EV/GEO and $/GEO: Judging Royalty Deal Value

Use EV/GEO for Accretion, Not Ranking

EV/GEO answers one question well: did a company just buy ounces more cheaply than the ounces the market already credits it with? Enterprise value over annual gold-equivalent ounces is the market’s price for each ounce of attributable production. Compare that with the implied $/GEO of a new deal and you have a fast accretion screen.

What EV/GEO does not do is rank companies by size. The conventional rule of thumb is a tier ladder, seniors at $50-100k per GEO, mid-caps at $35-60k, juniors at $15-35k. The June 2026 data refute it outright. The two largest names in the sector sit around $75-80k per trailing GEO, one mid-cap trades above them, and the juniors, supposedly the cheap end, screen at $137-184k. If you use EV/GEO as a “where should this company trade” band, the market will look permanently wrong to you. It isn’t: the multiple is carrying information about growth and portfolio quality.

What EV/GEO Actually Measures

EV/GEO is enterprise value (market capitalisation plus debt, less cash) divided by annual attributable gold-equivalent ounces. Because GEOs normalise silver, copper, and PGM revenue into gold terms, the metric works across portfolios with very different metal mixes.

Three things move the answer before you compare anything:

The denominator vintage. Trailing GEOs and forward guidance can be a long way apart. Wheaton produced 689,864 GEOs in FY2025, but its 2026 guidance is 860,000-940,000, including roughly 70,000 incremental ounces from the Antamina stream that completed on 1 April 2026. Divide the same enterprise value by the 2026 guidance midpoint instead of the trailing figure and the multiple falls by about a quarter. Same company, same day, same share price. Always state which denominator you are using.

The conversion prices behind the ounces. A GEO is not a physical unit. It is non-gold revenue restated at whatever gold and silver prices the company chose, so the denominator moves when those prices move. At Selborne’s planning deck of $3,500 gold and $48 silver, one GEO is about 73 ounces of silver; at Wheaton’s 2026 guidance prices of $4,800 and $80 it is 60. The same physical silver converts into roughly a fifth more ounces on the second deck, and the multiple falls with it. That is the metric’s central weakness: two companies converting at different prices are not measured on the same ruler.

What EV includes. Wheaton funded Antamina with cash on hand, a new $1.5B term loan and a draw on its revolver, leaving about $2.1B of pro forma net debt, so its enterprise value ($54.2B) runs ahead of its market cap ($52.1B). Franco-Nevada is debt-free, so cap and EV are close to interchangeable. Miss the debt and you flatter the acquirer.

Where the Tiers Sit (9 June 2026)

At the 9 June 2026 close, against FY2025 GEO denominators, the six listed royalty and streaming companies we follow fall into these ranges:

Market data at the 9 June 2026 close, with one mid-cap on the issuer’s own 31 March 2026 market cap.

Check the denominator before comparing any two readings. Franco-Nevada counts ounces sold, Wheaton ounces produced, OR Royalties ounces earned, and each sets its own conversion prices. The denominators are close cousins, so treat small gaps between neighbours as noise.

The ladder, though, is gone by more than noise. The seniors form a working band around $75-80k per trailing GEO. The mid-caps do not sit below it: they straddle it, one below the band and one above. A higher multiple per ounce reflects how the market weighs a portfolio’s duration and growth pipeline. And the juniors, on GEO bases of a few thousand ounces, trade at multiples no producing senior could ever carry.

On forward denominators the picture tightens but does not reorder much, and the companies with the most guided growth see their multiples fall furthest. Check what guidance leaves out, too: Franco-Nevada’s 2026 range of 510,000-570,000 GEOs excludes Cobre Panama entirely.

Why Juniors Screen Highest

A trailing production multiple on a tiny GEO base mostly measures optionality. The juniors sold a few thousand GEOs each in FY2025, and almost all of a junior’s enterprise value sits in royalties on development and exploration assets that contribute nothing to the denominator yet. So the trailing multiple looks absurd, and partly is, but it prices the portfolio that exists on paper.

Guidance shows the compression mechanism working. A junior that guides to nearly doubling its ounces sees its forward multiple fall by around 40% on the same enterprise value; one that guides to modest growth barely moves. The practical reading: for juniors, EV/GEO is a growth-delivery scorecard. Track whether the multiple compresses as guided ounces land. It is useless as a cheap/expensive screen against the seniors.

The Accretion Test

This is the use the metric was built for. A royalty or streaming company trades at some EV per annual GEO. When it buys a new stream, the deal has an implied price per annual GEO: upfront payment divided by annual ounces delivered. Buy below your own multiple and each acquired ounce costs less than the market pays for your existing ounces; the deal screens accretive. Buy above it and you are diluting your own per-ounce valuation.

Antamina gives the deal side of the worked example. Wheaton paid $4.3B for roughly 70,000 GEOs a year, an implied ~$61k per annual GEO, the largest upfront payment in streaming history. Now set that price against two hypothetical buyers. One whose own shares carry $80k per trailing GEO is buying ounces at a discount to its own paper: accretive on the screen. One whose shares carry $50k is paying above its own paper, and the same deal dilutes. The deal price is identical; the verdict depends on the buyer’s multiple.

Two cautions before the screen becomes a recommendation. First, $/GEO ignores duration: a five-year fixed entitlement and a multi-decade stream can carry the same headline number and very different values, which is why the screen only settles anything once a full stream model has run behind it. Second, an ounce is worth only its after-tax margin. Consolidated tax rates across the sector run from about 11% to 22%, so the same gross ounce is worth roughly 14% more at the bottom of that range than at the top, and modelling any of these companies at a miner’s 30% understates NAV by 20-30%.

Why Implied $/GEO Ranges So Widely

There is no going rate for an ounce. Two streams signed in the same year can price four times apart, which is what makes a comp table of headline $/GEO numbers close to useless on its own. Three filed deals show the spread:

Asset stage does most of the work. Kurmuk was an undeveloped project in Ethiopia when Wheaton committed in December 2024, with construction still to come and first gold not expected until 2026, so the ounces came cheap at ~$10.9k. Antamina is one of the largest operating mines in the world, already producing, and Wheaton paid nearly six times as much per ounce for it. Some of the gap is the fourteen months between the two deals, and metal prices that ran hard in between. Most of it is that one buyer took construction risk and the other did not.

Casa Berardi shows the second lever. It is a 2026 deal like Antamina, yet it sits near the bottom at ~$15.4k, because what Franco-Nevada bought is 6,500 fixed ounces a year for five years before the entitlement converts to a small percentage of production. Antamina’s runs until 100 million ounces have been delivered and then for the life of the mine. A cheap headline number usually means a short entitlement. Read the term before reading the multiple.

Feeding the Screen into a Valuation

The multiple is a fast accretion check. It flags whether a buyer paid above or below its own per-ounce paper and whether deal pricing has shifted since the comp you are using; for juniors, watch whether the premium compresses as guided ounces land. Sum-of-parts NAV work means modelling each royalty and stream after tax and discounting it.

The Royalty & Streaming Sector Primer covers that full framework, including the deal-evaluation method that pairs implied $/GEO with IRR so a cheap-looking headline number cannot smuggle in a bad deal.