Most gold price predictions for 2026 getting attention right now come in one of two forms: a bank's research note, or a chart. The widely shared S&P Global major-discoveries chart is the one getting attention, and it shows that the world found zero major gold deposits in 2023 and 2024. Meanwhile, there is a place where "where will gold end 2026" is expressed as live, tradeable prices: Kalshi's year-end gold ladder, 13 contracts on whether gold finishes above strikes from $4,300 to $5,500 on December 31. This piece reports what that ladder prices today and which rows of it actually carry an opinion, and then explains why the chart is accurate as published and still cannot settle the year-end question in either direction. The reason is a lag built into how deposits get counted, and it is worth understanding before you read any scarcity argument about any commodity.
The Quick Answer
As of the afternoon of August 18, 2026, Kalshi's year-end gold market prices the $4,600 strike as a coin flip (48 cents bid, 50 cents ask on the YES side), makes $4,300 the likely floor (last trade 69 cents), and shows the highest listed strike, $5,500, last trading at 13 cents with no fills in the prior 24 hours. Every market price on this page is a snapshot of a live board, dated where it appears, and will have moved by the time you read it. The full ladder, which strikes are actually changing hands versus merely listed, and why the viral "no new gold discoveries" chart should not, by itself, settle the year-end question, are all below.
What The Year-End Ladder Prices Right Now
The market is a ladder of yes/no contracts, one per strike, each asking whether the price of gold will be above that level at 5:00 PM EST on December 31, 2026. Each contract settles at $1 if the answer is yes and $0 if it is no, so a price in cents reads like a chance out of 100, with the caveats that a price is not exactly a probability, and that fees sit on top of it. If event contracts themselves are new to you, our primer on how prediction markets work covers the plumbing. Here is the board, quoted as the YES side's bid and ask, with the last trade and how many contracts changed hands in the 24 hours before the snapshot:
| Finishes Above | YES bid | YES ask | Last trade | Traded, prior 24h | Lifetime volume (contracts) |
|---|---|---|---|---|---|
| $4,300 | 58¢ | 70¢ | 69¢ | 232 | 28,481 |
| $4,400 | 53¢ | 57¢ | 55¢ | 125 | 16,881 |
| $4,500 | 41¢ | 54¢ | 41¢ | 97 | 12,063 |
| $4,600 | 48¢ | 50¢ | 50¢ | 290 | 9,500 |
| $4,700 | 41¢ | 44¢ | 41¢ | 70 | 2,071 |
| $4,800 | 37¢ | 38¢ | 38¢ | 51 | 10,376 |
| $4,900 | 33¢ | 34¢ | 34¢ | 0 | 9,439 |
| $5,000 | 29¢ | 30¢ | 30¢ | 0 | 8,410 |
| $5,100 | 24¢ | 27¢ | 26¢ | 1 | 4,289 |
| $5,200 | 15¢ | 25¢ | 25¢ | 8 | 4,708 |
| $5,300 | 20¢ | 24¢ | 24¢ | 0 | 3,648 |
| $5,400 | 13¢ | 16¢ | 16¢ | 0 | 3,360 |
| $5,500 | 13¢ | 16¢ | 13¢ | 0 | 15,533 |
Prices and volume from Kalshi's public market data, the afternoon of August 18, 2026 (Eastern). The ladder closes December 31, 2026.
The row to start with is $4,600, and not because of the price alone. It is the busiest strike on the board that day, 290 contracts, and it quotes 48 bid against a 50 ask: a two-cent spread on a coin flip is the closest thing this ladder has to a firm opinion. The two busy rows below it price as more likely than not (the $4,400 last trade at 55 cents, the $4,300 last trade at 69); $4,500 is the exception, and I take it apart in the next section. Everything above $4,600 prices as less likely than not, sliding through the $4,800 row at 37 to 38 cents down to the low-to-mid teens at $5,400 and $5,500. So the market's median for end-of-2026 sits right around $4,600, with the bulk of the traded opinion saying gold finishes the year in the mid-$4,000s and does not run away to the upside. Hold the $5,000 row in mind, though: it quotes a tidy 29 to 30 cents, and it is one of five strikes on which not a single contract changed hands in the prior 24 hours. The gap between a listed price and a transacted one is this board's most instructive lesson.
Notice, too, that the numbers do not descend as neatly as a probability curve should. The $4,600 bid (48) sits seven cents above the $4,500 bid (41), and $5,300's bid sits above $5,200's; read as clean probabilities that is impossible, because gold cannot be more likely to clear the higher bar than the lower one. That is not the market being stupid. It is the next section.
Worked Example: A Listed Strike Is Not A Traded One
A quote on a screen is an invitation. It only becomes a verdict when someone trades against it. When I read this ladder, I set five rows aside before I take anything else seriously: $4,900, $5,000, $5,300, $5,400 and $5,500 saw no contracts change hands in the 24 hours before this snapshot, and $5,100 saw exactly one. That also dissolves the second "impossible" quote pair above: $5,300's 20-cent bid sits above $5,200's 15-cent bid, but $5,300 had no fills at all and $5,200 had eight, so that is a dead row's stale bid sitting next to a thin-but-live row's, not the market contradicting itself. The first inversion has a different mechanism: $4,500 is a live row, 97 contracts that day, but its 13-cent spread (41 bid, 54 ask) is wide enough that the "price" is really a range, and its neighbor $4,600, three times busier and quoted two cents wide, is the one carrying the actual opinion. Read together they say the same thing: the coin flip lives in the mid-$4,000s, and $4,500's low bid is a resting order nobody has bothered to lift, not a considered view that gold is likelier to clear $4,600 than $4,500. The $5,200 row shows what a quiet day does to a quote: eight contracts traded and its spread is ten cents wide (15 bid, 25 ask), so anyone quoting the midpoint as the market's opinion is inventing precision that no one is willing to trade at.
Then there is $5,500, which holds 15,533 lifetime contracts, third-heaviest on the board; the round-number strikes on this ladder carry outsized lifetime volume, whatever the reason. It went the full 24 hours without a fill, so its 16-cent ask is closer to a stale resting offer than a live opinion, and its 13-cent last trade is the more honest number. And lifetime volume answers a different question than recent volume: only 2,071 contracts have ever traded at $4,700, against 28,481 at $4,300, yet $4,700's 3-cent spread and 70 contracts of business that day make it a young strike, not a dead one. Across all 13 strikes the market has traded about 129,000 contracts lifetime, and more than a fifth of that sits in the single $4,300 contract. So when you read this ladder, weight the tight, busy rows and treat the wide, quiet ones as rough sketches. This is the standard liquidity lesson of every event market, and it matters double on a board where a few strikes do most of the business. Kalshi also runs monthly and weekly gold ladders that settle sooner; the mechanics are the same, and our commodity markets explainer covers how those shorter-dated ladders work.
My Read: What The Active Rows Actually Say
I hold no view on where gold goes, and this page will not create one: it reports the ladder and stops. What I will own is a read of the board itself, which is a different thing. If you asked me which rows to quote to a friend, I would quote three: $4,300, $4,400 and $4,600. Each traded more than a hundred contracts that day, and between them they say gold finishes the year in the mid-$4,000s with a real, roughly even chance of clearing $4,600. $4,400 and $4,600 quote within a few cents of their own last trades. $4,300 does not, and I want to be honest about that: its 58-to-70 spread is as wide as $4,500's, but unlike $4,500 its last trade printed at 69, right at the ask, on 232 contracts, which tells you which side was doing the lifting. I would not quote $4,500, because a 13-cent spread with the last trade sitting on the bid is a shrug, and I would not quote anything from $4,900 up as a settled opinion, because those rows are quiet enough that their prices are mostly the last person to leave the room. My read, then, is that the active rows put the useful dividing line at $4,600 and the useful floor at $4,300, and that the upper strikes are too quiet to treat as a strong signal either way. That is a read of liquidity, not of gold.
The one row I keep coming back to is $4,300, and it is worth saying why. It has done more business than any other strike over the life of the market, its last trade sits at 69 cents, and its 24-hour tally of 232 contracts is second only to $4,600. When the most-traded contract on a ladder prices "above the lowest strike" at better than two-to-one on, the market is telling you where it thinks the floor is with more conviction than it tells you anything about the ceiling. The honest counterpoint is that four months is a long time and a ladder like this can move ten cents in a week; a read of today's board is a read of today's board. Which brings us to the chart, because the chart is an extreme version of the same mistake the idle rows invite: a number that is technically accurate, read as if it says more than it does.
The Viral Scarcity Chart Is Accurate, And It Still Cannot Answer This
The chart circulating right now comes from S&P Global's major-discoveries data, and its striking feature is real: 2023 and 2024 show zero major gold discoveries. Nobody faked anything. The catch is in how a "discovery" gets on the list. A deposit is credited to the year it was first drilled, but it only qualifies as major once it is proven to hold 2 million ounces, and that threshold is only confirmed by later drilling. So the series back-fills: past years keep gaining discoveries long after they end, and the most recent years always look emptiest before revisions arrive. S&P itself has said that almost all new additions to its list were discovered decades ago and have only recently met the 2-million-ounce criteria. The zeros are published, and the zeros are provisional. Both things are true at once.
The fact that actually answers the year-end question, and it points the other way from the chart's implication: 2025 was an all-time high for mine production at the same time as those blank discovery years. There is no contradiction, because a deposit takes roughly 10 to 30 years to go from first drill hole to producing mine. The gold that miners are pulling out of the ground right now was found in the 1990s and 2000s. Whatever the discovery data of 2023-24 ultimately shows after revisions, it is a question about the 2040s. A contract that settles on December 31, 2026 should not be read as a bet on a 2040s supply problem, which is why I do not treat the chart as evidence for any strike on the ladder above, in either direction.
"Record Production" Is Also Doing Less Work Than It Sounds
The bullish counter-meme, that miners are producing more than ever, deserves the same precision. 2025's output of 3,671.6 tonnes beat the previous high, 2018's 3,663 tonnes, by 8.6 tonnes. The margin is 0.23%, a record by a hair, and the World Gold Council's own forward view is that output will "gradually plateau over the next few years, rather than peak and then fall." The honest description is flat at an all-time high, not accelerating. And parts of the bear case survive the back-fill caveat cleanly, because they are not discovery-count statistics: exploration spending fell 15% in 2023 and another 7% in 2024, ending an uptrend that began in 2017, and the average new deposit found in the last five years holds 4.4 million ounces, down from 7.7 million the decade before. The one-sentence version of the whole argument: the trend is real, the endpoint is an artifact.
How Much Gold Exists, And Which "Years Of Supply" You Mean
Scarcity talk eventually lands on "years of gold left," and that number swings wildly depending on what you divide. The World Gold Council's end-2025 accounting says 219,891 tonnes of gold have ever been mined, a stock that would fit in a cube about 22 meters on a side. It splits into 97,645 tonnes of jewellery, 50,978 in bars, coins and ETFs, 38,666 held by central banks, and 32,602 in other uses. Against that stock, 2025's mine supply of 3,671.6 tonnes added about 1.67%, and recycling added another 1,404.3 tonnes, which is 28% of total supply and rises when the price rises, a built-in shock absorber that most scarcity charts ignore.
Now the division. Total stock over annual mine supply gives 59.9 years, but that treats every wedding ring on earth as sellable float. The strict version counts only the investable stock: bars, coins, ETFs and central bank holdings. That is 89,644 tonnes, and it gives 24.4 years. The 24-year figure is the one that survives scrutiny, because it divides only the investable stock, and it is the number that gives a scarcity argument its teeth; the 60-year figure counts all the metal in existence and is what makes the same argument look overdone. Most scarcity posts quote the second while implying the first. And neither number bears on a contract that settles on December 31: a 24-year float is a story about the 2040s, the same decade the discovery chart is really about, which is why nothing in this section moves any row on the ladder above.
More on this: Where Does Bitcoin End 2026? Kalshi's Price-Band Odds · How Bitcoin Price Contracts Work On Kalshi · The Forecast Is Not The Price On A Kalshi Weather Market · How Much Gold Is Left? What The Numbers Actually Say · Solana Price Prediction 2026: How High Can SOL Go?
How The Contract Settles, And The Shape Of The Risk
Settlement is mechanical: each strike resolves YES if the close of the one-minute candlestick for gold at 5:00 PM EST on December 31, 2026 is above that strike in dollars per troy ounce, and settlement then pays $1 or $0. The rules name the underlying simply as gold, in US dollars per troy ounce; check the market's rulebook on Kalshi for the reference feed before you trade it. There is no committee and no judgment call, which is exactly what you want in a market that will be argued about all year.
Two structural points for anyone tempted to trade a ladder like this rather than just read it. First, the asymmetry: selling an unlikely outcome, say selling the YES on a low-teens strike, collects a small premium and risks most of a dollar. Sold at 13 cents, one loss wipes out roughly 6.7 winning 13-cent premiums before fees, and the cheaper the contract, the worse that ratio gets. That arithmetic, not any hit rate, is what makes position sizing the entire game. Second, execution: recall the $5,200 row's ten-cent spread. In a contract where the whole question is worth a few cents, crossing a spread like that to get filled instantly can cost more than the view is worth; an instant fill usually means you paid for it. If prices quoted in cents are new to you, our primer on how Kalshi prices convert to American odds is the place to start before the ladder.
To zoom back out: neither the viral chart nor the record-production headline settles a contract that expires on December 31. The ladder, the one version of the forecast anyone is willing to back at a firm price, will settle itself, and the busy rows will tell you first.
A few disclosures, plainly: Kalshi's gold contracts are CFTC-regulated event derivatives traded on a designated contract market, not sportsbook wagers, and they can lose their full value. Stokastic trades prediction markets; nothing on this page describes our own activity, and we do not publish performance figures. We have no affiliate or commercial relationship with Kalshi. This page reports what the market prices and holds no view on where gold is going; nothing here is a recommendation to buy or sell anything. 18+, available where Kalshi operates, and the risk of loss is real.
The listed-versus-traded lesson is the same one our analysts apply to thin sportsbook lines in their free expert picks today, and every mechanic in this piece (the bid-ask spreads, the thin strikes, the settlement clock) runs the same way across Kalshi's other event ladders; to see the same reading applied to a board that resets every day, our Kalshi weather markets hub is the worked version.
