Silver Market Structure Monitor

THIS WEEK▲Fund net 19.9%→20.5% OI▼SHFE warrants +6.6% w/w▲Shanghai premium +0.29 to $-0.13vs prior week

§01

Weekly Read

THE READ

SILVER · AG / SISqueeze-risk level: LowSep First Notice Day · 7 trading days

After the fund unwind, silver remains structurally loose—with dry powder still sidelined

Managed-money net length is 20.5% of OI (41st percentile, down 3.3pp over four weeks) and still 16pp below the historical 85th-percentile launch zone. Top-four net short concentration is 28.3% (27th percentile), so there is little trapped-short fuel. SHFE silver warrants are +39.9% over four weeks, thickening China's deliverable pool. The Shanghai premium is $-0.13/oz (+0.40 over four weeks). A sustained widening would be the first reversal signal. The London 1M implied lease rate is 0.03% as of 31 Jul 2026 (past the freshness limit), in the normal band.

Managed-money net · % OI

20.5%

4w -3.3pp · 41st percentile

Top-four net shorts

28.3%

27th percentile

SHFE silver warrants

1,369 t

4w +39.9%

Shanghai premium

$-0.13/oz

4w +0.40

London 1M lease rate

0.03%

31 Jul 2026 · Normal · 20d old

§02

Structural Gauges

Five public states · thresholds and weights remain internalGAUGES

Five-factor state

Short structureRelaxed

Top-four net shorts 28.3% (27th percentile); the 2021 and 2024 episodes both moved above the 37% crowding area.

Fund crowdingRelaxed

Net length 20.5% of OI (41st percentile); past events launched around the 85th-percentile zone (~36%).

Delivery coverageWatch

Total futures claims equal 5.8× registered inventory.

Inventory momentumRelaxed

SHFE four weeks +39.9% · LBMA three months +2.8%; sustained draws are the entry ticket for a physical squeeze.

Spread confirmationRelaxed

Shanghai premium $-0.13/oz · four weeks +0.40; a widening premium signals Chinese demand pulling global metal.

Deliverable pool

COMEX registered · 2026-08-19

3,096 t

eligible 7,396 t · 13 depositories

Coverage compares the silver equivalent of total open interest with COMEX registered inventory; each SI contract represents 5,000 oz.

Inside the internal edition

Silver desk viewThe week's front-line read, risk flags and monitoring orderINTERNAL

Score breakdown and trigger levelsFive-factor construction and the lease-rate hard overlay behind the Low ratingINTERNAL

Strategy frameworkDirectional, structural and execution guidance for the current regimeINTERNAL

Unlock the internal edition →

Futures claims vs deliverable inventory

The multiple between paper claims and registered metal is the system's stored squeeze potential

§03

Lease-Rate Stress

Publicly reported implied 1M · hard squeeze overlayLONDON BORROWING

Current borrowing temperature

1M0.03%Normal · No acute borrowing stress

31 Jul 2026 · 20 days old · stale, reference only

2019–2024 normal-regime average -0.09%

Tier B · dated public quote · source register in internal edition

Why this is a hard squeeze indicator

The lease rate measures the marginal cost of obtaining deliverable bars. When spot metal is tight and lendable free float dries up, short-tenor rates can jump before monthly vault data reacts; 5% marks acute stress and 10% an extreme squeeze.

A fresh tight, acute or extreme reading can only lift the five-factor risk regime; it can never lower it. After 14 days the quote becomes reference-only and the score overlay switches off.

Specific sources and original links are available in the internal edition

Six latest verifiable direct quotes

Tier BLatest31 Jul 2026

0.03%Source register in internal edition

Tier B24 Jul 2026

0.00%Source register in internal edition

Tier B17 Jul 2026

-0.17%Source register in internal edition

Tier B10 Jul 2026

0.00%Source register in internal edition

Tier B3 Jul 2026

0.10%Source register in internal edition

Tier B19 Jun 2026

0.23%Source register in internal edition

Public 1M quote trail in 2026

19 dated direct quotes · observations are not interpolated · latest point in navy

Data quality and search boundary

Tier A

Official daily subscription data

Daily T+1 lease, loan and deposit data covers standard tenors. Exact readings are institutional subscription data, not a free continuous public series.

Tier B

Direct public quote chain

The archive contains 19 dated 1M public observations for 2026. The latest is 0.03% on 31 Jul 2026; the itemised source register remains in the internal edition.

Tier A

Public physical corroboration

Vault totals and industry surveys cross-check physical tightness. Headline inventory is not lendable free float and does not replace the lease rate.

Tier B

Institutional context

Institutional commentary is used only when tenor and measurement basis can be assessed; unspecified-tenor observations are excluded from scoring.

Tier C

Term-curve cross-check

Multi-tenor curves cross-check front-end and long-end structure but cannot override a dated direct quote for the same tenor.

Tier B

Domestic specialist channels

Domestic futures, precious-metals and research observations remain a separate parallel lineage and do not enter the canonical trail or score.

Tier C

Historical archive recovery

Term curves recovered from historical pages and PDFs preserve source order and are not interpolated into a continuous series.

Web search completed 2026-08-12; institution names, report titles and source links are available only in the internal source register.

Domestic cross-check archiveParallel evidence only; excluded from the canonical series and score31 modern72 archive15 curvesView audit samplesCollapse▾

China-channel public observations

These observations are parallel lineage and cross-checks; they do not replace the canonical direct quote trail or scoring.

Domestic specialist-channel samples are an independent parallel data layer for cross-checks, same-day dispersion and archive recovery. They do not enter recent_1m_quotes, latest or scoring; every observation is canonical_eligible=false.

Same-day dispersion example (2026-01-09): a parallel derived reading of 3.76% is shown alongside the 4.46% canonical public quote; neither replaces the other.

Modern observations

Date / as-of	Tenor · reading	Evidence type	Lineage / role
2026-07-21 / published 2026-07-21	1M <1%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-04-17 / published 2026-04-19	1M 0.73%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-04-10 / published 2026-04-10	— 0.07%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-03-27 / published 2026-03-29	1M 0.49%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-02-06 / published 2026-02-06	1M 4.5%, 3M 4.6%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-30 / published 2026-02-02	1M 1.88%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-28 / published 2026-02-05	1M 0.6%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-27 / published 2026-01-27	1M 2.3%, 3M 3.4%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-25 / published 2026-01-25	1M 8%-12% range	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-15 / published 2026-01-16	1M 4%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-14 / published 2026-01-14	1M 2.8%, 3M 4.8%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-09 / published 2026-01-12	1M 3.76%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-07 / published 2026-01-08	1M 5.89%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-05 / published 2026-01-05	1M 8.4%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-04 / published 2026-01-05	1M 10.25%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2026-01-04 / published 2026-01-04	3M 8.4%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-12-31 / published 2026-01-05	3M 8.07%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-12-31 / published 2025-12-31	1M 9.01%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-12-19 / published 2025-12-22	— 8.1%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-11-20 / published 2025-11-20	1M 5.8%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-11-17 / published 2025-11-19	1M 6.07%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-11-10 / published 2025-11-10	1M 5.3895%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-10-22 / published 2025-10-22	1M 16.39%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-10-17 / published 2025-10-20	1M 15.9%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-10-16 / published 2025-10-17	1M 25.9%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-10-13 / published 2025-10-14	1M 42.72%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-10-10 / published 2025-10-20	1M 40.3%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-09-22 / published 2025-09-22	1M 4.9046%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-07-14 / published 2025-07-14	1M 6%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.
2025-07-10 / published 2025-07-11	1M 4.5464%	Tier B Specialist-channel sample	Parallel-channel observation Observation date and original tenor are preserved; excluded from the canonical trail and score.

Recovered archive term curves

Date	All tenor readings	Evidence type	Caveat
2015-02-13	1M -0.41%3M -0.3%6M -0.07%1Y 0%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2013-07-29	1M -0.1936%2M -0.2725%3M -0.234%6M -0.1345%1Y 0.1332%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2012-06-26	1M -0.26642%2M -0.14125%3M -0.01507%6M 0.25773%1Y 0.65517%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2012-04-20	1M -0.2762%2M -0.1684%3M -0.0483%6M 0.2224%1Y 0.5657%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2011-12-19	1M -0.1976%2M -0.0662%3M 0.087%6M 0.3012%1Y 0.6027%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2011-06-15	1M 0.1428%2M 0.1945%3M 0.22%6M 0.38075%1Y 0.79375%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2011-02-21	1M 0.8745%2M 0.899%3M 0.925%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2010-12-09	1M -0.30031%2M -0.27312%3M -0.27281%6M -0.15531%1Y 0.1075%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2010-07-13	1M -0.37506%2M -0.29469%3M -0.19406%6M -0.03687%1Y 0.35%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2010-04-23	1M -0.24229%2M -0.21667%3M -0.2027%6M -0.05542%1Y 0.41208%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2010-03-22	1M -0.3078%2M -0.2947%3M -0.2881%6M -0.1334%1Y 0.2747%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2006-07-21	1M 0.4%2M 0.6%3M 0.7%6M 1%1Y 2%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2006-04-07	1M 1.5%2M 2.5%3M 2.6%6M 3.9%1Y 4.8%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2005-11-29	1M 0.6%2M 0.8%3M 0.9%6M 1.5%1Y 2.7%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.
2005-03-22	1M 0.16%2M 0.28%3M 0.35%6M 0.56%1Y 1.11%	Tier C Specialist-channel sample	Recovered in original tenor order; excluded from the continuous canonical series.

Excluded example: SIFO-only · The page provides SIFO only, not a LIBOR-SIFO lease rate.

2002–2026 lease-rate historical anchors

Long-run context on the left · expanded 2025–26 window on the right · blue=3M+, orange=1M · no interpolation

Mobile opens on the recent window · swipe horizontally for full history

Verifiable historical milestones

Date	Tenor	Reading	Market context	Verify
Dec 2002	1M	0.35%	An industry archive records one-month silver at 0.35%; it also mentions 0.7% for one year, but only one numeric row is added here.	Tier A
Spring 2006	3M	about 5%	Precautionary borrowing around the first silver ETF launch lifted the 3M rate to about 5%	Tier B
Sep 2008	curve	all tenors stayed below 2.5%	A jump in Indian bullion demand lifted the curve, but London liquidity absorbed the shock	Tier B
Jan–Feb 2011	curve	sharp rise across tenors	Lease-rate spikes helped push the market into backwardation	Tier B
2018 Q4	3M	-0.78%	The 2016–2018 quarterly archive shows the normal regime oscillating around zero	Tier B
2019–2024	1M	-0.09% average	Six-year pre-squeeze baseline	Tier B
9 Oct 2025	1M	34.9%	Peak of the London liquidity squeeze; overnight rates also briefly exceeded 200%	Tier B
27 Oct 2025	1M	5.6%	The squeeze eased as metal returned, but borrowing cost remained in the acute-stress band	Tier B
31 Oct 2025	—	about 2.6%	Institutional commentary put the late-October rate at about 2.6%; tenor unspecified, so this is context only.	Tier B
8 Dec 2025	—	around 6.0%	Institutional commentary cited around 6%; tenor unspecified, so this is context only.	Tier B
15 Jan 2026	—	subdued rates; near-term contango	An institution described subdued lease rates and near-term curve contango; no exact tenor or numeric value was given.	Tier B
Early Feb 2026	1M	6.3%	Post-squeeze tightness persisted, with 1M back in the acute-stress band	Tier B
11 May 2026	12M	1.63%	The 12M point remained above its five-year average, showing residual term risk	Tier C
31 Jul 2026	1M	0.03%	Latest verifiable public direct quote; no immediate borrowing stress	Tier B

Measurement boundary · Actual London OTC metal loans are bilaterally negotiated and normally private. This dataset contains only verifiable public implied or compiled rates and official-industry-report anchors. An implied rate can be negative without implying that a lender would lend metal at a negative negotiated rate. Historical observations preserve the source tenor. They are event anchors, not an interpolated or continuous daily series.

§04

Historical Context & Triggers

Place today's readings inside the long-run sampleCONTEXT

Structure map · how far is this week from the historical launch zone?

Grey bubbles = 17-year sample density · coloured paths = direction and phase of past events · navy = this week · upper-right = dual-85th-percentile launch zone

Mobile centres on this week · swipe horizontally for the full landscape

This week vs past squeeze episodes

Event window	Fund net peak	Top-four short peak	Backdrop
This week	20.5%	28.3%	—
Oct 2025 · London squeeze	29%	32%	Lease rates broke 30%; silver first cleared $50
2024–25 · London tightness	42%	39%	Free float tightened as metal moved across regions
Feb 2021 · SilverSqueeze	32%	37%	Retail campaign; record weekly ETF inflow
Nov 2010 · Silver peak	34%	37%	$49 peak; five margin hikes ended the run
Jul 2020 · Delivery squeeze	25%	33%	Pandemic logistics break; COMEX delivery spike

Peaks are measured inside each event window. Since 1986, silver squeezes have not required a single dominant long; crowded funds and trapped shorts have been the recurring launch signature.

What would change the call

▸ Managed-money net length crosses 36% of OI—the historical 85th-percentile launch zone.

▸ Top-four net short concentration approaches 38%, signalling that shorts are becoming crowded.

▸ SHFE warrants turn from accumulation to a four-week draw while the Shanghai premium holds above $0.50/oz.

▸ A fresh London 1M implied lease rate crosses 3%, then enters acute stress above 5%.

These are observable public-data conditions. Exact factor weights and scoring triggers remain internal.

§05

Positioning

CFTC publishes Friday · positions as of TuesdayPOSITIONING

Participant flows · who is buying and who is selling

Participant	Net position	WoW	% OI	1y percentile
Managed money	+23,646	▲ +1,366	+20.5%	21
Commercials	-43,974	▼ -3,552	-38.2%	48
Non-reportables	+20,328	▲ +2,186	+17.7%	79

Positive means net long. The combination of crowded funds and trapped commercial shorts is the structural backdrop common to past silver events.

Managed-money net length (% OI, six years)

The most responsive positioning signal across past silver events · shading marks event windows

Top-four net short concentration (six years)

Trapped shorts provide squeeze fuel · shading marks event windows

§06

Shanghai & Cross-Market

The layer many global silver screens missSHANGHAI

SHFE silver warrants (tonnes)

The daily pulse of China's deliverable pool

Shanghai premium to COMEX ($/oz equivalent)

A widening premium means China is pulling metal from the global pool

Historical anchors · 2026-03-31 · background only, not current

Multi-source/institutional cross-check

2026-03-31

Import margin-¥39.3/kgWindow closed

COMEX silver inventory327.821 Moz+0.231 Moz d/d

SLV holdings15,274.28 t-14 t d/d

May-Jul spread-$0.560/ozContango

Jul-Sep spread was -$0.554/oz; the pair was in contango, with no nearby squeeze. SHFE silver open interest was in a downtrend and institutional precious-metals positioning was described as very light. These anchors do not replace newer current readings and do not enter scoring.

§07

Gold Check & Gold/Silver Ratio

The other side of the precious-metals complexGOLD CHECK

Gold/silver ratio · COMEX front contracts

A falling ratio means silver is outperforming

Gold structure · cross-check

Gold/silver ratio · COMEX

69.4

A falling ratio means silver is outperforming

Gold fund net · % OI

54.4%

94th percentile over five years

Gold registered

451 t

eligible 377 t

Gold funds are crowded as well, raising precious-metals complex risk.

§08

London Vaults & Free Float

LBMA monthly · roughly one-month lagLONDON FLOAT

Headline inventory vs silver that may actually be lendable

The 2024–25 London stress was not simply about total ounces; it was about the pool of ounces available to borrow

LBMA silver holdings (million ounces, two years)

A lower-bound view of the OTC shadow inventory

Weekly Read

After the fund unwind, silver remains structurally loose—with dry powder still sidelined

Structural Gauges

Five-factor state

Deliverable pool

Lease-Rate Stress

Current borrowing temperature

Why this is a hard squeeze indicator

Six latest verifiable direct quotes

Data quality and search boundary

Official daily subscription data

Direct public quote chain

Public physical corroboration

Institutional context

Term-curve cross-check

Domestic specialist channels

Historical archive recovery

China-channel public observations

Modern observations

Recovered archive term curves

Verifiable historical milestones

Historical Context & Triggers

This week vs past squeeze episodes

What would change the call

Positioning

Participant flows · who is buying and who is selling

Shanghai & Cross-Market

Historical anchors · 2026-03-31 · background only, not current

Gold Check & Gold/Silver Ratio

Gold structure · cross-check

London Vaults & Free Float

Methodology

Risk framework

Sources and release lags

Data sentinels

Disclaimer & copyright