here as a new purchase unless you are specifically a high-bracket US taxable buy-and-hold holder — in which case, accumulate only on a widening discount. Not-a-short, ever. PHYS is not a business and there is no moat to underwrite. It is 3.67 million ounces of unencumbered, fully-allocated London Good Delivery gold sitting in the Royal Canadian Mint, wrapped in the best custody arrangement in the category and a genuinely valuable US tax election — and it charges roughly four times what the cheapest identical alternative charges. That is the whole argument, and the arithmetic is unforgiving. The custody is beyond reproach: the Mint is a Crown corporation whose obligations are effectively obligations of the Canadian Government, and it contractually bears the risk of loss. The tax edge is real and rules-based: a timely QEF election on Form 8621 converts what would be a 28% collectibles rate (what you pay on GLD, IAU, GLDM) into a ~20% long-term capital-gains rate — roughly 8 points on realised gains after the NIIT. But PHYS costs 0.39% against GLDM’s 0.10% and IAUM’s 0.09%, and its realised all-in drag has run ~0.60%/yr — and since inception in March 2010 the unit has compounded 1.06%/yr behind spot gold, a cumulative ~15.5% shortfall against simply owning the metal, of which fees explain only about 6.4 points. Fifteen years of owning gold and trailing it.

So the honest framing is a narrow window, not a recommendation. The ~8-point tax saving is banked once, on a realised gain; the ~0.50%/yr all-in cost gap versus GLDM compounds every year and consumes that saving in roughly 16 years. PHYS therefore wins for a large-gain, high-bracket, US-taxable holder on a long-but-finite horizon who will actually file Form 8621 annually — and loses, plainly, for an IRA/401(k), a non-US holder, a tax-indifferent institution, or anyone who might sell inside a decade. My accumulation zone is defined by the spread, not the metal: add only at a discount of ~3.5% or wider (roughly ≤$30.35 against today’s ~$31.44 NAV), and stop adding at anything inside ~1.5%, because at parity you are simply paying a 4x fee for a tax option you may never exercise. Framing: a decelerating falling knife off a genuine blow-off top — GoldPrice beta 1.63 with an R² of 0.89 means ~89% of what happens here is the gold price, the three-month Sharpe is −1.96, and the unit sits 25% below its January peak — but the long-horizon record is intact (three-year Sharpe 1.14) and the drawdown is entirely a metal event, not a wrapper failure. Conviction: medium. One fact flips me more constructive: the discount blowing out past ~5% (free gold, as at sibling PSLV), or Sprott finally cutting the management fee to defend share. One fact flips me negative: evidence the central-bank bid is breaking — sustained official-sector selling rather than the ~60 tonnes/month accumulation that is currently the floor under the whole complex. Tag: “Perfect bars, and a meter that never stops running.”

📈 Stock Price Action — Five-Year Event Map

Factual price history, not a recommendation. A PHYS unit is ~0.00774 oz of gold × spot, less a persistent discount to NAV — so every move below is, to first order, a gold move. Price moves are Fact; attributed drivers are Interpretation.

PHYS has run a five-year round-trip of unusual amplitude for a “boring” asset. From a cycle low of $12.52 (26 Sep 2022) the unit rose to an all-time closing high of $40.93 (29 Jan 2026) — an intraday $42.07 — a 3.3x move, before collapsing to $29.98 (24 Jun 2026) and closing at $30.64 on 24 Jul 2026. The unit now sits ~25.1% below its January peak, against a 52-week range of $25.02–$40.93, with NAV at $31.44/unit — a ~2.54% discount to the gold it owns.

Cycle narrative. (1) PHYS entered the window into the teeth of the most aggressive Fed tightening in four decades; with real yields rising and the dollar surging, a zero-coupon monetary asset did exactly what theory says and fell to its cycle low of $12.52. (2) The recovery began with pivot expectations and was accelerated by the March-2023 regional-banking stress, which re-established gold’s crisis bid, while central banks bought at record pace. (3) 2024 delivered a clean +27.2% year in the metal as the easing cycle began. (4) 2025 was the extraordinary year — gold rose 64.6%, from $2,624.50 to $4,319.37/oz, breaching $3,000 and then $4,000 for the first time in history, on what the market christened the “debasement trade”: fiscal-deficit anxiety plus relentless official-sector accumulation. (5) That tightened into a January-2026 parabola, gold printing an all-time high near $5,589–5,595/oz on 28–29 January. (6) The top broke violently and datably: on 30 January 2026 President Trump nominated Kevin Warsh — a well-known inflation hawk — as Fed Chair; Treasury yields and the dollar jumped, and CME margin increases compounded the move into forced liquidation. Bullion fell more than 9.8% in a single session, its sharpest one-day drop since 1983; PHYS fell 10.92% on 48.4 million units, more than twelve times its normal volume and the largest single-day move in its five-year history. (7) February and early March saw a genuine retest — PHYS back to $40.70 on 2 March, with gold at $4,668.06/oz at the end of Q1 — but the January high was never exceeded. (8) The subsequent five months have been a grinding unwind rather than a crash: the dollar reached its strongest in over a year, the Fed under Warsh turned hawkish, and the geopolitical premium drained as the US–Iran framework de-escalated and the Strait of Hormuz reopened. Gold broke below $4,000 on 24 June 2026, its first sustained move under that level since November 2025 and ~29% below the January high. PHYS bottomed the same day at $29.98 and has stabilised near $30.64, with the discount to NAV widening from ~1.3% (Q1-2025) to ~2.3% (Q1-2026) to ~2.54% today. Every price figure ties to the AZI five-year series; drivers are corroborated by contemporaneous reporting and the Trust’s own filings (see the Source Appendix).

1. Executive Summary

Sprott Physical Gold Trust is a closed-end mutual fund trust organised under the laws of Ontario, Canada, listed on the NYSE Arca and the Toronto Stock Exchange (both in USD), that holds 3,671,782 troy ounces of unencumbered, fully-allocated London Good Delivery gold bullion — roughly $14.9 billion of net assets — at the Royal Canadian Mint on behalf of unitholders. It is not an operating company. It has no revenue in any of the last five fiscal years, no earnings power, no employees, no capital to reinvest and no business to compound. A unit is, to first order, ~0.00774 ounces of gold × spot, less a ~0.39% headline fee and a fluctuating discount to net asset value. Ninety percent of the analysis of PHYS is therefore the analysis of gold; the residual — the wrapper — is where PHYS differs from iShares’ IAU, State Street’s GLD and GLDM, abrdn’s SGOL and Goldman’s AAAU, and it is where this article spends its effort.

On the metal (the 90%). Gold has just completed a genuine round-trip. Spot rose 64.6% in 2025 to $4,319.37/oz and peaked near $5,589–5,595/oz on 28–29 January 2026 before the Warsh nomination and CME margin hikes broke the move; bullion fell 9.8% in a single session on 30 January, its worst day since 1983, and has since ground down to roughly $4,061/oz as the dollar strengthened, the Fed turned hawkish and the Middle East risk premium drained. This is not a broken thesis — central banks are reported buying ~60 tonnes per month and remain the price-insensitive marginal bid — but it is a market whose sell-side targets were cut hard in June–July 2026 (Goldman $5,400→$4,900; JPMorgan $6,000→$4,500) and which now trades roughly 11–21% below a reduced consensus. Gold at $4,061 is off its mania and not obviously cheap.

On the wrapper (the 10% that is PHYS-specific). PHYS is differentiated on three axes and disadvantaged on two, and the balance is closer than its marketing suggests. In its favour: custody integrity — allocated, unencumbered LGD bars at a Canadian Crown corporation that contractually bears the risk of loss, with no lending or rehypothecation; issuance discipline — a hard contractual prohibition on issuing units below 100% of NAV, so the at-the-market program is structurally incapable of diluting gold-per-unit; and the tax edge — a PFIC/QEF election that converts the 28% US collectibles rate into a ~20% long-term capital-gains rate, worth roughly 8 percentage points on realised gains. Against it: cost — a 0.39% MER against GLDM at 0.10% and IAUM at 0.09%, roughly a 4x gap on an identical underlying; and the discount — PHYS has closed below NAV in every one of the last five years (−1.58%, −2.49%, −2.15%, −2.19%, −1.73%) and sits at −2.54% today, a structural feature the open-ended competitors, kept tight by authorised-participant arbitrage, simply do not carry.

The finding that matters most. Two independent calculations — gold-per-unit derived from ounce and unit counts, and the Trust’s own NAV-versus-spot compound returns — agree that PHYS’s all-in drag has run ~0.59–0.62% per year against a stated MER of 0.40–0.42%. That is roughly 20 basis points a year of leakage beyond the headline fee, mechanically attributable to the up-to-3% cash reserve the Trust retains from offering proceeds, plus commissions and issuance timing. Extend it: since inception in March 2010 the unit has compounded 1.06%/yr behind spot gold — a cumulative shortfall of about 15.5% — of which fees explain only 6.4 points. A holder who bought PHYS at inception to own gold has, fifteen years later, roughly 15% less gold-equivalent wealth than someone who owned the metal.

The embedded expectation. Because PHYS is its gold, there is no multiple to re-rate and no earnings to beat; P/NAV is the only valuation metric that exists and the discount is its only input. A buyer at $30.64 is making two separable bets: on the metal at an implied ~$4,061/oz, and on a wrapper that costs ~0.60%/yr all-in, delivers ~8 points of tax relief once on realised gains, and currently offers a ~2.5% entry spread. The crossover is arithmetic: against GLDM the cost gap consumes the tax saving in roughly sixteen years. That defines a real but narrow window of holders for whom this vehicle is the right answer — and a large majority for whom it is not.

Neither a recommendation nor a price target appears below; see Claude’s Take above for the one place a view is expressed.

2. Business Overview

What PHYS is. Sprott Physical Gold Trust is a closed-end mutual fund trust established under the laws of the Province of Ontario, Canada, which commenced operations on 3 March 2010. Its units trade on the NYSE Arca under “PHYS” and on the Toronto Stock Exchange under “PHYS” and “PHYS.U”, denominated in US dollars on both exchanges. Its registered office is Royal Bank Plaza, South Tower, 200 Bay Street, Toronto. It is a Canadian MJDS filer with the SEC (CIK 0001477049, commission file 001-34638) and therefore files Form 40-F annually and Form 6-K for interim periods — not 10-K/10-Q.

The stated objective, verbatim. From the FY2025 Annual Management Report of Fund Performance: the Trust “was created to invest and hold substantially all of its assets in physical gold bullion. The Trust seeks to provide a secure, convenient and exchange-traded investment alternative for investors interested in holding physical gold bullion without the inconvenience that is typical of a direct investment in physical gold bullion. The Trust intends to achieve its objective by investing primarily in long-term holdings of unencumbered, fully allocated, physical gold bullion and does not speculate with regard to short-term changes in gold prices.” (Fact.)

That last clause is not boilerplate — it is the operating constraint. There is no trading, no hedging, no overlay, no lending. Portfolio turnover has been 2.24% (2025), 2.82% (2024), 3.05% (2023), 0.04% (2022) and 0.71% (2021), and what little turnover exists is the sale of bullion to meet redemptions and pay expenses, not investment activity. The Trust’s own disclosure notes that the trading expense ratio is “N/A” because “there are no direct trading costs associated with physical bullion trades.” (Fact.)

What it owns. The Trust “has only purchased and expects only to own ‘London Good Delivery’ bars as defined by the London Bullion Market Association, with each bar purchased being verified against the LBMA source.” As of 31 March 2026 it held 3,688,439 troy ounces; as of 24 July 2026, 3,671,782 ounces. The Trust is expressly prohibited from investing in units or shares of other investment funds other than money-market mutual funds, and holds no assets subject to special arrangements arising from illiquidity. (Fact.)

How it makes — and does not make — money. PHYS has reported total revenue of nil in each of the last five fiscal years. Its entire statement of comprehensive income is a mark-to-market on one asset. For FY2025: unrealised gains on gold bullion of $5,849.3 million, realised gains of $147.6 million, against management fees of $43.2 million, sales tax of $2.7 million and operating expenses of $4.1 million — producing net income of $5,948.6 million. For Q1-2026: unrealised gains of $1,004.7m and realised gains of $322.7m against $15.7m of management fees. There is no operating leverage to analyse because there are no operations. (Fact — FY2025 40-F Ex-99.6; Q1-2026 6-K Ex-99.1.)

The unit. Each unit represents an equal, fractional, undivided beneficial ownership interest in the net assets of the Trust. NAV per unit is computed as (gold ounces × spot, plus other assets, less liabilities) ÷ units outstanding, and was $36.24 at 31 March 2026 and $31.44 at 24 July 2026. Because the vehicle is closed-end, the exchange price is set by supply and demand and trades at a premium or discount to NAV — the central analytical fact about this security and the subject of the Valuation section. Gold-per-unit was 0.00774 oz at Q1-2026. There are no distributions: the Trust holds a non-income-producing asset and “does not anticipate making regular cash distributions to unitholders,” so total return is purely price change — gold, plus or minus the change in the discount, less costs. (Fact.)

Governance and the parties. Four parties matter. Sprott Asset Management LP is the Manager, under a management agreement dated 24 February 2010; it is the only party with economic interest in the Trust’s growth. The Trustee, a trust company organised under the federal laws of Canada, holds title to the Trust’s assets, shares exclusive authority over the assets and affairs of the Trust with the Manager, and “has a fiduciary responsibility to act in the best interest of the unitholders.” The Royal Canadian Mint — a Canadian Crown corporation acting as agent of the Canadian Government, “whose obligations generally constitute unconditional obligations of the Canadian Government” — is custodian of the physical gold under a precious-metals storage and custody agreement, and “is responsible for and bears all risk of the loss of, and damage to, the Trust’s physical gold bullion that is in the Mint’s custody,” subject to limitations including events beyond the Mint’s control. The Trustee separately custodies all non-bullion assets. There is also an Independent Review Committee, whose expenses the Trust bears. (Fact.)

Recurring versus non-recurring revenue. The distinction does not map and it would be dishonest to force it. There is no revenue to be recurring. What is durable is the structure: a standing pool of allocated metal, a contractually fixed fee, and an issuance/redemption machinery that scales the pool. What is not durable is the fee level, which is under sustained competitive attack (see Industry Dynamics and Competitive Position).

Verdict. PHYS does precisely what it says on the tin, and does it cleanly: it holds allocated gold in a sovereign vault, publishes a daily NAV, and lets an exchange investor own metal without a safe-deposit box. Judged as a wrapper, the construction is careful and the disclosure is honest — the Trust volunteers its own tracking shortfall against spot in every MRFP, which many sponsors would bury. Judged as a business, there is none: no revenue, no reinvestment, nothing to compound. The analytical work is therefore entirely about cost, spread, and structure, and this article treats it that way.

3. Industry Dynamics

Two distinct “industries” bear on PHYS and conflating them is the most common analytical error made about these vehicles. The first is the gold market, which determines essentially all of the unit’s return. The second is the gold-ETP industry, in which PHYS competes for the investor’s dollar and which determines whether the wrapper is a good deal. The first is cyclical and currently mid-correction; the second is structurally hostile to PHYS and getting worse.

3.1 The gold market

Where the price has been. The Trust’s own reported spot marks give a clean series: $2,624.50/oz at 31 December 2024, $4,319.37/oz at 31 December 2025 (+64.6%), $4,668.06/oz at 31 March 2026. Gold reached an all-time high near $5,589–5,595/oz intraday on 28–29 January 2026, then broke below $4,000 on 24 June 2026 — its first sustained move under that level since November 2025 and roughly 29% below the January peak. Backing out from the Trust’s current disclosures ($14.91bn of net assets on 3,671,782 ounces), spot sits at approximately $4,061/oz as of 24 July 2026. (Fact.)

What broke the top, precisely. This matters because it tells you whether the correction is structural or positional. The proximate trigger on 30 January 2026 was the nomination of Kevin Warsh as Federal Reserve Chair — a former governor with a long public record of inflation hawkishness and balance-sheet-reduction advocacy. Treasury yields rose, the dollar index strengthened, and equities and precious metals sold off together. CME Group margin increases then amplified the move, forcing leveraged futures longs to post capital or liquidate. Bullion shed more than 9.8% on 30 January — its sharpest one-day decline since 1983. (Fact — contemporaneous reporting, the Source Appendix.) Interpretation: this is a positioning event, not a demand event. Margin hikes broke the 2011 and 2021 precious-metals spikes in exactly the same way. Nothing about the official-sector bid changed on 30 January; what changed was the cost of carrying leverage and the expected path of real rates.

What has driven the subsequent five-month grind. Three identifiable forces: a dollar at its strongest in over a year; a genuinely hawkish Fed under Warsh, with markets pricing tighter-for-longer; and the draining of a geopolitical risk premium as the US–Iran framework de-escalated and the Strait of Hormuz reopened, with crude retreating toward pre-conflict levels. (Fact.) Interpretation: two of these three (the dollar and real rates) are the classic gold headwinds and are cyclical; the third (geopolitics) was always a temporary premium that was correctly given back.

The demand floor. Central-bank purchases are reported running in the region of 60 tonnes per month, and the “debasement trade” — institutional demand driven by fiscal-deficit concern — represents a category of buyer largely absent from prior gold cycles. Interpretation: this is the single most important structural change in the gold market and it genuinely does alter the correction/structural-demand relationship. Central banks are price-insensitive, do not mark to market for career risk, and do not sell into weakness. A price-insensitive marginal buyer of that size puts a real floor under the market — but a floor is not a forecast, and it did not prevent a 29% drawdown.

Where the sell-side now sits. Targets were cut hard and recently. Goldman Sachs cut its year-end-2026 target from $5,400 to $4,900 in June 2026, removing all remaining 2026 rate cuts from its forecast and pushing easing to June and December 2027; it quantifies roughly $120/oz of support per 50bp of Fed easing. JPMorgan cut its Q4-2026 target from $6,000 to $4,500 on 3 July 2026 — a 25% reduction — citing softer demand from key buying sectors and heightened sensitivity to real rates. Deutsche Bank expects roughly $4,300 average in Q3-2026 and $4,800 by Q4. Consensus year-end sits in a $4,500–4,900 band. (Fact.) Interpretation: even after aggressive cuts, consensus is 11–21% above spot. That is a market where the analyst community has capitulated on the mania but not on the thesis — a configuration that historically offers less asymmetry than it appears, because the cuts tend to lag the price rather than lead it.

The capital-cycle lens and why it under-performs here. Marathon’s framework asks where capital is flowing and whether high returns are attracting supply that will mean-revert them. Applied to gold, the framework is unusually weak, and it is worth being explicit about why rather than forcing it. Above-ground gold stocks are on the order of 215,000 tonnes against annual mine supply of roughly 3,600 tonnes — a ratio near 60:1. Essentially all gold ever mined still exists and is available to the market at a price. Mine supply is therefore close to irrelevant to price formation on any investable horizon: a doubling of the gold price cannot call forth enough new metal to matter against the existing stock, and a halving cannot destroy meaningful supply. Gold does not clear like copper. It clears like a currency, on the relative attractiveness of holding a zero-coupon monetary asset versus an interest-bearing one — which is why real yields and the dollar explain so much of it. Interpretation: there is no industry structure here that will mean-revert the price, only a monetary regime that will or will not persist. Analysts who apply supply-demand-deficit reasoning to gold (as is legitimately done for silver, where industrial demand and by-product supply genuinely bind) are using the wrong tool.

Verdict on the metal. Structurally supported, cyclically corrected, not cheap. The official-sector bid is a real and novel floor; the drawdown was a leverage-and-real-rates event rather than a demand failure; but the metal has doubled in under two years and sits above a reduced consensus’ implied entry point only if you believe the sell-side, which just cut by 9–25%. For PHYS the relevant conclusion is narrower and more useful: this is an asset class that delivers amplitude, and amplitude is precisely what makes an extra 20–50bp of annual drag and a floating 2.5% spread easy to overlook and expensive to bear.

3.2 The gold-ETP industry — the one that decides the wrapper

This is a textbook commoditised fee war, and it is the single most important structural fact in this article.

(Fact — sponsor disclosures and industry compilations, accessed 2026-07-25; PHYS MER per sprott.com.)

Every one of these delivers the same thing: allocated bullion priced off the same LBMA benchmark. Industry commentary observes plainly that GLDM at 0.10% and IAU at 0.25% “delivered identical 22% trailing returns, making the fee gap the only real difference.” There is no product differentiation on the underlying, the marginal cost to a competitor of cutting price is close to zero, and the sponsors cutting price are the largest asset managers on earth, for whom a gold ETP is a loss-leader in a broader relationship.

Assess this with Greenwald’s structural test . Is there a barrier to entry? No — GLDM and AAAU are recent entrants that took tens of billions of dollars precisely by undercutting. Is there customer captivity? Almost none — switching costs are a brokerage commission and, for a taxable holder, a capital-gains realisation, which is a genuine but one-directional lock (it deters leaving, it does nothing to attract new money). Are there economies of scale with captivity? Scale exists — PHYS is a $15bn pool with a fixed cost base of only $4.1m of operating expenses — but it confers no pricing power, as the fee history demonstrates conclusively.

Verdict on the industry. For a gold holder, this industry is excellent and improving: you can now own allocated bullion for 9 basis points. For a gold-ETP sponsor charging 39 basis points, it is structurally bad and deteriorating, and there is no visible mechanism by which it improves. PHYS competes in it on a feature set — tax treatment and redeemability — that appeals to a real but bounded segment. That is a defensible niche, not a good industry.

4. Competitive Position

For a passive trust, “competitive position” means one question: against every other way to own gold, why this one? The honest answer is a narrow, conditional, but genuine edge — decisive for one clearly-defined investor, irrelevant or negative for everyone else. This section makes the arithmetic explicit rather than asserting the conclusion.

4.1 The comparison set

4.2 Edge #1 — the tax structure (the real one, and the only durable one)

For US individuals, gold held through a US grantor trust — GLD, IAU, GLDM, SGOL, AAAU — or held directly, is a “collectible” under IRC §408(m), and long-term gains are taxed at the 28% maximum collectibles rate. PHYS, being a Canadian trust holding passive assets, is a Passive Foreign Investment Company (PFIC): at least 75% of its gross income is passive, and gains from the sale of commodities held as investment property are passive income for these purposes. Sprott makes a Qualified Electing Fund (QEF) election available by furnishing a PFIC Annual Information Statement, and a US non-corporate holder who makes a timely QEF election on IRS Form 8621 receives ordinary long-term capital-gains treatment (top 20%) instead. With the 3.8% net investment income tax, the top all-in rate is approximately 23.8% for PHYS versus ~31.8% for the grantor trusts — an ~8-percentage-point after-tax advantage on realised gains. (Fact — prospectus supplement 2025-10-29, “PFIC Status and Tax Considerations”; Sprott’s published US tax guide.)

The qualifiers are not decorative and a memo that omits them is selling rather than analysing:

• The election must be made in the first taxable year the holder owns units, and Form 8621 must be filed annually thereafter for as long as the units are held, including years with no distributions.
• A holder who fails to elect falls into the default §1291 excess-distribution regime, under which gain is allocated ratably over the holding period, taxed at the highest ordinary rate for each prior year, plus an interest charge. That outcome is materially worse than the 28% collectibles rate the holder was trying to avoid. The tax edge is thus a conditional benefit with a punitive failure mode.
• The edge is worth exactly zero in an IRA, 401(k) or other tax-deferred account; zero for a non-US holder; zero for a tax-exempt institution; and near-zero for a holder with a small gain or a short horizon.

4.3 Edge #2 — custody integrity and the redemption right

PHYS’s bullion sits at the Royal Canadian Mint, a Crown corporation and agent of the Canadian Government whose obligations “generally constitute unconditional obligations of the Canadian Government,” and which contractually “bears all risk of the loss of, and damage to” the metal in its custody. The bars are fully allocated and unencumbered — specific bars, not a pooled claim — and are not lent or rehypothecated. This is, on the documentation, the strongest custody arrangement in the category, and it is the core appeal to the constituency that distrusts paper claims on metal. (Fact.)

Units are redeemable for physical gold monthly at 100% of NAV. But read the mechanics before crediting the feature: redemption requests “must be for amounts that are at least equivalent in value to one London Good Delivery bar, or an integral multiple thereof, plus applicable expenses,” where an LGD bar weighs 350–430 troy ounces (typically 390–410). At ~$4,061/oz that is a minimum redemption of roughly US$1.6 million of metal, with handling, delivery and insurance borne by the unitholder, and subject to the sizes of bars the Trust happens to hold. There is also a cash redemption route, but it carries a 5% haircut — the FY2025 notes describe “the 5% reduction to the redemption price for Units redeemed for cash,” which is credited to Unit premiums and reserves and therefore accrues to the remaining unitholders. (Fact.)

Interpretation — and this is the correct way to understand the feature. The physical redemption right is not a retail benefit. At a $1.6 million minimum it is unusable by all but institutional holders, and the marketing framing that presents it as a consumer feature is misleading. Its actual function is as the discount-control valve: when the discount widens far enough to exceed the frictional cost of taking delivery, an institution can buy units on-exchange below NAV, redeem for bars at 100% of NAV, and pocket the spread. That arbitrage is what caps the discount — and the Capital Allocation section shows it firing hard in Q1-2026. It is a genuine structural benefit to all holders, but it operates indirectly, and it is worth being precise about that rather than repeating the brochure.

It is also worth correcting a persistent myth in this category: GLD does not lend its gold either. The “our metal isn’t lent out” claim is a real differentiator against unallocated pooled products and against unsecured metal accounts; against GLD, IAU or GLDM specifically, it is largely a branding edge rather than a substantive one. The genuine custody differences are the sovereign counterparty and the contractual assumption of loss risk by the Mint, which are real but which have never been tested by an actual loss at any of these vehicles.

4.4 Edge #3 — issuance discipline

The trust agreement provides, verbatim: “The Trust may not issue trust units except (i) if the net proceeds per trust unit to be received by the Trust are not less than 100% of the most recently calculated NAV per trust unit immediately prior to, or upon, the determination of the pricing of such issuance.” This is reinforced by National Instrument 81-102 paragraph 9.3(2), under which the issue price must not “cause dilution of the NAV of the Trust’s other outstanding securities at the time of issue” and must not “be a price that is less than the most recently calculated NAV per trust unit.” (Fact.)

Interpretation: this is genuinely excellent structural alignment and it is the best feature of the wrapper after custody. The at-the-market program is contractually incapable of diluting gold-per-unit. It can only print units when the market is paying a premium; at a discount it simply pauses. Very few equity issuers of any kind operate under a comparable hard constraint, and unitholders should value it. the Capital Allocation section shows it working.

4.5 The offsetting disadvantages

Cost. At 0.39% MER, PHYS is roughly 4.3x IAUM (0.09%) and 3.9x GLDM (0.10%), 2.3x SGOL (0.17%), 1.6x IAU (0.25%), and level with the legacy GLD (0.40%). On a $100,000 position the annual gap versus GLDM is $290 — every year, compounding, forever, for an identical claim on identical metal.

The discount. PHYS closed below NAV in every one of the last five fiscal years: −1.58% (2021), −2.49% (2022), −2.15% (2023), −2.19% (2024), −1.73% (2025), −2.21% (Q1-2026), and −2.54% at 24 July 2026. The Trust’s own reported average discounts on the NYSE Arca were ~1.2% (2021), ~1.6% (2022), ~1.6% (2023), ~1.6% (2024), ~1.8% (2025) and ~2.3% in Q1-2026 against ~1.3% in Q1-2025. The open-ended competitors, kept tight to NAV by continuous authorised-participant creation and redemption, carry no such spread. Interpretation: the discount is genuinely double-edged and must be presented as such. It is a realised loss for a holder who bought near parity and sells at a discount; it is a benefit to a new buyer, who acquires gold 2.5% below spot and loses only if the discount widens further. What it is unambiguously, for everyone, is an additional source of volatility that the alternatives do not have — and that shows up in the data (see Financial Quality and Valuation).

The trend in the discount is the concerning part. Average discount has widened in each of the last two annual observations and nearly doubled year-over-year in the most recent quarterly comparison (1.3% → 2.3%).

4.6 The crossover arithmetic

This is where the section has to land, because everything above is qualitative until it is priced.

• Cost gap versus GLDM: ~0.29%/yr on headline fees; ~0.50%/yr on realised all-in drag (the Financial Quality section shows PHYS’s actual drag at ~0.60%/yr against GLDM’s ~0.10–0.15%). Compounds annually.
• Tax benefit: ~8 percentage points of the realised gain, banked once, on exit.
• Crossover: a holder sitting on a 100% cumulative gain banks roughly 8 points of NAV on exit. At ~0.50%/yr the cost gap consumes that in approximately 16 years. At a 50% gain (~4 points of benefit) it consumes it in about 8 years.

Verdict. PHYS possesses a real, rules-based edge, but it is segmented, conditional and time-limited, not a moat. It is the right vehicle for a high-bracket, US-taxable, buy-and-hold holder with a large expected gain, a horizon measured in years but not decades, and the discipline to file Form 8621 every year — a genuine constituency, and one large enough to sustain a $15bn trust. For a tax-deferred account, a non-US holder, a tax-indifferent institution, a fee-sensitive allocator or anyone with a short horizon, GLDM or IAUM dominate on every axis that matters and the analysis is not close. Applying the standard directly: PHYS’s alleged moat cannot be tied to a financial outcome that would deteriorate without it — because it has already failed to produce the one outcome a moat would produce, which is pricing power. That is the subject of the next-but-one section.

5. Growth History and Forward Opportunities

For a bullion trust, “growth” has two meanings that must never be conflated: growth in the gold price, which drives NAV and the unit price but is not something the Trust does; and growth in the ounce count and unit count, which is the Trust’s only actual activity. Both have been substantial, for entirely different reasons and with entirely different significance.

AUM growth — overwhelmingly price. Total net assets: $5,008.0m (2021) → $5,746.1m (2022) → $6,531.7m (2023) → $8,608.2m (2024) → $15,976.5m (2025) → $17,275.6m (Q1-2026), and back to ~$14,910m at 24 July 2026 as gold corrected. That is a 3.2x increase from end-2021 to the Q1-2026 peak. Over the same period gold went from roughly $1,829/oz to $4,668/oz — a 2.55x move. Interpretation: the overwhelming majority of AUM growth is the gold price, not distribution success. This distinction matters because a fee earned on price appreciation is not evidence of a product winning share.

Ounce growth — the Trust’s actual activity. Gold held: 2,734,026 oz (2021) → 3,153,868 oz (2023) → 3,277,693 oz (2024) → 3,693,298 oz (2025) → 3,688,439 oz (Q1-2026) → 3,671,782 oz (24 Jul 2026). That is +35.0% over four and a half years, or roughly 7%/yr — a genuinely respectable accumulation of metal, and materially better than the flat-to-shrinking ounce counts many gold vehicles posted through the 2022–23 rate-hiking cycle.

But note the inflection. Ounces peaked at 3,693,298 on 31 December 2025 and have declined in each subsequent observation — to 3,688,439 at 31 March 2026 and 3,671,782 at 24 July 2026. The Trust is now shrinking in metal terms, by roughly 0.6% since the December peak. That is not a crisis, but it is a turn: for the first time in the window, redemptions plus fee-driven sales are outrunning creations. (Fact.)

Unit growth — the funding. Units outstanding: 343,316,087 (2021) → 397,346,037 (2022) → 401,306,562 (2023) → 418,098,502 (2024) → 475,484,857 (2025), roughly 476.7 million at Q1-2026 — +38.5% over four years. This was funded entirely by the at-the-market program. Because issuance is NAV-floored, unit growth cannot by itself dilute gold-per-unit.

The number that reconciles the two — and the one that matters. Gold per unit:

Cumulative 2021→2025: −2.463%, or −0.62% per year compound. (Fact — computed from the Trust’s disclosed ounce and unit counts.)

Interpretation: this is the honest scorecard for a bullion trust, and it says the Trust’s gold-per-unit has eroded roughly 21 basis points a year faster than its stated MER of ~0.41%. Every ounce of “growth” in the metal pile was funded by new unitholders; the metal backing each existing unit has gone down every single year. That is not a criticism of the ATM — which, being NAV-floored, is not the cause — but it does dispose of any framing of PHYS as a vehicle that accumulates gold for you. It accumulates gold for the pool, and it charges each unit for the privilege of being in it. The Financial Quality section decomposes where the extra 21bp goes.

Forward opportunities — and there are genuinely few. This is a passive single-asset trust with no product roadmap. The realistic levers are: (i) more of the same — ATM issuance into premium windows, which requires premiums, which requires investor enthusiasm PHYS has not enjoyed in five years; (ii) fee competition — Sprott could cut the 0.35% management fee to defend share, which would be excellent for unitholders and directly contrary to the Manager’s economics; (iii) distribution — the Manager’s stated rationale for the ATM is that it “may increase liquidity for the trust units with the goal to make the Trust more available for institutional investors,” and that scale “may result in economies of scale which may lead to an ultimate decrease of expenses on a per trust unit basis.” Interpretation: that last promise is falsifiable and has been falsified. Net assets tripled from 2021 to 2025; the MER fell from 0.42% to 0.40%. Two basis points. Unitholders captured essentially none of the scale economies, because the dominant cost is a fixed 0.35% ad valorem management fee that does not step down with size — operating expenses, the only genuinely scale-sensitive component, are already down to 0.03% of average net assets. There is no mechanism by which further growth materially reduces the holder’s cost.

Verdict — low-quality growth in the only sense that matters to an existing holder. The Trust has grown its metal pile creditably (+35% in ounces over four and a half years) and has done so without diluting anyone, which is genuinely well-constructed. But growth has produced no fee reduction, no per-unit metal accumulation, and no competitive repositioning — and it has now stalled, with ounces down 0.6% from the December 2025 peak and quarterly redemptions running at 91% of quarterly creations. Growth in a vehicle whose per-unit economics deteriorate slightly every year is growth for the Manager, not for the unitholder. The distinction is the whole point.

6. Financial Quality

There is no income statement to analyse in any conventional sense, so this section does the analysis that actually determines a holder’s outcome: how faithfully does the vehicle deliver gold, and what does the gap cost? The answer is the most important finding in this article, and it is established two independent ways.

6.1 The five-year audited series

(Fact — Ratios and Supplemental Data, FY2025/FY2023/FY2021 40-F Ex-99.6; Q1-2026 6-K Ex-99.1. Year-end discount computed as closing price ÷ NAV per unit − 1.)

Balance sheet. The Trust is approximately 100% one asset: allocated gold bullion carried at fair value, plus a small cash sleeve, with no debt, no leverage, no covenants and no financing structure. There is nothing to stress-test. Liquidity risk is genuinely negligible — gold is among the most liquid assets on earth, and physical redemptions are settled in the metal itself rather than requiring a sale. This is a real strength and should be said plainly: whatever else is wrong with PHYS, it cannot blow up in the way a leveraged fund can.

6.2 The tracking analysis — method one: gold per unit

Established above: gold-per-unit fell from 0.00796358 oz (2021) to 0.00776744 oz (2025) — −2.463% cumulative, −0.62%/yr compound — against a stated MER averaging ~0.41% over the same period. That is roughly 21 basis points a year of erosion beyond the headline fee.

6.3 The tracking analysis — method two: the Trust’s own return table

The FY2025 annual report publishes annual compound returns to 31 December 2025 alongside spot gold. Computing the drag geometrically (the correct method; simple subtraction overstates it at high return levels):

(Fact — returns as published by the Trust; drags computed here.)

The corroboration is what makes this finding solid. Two methods that share no inputs — one built from ounce and unit counts, the other from published return series — both put all-in NAV drag at ~0.59–0.62%/yr against a 0.40–0.42% stated MER. The gap is real and it is roughly 20 basis points a year.

Where the extra 20bp goes. The mechanically identified source is the cash reserve. The prospectus provides that the Trust “will retain cash from the net proceeds of each of its offerings of trust units in an amount not exceeding 3% of the net proceeds of each such offering,” held to fund ongoing expenses and cash redemptions, replenished from time to time by selling bullion. In a year when gold rose 64.6%, a non-appreciating cash sleeve of up to 3% of recent proceeds is a material drag on a vehicle whose entire job is to track the metal. Add underwriting commissions ($5,112k in FY2025) and issuance/redemption timing effects, and the 20bp is accounted for. ASSUMPTION: this attribution is inferred from disclosed mechanics — the Trust does not publish a drag bridge, and a reader should treat the decomposition, though not the total, as inference.

6.4 The number that should govern the decision

Extend the since-inception figures. Over the 15.83 years from 3 March 2010 to 31 December 2025, the market return compounded 1.055%/yr behind spot gold. Compounded, that is a cumulative shortfall of approximately −15.5% versus simply owning the metal. Fees at ~0.42%/yr explain only about −6.4% of it. The remaining ~9 percentage points is non-fee drag — the cash sleeve, and the de-rating from PHYS’s early-life premium (the vehicle launched in 2010 into enthusiastic demand and traded at a premium) to its present persistent discount.

Interpretation: a holder who bought PHYS at inception in order to own gold has, fifteen years later, roughly 15% less gold-equivalent wealth than a holder who simply owned the metal — and only about 40% of that shortfall is the fee they agreed to pay. This is the central empirical fact about this security, it is derived entirely from the Trust’s own disclosures, and it is not disclosed in this form anywhere in the marketing.

In fairness, the counter-argument. The since-inception window begins at a premium and ends at a discount, so it embeds a one-time de-rating that cannot repeat indefinitely — the discount cannot widen forever. The 5- and 10-year market drags of 0.73–0.78%/yr are the more representative forward figures, and even those are roughly double the headline MER. Nothing in the honest version of this argument requires the 15-year number; the 10-year number is damning enough.

6.5 Expenses and quality of disclosure

FY2025 expenses: management fees $43.2m, sales tax $2.7m, operating expenses $4.1m (0.03% of average net assets). Q1-2026 management fees $15.7m, annualising to roughly $62.8m. Operating expenses at 3 basis points are genuinely low and reflect real scale — the problem is not the Trust’s cost base but the ad valorem fee sitting on top of it.

On disclosure quality — a point in the Trust’s favour. PHYS reports its own tracking shortfall against spot gold in every MRFP (“The Trust returned 63.2% compared to the return on spot gold of 64.6%”), and publishes its average discount to NAV every period. Many sponsors would omit both. The accounting is IFRS, audited, with the physical bullion count treated as a critical audit matter in the sibling trust’s audit. There is no aggressive accounting here, no non-GAAP gymnastics beyond a clearly-footnoted “operating expenses” measure, and no revenue-recognition judgement to worry about. The numbers in this article are damaging to the product, and every one of them came from the Trust’s own filings.

Verdict — do the economics improve with scale? No, and that is the finding. Net assets tripled between 2021 and 2025 while the MER fell two basis points, because the dominant cost is a fixed-rate ad valorem fee that does not step down. The genuinely scale-sensitive line — operating expenses — is already at 0.03% and cannot fall much further. Meanwhile the realised cost to a holder is not the 0.39% headline but roughly 0.60%/yr, and has been persistently so across every measurement horizon and by two independent methods. Financial quality, judged by the only standard that applies to a tracking vehicle — faithfulness to the underlying — is mediocre and structurally so.

7. Capital Allocation

There is no management team allocating capital here, which means this section must be redefined rather than skipped. For a closed-end bullion trust there are exactly two capital-allocation levers — the at-the-market issuance program and the redemption window — plus one question of incentive alignment between the Manager and the unitholders. The first is well-designed and well-executed. The second is working. The third is the clearest structural problem in the vehicle.

7.1 The ATM — well-constructed and used well

The mechanics were established above: units may not be issued unless net proceeds per unit are at least 100% of the most recently calculated NAV per unit, reinforced by NI 81-102 9.3(2). The current shelf, filed 28 October 2025 and supplemented 29 October 2025, authorises up to US$2,000,000,000 of units. The sales agreement permits commissions of “up to 3.0% of the aggregate gross proceeds.”

The history:

(Fact — Results of Operations, FY2025/FY2023/FY2021 40-F Ex-99.6 and Q1-2026 6-K Ex-99.1.)

Actual commissions were far below the cap: $5,112k on $1,702,893k of FY2025 gross proceeds = 0.30% of gross; $854k on $564,199k = 0.15% in FY2024. The Manager is not extracting value through the underwriting channel.

The apparent paradox, resolved. The Trust issued $1.70 billion of units in 2025 while reporting a 1.8% average discount for the year. Both are true. The NAV floor binds per issuance, not on average, and the reported discount is a mean over a highly volatile series. Issuance clusters in the premium windows — which, in 2025, coincided with gold’s near-vertical advance from September to January, when enthusiasm pushed PHYS to premiums for meaningful stretches. Interpretation: the Manager is effectively running a systematic “issue units into euphoria, convert the proceeds into bullion” program. For continuing unitholders that is accretive and it is exactly the right behaviour. Contrast it with the typical listed company, which reliably issues equity at cyclical lows and buys it back at highs; PHYS is structurally forbidden from doing the equivalent damage.

Verdict on the ATM: good. Correctly constrained, cheaply executed, used opportunistically. This is the strongest single element of the vehicle’s design.

7.2 The redemption valve — and it just fired

Compare Q1-2026 with Q1-2025. In Q1-2025 the Trust redeemed 53,000 units (plus 270 for cash) — essentially nothing. In Q1-2026 it redeemed 12,548,521 units for $514.4 million of gold bullion — against 13,746,168 units issued for $504.6 million. Redemptions rose roughly 237x year-over-year in unit terms and nearly offset creations entirely. Over the same comparison, the average discount widened from ~1.3% to ~2.3%. (Fact.)

Interpretation: this is the discount-control valve operating precisely as designed, and it is the best available evidence that the redemption right — however unusable at retail — is economically live. As the discount widened past roughly 2%, institutional holders found it profitable to buy units on-exchange below NAV and redeem them for physical bars at 100% of NAV, capturing the spread net of delivery and handling costs. That arbitrage consumes units and metal but it puts a floor under how wide the discount can go — and it is why PHYS’s discount sits at ~2.5% while its silver sibling PSLV’s ran to ~6.5% in July 2026. Gold’s far higher value density (an LGD gold bar is ~$1.6m versus a silver bar’s far lower value) means the fixed costs of taking delivery are a much smaller percentage of the metal collected, so the arbitrage bites at a much narrower spread.

The cash-redemption route reinforces it. Cash redemptions carry a 5% haircut, and that haircut is credited to Unit premiums and reserves — it accrues to the remaining unitholders. A holder who exits inefficiently subsidises those who stay. That is well-designed.

7.3 Incentive alignment — the structural problem

The Manager, Sprott Asset Management LP, earns 1/12 of 0.35% of net assets monthly, plus applicable sales taxes. At current NAV that is a run-rate of roughly $62.8 million per year. PHYS’s $17.3 billion at 31 March 2026 represented approximately 26.5% of Sprott Inc.'s $65.1 billion of total AUM — the single largest product in the group. (Fact — Sprott Inc. Q1-2026 results.)

Now state the alignment plainly:

• The Manager’s economics are levered entirely to AUM, which is levered entirely to the gold price and secondarily to unit issuance.
• There is no performance fee, no high-water mark, and — critically — no breakpoint or fee step-down at scale.
• The Manager’s revenue is completely insensitive to the two variables that determine a unitholder’s relative outcome: the fee level and the discount to NAV.

Interpretation: this is not misconduct and there is no evidence of any. It is a structural misalignment, and its consequence is visible in the fee history. Between 2021 and 2025 the Trust’s net assets tripled; the MER fell from 0.42% to 0.40% — two basis points, and a further one basis point since. Over the same window the competitive frontier moved from roughly 0.25% to 0.09%. Unitholders captured essentially none of the scale economies, and the Manager had no economic reason to hand them over. A fee cut from 0.35% to, say, 0.15% would transfer roughly $36 million a year from Sprott Inc. — whose 2025 revenue was $285 million — directly to unitholders. Nothing compels it, and nothing in the structure ever will.

The absent governance signal. It is worth noting what an analyst cannot see here. There are no Forms 3/4/5 and no DEF 14A for this issuer — no insider-transaction record, no say-on-pay, no director election, no compensation committee report. The insider-transaction read has no analog. Governance sits with the Manager, the Trustee (which owes a fiduciary duty to unitholders) and an Independent Review Committee. That is a legitimate structure for a passive trust, but it means the ordinary alignment diagnostics — is management buying? does compensation track shareholder outcomes? — simply do not exist. The absence is itself a finding.

Verdict — has capital been allocated intelligently? Mixed, and the split is instructive. The mechanisms are excellent: an ATM that cannot dilute you, used opportunistically into strength; a redemption valve that caps the discount and just proved it works; a cash-redemption haircut that pays the patient. Judged on mechanism design, this is better than the great majority of listed issuers. But the fee — the one lever that would return the benefits of scale to unitholders — has not moved in any meaningful way in five years while the competitive frontier collapsed, and the structure gives the Manager no reason to move it. Good machinery, wrong incentive on the one variable that matters most.

8. Changes and Headwinds — Last Two Years

The corporate record is genuinely, appropriately empty. Both the FY2025 annual MD&A and the Q1-2026 interim MD&A state verbatim: “There were no changes to the Manager of the Trust, nor were there any material changes to the investment objective or processes.” The FY2025 filing further confirms disclosure controls were evaluated and found effective, with no changes to internal control over financial reporting. Across the full trailing-60-month corpus — 21 Forms 6-K, 5 Forms 40-F, 3 Forms F-10, 5 prospectus supplements and 6 Forms F-X — there is no M&A, no divestiture, no partnership, no product launch, no litigation, no leadership change and no accounting-policy change to report. (Fact.) For a passive trust this is structurally correct and should not be mistaken for a lack of diligence in reporting it.

The changes that matter are therefore all market changes, and there are five.

1. Gold’s parabola and its break (Sept 2025 – Jan 2026). Gold rose 64.6% in calendar 2025, breaching $3,000 and $4,000 for the first time, and peaked near $5,589–5,595/oz on 28–29 January 2026. The top broke on 30 January on the Warsh Fed nomination plus CME margin increases; bullion fell more than 9.8% in a session, the worst day since 1983. Effect on thesis: neutral to the wrapper, everything to the holder. It re-established that gold is a high-amplitude asset in which a 20bp annual drag is easy to ignore and expensive to bear.

2. The five-month unwind and the sub-$4,000 break (Mar – Jun 2026). A dollar at a one-year high, a hawkish Fed under Chair Warsh, and the draining of the Middle East risk premium after US–Iran de-escalation took gold below $4,000 on 24 June 2026, ~29% off the high. Effect: PHYS set its maximum five-year drawdown of −26.75% the same day.

3. Sell-side capitulation on targets (Jun – Jul 2026). Goldman $5,400 → $4,900; JPMorgan $6,000 → $4,500 (−25%, on 3 July 2026); Deutsche Bank ~$4,300 average Q3. Effect: consensus remains 11–21% above spot, but the direction of revision is down and recent.

4. The discount has widened — the wrapper-specific headwind. Average NYSE Arca discount went from ~1.3% in Q1-2025 to ~2.3% in Q1-2026, and stands at −2.54% at 24 July 2026. Reported year-end discounts have been negative for five consecutive years. Effect on thesis: directly negative for existing holders, modestly positive for new buyers. This is the one genuinely PHYS-specific deterioration in the window and it deserves more weight than the gold noise.

5. Flows have turned — ounces are now shrinking. Gold held peaked at 3,693,298 oz on 31 December 2025 and has fallen to 3,671,782 oz — down 0.6%. Q1-2026 redemptions (12.5m units) reached 91% of creations (13.7m units), against effectively zero redemptions in Q1-2025. Effect: the arbitrage valve is doing its job, but the Trust has stopped growing in metal terms for the first time in the window.

A structural headwind that predates the window and will outlast it. The competitive frontier in gold ETPs continued to move against PHYS. GLDM (0.10%) has grown past $25 billion and IAUM sits at 0.09%, while PHYS’s MER moved from 0.41% to 0.39%. Nothing in the last two years suggests this reverses.

Verdict — do these changes strengthen or weaken the thesis? They weaken it, though less dramatically than the price chart implies. The gold drawdown is cyclical and, on the evidence, positional rather than structural — the central-bank bid that underpinned the move is intact. But the three PHYS-specific developments all point one way: the discount widened, ounces have begun to shrink, and the fee gap versus the cheapest alternatives widened further. None is individually alarming; together they describe a vehicle losing relative ground in a market that has stopped rewarding it.

9. Risk Analysis

The risk profile of PHYS is unusual: the catastrophic-loss risks that dominate a normal equity analysis are close to absent, while a set of quiet, chronic, low-drama risks do essentially all of the damage. The matrix reflects that.

On the shape of the risk profile. Two observations that a matrix alone obscures.

First, the tail risks here are unusually benign and the chronic risks unusually corrosive. There is no leverage, no debt maturity, no covenant, no customer concentration, no technological obsolescence, no key-person dependency and no realistic path to a zero. What there is instead is a set of small, certain, compounding costs — 20bp of excess drag, 29bp of excess fee, a 2.5% floating spread — that have already cost a since-inception holder about 15% of their gold. Investors systematically over-weight the first category and under-weight the second, and PHYS is a vehicle where that bias is expensive.

Second, risk #5 deserves more attention than it usually gets. The QEF election is marketed as PHYS’s principal advantage, but it is a conditional advantage with a punitive failure mode. A holder who buys PHYS for the tax treatment and then fails to file a timely Form 8621 in year one lands in the default §1291 regime — ratable allocation of gain over the holding period, taxed at the highest ordinary rate applicable in each prior year, plus an interest charge on the deferred tax. That is materially worse than the 28% collectibles rate they were trying to escape. The advantage is real, but it is an advantage that must be executed, annually, correctly, and the marketing does not lead with that.

10. Valuation Discussion — Embedded Expectations

Conventional valuation does not apply, and saying so is the honest answer rather than an evasion. PHYS has no earnings, no cash flow, no book value in the equity sense, no revenue, no terminal value and no reinvestment rate. P/E, EV/EBITDA, EV/Sales, P/B, DCF and Greenwald’s earnings-power value are all undefined or degenerate for this issuer. Even the repo’s usual highest-signal valuation datum — AZI’s own-history valuation_index percentile ranks — returned no usable payload for PHYS, which is correct behaviour rather than a data gap: a trust with no earnings, no equity book and no sales has no P/E, P/B or P/S history to rank against itself.

There is exactly one valuation metric: price to NAV. Its only input is the discount, and its history is short and consistent.

At −2.54%, today’s discount is at the wide end of its own five-year year-end range (−1.58% to −2.49%). Interpretation: for a new buyer this is a modest positive — you are acquiring gold roughly 2.5% below spot, and history suggests the redemption arbitrage caps how much further that can widen. For an existing holder it is a modest realised negative.

10.1 What the price embeds — decomposed

A buyer at $30.64 is making two independent bets that are routinely conflated and should not be.

Bet 1 — the metal (~97.5% of the position). You are buying 0.00774 oz per unit at an implied spot of roughly $4,061/oz. What must be true for that to work out: the debasement trade survives a hawkish Fed; central banks keep absorbing ~60 tonnes a month; and real rates do not grind higher for years. What the market is not pricing: either the January mania ($5,589) or a genuine bear case. Reduced sell-side consensus of $4,500–4,900 sits 11–21% above spot — but those targets were cut by 9–25% within the last six weeks, and target revisions in commodities lag price rather than lead it.

Bet 2 — the wrapper (~2.5%). You are buying that gold 2.5% below NAV, paying ~0.60%/yr of realised all-in drag to hold it, and receiving an ~8-point tax option on eventual realised gains.

10.2 Scenario analysis — the wrapper only

The following deliberately holds gold flat so the reader can see what the structure alone contributes over a five-year hold. This isolates the only thing an analyst can actually assess about PHYS as distinct from gold.

ASSUMPTION: drag of 0.60%/yr — the observed 5- and 10-year figure, not the 0.39% headline MER. Using the headline would flatter every row by roughly 20bp/yr and would be the more flattering and less accurate choice.

Interpretation: in the base case, the wrapper costs a holder roughly 2.5 percentage points relative to GLDM over five years before tax, and the QEF election repays that only if a substantial gain is realised. In the bull case — discount closes to parity — PHYS roughly matches GLDM before tax and beats it decisively after tax. In the bear case it loses on every axis. The distribution is not symmetric in PHYS’s favour, and the base case is not a win.

10.3 The crossover — the single most decision-relevant calculation

• Annual cost gap vs. GLDM: ~0.29%/yr on headline fees; ~0.50%/yr on realised drag. Compounds.
• Tax benefit: ~8 percentage points of the realised gain. Banked once, on exit.
• Crossover at a 100% cumulative gain: ~8 points of benefit ÷ ~0.50%/yr ≈ 16 years.
• Crossover at a 50% cumulative gain: ~4 points ÷ ~0.50%/yr ≈ 8 years.

Interpretation: PHYS is the correct vehicle inside a genuinely narrow window — a large expected gain, a high US tax bracket, a taxable account, and a horizon long enough to accumulate the gain but short enough that the fee gap has not eaten it. Outside that window — tax-deferred accounts, non-US holders, tax-exempt institutions, small gains, or holding periods beyond roughly two decades — the arithmetic favours the 0.09–0.10% alternatives, and it is not close.

10.4 The volatility the fee schedule does not disclose

One further cost belongs in a valuation section because it is invisible in the MER. Idiosyncratic volatility, from the same factor model, annualised:

Interpretation: PHYS carries measurably more non-gold volatility than every plain-vanilla peer. Netting the common component in variance terms, the excess is roughly 3.7% annualised of PHYS-specific noise that a GLD or IAU holder does not bear. Since PHYS holds the identical asset in an arguably safer vault, that excess is the premium/discount channel: the closed-end structure converts a tightly-tracking product into one with a floating, mean-reverting but unpredictable spread. It is a real cost of the structure, it does not appear in any fee table, and it is the quantitative expression of the qualitative point made throughout the Competitive Position section. These are third-party statistical estimates, not primary data, and the peer-relative comparison is more reliable than any single absolute level.

Verdict. There is nothing to re-rate and no multiple to expand. The entire PHYS-specific value question is whether a ~2.5% discount narrows or widens, and whether an ~8-point tax option is worth a ~0.50%/yr compounding cost gap. At today’s discount — the wide end of its own range — the entry is modestly favourable, and the vehicle is defensible for a specific, identifiable, minority holder type. For the median gold buyer, the embedded expectation required to justify PHYS over GLDM is one most buyers cannot satisfy: that they are in the top US bracket, in a taxable account, and will realise a large gain on a well-defined horizon.

11. Variant Perception

The consensus belief. PHYS is widely understood — and marketed — as “the safe way to own gold”: allocated bars in a sovereign vault, redeemable for physical, with a tax advantage over the ETFs. The implicit consensus conclusion is that these features justify the fee, and that the discount is a curiosity or even a free lunch for patient buyers. Sprott’s own product page leans on this directly, framing PHYS as the way not to “overpay for gold.”

Where consensus is right. The custody is best-in-class on the documentation. The NAV floor on issuance is genuinely superior alignment. The redemption valve does work, and Q1-2026’s 12.5 million redeemed units prove it empirically. The tax edge is real and worth roughly 8 points. None of this is marketing fiction and a sceptical memo should say so.

Where consensus is wrong — the variant perception. The consensus prices the features and ignores the arithmetic. Three specific errors:

1. It anchors on the 0.39% MER when the realised cost is ~0.60%/yr. Two independent methods — gold-per-unit from ounce counts, and the Trust’s own NAV-versus-spot return table — put all-in drag at 0.59–0.62%/yr across every horizon measured. Roughly 20bp a year of that is not in the fee table. Almost no one holding PHYS knows this, because it requires computing gold-per-unit across years, which nobody does.

2. It treats the discount as a static feature rather than a compounding de-rating. Since inception the market return has trailed spot gold by 1.055%/yr — a cumulative ~15.5% — of which fees explain only 6.4 points. The vehicle launched at a premium and now sits at a persistent discount, and that de-rating was a permanent, one-way transfer away from early holders. The 10-year market drag of 0.73%/yr is nearly double the headline fee.

3. It over-weights an 8-point one-time tax benefit against a 0.50%/yr perpetual cost, without doing the crossover. The crossover is roughly 16 years at a 100% gain and 8 years at a 50% gain. Very few holders have done that calculation, and the calculation is the entire decision.

The strongest bull case. For a specific and real constituency, PHYS is genuinely the best available vehicle. A US taxable investor in the top bracket, allocating 3–5% to gold as a strategic monetary hedge, expecting to hold for a decade and to realise a large gain, gets: allocated sovereign-vaulted bars with contractual loss-bearing by a Crown corporation; a structural guarantee against dilution that no open-ended competitor offers; an arbitrage-capped discount that today lets them buy metal 2.5% under spot; and roughly 8 percentage points of after-tax advantage on exit. Against GLDM, that package wins on a 10-year horizon with a large gain. Add the possibility that the discount closes toward parity — which would add another 2.5 points — and the bull case is coherent, quantified, and not merely a story.

The strongest bear case. PHYS is a fee-uncompetitive, structurally-leaking wrapper on a commodity that has just fallen 29% from a leverage-driven blow-off. It charges four times the cheapest identical alternative for a benefit most of its holders cannot use — because a large share of gold ETP money sits in tax-deferred accounts or non-US hands, where the QEF edge is worth exactly zero. It has trailed the metal it holds by more than its fee across every measurement horizon in its history. Its per-unit gold has fallen every single year. Its discount is widening, its ounce count has started shrinking, and its Manager — who collects ~$63m a year and has no breakpoint, no performance fee and no exposure to the discount — has cut the fee by three basis points in five years while competitors went to nine. There is no mechanism in the structure that fixes any of this, because the entity that would have to fix it is the one being paid.

The 3–5 assumptions that actually matter.

1. That the holder is in the narrow window where the tax edge exceeds the fee gap. This is the single highest-leverage assumption in the entire analysis and it is a fact about the holder, not the security. Get it wrong and everything else is irrelevant.
2. That the discount does not structurally widen. The redemption arbitrage caps it, but the cap is set by the frictional cost of taking $1.6m of physical delivery, not by any contractual floor. If fee-aware money continues to leave for GLDM, the marginal buyer thins and the cap loosens.
3. That realised drag stays near 0.60%/yr and does not deteriorate further. It has been remarkably stable at 0.59–0.62% across 5- and 10-year horizons, which is reassuring.
4. That the central-bank bid holds. ~60 tonnes a month of price-insensitive buying is the floor under the entire gold complex and therefore under ~97.5% of a PHYS position.
5. That US collectibles and PFIC/QEF tax law is unchanged. The entire differentiator is a legislative artifact. It has been stable for decades, but it is not a property of the business — because there is no business.

The factor-positioning read, and what it says about where consensus may be offside. The empirical picture is unambiguous and worth stating precisely because it disciplines the narrative. PHYS’s dominant loading is GoldPrice at +1.63, with R² 0.887; every other exposure is an order of magnitude smaller (Momentum +0.078, Sector:Materials +0.106, USDollar −0.048, Liquidity −0.062, BetaFactor −0.027 in the All-Factors model). The factor model’s own “related stocks” list returns UGL, SGOL, OUNZ, IAU, GLDM, AAAU, GLD and IAUM at similarities of 0.997–0.998 — an independent confirmation that the market treats PHYS as interchangeable with vehicles charging a quarter of its fee. On the tape: the three-month Sharpe is −1.96 and the six-month −1.08 (actual moves of −14.3% and −19.5% respectively, de-annualised), the unit trades below its 21-, 50- and 200-day moving averages, and the maximum drawdown of −26.75% was set on 24 June 2026. Against that, the three-year Sharpe is 1.14 and the ten-year 0.53.

Interpretation: this is a decelerating falling knife off a genuine blow-off top, not a broken thesis. The long-horizon risk-adjusted record is intact; the damage is entirely a gold-price event; and the wrapper performed exactly as designed through the drawdown — the redemption valve fired and capped the discount at ~2.5% rather than letting it blow out to PSLV’s ~6.5%. Where consensus may be offside is not on the direction of gold — that debate is well-populated and the sell-side has already capitulated on the mania. It is on the cost of the wrapper: a market that prices PHYS as interchangeable with GLDM (similarity 0.997) while PHYS charges 3.9x the fee and delivers ~3.7pp more idiosyncratic volatility is a market that has not done the arithmetic. Regime caveat: these are in-sample statistical estimates over a 756-day window, and the loadings describe what PHYS has been, not what it will be.

12. Fact vs. Interpretation

13. Open Questions

1. What is the exact composition of the ~20bp of annual drag beyond the MER? The Trust discloses the ≤3% cash-retention policy and the commission line, but publishes no bridge from MER to realised tracking difference. The total is established by two independent methods; the decomposition is inference. Resolvable only by the Manager.

2. What are the intra-period extremes of the premium/discount? Sprott publishes daily NAV but no accessible machine-readable historical spread series. This memo uses the Trust’s own disclosed year-end and quarterly-average figures, which are primary and reliable — but the intra-year maxima and minima, and specifically how high the premium ran during the Sept-2025-to-Jan-2026 ATM window, are not established here.

3. At what discount, precisely, does the physical-redemption arbitrage become economic? The Q1-2026 data show it firing between ~1.3% and ~2.3%, but the full cost stack for an institutional redeemer — handling, freight, insurance, assay, the bid/offer on disposing of LGD bars — is not disclosed. This threshold is the effective floor under the discount, so it is the single most valuable unknown in the analysis.

4. What proportion of PHYS’s holder base is actually in the tax window? The entire bull case rests on holders being US taxable, top-bracket, long-horizon and QEF-electing. If a large share of the $14.9bn sits in IRAs or non-US accounts, those holders are paying a 29bp premium for nothing, and are the most likely marginal sellers to GLDM. Not disclosed anywhere.

5. Will Sprott ever cut the management fee? The competitive frontier moved from ~0.25% to 0.09% while PHYS moved 0.42% to 0.39%. Is there an internal AUM or outflow threshold that triggers a cut? Nothing in the filings or Sprott Inc.'s disclosure addresses it.

6. How would a genuine custody stress test resolve? The Mint “bears all risk of the loss of, and damage to” the bullion, “subject to certain limitations, including events beyond the Mint’s control.” Those carve-outs have never been tested at any bullion trust. The strength of the guarantee is documentary, not empirical.

7. Does the ounce-count decline since December 2025 mark a turn or a blip? Ounces are down 0.6% from peak and Q1-2026 redemptions hit 91% of creations. One quarter is not a trend, but it is the first such quarter in the window.

14. What Must Be True

14.1 For the bull case

14.2 For the bear case

The asymmetry worth naming. Bull assumption #4 is unlike every other line in both tables: it is not a forecast about the world but a verifiable fact about the buyer, knowable with certainty before any capital is committed. That is unusual and it is the practical takeaway from this entire memo. Most investment theses require you to be right about the future. This one primarily requires you to be honest about your own tax situation — and if you are not in the window, no amount of being right about gold rescues the decision, because you could have owned the same metal in the same form for a quarter of the price.

15. Source Appendix

The full source appendix is reproduced as Appendix B below. Primary sources are the Trust’s own SEC filings (Form 40-F for FY2021, FY2023, FY2024 and FY2025; Form 6-K for Q1-2026 and prior interims; the Form F-10 shelf and the October-2025 prospectus supplement), Sprott’s published product and tax-guide disclosures, the AZI five-year price series, the FactorsToday factor model, and contemporaneous financial reporting for gold-market events. Every quantitative claim in this article traces to one of those, or to a computation on them that is shown explicitly in the text.

This article contains no recommendation and no price target outside the clearly-labelled Claude's Take block at the top, which is the author’s own subjective view. Nothing herein should be construed as investment advice, and no statement should be read as indicating that the author holds, has held, or intends to hold any position in the security discussed. Readers should do their own research and consult a qualified tax adviser before acting on anything relating to the PFIC/QEF election discussed below.

APPENDIX A — Standard Diligence Questionnaire

Sprott Physical Gold Trust (NYSE Arca / TSX: PHYS) — 2026-07-25

A standard diligence checklist applied to PHYS. Where a question does not map to a passive bullion trust, that is stated plainly and the correct sector analog is given rather than the question being skipped.

General

What thoughtful questions have other investors asked about this company?

The serious questions asked about PHYS cluster into five, and they are the right five:

1. “Is the gold actually there?” — the foundational question for any bullion vehicle, and the reason PHYS exists as a product. Answer: the bars are fully allocated, unencumbered LGD bars verified against the LBMA source, held at the Royal Canadian Mint, a Crown corporation that contractually bears the risk of loss. This is as strong as the category offers.
2. “Is the tax election worth the fee?” — the actual investment question, and the one most often asserted rather than computed. The crossover is roughly 16 years at a 100% gain and 8 years at a 50% gain (see the Valuation section).
3. “Why does it trade at a discount, and will it close?” — PHYS has closed below NAV in each of the last five years. The redemption arbitrage caps it near 2–3%; nothing forces it to parity.
4. “Can the ATM dilute me?” — no. The trust agreement prohibits issuance below 100% of the most recent NAV per unit. This is the question with the most reassuring answer.
5. “Can I actually redeem for metal?” — technically yes, practically no for retail: the minimum is one London Good Delivery bar, roughly US$1.6 million of gold at current prices, plus costs. Its real function is as the institutional arbitrage that caps the discount.

A sixth question that is not commonly asked, and should be: “How much has this vehicle actually trailed gold, over its life, versus what it says it charges?” The answer — ~1.06%/yr since inception, ~15.5% cumulative, against a ~0.42% fee — is derivable entirely from the Trust’s own filings and is, in my view, the most decision-relevant fact about the security.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low?

Does not map. PHYS has no earnings — total revenue was nil in each of the last five fiscal years, and its entire reported “net income” is the mark-to-market on gold (FY2025: $5,948.6m, of which $5,849.3m was unrealised). The correct analog is: is the gold price at a cyclical high or low? Gold peaked near $5,589–5,595/oz on 28–29 January 2026 and now sits near $4,061/oz — roughly 27% below the peak but still ~55% above the end-2024 level of $2,624.50 and roughly double the 2023 level. So: off a cyclical high, not at a cyclical low.

Driven by the external environment or internal actions?

Entirely external. There are no internal actions. The factor model puts the GoldPrice loading at +1.63 with an R² of 0.887 — ~89% of return variance is one exogenous variable. The Manager cannot influence it and does not try to; the mandate expressly forbids speculating on short-term gold moves.

How stable are revenues?

Does not map — there are none. The correct analog is the stability of the Manager’s fee base, which is 0.35% of NAV and therefore as volatile as gold itself. Management fees went $26.8m (FY2024) → $43.2m (FY2025) → $15.7m in Q1-2026 alone (a ~$62.8m run-rate). That is a +61% year, driven purely by the gold price. From a unitholder’s perspective the relevant stability question is the cost, which is genuinely stable: MER has sat in a 0.39–0.42% band for five years.

Outlook for products/services?

There is one product and it will not change: allocated gold bullion. There is no roadmap, no R&D, no line extension. The only variable within the Manager’s control is the fee, and it has moved three basis points in five years.

How big will this market be — growing, shrinking, domestic or international?

Two markets. Gold is a ~$30-trillion-plus above-ground stock (~215,000 tonnes) against ~3,600 tonnes of annual mine supply — a global, deeply liquid, non-growing-in-the-usual-sense monetary asset. Gold ETPs are growing but the growth is accruing to the cheap end: GLD holds $168bn at 0.40%, IAU >$61bn at 0.25%, GLDM >$25bn at 0.10%. PHYS at ~$14.9bn is mid-sized and competing at the expensive end of a market whose flows favour the cheap end. The market is international; PHYS is Canadian-domiciled, US- and Canada-listed, and its differentiating feature is specific to US taxable investors.

Business Quality & Competitive Moat

Is the industry getting more or less competitive?

Decisively more. Over the five years in which PHYS’s MER went 0.42% → 0.39%, the competitive frontier went from roughly 0.25% to 0.09% (IAUM) and 0.10% (GLDM). GLDM has grown past $25bn and AAAU (Goldman, 0.18%) entered and reached ~$2.4bn. Every one of these holds allocated bullion priced off the same LBMA benchmark. Industry commentary notes GLDM and IAU delivered “identical 22% trailing returns, making the fee gap the only real difference.” This is a commoditised fee war and PHYS is on the wrong side of it.

How profitable is the business (ROIC, ROE)?

Does not map. ROIC and ROE are undefined for PHYS: there is no invested capital earning a return, no operating profit, and the “equity” is simply the gold. Reporting a ratio here would be meaningless. The correct analogs:

• For the unitholder: the return is gold’s return, less realised drag of ~0.60%/yr and the change in the discount.
• For the Manager: the economics are outstanding — 0.35% of ~$15bn (~$63m/yr run-rate) against a Trust cost base of $4.1m of operating expenses (0.03% of average net assets). PHYS is ~26.5% of Sprott Inc.'s $65.1bn AUM. The profitable business here is the management contract, not the trust.

How profitable is the industry — how many competitors, what barriers to entry?

At least eight credible US-listed gold vehicles (GLD, IAU, GLDM, IAUM, SGOL, AAAU, OUNZ, PHYS) plus Sprott’s own CEF. Barriers to entry are effectively nil — GLDM and AAAU are recent entrants that took tens of billions by undercutting. Under Greenwald’s taxonomy there is no supply/cost advantage, no demand-side captivity beyond the one-directional lock of an unrealised capital gain, and no economies of scale with captivity: PHYS has scale ($15bn) but it confers no pricing power, which the fee history proves conclusively. Industry profitability is migrating to the two largest sponsors on earth, for whom a gold ETP is a relationship loss-leader.

Can the business be easily understood?

Yes — exceptionally. This is one of the simplest securities in existence: ounces × spot ÷ units, less a fee, plus or minus a spread. The complexity is entirely in the tax and in the drag arithmetic, and both are tractable. That said, the simplicity is deceptive: the 20bp of drag beyond the headline fee, and the 15-year cumulative shortfall versus the metal, are invisible unless you compute gold-per-unit across years — which almost no holder does.

Can it be undermined by foreign low-cost labour?

Not applicable. There is no labour input. The nearest analog — undermined by lower-cost capital — is not only possible but actively happening: IAUM and GLDM at 0.09–0.10% are precisely a low-cost substitute doing to PHYS what low-cost labour does to a manufacturer.

Do brands matter?

Marginally, and less than Sprott would like. “Sprott” carries genuine standing with the hard-money constituency, and that brand plus the Royal Canadian Mint association is a real reason some holders pay 4x. But the factor model’s own related-stock list puts PHYS at 0.997–0.998 similarity to GLD, IAU, GLDM, SGOL, OUNZ and AAAU — the market treats these as interchangeable. A brand that cannot sustain a price premium against an identical substitute is a preference, not a franchise.

What is the nature of competition?

Almost purely price, secondarily structure (tax treatment, redeemability, custody domicile) and liquidity. PHYS competes on structure because it cannot compete on price; GLD competes on liquidity because it cannot compete on price either; GLDM and IAUM compete on price and are winning the flows.

Customers’ switching costs?

Low in one direction, high in the other — and this asymmetry is important. Switching into PHYS costs a brokerage commission. Switching out costs a realised capital gain, which for a holder with a large embedded gain is a powerful lock-in. That lock-in is real but it is entirely one-directional: it retains existing holders, it does nothing to attract new ones. A business whose only customer captivity is the tax cost of leaving is not a business with pricing power — it is a business with a decaying installed base.

Financial Condition & Balance Sheet

Assets not fully recognised on the balance sheet?

No. The balance sheet is ~100% one asset — allocated gold bullion at fair value — plus a small cash sleeve. Everything is marked and recognised. There is no goodwill, no intangible, no equity-method investment, no deferred cost. If anything the reverse is true: the balance sheet carries a very large unrealised gain (gold bought far below current prices) that is fully recognised at fair value under IFRS but represents an embedded deferred tax liability at the holder level, not the Trust level — which is exactly what the QEF election is designed to manage.

Off-balance-sheet liabilities?

None identified. No debt, no leases, no pension, no guarantees, no derivatives, no securities lending, no rehypothecation. The bullion is explicitly “unencumbered.” The nearest thing to an off-balance-sheet obligation is the monthly redemption right, which is a claim on the Trust’s own assets settled in kind rather than a liability requiring external funding — and it is settled in the very asset the Trust holds, which is why liquidity risk is genuinely negligible.

How conservative is the accounting?

Conservative and, notably, self-critical. IFRS, audited, single-asset fair value with a Level 1 input (spot gold) — there is essentially no accounting judgement to exercise. More to the point on integrity: the Trust discloses its own underperformance versus spot gold in every MRFP (“The Trust returned 63.2% compared to the return on spot gold of 64.6%”) and publishes its average discount to NAV each period. Many sponsors would omit both. The only non-GAAP measure is a clearly-footnoted “operating expenses” figure that reconciles explicitly to total expenses less management fees, sales tax and FX. Every unflattering number here came from the Trust’s own filings.

How CapEx-hungry is the business?

Does not map — there is no capital expenditure. The analog is storage and custody cost, which is embedded in operating expenses of $4.1m, or 0.03% of average net assets. That is genuinely low and reflects real scale. The Trust’s cost problem is not its cost base; it is the 0.35% ad valorem management fee sitting on top of a 3bp cost base.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy?

Does not map. PHYS generates no free cash flow — it holds a non-income-producing asset and pays no distributions. It is cash-consuming: it periodically sells bullion to fund expenses and cash redemptions, which is precisely why gold-per-unit declines every year. The philosophy analog is the issuance/redemption discipline, and here the Trust scores well: units may not be issued below 100% of the most recent NAV per unit (trust agreement, reinforced by NI 81-102 9.3(2)), so the ATM is contractually incapable of diluting gold-per-unit. It can only print into a premium.

Significant acquisitions recently?

None, and none possible. The Trust is prohibited from investing in other funds except money-market funds. Its only “acquisition” is buying more gold with ATM proceeds — $1,702.9m of gross proceeds deployed into bullion in FY2025 and $504.6m in Q1-2026.

Buying back shares?

No, and it structurally cannot in the usual sense. The economic equivalent is the redemption window, and it is currently very active: Q1-2026 saw 12,548,521 units redeemed for $514.4m of gold, against just 53,000 units in Q1-2025. Because redemptions occur at 100% of NAV while the units trade at a discount, this is economically similar to a buyback below intrinsic value — the redeeming holder captures the spread, and the Trust shrinks. Additionally, cash redemptions carry a 5% haircut that is credited to Unit premiums and reserves, accruing to the remaining unitholders. The patient are subsidised by the impatient, which is well-designed.

Issuing large amounts of new shares to insiders?

No. There is no stock-based compensation, no option pool, no insider grant mechanism — the Trust has no employees. Units issued in FY2025 (66,426,933 for $1,702.9m gross) went through an at-the-market program to public buyers at or above NAV, with commissions of just 0.30% of gross proceeds (well under the 3.0% contractual cap).

Compensation policy of directors/management?

Does not map, and the absence is itself a finding. There is no board of directors elected by unitholders, no DEF 14A, no say-on-pay, and no Forms 3/4/5 in the entire 60-month SEC corpus. Governance sits with the Manager, the Trustee (which owes a fiduciary duty to unitholders) and an Independent Review Committee. Compensation is a single line: the Manager receives 1/12 of 0.35% of NAV monthly, plus sales taxes. There is no performance fee, no high-water mark and no AUM breakpoint. The ordinary alignment diagnostics an analyst would run — is management buying? does pay track shareholder outcomes? — simply do not exist here.

Motivations of management?

State it plainly: the Manager is paid on AUM and nothing else. Sprott Asset Management earns ~$62.8m/yr at current NAV; PHYS is ~26.5% of Sprott Inc.'s $65.1bn total AUM and Sprott’s 2025 revenue was $285m. The Manager’s revenue is completely insensitive to the two variables that determine a unitholder’s relative outcome — the fee level and the discount to NAV. The consequence is visible and quantified: net assets tripled from $5.0bn (2021) to $16.0bn (2025) while the MER fell two basis points, and unitholders captured essentially none of the scale economies. A cut from 0.35% to 0.15% would transfer ~$36m/yr from Sprott Inc. to unitholders. Nothing in the structure compels it. This is not misconduct — there is no evidence of any — it is structural misalignment on the one lever that matters most.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer?

None of the three — but the tax classification is more consequential than any of them. PHYS is a Canadian closed-end mutual fund trust whose units (not shares, not ADRs) list directly on the NYSE Arca and the TSX. It issues no K-1. It is, however, a Passive Foreign Investment Company (PFIC) for US holders, which means:

• A US non-corporate holder should make a timely QEF election on IRS Form 8621 in the first taxable year of ownership, and must file Form 8621 annually thereafter for as long as the units are held, even in years with no distributions.
• Doing so converts what would be a 28% collectibles rate (IRC §408(m), applicable to GLD/IAU/GLDM/SGOL/AAAU and to direct bullion) into long-term capital-gains treatment (top 20%) — roughly 23.8% vs 31.8% all-in with the 3.8% NIIT, an ~8-percentage-point advantage on realised gains.
• Failing to elect is worse than not bothering: the default §1291 excess-distribution regime allocates gain ratably over the holding period, taxes it at the highest ordinary rate for each prior year, and adds an interest charge. This is materially worse than the 28% the holder was trying to avoid.

This is the single most important operational fact for a prospective US holder, and it is a compliance obligation, not a passive benefit.

Dividend policy?

No distributions. The Trust holds a non-income-producing asset and “does not anticipate making regular cash distributions to unitholders.” Trailing dividend yield is 0%. Total return is purely price change.

How profitable is the business?

Covered above: undefined for the Trust, excellent for the Manager. Worth restating because it is the crux — the profitable entity in this structure is Sprott Asset Management, not Sprott Physical Gold Trust.

Is net income diverging from cash from operations?

Yes, and it is by design rather than a red flag. FY2025 net income of $5,948.6m was almost entirely unrealised ($5,849.3m of unrealised gains vs $147.6m realised). Cash from operations is structurally negative — the Trust sells bullion to fund expenses. In an operating company this divergence would be a serious quality-of-earnings warning; here it is the arithmetically necessary consequence of holding an appreciating non-income asset and paying fees out of it. The legitimate concern is not the divergence but its cumulative effect: paying fees out of principal is exactly why gold-per-unit fell every year (0.00796358 oz in 2021 → 0.00776744 oz in 2025), and why the vehicle has trailed the metal it holds by more than its stated fee across every horizon in its history.

Risks & Downside

What factors would cause the stock to decline?

In descending order of expected contribution:

1. A falling gold price — ~89% of return variance (GoldPrice beta +1.63, R² 0.887). Realised: −26.75% maximum drawdown, set 24 June 2026, as gold fell ~29% from its January peak. Drivers currently in force: a dollar at a one-year high, a hawkish Fed under Chair Warsh, and a drained geopolitical premium after US–Iran de-escalation.
2. A widening discount to NAV — currently −2.54%, the wide end of its five-year year-end range. The sibling PSLV reached ~6.5%.
3. Continued fee-driven outflows to GLDM/IAUM, thinning the marginal buyer and loosening the effective cap on the discount.
4. The chronic ~0.60%/yr drag, which is not a “decline” event but has already cost a since-inception holder ~15.5% relative to the metal.
5. An adverse change to PFIC/QEF or collectibles tax law, which would eliminate the only durable differentiator overnight.

Risk of a catastrophic loss?

Very low, and this is genuinely the vehicle’s strongest attribute. There is no leverage, no debt, no covenant, no counterparty derivative exposure, no customer concentration, no key-person risk and no technology risk. A catastrophic loss requires failure of custody at the Royal Canadian Mint — a Crown corporation whose obligations effectively constitute unconditional obligations of the Canadian Government, and which contractually bears all risk of loss and damage to the bullion in its custody. The honest caveat: that guarantee is subject to limitations including events beyond the Mint’s control, and it has never been tested by an actual loss at any bullion trust. The strength is documentary, not empirical. A further, non-trivial tail: sovereign confiscation or an adverse change in Canadian treatment of foreign-owned vaulted bullion — historically remote, but it is the reason the exposure is not literally riskless.

Chance of a total loss?

Negligible — effectively zero. It would require the simultaneous destruction or confiscation of allocated, sovereign-vaulted bullion and the failure of the Canadian Government’s backing. Gold cannot go to zero; it has been a monetary asset for five millennia. The realistic worst case is not loss of capital but a long, quiet, compounding underperformance of the very asset the vehicle exists to track — which is not hypothetical, because it has already happened to the tune of ~15.5% since inception.

Recent News & Events

Has the business environment changed recently?

Yes — materially, though almost entirely at the market level rather than the Trust level. Five changes in the last two years:

1. Gold’s parabola and its break. Gold rose 64.6% in 2025 to $4,319.37/oz and peaked near $5,589–5,595/oz on 28–29 January 2026. On 30 January the nomination of Kevin Warsh as Fed Chair, amplified by CME margin increases, broke the move — bullion fell >9.8% in a single session, its worst day since 1983; PHYS fell 10.92% on 48.4 million units.
2. The unwind and the sub-$4,000 break. A one-year-high dollar, a hawkish Fed and US–Iran de-escalation took gold below $4,000 on 24 June 2026, ~29% off the high.
3. Sell-side capitulation. Goldman cut year-end 2026 from $5,400 → $4,900 (June); JPMorgan cut Q4-2026 from $6,000 → $4,500 (3 July, −25%); Deutsche Bank ~$4,300 average Q3.
4. The discount widened — average NYSE Arca discount ~1.3% (Q1-2025) → ~2.3% (Q1-2026), and −2.54% at 24 July 2026.
5. Flows turned. Ounces peaked at 3,693,298 (31 Dec 2025) and have fallen to 3,671,782 — the Trust is now shrinking in metal terms for the first time in the window, with Q1-2026 redemptions at 91% of creations.

Significant acquisitions?

None (see Capital Allocation above).

Change in accounting policies?

None. No changes to internal control over financial reporting; disclosure controls evaluated and found effective in the FY2025 40-F.

Recent changes — new markets, facilities, management?

None whatsoever, and the filings say so verbatim. Both the FY2025 annual MD&A and the Q1-2026 interim MD&A state: “There were no changes to the Manager of the Trust, nor were there any material changes to the investment objective or processes.” Across the full trailing-60-month SEC corpus — 21 Forms 6-K, 5 Forms 40-F, 3 Forms F-10, 5 prospectus supplements, 6 Forms F-X, and zero Forms 3/4/5 or DEF 14A — there is no M&A, no divestiture, no partnership, no launch, no litigation and no leadership change. The one recurring corporate action is the shelf-registration cycle for the ATM: the current Form F-10 was filed 28 October 2025 and supplemented 29 October 2025, authorising up to US$2,000,000,000 of trust units.

For this issuer an empty event calendar is structurally correct and should not be mistaken for insufficient diligence. The events that matter to a PHYS holder are gold-price events and flow events, and both are covered above.

APPENDIX B — Source Appendix

Sprott Physical Gold Trust (NYSE Arca / TSX: PHYS) — 2026-07-25

All sources accessed 2026-07-25 or 2026-07-26 unless otherwise noted. Primary sources are listed first. Every quantitative claim in this article traces to a source below, or to a computation performed on one and shown explicitly in the text.

B.1 Primary — SEC filings by the Trust (CIK 0001477049)

The Trust is a Canadian MJDS filer and files Form 40-F (annual) and Form 6-K (interim), not 10-K/10-Q. The trailing-60-month corpus (since 2021-07-25) comprises 21 Forms 6-K, 6 Forms F-X, 5 SUPPL, 5 Forms 40-F and 3 Forms F-10. There are zero Forms 3/4/5 and zero DEF 14A — noted above as a finding in its own right.

B.2 Primary — Sponsor disclosures

B.3 Market and quantitative data

B.4 Competitive set — expense ratios and AUM

B.5 Gold market — events, drivers and forecasts

B.6 US tax treatment

B.7 Analytical frameworks

B.8 Sources deliberately not relied upon

• Sprott’s “Don’t Overpay for Gold” marketing page (sprott.com) — reviewed and not relied upon. It is promotional framing, and this article’s cost analysis reaches the opposite conclusion from the Trust’s own audited figures.
• companiesmarketcap.com PHYS premium/discount page — fetched but returned only a single data point (−2.58% at 2026-07-23) with no accessible history. This article’s discount series is instead computed from the Trust’s own disclosed year-end NAV and closing prices, which is primary and superior. Intra-period extremes are consequently not claimed anywhere in this article (noted as Open Question #2).
• Analyst price targets on PHYS — none exist, and none would be used. No third-party price target appears anywhere in this article.

B.9 Note on computed figures

Several of this article’s most important numbers are computations performed by us on primary data, not figures published by any source. They are flagged as computed in the Fact-vs-Interpretation table above and the arithmetic is shown in the text. Specifically:

• Gold per unit (0.00796358 oz in 2021 → 0.00776744 oz in 2025; −0.62%/yr) — computed from the Trust’s disclosed ounce counts ÷ units outstanding.
• NAV and market drag versus spot gold at each horizon — computed geometrically from the Trust’s published compound-return table, i.e. (1 + r_trust)/(1 + r_spot) − 1. Simple subtraction of the percentages, which is what a casual reader would do, materially overstates the drag at high return levels and is not used.
• Cumulative since-inception shortfall (~−15.5%) — (1.0782/1.0897)^15.83 − 1, using the Trust’s own since-inception market and spot returns over the 3 March 2010 to 31 December 2025 period.
• Year-end discounts — closing NYSE Arca price ÷ NAV per unit − 1, both as disclosed by the Trust.
• Excess idiosyncratic volatility (~3.7pp) — netted in variance terms from source 19: √(0.1036² − 0.0968²).
• Crossover horizons (~16 years at a 100% gain; ~8 years at a 50% gain) — the ~8pp one-time tax benefit divided by the ~0.50%/yr realised cost gap versus GLDM.
• Implied gold price (~$4,061/oz) — total net assets ÷ ounces held, from source 11.
• PHYS as ~26.5% of Sprott Inc. AUM — $17.2756bn ÷ $65.1bn, from sources 4 and 13.