Gold reaches $4,200/oz by September 30, 2026 in the base case, $4,650 in the bull case, and $3,850 in the bear case. This is a downgrade from our $4,400 call, and the reason is specific: the official-sector bid that anchors every bull forecast has already stalled in the revealed data, even as stated intent hits a record high.

Central banks bought a net 16 tonnes of gold in the first quarter of 2026 — against a run-rate that has averaged roughly 1,000 tonnes a year since 2022 — while selling 129 tonnes, headlined by Türkiye’s 60-tonne disposal in March. Gold has spent July testing $4,125 resistance with support at $4,025 (FX Leaders, July 10, 2026). Every $4,900-to-$6,000 forecast on the street rests on official-sector accumulation continuing. It has not. This call breaks if any of four signals in the Disconfirmation section fires.

Key Levels:

• Asset: Gold (XAU/USD), testing $4,125 with support at $4,025 — market levels, July 10, 2026
• Base case target: $4,200/oz by September 30, 2026 — consolidation, official bid absent
• Bull case target: $4,650 — requires a dovish turn from Chair Warsh and a resumption of central-bank buying
• Bear case target: $3,850 — real yields rise and official-sector selling continues
• Major support: $4,025, with the $4,000 round number beneath it
• Major resistance: $4,125; the $4,120 area is where bulls have repeatedly defended
• Invalidation: a monthly close above $4,400 or below $3,900 retires the consolidation thesis

Methodology

The estimate uses three inputs: official-sector demand as reported for Q1 2026 versus the 2022–2025 run-rate; the real-rate environment under Federal Reserve Chair Kevin Warsh; and the technical range that has contained price through July. Sell-side targets are a check, not an input. Lookback: January 2022 to July 13, 2026. The principal caveat is real — central-bank reporting lags, is incomplete, and some purchases are never disclosed, so 16 tonnes is a floor on true demand, not a measurement.

The data: intent and behaviour have diverged

The bull case for gold has had one load-bearing wall since 2022: emerging-market central banks diversifying reserves away from the dollar. That thesis produced roughly 1,000 tonnes of annual official demand and underwrote the entire re-rating.

Input	Reading	Period	Implication
Central-bank net purchases	16 tonnes	Q1 2026	Far below the ~250t quarterly run-rate
Central-bank gross sales	129 tonnes (Türkiye 60t in March)	Q1 2026	Official sector is a two-way market again
Survey intent	45% of 76 banks expect to add — a record	Feb–May 2026	Stated intent at an all-time high
Goldman Sachs target	$4,900 end-2026, cut from $5,400	2026	The bulls are already trimming
Spot	Testing $4,125, support $4,025	July 10, 2026	Range-bound, no trend

Sources: World Gold Council central-bank demand data and Central Bank Gold Reserves Survey (Feb–May 2026); published Goldman Sachs and J.P. Morgan targets, July 2026; market levels July 10, 2026.

Why does a 16-tonne quarter matter when the survey says 45%? Because a survey measures intent and a purchase measures behaviour, and gold is priced on behaviour. A record 45% of the 76 central banks surveyed by the World Gold Council between February and May 2026 said they expect to increase reserves over the next 12 months. In the same window, the official sector as a whole bought a net 16 tonnes and sold 129, with Türkiye disposing of 60 tonnes in March. The two facts are not contradictory — banks can intend to buy later — but they are not interchangeable. Every forecast in the $4,900-to-$6,000 range assumes accumulation at something close to the 2022–2025 pace. The first quarter of 2026 did not deliver it, and a thesis that requires 1,000 tonnes a year cannot be underwritten by a quarter that produced 16.

“Our rationale is that emerging market central banks remain significantly underweight gold compared to their developed market counterparts and are gradually increasing allocations as part of a broader diversification strategy.”

— Lina Thomas, Research Analyst, Goldman Sachs (Goldman Sachs)

The mechanism: what is actually setting the price now

With the official bid dormant, gold is left to trade on the thing it traded on before 2022: real yields. That is a worse environment than the one the bulls are modelling.

Chair Warsh’s Federal Reserve is guiding higher-for-longer, and a June non-farm payrolls print that shocked to the downside at 57,000 produced a squeeze above $4,120 rather than a trend — which is itself the tell. In a market with a structural bid underneath it, a dovish data surprise starts a move. In a market without one, it produces a short-covering pop into resistance that then fades. That is precisely what July has delivered.

The carry cost is the mechanism. Gold pays no yield, so when real rates are held high its opportunity cost rises and the marginal buyer needs a reason beyond momentum. From 2022 to 2025 that reason was a central bank with a mandate to diversify and no price sensitivity. Strip that buyer out and you are left with exchange-traded fund flows and speculative positioning — both rate-sensitive, both currently unenthusiastic.

The steelman should be taken seriously. Goldman Sachs still carries $4,900 for end-2026 and J.P. Morgan is near $6,000/oz across the fourth quarter. Both assume the official bid resumes, and both may be right — reserve diversification is a decade-long process, and one soft quarter proves little about a ten-year reallocation. If buying restarts at the old pace, $4,200 will look far too conservative.

What the model misses

Two things, and the first is serious. Central-bank gold reporting is lagged, partial and sometimes deliberately opaque; several of the largest historical buyers have accumulated without disclosing in real time. The true Q1 number is almost certainly higher than 16 tonnes. The argument is not that official demand is zero — it is that it is not visibly running at the pace the $6,000 forecasts require.

Second, the framework treats Türkiye’s 60-tonne sale as a signal when it may be idiosyncratic. Reserve managers sell for domestic liquidity reasons unrelated to a view on gold. One large seller can distort a quarterly net figure — and reading a regime change into a single disposal is the mirror error of the one that produced the $5,400 targets now being cut.

“The usual suspects which might provide support via investment demand are notably absent, for now.”

— Michael Hsueh, Strategist, Deutsche Bank (GoldSilver)

Disconfirmation

This downgrade is wrong if any of the following fires:

• Second-quarter official-sector demand prints above 200 tonnes. That would restore the run-rate and confirm the first quarter as an aberration. The entire basis for the downgrade disappears, and $4,650 becomes the base case.
• A monthly close above $4,400. This is the level of our previous call. Price reclaiming it says the market has found a buyer we cannot see in the reported data — which, given the reporting lag, is entirely possible.
• Chair Warsh signals cuts. The carry-cost mechanism runs in both directions. A dovish pivot compresses real yields, and gold does not need the official bid if the rate environment does the work instead.
• A monthly close below $3,900. This would say the bear case is running — real yields rising and official selling continuing — and the consolidation thesis is too optimistic, not too cautious.

What to watch next

The World Gold Council’s second-quarter demand data is the most important release for this call — the first clean read on whether the first quarter was a pause or a turn. Watch, in order: official-sector net purchases against the roughly 250-tonne run-rate; whether Türkiye keeps selling; the next Federal Open Market Committee decision and Warsh’s language on the real-rate path; and whether $4,125 caps another attempt or gives way.

A weekly close through $4,125 that then holds is the earliest signal that this downgrade is premature.

FAQ

What is the gold price forecast for Q3 2026?

Base case $4,200/oz by September 30, 2026, downgraded from $4,400. Bull case $4,650 on a Warsh pivot and resumed central-bank buying; bear case $3,850 if real yields rise and official selling continues. Gold has been testing $4,125 with support at $4,025.

Why downgrade gold now?

Because official-sector demand — the anchor of every bull forecast — printed a net 16 tonnes in the first quarter of 2026 against a run-rate near 1,000 tonnes a year, with 129 tonnes of gross sales. A thesis that needs 1,000 tonnes annually cannot rest on a quarter that produced 16.

Are central banks still buying gold?

Stated intent is at a record: 45% of the 76 central banks surveyed by the World Gold Council between February and May 2026 expect to add over the next 12 months. Revealed behaviour is weaker — net purchases of 16 tonnes in Q1 2026, with Türkiye selling 60 tonnes in March. Reporting lags mean the true figure is likely higher.

What do the banks forecast?

Goldman Sachs carries $4,900 for end-2026, cut from $5,400. J.P. Morgan is near $6,000/oz averaged across the fourth quarter. Both assume the official bid resumes at close to its 2022–2025 pace — the assumption this downgrade questions.

What would push gold back above $4,400?

Second-quarter official-sector demand above 200 tonnes, or a dovish turn from Chair Warsh that compresses real yields. Gold does not need the central-bank bid if the rate environment does the work instead.

This call revises our earlier Gold to $4,400 Warsh-pivot upgrade. See also our related calls on Silver to $70 and Platinum to $1,900. Central-bank demand data is published by the World Gold Council.

This article is informational analysis only and is not financial, investment, or trading advice. Foreign-exchange, commodity, and equity markets are highly volatile and can lose substantial value rapidly. Leveraged products carry total-loss risk and may exceed the initial margin posted. Past performance and historical correlations do not guarantee future results. Do your own research and consult a regulated financial adviser before making any investment decision.