Gold & Silver Outlook 2026: Next Leg of Rally or Correction?

Overview: Gold and silver may be positioned for another leg higher, although near-term volatility remains possible. According to Monarch PMS, the January-June 2026 correction was primarily driven by higher real yields, a stronger US dollar, margin-led liquidation and hawkish Fed expectations rather than a deterioration in the structural case for precious metals. Gold declined 28.7% from its January peak, while silver fell 52.6%.

EXECUTIVE SUMMARY

The January-June 2026 correction in gold (-29%) and silver (-53%) was a deleveraging event driven by higher real yields, a stronger dollar, margin-led liquidation, and hawkish Fed expectations rather than a deterioration in the long-term investment case for precious metals.

A. Anatomy of the Correction: What Broke, and What Did Not

• Gold hit an all-time high of $5,589/oz on 28 January 2026 (+60% in 2025, best year since 1979); silver peaked at $121.64/oz a day later (+147% in 2025, +68% in January alone).
• The fall was fast and mechanical: Gold fell $380 (~7%) in 28 minutes on 29 January; Silver lost 11% in the same window. By late June gold touched a low near $3,985 (-28.7%) and silver near $57.6 (-52.6%).
• Five causes: All are price-of-money or price-of-leverage stories, none is a store-of-value story:
  1. Fed leadership change — Kevin Warsh (from 13 May) repriced the terminal rate higher;
  2. CME margin hikes forced leveraged (esp. silver) liquidation;
  3. DXY hit a 13-month high;
  4. an oil shock (WTI $57→$113 peak) pushed core PCE to 3.4% and read as hawkish, not as debasement;
  5. crowded positioning — the gold-silver ratio compressed to 46x, tightest since 2011.
• Absent from that list: No deterioration in central bank demand, No fiscal repair, No real dollar-yield advantage restored, No resolution of the silver deficit.

B. Where the Market Stands Today

• The Fed held rates at 3.50–3.75% on 29 July (fifth consecutive hold); the 9-3 vote was the widest hawkish dissent since Sept-2016 (Hammack, Kashkari, Logan). Chair Warsh gave no forward guidance.
• September-hike odds have fallen from 78.8% to 25% post-meeting & benign July CPI numbers.
• The 10-yr Treasury yields 4.63%, the 30-yr 5.20%, and the 10-yr TIPS real yield sits at 2.41%. This is the single genuine headwind for a zero-coupon asset like Gold.
• DXY (Dollar Index) has slipped from its June peak below 100.
• Every macro variable that caused the correction has peaked / reversed except the absolute level of the real yield.

C. Why the Monetary Case Remains Intact!

The core gold thesis is fundamentally a sovereign balance-sheet thesis.

• US Federal debt has risen to nearly US$40 trillion, increasing by approximately US$10.7 trillion over the past five years, despite the absence of a major recession or large-scale stimulus cycle.
• Annual net interest payments are projected to exceed US$1 trillion, making interest expense one of the largest expenditure items within the federal budget.
• The weighted average cost of government debt continues to rise as low-cost debt issued during 2020-21 is refinanced at significantly higher rates.

Inference: Even if deficits do not widen further, interest costs are likely to continue increasing, reducing fiscal flexibility and increasing pressure on policymakers.

D. What Gold and Silver Are Actually Worth

• Above-ground gold stock (~221,700 t) if valued against reserve-currency M2 (US M2 fully + 50% euro-area M2) gives a modelled gold range of $3,248–$4,595, midpoint at $3,922.
• At $4,300, gold trades +10% above midpoint, and bottomed within 2% of fair value range in June’26.
• Silver, valued via the gold-silver ratio (21st-century average 69:1; normalised at 60:1), gives a modelled range of $54–$77, midpoint $65.
• At $62, silver trades 5% below midpoint.

E. The Official Sector: A Price-Insensitive Bid

• Central banks bought a net 288.9 Ton in Q2CY26 (+62% YoY) representing a record second quarter buying while gold fell ~16%.
• H1 CY26 totalled a weaker 345 Ton (lowest since 2022) after a Q1 downgrade from 244ton to 57ton; Russia and Turkey were liquidity-driven sellers, not reserve-allocators.
• 89% of reserve managers expect global official gold holdings to rise over 12 months. 74% expect to hold fewer dollars over five years (WGC 2026 survey).
• Recently, France has fully repatriated its US-held gold and Germany is debating the same for a third of its holdings highlighting a jurisdiction-risk signal.
• However, Private flows moved the opposite way: gold ETFs saw net outflows through May-June.

F. Silver: A Monetary Asset with an Industrial Floor

• The market is in a sixth consecutive annual deficit, widening to 46.3 Moz in 2026F from 40.3 Moz in 2025;
• Mine supply has been effectively flat (~830-850 Moz/yr) since 2015 because most silver is a copper/lead/zinc by-product with decade-long permitting timelines. At the same time demand has seen multi lever growth.
• Physical investment demand is forecast +20% to a 3-yr high of 227 Moz; US retail demand is set to rebound 57%. COMEX registered inventory is ~96 Moz (up from a 76 Moz March low); coverage is ~17.8% and paper leverage ~5.6x — thin deliverable metal makes upside moves prone to gapping.
• Key dates: Late-Aug Jackson Hole (Warsh's first); 15-16 Sept FOMC with new dot plot; ongoing Strait of Hormuz talks (cuts both ways — lowers hike odds, but removes geopolitical premium); late-Oct WGC Q3 Gold Demand Trends, testing whether Q2's central-bank record was catch-up or trend.

G. The Bear Case for Bullions

• Real yields are genuinely positive, 2.41% on 10-yr TIPS is the highest competing return in over a decade for a zero-coupon asset.
• H1 CY26 central-bank buying (345t) was the weakest since 2022, and the Q2 record may be partly catch-up; Deutsche Bank has modelled a slower forward pace.
• A visible Gold ETF overhang (~298tons of gold bought through ETFs are in losses). This caps rallies until the cloud clears.
• The bear case is predominantly about the next two quarters; the bull case is about the next three years. Both can be right — the answer would depend on horizon, not conviction.

H. Conclusion and What Would Change the View

• The correction did its job: It cleared ~30% from gold and ~50% from silver, unwound leverage, reset the gold-silver ratio from a stretched 46x to 69x, and pulled volatility back to investable levels — without repairing the fiscal position that underpins the structural case.
• The structural bid is intact and price-insensitive: Record Q2 central-bank buying, 89% of reserve managers expect higher official holdings, 74% expect fewer dollar holdings.
• Bearish signals to watch: Central-bank net buying below 150t in Q3 CY26; COMEX registered silver back above 130 Moz with sustained contango; 10-yr TIPS real yield sustained above 2.75% for a quarter.
• Bullish confirmation: A September hold by Fed, a soft August payroll print, and a 10-yr real yield below 2.20% would, in this framework, mark the start of the next leg. Near term, positioning should stay two-sided — a retest of the June lows (base case: a retest, not a break, given a published floor consensus near $3,900) remains possible.

Index

Setting the scene: A violent correction inside an intact bull market

• Gold set an all-time high of US$5,589/oz on 28 January 2026, capping a run that had already delivered a 60% gain in calendar 2025 - its best year since 1979. Silver peaked a day later at US$121.64/oz after a 147% gain in 2025 and a further 68% in January alone.
• What followed was brutal. By late June gold had traded below US$4,000 and printed a low near US$3,985, a drawdown of roughly 29%. Silver fell to around US$57.6, down 53% from its high. On the World Gold Council's numbers, Q2 CY26 was gold's worst quarter since 2013.
• As at 19 August 2026 gold trades near US$4,300/oz and silver near US$62.7/oz, both at seven-week highs, with the Strait of Hormuz partially reopening and July payroll data softening.
• The correction was severe, but nothing that drove the rally changed. It was caused by a hawkish change of Fed leadership, a margin-driven deleveraging of a crowded trade, a stronger dollar, and an oil shock that made inflation read as a rate-hike problem rather than a debasement problem.
• Underneath the price fall, the structural buyers did the opposite of what the price implied. Central banks bought a record 288.9 tonnes in Q2 CY26, up 62% y/y, into a quarter in which gold fell 16%. Silver entered a sixth consecutive year of physical deficit.

This note sets out what happened between January and July 2026, what the data says about the market today, and the conditions under which the next leg gets going.

Our conclusion in one line:

The January-June drawdown removed leverage and speculative froth from the price without removing a single one of the structural reasons the price went up in the first place. That is the shape of a mid-cycle reset, not a cycle top.

Key Questions

1. Was the 30% fall a break in the bull market, or a deleveraging event inside it?
2. Do the drivers that took gold from US$2,000 to US$5,589 still exist at today's price?
3. Which of gold and silver offers the better risk-reward from here, and what would prove the thesis wrong?

A. Anatomy of the correction: what broke, and what did not

• The move down was fast and it was mechanical. Both metals topped within 24 hours of each other. On 29 January gold fell US$380, close to 7%, in 28 minutes, while silver lost 11% in the same window. By 30 January gold was more than 10% off its high and silver more than 30% off its own.
• Realised volatility in gold went above 50%, against a 20-year average of about 17%. It has since settled back below 30%, which is still elevated but no longer disorderly.

Chart 1: Gold and silver, monthly, January 2024 to August 2026

Source: LBMA / COMEX front-month settlement; August 2026 figure is the 6 August intraday level. Monthly points are month-end closes and are rounded.

The five things that actually caused it:

• Leadership change at the Fed. Kevin Warsh took over as Chair on 13 May 2026 having described inflation as "a choice". Markets repriced the terminal rate upward and stripped out the cuts that had been embedded in the January price. A metal with no coupon carries that repricing directly.
• Margin and forced deleveraging. CME raised margin requirements into the spike. Silver's margin change in particular forced leveraged longs to liquidate. Because margin was being calculated against a rising contract value, the requirement rose with the price, which turns a normal correction into a cascade. The same mechanism had produced roughly US$6.5bn of forced selling in 2013.
• A dollar squeeze. DXY pushed to a 13-month high through May-June as the market moved from pricing cuts to pricing a hike. Gold is priced in dollars, so a firmer dollar mechanically suppresses the price for every non-dollar buyer.
• An oil shock that read as hawkish, not dovish. WTI went from about US$57/bbl at the start of the year to a peak of US$113 in April on the US-Iran conflict and the closure of the Strait of Hormuz. Energy pushed core PCE from 3.0% in December 2025 to 3.4% by May 2026. Rather than reading that as a debasement signal, the market read it as a reason the Fed would hike, and sold gold.
• Positioning. After a 290%-plus move in silver over two years and twelve all-time highs in gold, the trade was crowded. The gold-silver ratio had compressed to about 46x, its tightest since 2011, which is a reliable marker of silver overshoot. Speculative length and ETF holdings both had a long way to fall before they were merely neutral.

Note what is absent from that list. There is no deterioration in central bank demand, no repair of the fiscal position, no return of the dollar's real yield advantage in absolute terms, and no resolution of the silver supply deficit. Every one of the five causes is a price-of-money or price-of-leverage story. None of them is a store-of-value story.

Chart 2: Corrections of this size are normal inside gold bull markets

Source: Bloomberg, World Gold Council drawdown analysis (data 1 January 1971 to 26 June 2026); 2026 figure measured from the 28 January peak to the late-June low. The 1980-82 and 2011-15 episodes are shown for contrast: those were cycle terminations, and both were considerably deeper than the current move.

How the correction looked in numbers:

Source: LBMA, COMEX, GoldSilver price series. Ratio computed on the same-day fixes; 2024 and 2025 closes are London PM fixes.

B. Where the market is today: the macro tape

• The Federal Reserve held the funds rate at 3.50-3.75% on 29 July 2026, its fifth consecutive hold. The vote was 9-3, with Hammack, Kashkari and Logan dissenting in favour of a 25bp hike. That is the widest split in favour of tightening since September 2016.
• Chair Warsh described the meeting as the "family fight" he had asked for, declined to give forward guidance, and justified the hold on the grounds that market-driven tightening in nominal and real rates had already done some of the work.
• Immediately before the meeting, CME FedWatch put the odds of a September hike at 78.8%. After Warsh's press conference they fell to 60.1%. As of early August, with oil falling and July ADP hiring slowing, they have eased further. Gold rose from US$4,029 to US$4,100 on the day of the decision and has since added a further 3.5%.
• Inflation is coming down slowly rather than not at all. Headline CPI eased to 3.5% in June, its first decline in five months. Core PCE was 3.4% in May against a 2% target. The June and July energy pullback should show through in the next two prints.
• The rate structure remains restrictive in absolute terms. The 10-year Treasury yields 4.63%, the 30-year 5.20%, and the 10-year TIPS real yield sits at 2.41% - up 0.53pp year-on-year. That is a genuine headwind for a zero-coupon asset and it deserves to be stated plainly.
• The dollar has stopped rising. DXY has slipped from its June high back to roughly 100, having failed to hold above 101.5. Momentum, not level, is what matters for metals.

Chart 3: The correction tracks the real yield, which is exactly what should happen

Source: US Treasury daily TIPS yield curve, LBMA. The two series moved together through the drawdown; the question for the next twelve months is what happens to the real yield, not whether the relationship is intact.

Key macro markers relevant for bullion and their trajectory:

Source: Federal Reserve, US Treasury, BEA, CME FedWatch, ICE, LBMA. January real-yield and DXY figures are approximate month-average levels.

The single most important observation in this section:

Every macro variable that caused the correction has either peaked or begun to reverse, with one exception - the absolute level of the real yield. That exception is the whole bear case, and this is captured exhaustively in Section H.

C. The monetary case: the arithmetic has not changed

The reason gold went from US$2,000 to US$5,589 in twenty-eight months was never the Fed's next 25bp. It was the growing suspicion that the sovereign balance sheet behind the world's reserve currency cannot be repaired through growth or taxation, and will therefore be repaired through the currency. That suspicion is arithmetic, not sentiment, and the arithmetic has got worse during the correction, not better.

C.1 Gold against the money supply

• Valuing the entire above-ground gold stock against US M2 gives a long-run measure of how much monetary claim is outstanding relative to the anchor asset. At the January peak that ratio reached about 171%, the highest reading in the hundred-year series.
• It has crossed 120% only twice before. In 1934, when Washington devalued the dollar by 40% overnight because it had issued far more dollars than it held gold against. And heading into 1980, after a decade in which the dollar had already been devalued by 11% and the Bretton Woods link had been abandoned altogether.
• On both prior occasions the ratio was signalling the same thing: the stock of paper claims had run too far ahead of the anchor, and something had to give. On both occasions what gave was the currency, not the gold price.
• Today's correction has taken the ratio back to roughly 131%. That is still above the threshold that has historically preceded a monetary reset, and it is still the third-highest reading on record.

Chart 4: Value of the above-ground gold stock as a percentage of US M2, 1920-2026

Source: reconstructed by me from Metals Focus / World Gold Council above-ground stock estimates, the LBMA gold price, and US M2 (FRED series M2SL and its historical antecedents). Pre-1959 M2 is spliced from Friedman-Schwartz monetary aggregates and the series is indicative rather than exact. Shaded segments mark the three episodes above the 120% threshold.

C.2 The fiscal position, which is the actual driver

• US gross federal debt crossed US$39 trillion on 17 March 2026 and stood at US$39.84 trillion on 30 July. It has risen by roughly US$2.7 trillion year-on-year and US$10.75 trillion over five years, without a recession, a stimulus package or a full-scale war to explain it.
• The CBO puts net interest at about US$1.04 trillion in FY2026, up from US$345bn in FY2020. Interest is now a larger line item than national defence and is projected to overtake Medicare around 2027. It absorbs an estimated 13.95% of federal outlays this year, rising to 14.94% by FY2028.
• The weighted-average coupon on outstanding Treasury debt is 3.41% as at 30 June 2026, against roughly 1.5% in 2021. That number is still climbing, because low-coupon paper issued in 2020-21 keeps maturing into a 4-5% curve. The interest bill rises even if the Fed does nothing and the deficit does not widen.
• Debt held by the public is US$32.04 trillion, and the July quarterly refunding statement pointed to elevated issuance across the curve with a particular expansion at the long end. Larger auctions into a market that is already absorbing record supply is how term premium gets rebuilt.

Debt and the cost of carrying it

Source: US Treasury Debt to the Penny, Congressional Budget Office, Peter G. Peterson Foundation monthly interest tracker. FY26E onward are CBO projections.

C.3 The trap this creates:

• Raise rates to break inflation, and the marginal cost of rolling US$32 trillion of public debt rises against an interest bill that already exceeds a trillion dollars. Fiscal dominance stops being a theoretical term.
• Cut rates to protect growth, and money supply expansion outruns the compensation bondholders receive for it. The interest-adjusted debasement rate - the 10-year yield less the growth rate of money supply - turns negative, which is precisely the condition under which gold outperformed for the whole of the 1970s.
• The historical precedent is instructive and counterintuitive. Through the 1970s the Fed took the funds rate to 13%, and then to nearly 20%. Gold still multiplied roughly twentyfold over the decade, from US$35 to over US$800. Rising nominal rates did not defeat gold, because inflation and money growth were rising faster.

This is the pivot of the whole argument. A hawkish Fed is bearish for gold when it is credible and when the sovereign can afford the policy. In 2026 the Fed is credible but the sovereign is increasingly constrained, and the market can see both facts at once.

D. What gold and silver are actually worth

Everything in Section C says the monetary case is intact. It does not say what the metal is worth. For that we need a valuation anchor, and gold is famously hard to value because it produces no cash flow. The most useful framework we have found takes the assumption in Section C literally: if gold is money, its price should track the money it is measured against.

D.1 How the framework works

• Take the entire above-ground stock of gold ever mined, and divide the money supply of the reserve-currency blocs by it. The result is what an ounce would be worth if that money were fully backed by that gold.
• The lower bound uses US M2 alone. The upper bound adds half of euro area M2. Half, rather than all, because the euro area itself holds gold and US dollars as reserves, so counting the whole of EU M2 would double-count the same monetary claim.
• The midpoint of those two bounds is the reference price. It is a slow-moving anchor: it changes only as money supply grows or as miners add to the above-ground stock, which is roughly 1.7% a year.
• What the framework does not do is tell you when. It is a statement about value, not about timing, and we will deal with this in Section D.4 in detail.

The framework as of August 2026 available data

D.2 What the framework says today, and what it said in June

• On these inputs gold's modelled range is US$3,248 to US$4,595, with a midpoint of US$3,922. At US$4,242, gold trades about 8% above the midpoint and about 8% below the upper band.
• The interesting part is not where gold is. It is where gold went. At the January peak of US$5,589 the metal was trading 22% above the upper band - not above the midpoint, above the ceiling of the entire range. That is the definition of an overshoot, and it is the cleanest single piece of evidence that the January top was speculative rather than fundamental.
• Then look at the June low. Gold bottomed at US$3,985, which is within 2% of the modelled midpoint of US$3,922. The correction stopped almost exactly at fair value and turned.

I do not think that is a coincidence. It is what you would expect if the marginal seller through the drawdown was leveraged and speculative, and the marginal buyer at the bottom was valuation-driven - which is precisely what the central bank data in Section E shows.

Lets understand how historically Gold has been against its modelled value, 2000 to 2026

Source: Federal Reserve H.6 (US M2), ECB (euro area M2), Metals Focus and World Gold Council (above-ground stock), LBMA (price); framework after DSP Netra. Historical money supply for the euro area converted at prevailing year-end exchange rates. Actual gold plotted at year-end except 2026, which is spot at 6 August.

Read plainly: Gold spent twenty-five years below its modelled value, closed the gap in 2025, overshot the entire range in January 2026, and has now settled between the midpoint and the upper band. That is a market that has repriced to fair value and is deciding what to do next - not one that has exhausted itself.

D.3 Silver, valued through the ratio

• Silver has no independent monetary anchor, so the framework values it through its ratio to gold. History gives a wide but instructive range. Under the Roman Empire the ratio was roughly 12:1; in medieval Europe 9.4:1; the US Coinage Act of 1792 fixed it at 15:1. Once the monetary link was cut the ratio blew out - to 98:1 when the US revalued gold to $35 in 1939, and to 97.5:1 after the gold standard was abandoned altogether.
• The 21st century average is 69:1. Using a normalised 60:1 against our gold bands gives silver a modelled range of US$54 to US$77 with a midpoint of US$65.
• At US$61.7, silver trades about 6% below its modelled midpoint. Gold trades 8% above. On this framework silver is the cheaper of the two, which is the opposite of what the January price action implied.
• The ratio itself tells the same story. At the January peak the ratio compressed to 46x, tighter than at any point since 2011 and far through the 60:1 normalisation assumption. Today it sits at 69x, almost exactly the 21st century average. Silver has given back its entire relative overshoot while gold has given back only part of its absolute one.

The gold-silver ratio through five decades

Source: LBMA daily fixes. Marked extremes follow the DSP Netra annotation of the same series. The 18x low of January 1980 is the Hunt brothers' corner; the 119x spike of March 2020 is the COVID liquidity event. The long-run average of roughly 60x is the level the framework assumes for normalisation.

D.4 What this framework gets right, and where it fails

• It called the June low. A model built on money supply and above-ground stock, with no knowledge of the Fed, the Strait of Hormuz or CME margin policy, put fair value at US$3,922, and the market turned at US$3,985. That is a strong result.
• It also failed as a timing tool, and we want to be explicit about that. If we run this framework in October 2025 with gold at US$3,895 and fair value midpoint of US$3,825, and conclude that it was time to turn conservative and trim into strength between US$3,860 and US$4,000. Then we would be grossly wrong as Gold went on to trade at US$5,589 before it actually corrected. Anyone who sold the whole position at US$4,000 missed a 40% move.
• The lesson is not that the framework is wrong. It is that money supply sets the anchor while flows set the price, and flows can hold a market far above its anchor for a long time. Gold spent the whole of 1980-2010 below its modelled value; there is no rule that says it cannot spend years above it.
• There are three structural limitations worth stating. The model treats all above-ground gold as monetary, when roughly 45% of it sits in jewellery form and is not realistically available. It uses M2 rather than a broader measure of claims on the state, which understates the problem described in Section C.2. And it carries no term for official sector (Central bank) demand, which is the single largest change in this market since 2022.

So we will use it as a valuation anchor and not as a signal. It tells us that at US$4,242 gold is no longer cheap, that at US$61.7 silver still is, and that the January price was indefensible on any monetary measure. It does not tell me what happens next quarter.

D.5 The record this framework has to explain

• Gold returned 60% in 2025, its best calendar year since 1979. Silver returned 147%. Those are not normal numbers, and it is worth seeing them in the context of a fifty-five year series.
• Note what the same chart shows for 2026: after a 34% run to the January peak and a 29% collapse, gold is up roughly 1% year to date. The entire drama of this year has produced almost no net movement in price, which is exactly what a reset looks like from a distance.

Annual gold returns, 1971 to 2026 YTD

Source: LBMA annual returns in US dollars. 2026 is year-to-date at 6 August, measured from the 2025 close of US$4,190.

• The longer record is more striking: Over the 21st century to date, gold has outperformed the domestic equity market of every major country measured in local currency. Not most. Every one.
• In developed markets the excess runs from 0.7 percentage points a year in Australia to 7.3 in Japan. In emerging markets it runs from 0.3 points in India to 9.6 in Turkey. A non-yielding asset with no earnings, no management and no buybacks has beaten the equity index of every one of these economies over twenty-five years.
• The stock-level data sharpens it further. Over twenty years only 7% of S&P 500 constituents and 3% of FTSE 100 constituents outperformed gold. India is the notable exception at 35% of the NSE 500, which is the arithmetic behind gold's unusually thin 0.3 point margin over Indian equities.

Gold's excess return over local equity markets, 21st century to date

Source: DSP Netra, October 2025, data since 31 December 1999; all market returns are total return indices in local currency. Gold's 2026 drawdown reduces these figures at the margin but does not change the ranking or the conclusion.

Table 4: Gold through previous bull phases, versus equities and bonds

Source: DSP Netra for the first five episodes (S&P 500 returns are total return; for periods prior to 1989 the TRI is estimated using the average dividend yield from 1975 to 1989). The final row extends the DSP series from September 2022 to the January 2026 peak using LBMA and Bloomberg data. The current episode is already the second-strongest of the six on gold's CAGR.

The pattern across those six episodes is that gold's outperformance has been front-loaded and violent, and that it has ended when the real rate structure turned decisively positive and stayed there. That is the single condition we are watching, and it is the subject of Section H.

E. The official (Central Banks) sector: the bid that does not care about the price

• On 30 July the World Gold Council reported that central banks bought a net 288.9 tonnes in Q2 CY26, up 62% on the 177.9 tonnes of Q2 CY25 and the strongest second quarter in the data series. They did that during a quarter in which the gold price fell around 16%.
• The National Bank of Poland was the largest buyer at 51 tonnes, taking reserves to 632 tonnes against a stated target of 700. The People's Bank of China added 33 tonnes, its largest quarterly addition since Q4 2023, lifting holdings to 2,346 tonnes. Uzbekistan added 16 tonnes, Kazakhstan 15 tonnes, and Jordan and the Czech Republic six tonnes each.
• The honest counterpoint is that first-half demand was weak. Metals Focus revised Q1 down sharply, from an initially reported 244 tonnes to just 57 tonnes, which puts H1 CY26 at about 345 tonnes - the lowest half-year since 2022. Turkey, Russia and Azerbaijan were meaningful sellers.
• The sellers are worth understanding, because their motives are the opposite of a loss of faith in the asset. Russia sold 22 tonnes in the quarter and 43.5 tonnes in H1 to plug a federal budget deficit approaching six trillion roubles. Turkey has been swapping gold for dollars to meet an energy import bill inflated by a 40% rise in crude. These are liquidity-driven sales by constrained sovereigns, not reserve-allocation decisions.
• The stated intentions are unambiguous. The WGC's 2026 Central Bank Gold Reserves Survey found 89% of reserve managers expect global central bank gold holdings to rise over the next twelve months, and 74% expect to hold fewer US dollars over the next five years.

Official sector demand - annual and quarterly

Source: World Gold Council, Gold Demand Trends Q2 2026 (published 30 July 2026) and prior editions; Metals Focus. Q1 CY26 shown at the revised 57t. Annual purchases have averaged roughly 1,000 tonnes a year since 2022, against 400-600 tonnes in the decade before.

E.1 Custody is the second signal

• Volume is one thing; location is another. France, the fourth-largest sovereign holder, has repatriated the entirety of the physical gold it held in the United States. Germany, the second-largest, is debating the same for the roughly one third of its holdings still vaulted at the Federal Reserve.
• When close allies move from holding a claim on metal to holding the metal itself, the concern being expressed is counterparty and jurisdiction risk, not price. It is a slow-moving signal and it does not show up in any daily print, but it is the same signal the 1960s reserve diversification cycle produced ahead of the collapse of Bretton Woods.

E.2 The divergence that defines this market

• Private capital did the exact opposite of the official sector through the drawdown. Gold ETFs recorded net outflows of 16 tonnes in May and continued bleeding into June before a US$1.1bn weekly inflow broke a four-week run of redemptions.
• Standard Chartered estimated in late June that around 298 tonnes of gold held inside ETFs was underwater at a US$4,000 price, up from 270 tonnes when gold was above US$4,250. Those are traders waiting for a level at which to exit, and they are a genuine supply overhang on the way back up.
• The distinction matters. ETF holders sell on drawdown; central banks buy on drawdown. One of those two groups sets the price on a three-month view and the other sets the floor on a three-year view.

Two buyers, opposite behaviour

Source: World Gold Council ETF flow data; CME Group COMEX warehouse and open interest reports. Silver paper leverage is computed as open interest (contracts x 5,000 oz) divided by registered inventory.

Q2 CY26 official sector activity

Source: World Gold Council, Gold Demand Trends Q2 2026. Turkey and Azerbaijan quantities not separately disclosed in the summary release.

F. Silver: A monetary asset with an industrial floor under it

Silver fell twice as hard as gold and for good reason. The market is roughly one tenth the size, about 60% of demand is industrial, and it carries far more speculative leverage per dollar of physical metal. That is the risk. The same characteristics are what create the asymmetry on the way back.

F.1 The sixth consecutive deficit

• The Silver Institute's World Silver Survey 2026, produced with Metals Focus, puts the market in deficit for a sixth straight year, widening to 46.3 Moz in 2026 from 40.3 Moz in 2025.
• Cumulatively, 762 Moz has been drawn out of above-ground stocks since 2021 - close to a full year of global mine production, gone.
• The deficit is widening for an uncomfortable reason: supply is contracting faster than demand. Global mine supply grew 3% to 846.6 Moz in 2025, but production has been effectively flat near 830 Moz a year since 2015. Primary silver mines are rare, most silver arrives as a by-product of copper, lead and zinc, and permitting-to-production timelines run close to a decade. Supply cannot respond to price.

Silver market balance and cumulative inventory drawdown

Source: Silver Institute / Metals Focus, World Silver Survey 2026 (15 April 2026). 2026 figures are forecasts. Cumulative drawdown is the running sum of annual deficits from 2021.

F.2 The solar bear argument, and why it is weaker than it looks

• Photovoltaic silver demand fell 6% to 186.6 Moz in 2025 and is forecast to fall a further 19% in 2026 to about 151 Moz - the largest single-year reduction on record. Total industrial fabrication is expected to decline 2% to roughly 650 Moz, a four-year low.
• That sounds bearish until you look at the mechanism. The reduction is thrifting, meaning less silver per cell, not substitution, meaning silver replaced by something else. Thrifting has a floor: below a certain loading, cell efficiency and reliability degrade.
• Copper substitution in the dominant TOPCon architecture still has unresolved reliability problems, and mass adoption is estimated at 2028-2030, not now.
• And the deficit widened anyway. The largest recorded fall in the largest single source of industrial demand was not enough to balance the market. That is a stronger statement about the supply side than any bullish demand forecast.

F.3 Where the marginal demand is coming from instead

• Physical investment demand is forecast to rise 20% to a three-year high of 227 Moz in 2026. US retail demand is forecast to rebound 57% after three consecutive years of decline.
• Industrial applications still account for roughly 60% of total consumption, up from about 50% a decade ago, across solar, electronics, EVs, data-centre infrastructure and 5G.
• The composition of the deficit has shifted from an industrial crunch to a retail and investment squeeze. For price formation that is arguably more powerful, because investment demand is price-insensitive in a way that a solar module bill of materials never is.

F.4 Market structure: the leverage that cuts both ways

• COMEX registered silver inventory sits near 96 Moz, having recovered from a March low of 76 Moz. Against outstanding paper claims, the coverage ratio is roughly 17.8% and paper leverage roughly 5.6x.
• In the first week of January 2026, 33.45 Moz - about 26% of registered inventory - left COMEX warehouses in seven days. Over one four-week stretch roughly 90 Moz left COMEX and Western ETFs combined, with India the primary destination.
• LBMA vault holdings are stable but not expanding, and intermittent backwardation in short-dated contracts has reappeared, which is the market's way of saying that metal for immediate delivery is worth more than metal later.
• This is why silver's volatility is asymmetric. Downside moves grind lower under hedging pressure. Upside moves can gap, because when sentiment turns there is very little deliverable metal standing between a bid and a much higher price.

Silver supply and demand, 2021-2026F

Source: Silver Institute / Metals Focus, World Silver Survey 2026.

G. Where the market is and where it goes next: a phase framework

It helps to think of monetary metals repricing in phases, in the same way that grid operators think about renewable penetration. Each phase has a different marginal buyer and a different binding constraint. The market is currently sitting on the boundary between phase three and phase four.

• Through 2024 and 2025 the market moved from phase two to phase three. The official sector took over as the marginal buyer, and it bought roughly 1,000 tonnes a year for four consecutive years against 400-600 tonnes a year in the preceding decade.
• The January 2026 top was an attempt to jump straight to phase four on retail and momentum flows rather than on institutional reallocation. It failed, and the failure is what the 30% drawdown was.
• Phase four requires a different buyer: pension funds, insurers, multi-asset allocators and sovereign wealth funds moving from a 0-1% allocation to a 3-5% allocation. Those flows are slow, they respond to realised volatility falling rather than to price rising, and they are only now becoming possible as gold volatility drops back below 30%.

G.1 What the sell side thinks

Published 2026 targets against spot

Source: Goldman Sachs (year-end, revised 20 June 2026), Deutsche Bank (Q4 CY26 target and model fair value, 3 August 2026), RBC Capital Markets (2026 average), J.P. Morgan (Q4 CY26), Commerzbank (year-end), World Gold Council mid-year outlook 2026 (fair value estimate, +/-5%). Silver: Commerzbank year-end, ING, Reuters poll of 30 analysts (median), J.P. Morgan 2026 average, Bank of America bull-case range.

• Deutsche Bank's note of 3 August is the most analytically interesting. Michael Hsueh and Bryant Xu did not lead with a price target; they led with a statistical test. Using a BSADF measure of explosive price behaviour, they identify only five such episodes in gold since 1975. The current one began in August 2024 and is still running: the statistic peaked at 3.3 earlier this year and has moderated to 1.3, but remains above the 95% critical value.
• Their conclusion is that the correction has been muted relative to prior explosive phases, which they read as evidence of underlying demand rather than exhaustion. Their model fair value is US$4,700 against a maintained Q4 target of US$4,600, and they put the floor around US$3,900 rather than the US$2,600 that simple real-rate-and-dollar models produce, because those models do not carry central bank demand.
• The World Gold Council's mid-year outlook is the most conservative of the mainstream views, putting fair value near US$4,100 with a lower bound around US$3,895. It is worth noting that spot is already above that estimate.
• Silver forecasts are unusually dispersed even by silver's standards: Commerzbank at US$67, ING at US$78, a Reuters poll of 30 analysts with a median at US$79.50, J.P. Morgan at a US$81 full-year average, and a Bank of America bull case running from US$135 upward if physical shortages intensify. The width of that range is itself information about how binding the physical constraint could become.

G.2 Our scenario framework

Scenarios to end-2026

Probabilities are our own judgement and are not derived from an option-implied distribution. Silver ranges assume the gold-silver ratio trades between 50x and 72x.

G.3 The calendar between now and the decision point

• Mid-August 2026 - July CPI. The first clean read on whether the energy pullback is feeding through to core.
• Late August 2026 - the Jackson Hole symposium. Historically where a Fed chair signals a regime shift, and Warsh's first appearance in that setting.
• 15-16 September 2026 - FOMC with a full Summary of Economic Projections and a new dot plot. The June dots had nine members projecting at least one hike in 2026 against eight projecting none. The September dots resolve that.
• Ongoing - Strait of Hormuz negotiations. A durable reopening removes the energy-inflation impulse, which cuts both ways: it lowers the odds of a hike, which is bullish, while removing a geopolitical risk premium, which is not.
• Late October 2026 - WGC Gold Demand Trends Q3, which tells me whether the Q2 record in official sector buying was a catch-up after a weak Q1 or the resumption of trend.

H. The bear case, stated properly

A note that only argues one side is not research. The following are the strongest arguments against everything above, and none of them is trivial.

• Real yields are still positive in absolute terms. At 2.41% on the 10-year TIPS, an investor is being paid a real return to hold a government-guaranteed instrument. Gold pays nothing. Momentum matters, but a 2.4% real yield is genuine competition and it is the highest in over a decade.
• First-half official sector demand was the weakest since 2022. 345 tonnes for H1 is a fact, and the Q1 revision from 244 tonnes to 57 tonnes was a downgrade of more than three quarters. Read soberly, the record Q2 was partly catch-up. Deutsche Bank itself built a slower central bank pace into its fair value model.
• There is a visible supply overhang in the ETF complex. Roughly 298 tonnes sits at a loss around US$4,000. Those holders sell into strength, which caps rallies until the overhang clears.
• A September hike is a live possibility. Three FOMC dissents is the most hawkish dissent count since 2016, and one to two hikes by end-2026 was still being priced in the days after the July meeting. A delivered hike would strengthen the dollar and lift real yields further.
• Peace is bearish in the near term. The reopening of the Strait of Hormuz has already knocked crude from US$113 to the mid-US$80s. Lower energy prices lower headline inflation, which is helpful for rate expectations, but they also remove the geopolitical bid that took gold to US$5,589 in the first place.
• Silver's industrial floor is being eroded at the margin. Solar thrifting is real, the 19% reduction is the largest on record, and copper substitution becomes commercially viable somewhere in 2028-2030. The deficit narrative has a shelf life.
• Risk assets are behaving well. Equity markets remain resilient, which reduces the urgency of defensive allocation. Gold does its best work when investors are frightened, and at present they are not.

On the honest arithmetic, the bear case is mainly a case about the next two quarters. The bull case is mainly a case about the next three years. Both can be right, and an investor's answer depends far more on horizon than on conviction.

I. Conclusion and way forward

• The correction did its job. It cleared roughly 30% of price from gold and 50% from silver, removed the leverage that CME's margin regime had made unstable, reset the gold-silver ratio from a stretched 46x to a neutral 69x, and pushed realised volatility back toward a level at which institutional allocators can underwrite a position. Nothing was repaired on the fiscal side while that happened.
• The structural bid is intact and is now demonstrably price-insensitive. Central banks bought a record second quarter into a 16% price decline. Eighty-nine per cent of reserve managers expect global official gold holdings to rise, and seventy-four per cent expect to hold fewer dollars in five years. That is not a trade; it is a change in what sovereigns consider a safe asset.
• Silver is the higher-beta expression of the same view with an additional, independent support. A sixth consecutive physical deficit, 762 Moz drawn from above-ground stocks since 2021, flat mine supply for a decade, and a paper-to-physical leverage of roughly 5.6x on COMEX is a configuration that resolves violently when it resolves. It may not resolve this quarter.
• The direction of the real yield, not its level, is the variable to track. Every prior turn in this cycle - late 2007, mid-2019, 2002-03 on the dollar - saw metals begin to reprice before the policy pivot was announced, not after.
• However, in short term – position sizing should reflect that the path is genuinely two-sided. A September hike, a durable Hormuz settlement and a further leg down in oil would produce a retest of the June lows. Our base case is that it would be a retest rather than a break, given a published floor consensus clustered around US$3,900.

What we would watch to change our mind

• Central bank net purchases falling below 150 tonnes in Q3 CY26 would be the single most damaging data point to the structural case.
• COMEX registered silver climbing back above 130 Moz with sustained contango would remove the physical tightness argument.
• A move in the 10-year TIPS real yield above 2.75% sustained for a quarter would make the opportunity cost argument decisive.
• Confirmation of commercial copper metallisation in TOPCon cells ahead of 2028 would materially shorten the shelf life of silver's deficit.

Conversely, a September hold combined with a soft August payroll print and a 10-year real yield below 2.20% would, in our reading, mark the start of the next leg.

Annexure I: What changed between the pre-correction and post-correction market

Annexure II: Data appendix

Definitions. Real yield is the yield on Treasury Inflation-Protected Securities, i.e. the nominal yield less market-implied inflation. Structural deficit in silver means annual consumption exceeds mine supply plus recycling, with the gap met from above-ground stocks. Registered COMEX inventory is metal available for delivery against futures; eligible metal meets exchange standards but is not offered for delivery. Thrifting means using less of a metal per unit of output; substitution means replacing it. BSADF is the backward supremum augmented Dickey-Fuller statistic, used to test for explosive rather than merely trending price behaviour.

Turn insights into action — invest in Index ETFs and Derivatives now.  
Open your MNCL Demat Account today!

Frequently Asked Questions

Is the 2026 gold correction a sign that the bull market is over?

Monarch PMS views the January-June correction as a deleveraging event within the broader bull market rather than a confirmed cycle top. The report notes that the correction removed leverage and speculative excess while structural drivers remained in place.

What is the gold outlook for 2026?

The report's base scenario sees gold stabilising and grinding higher toward US$4,300–4,700 by end-2026, subject to the stated macro triggers. This is a scenario framework, not a guaranteed price target.

What is the silver outlook for 2026?

The report's base scenario places silver at US$70–85, while its bull scenario reaches US$95–120 if physical tightness and institutional demand strengthen.

Why is silver considered relatively attractive?

The report's valuation framework places silver's midpoint at US$65 compared with a 6 August spot price of US$61.7. The sixth consecutive projected market deficit and constrained mine supply provide additional structural support.

What is the biggest risk to gold?

The key risk identified is persistently high real yields. A 10-year TIPS real yield sustained above 2.75% for a quarter would make the opportunity-cost argument materially more challenging for gold.

Should investors buy gold or silver based on this report?

The report provides a market framework and scenario analysis rather than a personalised investment recommendation. Investors should consider their objectives, risk profile, investment horizon and suitability before making investment decisions.

Disclaimer: Monarch Networth Capital Limited (MNCL) is a SEBI-registered Portfolio Manager (Registration No. INP000006059). Investments in securities market are subject to market risks. This content is for educational purposes only. Read full disclaimer