What do analysts expect gold to do in 2027?
The published forecasts point higher, and they disagree by roughly a quarter on how much higher. J.P. Morgan has an average above $6,200 an ounce for the year while Morgan Stanley describes a path above $5,000, and several banks have already revised their own numbers. The table below pairs each forecast with what it assumes and what would break it.
Key Takeaways
- The direction is agreed on and the magnitude is not. J.P. Morgan Global Research published a 2027 average of $6,263 an ounce in June 2026, against Morgan Stanley's path above $5,000 for the same year.
- Forecasters revise as fast as they disagree. J.P. Morgan's own 2027 average came down from $6,550 in February 2026 to $6,263 in June, and Bank of America cut its 2026 average by 14% in a single revision.
- Almost every forecast rests on the same handful of conditions: central bank buying continuing, the Federal Reserve staying on hold, and investment demand returning through ETFs.
- Not every forecaster publishes a number at all. The World Gold Council sets out conditional scenarios instead of price targets, which is a more honest shape for a figure that depends entirely on which macro path arrives.
- Gold trades on the Entry.fi Terminal as a USDC-margined perpetual, which gives exposure to the price without ownership of metal and carries a funding cost for as long as the position stays open.
In This Article
- What the published 2027 forecasts say
- Why the spread matters more than the midpoint
- What every gold forecast is really a bet on
- Why a perpetual is an awkward way to hold a 2027 view
- Trading gold on the Entry.fi Terminal
- FAQ
What the published 2027 forecasts say
| Forecaster | Published 2027 view | As published | What it assumes | What would break it |
|---|---|---|---|---|
| J.P. Morgan | $6,263 average, with quarters of $6,200 to $6,300 | 9 June 2026 | Central bank buying continuing, Chinese demand holding, and possible allocations by Chinese insurers starting near 1% of assets | Strong US growth with accelerating inflation and a Fed hiking cycle, which the bank itself calls a high bar |
| Morgan Stanley | A path above $5,000 | 20 August 2026 | The Fed on hold through 2026, ETF demand recovering, and gold pricing fiscal concerns more than the level of yields | COMEX short positioning already near its lowest since April 2020, leaving little room for further short covering |
| Bank of America | $5,000 within reach once Fed tightening ends | 7 July 2026 | An end to the tightening cycle | A more hawkish Fed, the stated reason it cut its 2026 average by 14% |
| ANZ | $6,000, deferred to mid-2027 | 15 May 2026 | No driver published alongside the target | The target has already slipped once, having previously been set for early 2027 |
Figures as published on the dates shown and gathered on 11 October 2026. Several of these have been revised since they were first issued, and all of them may be revised again. None is a statement of what gold will do.
Why the spread matters more than the midpoint
The instinct with a table like that is to average it and call the result a consensus. The gap between the top and bottom of the range is around 25%, which is larger than most of the annual moves these same desks are forecasting. An average of numbers that far apart describes nobody's actual view.
The revision history is the more useful read. J.P. Morgan lowered its 2027 average by 4% between February and June 2026, having also lowered every 2026 quarter by 4% to 13% in the same update. Bank of America took 14% off its 2026 average in one move and attributed it to a more hawkish Fed. ANZ pushed a $6,000 target back by a quarter or two without changing the number. All three were revising a view about a year that had not started yet.
Read together, that says something specific. These are not predictions in the ordinary sense. They are the current output of a model whose inputs move monthly, published by institutions that update them without embarrassment. A reader who anchors a position on one figure is anchoring on something its own author has moved by double digits inside six months.
What every gold forecast is really a bet on
Strip the targets away and the same conditions appear underneath all of them. Central banks continuing to accumulate. The Federal Reserve holding rather than hiking. Investment demand returning through ETFs after a period of outflows. Gold responding to fiscal anxiety instead of to the level of real yields. Each forecast is a view on those conditions, converted into a dollar figure at the end.
Those conditions are observable, which makes them the part a reader can actually track. On central banks, J.P. Morgan noted in June 2026 that reported purchases had slowed during the first quarter while the World Gold Council estimated unreported buying ran higher than in late 2025. Morgan Stanley counted China adding 60 tons through August, its most since 2023, and Poland adding 82 tons toward a stated 700-ton target. On investment demand, the same note recorded ETFs taking in 70 tons across July and August after 93 tons of outflows in May and June.
None of that settles 2027, but it is the right place to be looking. A target is downstream of its conditions, so the question to ask of a forecast is not whether the figure looks plausible. It is whether you believe the conditions the forecaster listed, and what you would do if the opposite arrived instead.
Why a perpetual is an awkward way to hold a 2027 view
A 2027 forecast describes a horizon more than a year out. A perpetual futures contract is built for much shorter holding periods, and the mismatch is worth understanding before using one to express a long-dated view.
Funding is the reason. A perpetual has no expiry, so instead of settling it exchanges a recurring payment between the two sides to keep the contract near spot. Hold the paying side of that for fifteen months and the carry accumulates the whole time, charged against position size rather than against the margin behind it. A thesis that is eventually right can still lose money because the rent on the position outran the move.
Leverage compounds the mismatch. The further a position is levered, the nearer its liquidation price sits to the current mark, and a year is a long time to avoid a drawdown that a forecast of this kind does not rule out. The distinction between renting exposure and owning an asset matters most at exactly this horizon. Perpetuals suit trading a move. A multi-year view usually calls for an instrument that does not charge rent while you wait.
Trading gold on the Entry.fi Terminal
Gold appears on the Terminal as a perpetual market margined in USDC, in the same account as crypto perpetuals and spot markets. The position is exposure to the gold price and nothing more. There is no bar, no vault, no delivery, and no storage cost, which also means none of the properties people hold physical metal for in the first place.
Two mechanics are specific to commodity markets of this kind. Leverage limits are set per market and shown in the interface, so the cap on a gold contract need not match a major crypto pair. And on markets deployed by third parties, the deployer defines the oracle and maintains its prices, so the market header is the place to read the current funding rate and reference price instead of assuming either behaves like BTC.
Everything in the interface is informational. The Terminal Terms state that charting data, analytics and routing suggestions do not constitute investment advice, and the same document excludes residents of the United States, the United Kingdom, Belgium and Ontario along with sanctioned jurisdictions. Nothing in this article is a recommendation to take a position in gold, and the forecasts above belong to the institutions that published them.
FAQ
What is the highest published gold forecast for 2027?
J.P. Morgan's average of $6,263 an ounce is the highest of the figures in the table, published in June 2026 with quarterly estimates between $6,200 and $6,300. ANZ's $6,000 target sits close to it on a mid-2027 timeframe. Both were issued before 2027 began and both may be revised.
Why do bank forecasts for the same year differ so much?
Because they are each the output of a different macro view. A desk expecting the Fed to hold produces a materially different gold number than one pricing a hiking cycle, and the conditions each desk lists in its own publication explain most of the gap between them.
Can I buy physical gold on an onchain terminal?
No. A gold perpetual tracks the price without conveying ownership, delivery or a claim on metal. It is a margined contract, so it carries funding costs and liquidation risk that holding bullion does not, and it lacks the properties that make physical metal attractive to some holders.
How is onchain gold priced if there is no metal behind it?
Through an oracle that reports a reference price to the contract, against which positions are marked and funding is calculated. On markets deployed by third parties the deployer maintains that oracle, which makes the reference source part of what you are taking on when you open the position.
What does a forecast revision tell me?
That the conditions behind it changed, and usually which ones. A revision note is often more informative than the target it replaces, because it names the input that moved. Several of the 2027 figures above were lowered during 2026 while the direction of the call stayed the same.
Is gold still considered a safe haven in 2026?
The forecasters above describe demand coming from central bank accumulation and fiscal anxiety, which is a different mechanism from the crisis hedging the phrase usually implies. Morgan Stanley has noted gold decoupling from long-term real yields and pricing fiscal concerns instead. Whatever the label, the drivers named in current research are specific and worth reading directly.
A forecast is a conditional statement wearing a dollar sign. The conditions are the part worth reading.
Open the Entry.fi Terminal and see how the gold market is quoted and funded.