GDX vs GLD vs Physical Gold: What Hedging Reveals About Real Exposure

Here is a quiz with a genuinely surprising answer. Two companies both live and die by the gold price. One is a jewellery giant with billions of dollars of metal sitting in shop windows. The other is a mining company pulling ounces out of the ground. Which one hedges its gold exposure, and which one refuses to?

Most people guess the miner hedges, because miners are industrial businesses with production schedules and debt. The truth is the reverse. The jeweller hedges — sometimes almost completely. The miner deliberately stays exposed, and its shareholders would revolt if it did anything else.

That inversion is not trivia. It is the cleanest way to answer a question StackFi readers ask constantly: GDX vs GLD vs physical gold — which one am I actually supposed to own? Most comparisons of gold stocks vs physical gold stop at fees and convenience. The real difference is upstream of all that, in a single policy decision almost no retail buyer checks before clicking buy: does this company want gold price risk, or has it engineered the risk away?

The one question that separates every gold company

Every business that touches gold has to answer the same thing: do we want gold price risk on our balance sheet, or do we want it gone?

• If the answer is “gone,” the company hedges. It is selling something other than gold — craftsmanship, retail footprint, brand.
• If the answer is “we want more of it,” the company stays unhedged. It is selling gold beta, and the metal in the ground is the product.

Once you know which answer a company gave, you know what you are actually buying when you buy its shares. And in 2026 this matters more than it has in years, because gold has been violent in both directions. Gold futures set a record at $5,542.40/oz on 29 January 2026, then spent the rest of the year retracing; spot sat near $4,472/oz on 20 August 2026, roughly 19% below the January peak. Every wrapper below behaved very differently across that round trip.

Jewellers hedge because gold is their raw material, not their product

A jewellery retailer’s profit comes from workmanship charges, design premium and brand markup. Gold price movement is not the upside — it is the thing that can destroy a season.

The structural problem is the inventory cycle. Metal is bought, sent to a factory, designed, finished, quality-checked and distributed to stores. That takes months. During those months the retail price of the finished piece is still pinned to spot. If gold falls 15% while the necklace is in transit, the retailer eats the difference on inventory it has already paid for.

Three tools solve this:

1. Gold loans / gold leases. Instead of buying metal with cash, the retailer borrows physical gold from a bank and pays a lease rate. The finished piece is sold, and gold is bought back at that moment to repay the loan. Buy and sell happen at effectively the same price, so the price risk lands on the market rather than the shareholder.
2. Futures and forwards. If cash purchase is unavoidable, a matching short position on an exchange offsets the inventory. Inventory falls, the short gains; inventory rises, the short loses. The workmanship margin survives either way.
3. Dynamic retail pricing. The tag price on the counter is repriced against spot through the day, and the back office hedges as it sells.

The real numbers are more interesting than the textbook. Chow Tai Fook, one of the largest jewellers on earth, disclosed a gold hedging ratio of 39% at the end of FY2026 (year to 31 March 2026), down from 55% a year earlier and 69% the year before that. Management described this explicitly as a “pragmatic” rather than rigid stance. In the same year, fair value losses on gold loans ran at 6.6% of group revenue while gold price fluctuation gains ran at 10.3% of revenue — the two legs of the hedge showing up on opposite sides of the P&L.

Two lessons for an investor. First, when a jeweller reports a large “unrealised loss on derivative financial instruments,” that is usually evidence the hedge worked, offset by inventory that appreciated. Second — and this is the part most write-ups miss — the hedge ratio is a discretionary number that moves. A jewellery equity is not a clean non-gold business. It is a retail business plus a management team’s active view on gold, resized every year.

Miners stay unhedged because unhedged is the product

A miner’s economics are brutally simple: profit ≈ (gold price − all-in cost) × ounces sold. Costs are largely fixed and sticky. That creates margin leverage. The standard illustration: with gold near $2,000/oz a producer might earn $600–800/oz of margin; at $3,000/oz that margin roughly doubles even though gold itself rose 50%.

Investors buy miners for that amplification. Hedge it away and the equity loses its reason to exist.

The industry learned this the hard way. In September 2009, Barrick Gold announced it would eliminate its entire gold hedge book, taking a $5.6 billion charge, calculated against a spot price of about $993/oz. It raised $3 billion in equity partly to buy its way out. Barrick said plainly that the hedges were hurting its appeal to investors and its share price. With gold near $4,470 today, the counterfactual is stark: those contracts would have locked away the entire move since.

The industry never went back. The aggregate global producer hedge book fell to 118 tonnes in Q4 2025 — down 74 tonnes year-on-year and the ninth consecutive quarterly decline — with a further ~20 tonne drop estimated for Q1 2026, according to World Gold Council data compiled by Metals Focus. Set that against record Q1 2026 mine production of 884.7 tonnes. The entire world’s forward-sold gold is a rounding error against a single quarter of output. Producers are, as a matter of policy, long gold.

GDX vs GLD: what that leverage actually looked like

Period	Gold / GLD	GDX (gold miners)
Calendar 2025	Gold ~+70% — best year in 45 years	+154.78% total return
12 months to 30 Mar 2026	GLD +52.15%	+109.63% (~2x)
12 months to Aug 2026	—	+57.06% (YTD +4.90%)
Historical maximum drawdown	GLD -45.56%	GDX -80.34%

That last row is the one to sit with. The same operating leverage that doubled gold’s best year in almost half a century is what produces an 80% peak-to-trough drawdown. Miner exposure is not “gold, but better.” It is gold, plus costs, plus jurisdiction risk, plus dilution, plus a management team, geared roughly 1.5–3x in both directions.

There is a middle path: royalty and streaming companies buy a fixed share of future production at a pre-agreed below-market price in exchange for upfront capital. That is not a hedge — it is a financing structure that keeps price exposure while removing cost inflation. We covered the 2026 case for that model in gold royalty and streaming stocks vs miners.

The crypto-native mirror: bitcoin miners taught this lesson first

If you came to gold from crypto, you have already run this experiment.

For years the pitch for MARA and Riot was identical to the pitch for GDX: leveraged exposure to the underlying, with operational upside. Many miners ran a strict HODL policy — mine coins, sell nothing, borrow against the stack. Where that broke, it broke completely. Core Scientific filed for Chapter 11 in December 2022 with roughly $4 million of cash, after a strategy one analyst summarised as building out capacity while “never selling Bitcoin on hand and never hedging prices.”

The 2026 picture is different again, and instructive. MARA sold 15,133 BTC between 4–25 March 2026 for about $1.1 billion (~$72,689 per coin) to retire convertible debt, with roughly 28% of remaining holdings loaned or pledged. Riot sold 3,778 BTC for $289.5 million. Across the sector, capital is being redirected into AI and high-performance computing leases, and miner equities have been rerated as data-centre developers rather than coin proxies.

So the “leveraged bitcoin proxy” stopped being a bitcoin proxy — not because bitcoin changed, but because the company changed. With bitcoin near $64,890 on StackFi’s 9 August 2026 snapshot, an investor who wanted bitcoin exposure and bought miner equity in 2026 got a balance-sheet story and an AI-infrastructure story instead.

This is the identical risk in gold equities, and it is the whole argument of this article: an operating company is a wrapper that can change its mind. Hedge ratios move. Treasuries get sold. Business models pivot. The metal does none of those things.

What you actually own, wrapper by wrapper

Wrapper	Gold price exposure	What else you are buying
Jewellery equity	Partial and discretionary (Chow Tai Fook 39% hedged, FY2026)	Retail demand, workmanship margin, consumer cycle, an active hedging view
Miner equity (GDX)	~1.5–3x, unhedged by policy	Operating costs, jurisdiction and permitting risk, dilution, execution
Royalty / streaming	High, with fixed cost base	Deal pipeline quality, counterparty mine performance
Gold ETF (GLD)	~1:1, less 0.40% annual fee	Brokerage rails, fund and custody structure, market hours
Physical bullion	1:1	Dealer premium, storage, insurance, resale friction
Tokenized gold (XAUT, PAXG)	~1:1	Issuer structure, redemption terms, chain and venue considerations

Three questions that pick your wrapper

1. Do you want gold, or a business that touches gold? If your thesis is “gold goes up,” every equity row above adds risks unrelated to that thesis. Only the bottom three track the metal itself.
2. Can you survive the drawdown your wrapper implies? An 80% historical drawdown is a different instrument from a 45% one, even when both are labelled “gold.”
3. Do you need it inside broker rails? If you already hold assets in a wallet rather than a brokerage account, the ETF row is the awkward one — that is exactly the gap tokenized gold was built for, and it is the path we map in how to buy gold without a brokerage account.

If you are choosing between the three roughly-1:1 wrappers, start with the full breakdown in physical gold vs gold ETF vs tokenized gold, then work through the gold ownership decision guide. Tokenized gold is not automatically the right answer — it carries issuer and venue considerations that bullion does not — but for a crypto-native holder it is usually the only wrapper that keeps gold in the same place as everything else you own.

FAQ

Do gold mining companies hedge the gold price?

Very few do, and only partially. The global aggregate producer hedge book fell to 118 tonnes in Q4 2025 — a ninth straight quarterly decline — against record Q1 2026 mine production of 884.7 tonnes. Producers stay unhedged deliberately, because shareholders buy mining equities for leveraged gold exposure.

Why do gold jewellery companies hedge when miners don’t?

Because gold is a jeweller’s raw material, not its product. Profit comes from workmanship and brand, while months of inventory sit exposed to spot. Gold loans, forwards and dynamic retail pricing strip that risk out. Chow Tai Fook reported a 39% hedge ratio at the end of FY2026, down from 55% and 69% in the two prior years.

Are gold mining stocks better than physical gold?

They are a different instrument, not a better one. GDX returned 154.78% in 2025 against gold’s roughly 70%, but GDX’s historical maximum drawdown is -80.34% versus -45.56% for GLD. Miners amplify gold in both directions and add cost, jurisdiction and management risk on top.

What is the difference between GDX and GLD?

GLD is a physically-backed ETF that holds gold bullion, so it tracks the metal roughly 1:1 less a 0.40% expense ratio. GDX is an equity basket of gold mining companies charging 0.51%, whose value depends on production, costs and capital allocation as well as the gold price. GLD is a gold position; GDX is a leveraged bet on miners’ margins.

Is GDX a substitute for owning gold?

No. GDX is an equity basket whose value depends on production, costs and capital decisions as well as the gold price. If the goal is exposure to the metal itself, bullion, a physical gold ETF, or tokenized gold such as XAUT or PAXG track it far more directly.

Do bitcoin miners hedge like gold miners?

More actively, and their 2026 behaviour shows why the analogy matters. After the 2022 cycle — when Core Scientific filed for Chapter 11 having neither sold nor hedged — public miners moved toward selling production and using derivatives. MARA sold 15,133 BTC in March 2026 and the sector has been rerating toward AI infrastructure, which means miner equity no longer tracks bitcoin cleanly.

What is the cleanest way to get 1:1 gold exposure?

Own the metal or a claim structured directly on it: physical bullion, a physically-backed gold ETF, or tokenized gold. Each has its own trade-offs around storage, fees, market hours and issuer risk, but none of them depends on an operating company’s hedging policy or capital allocation.