Gold investment is rebounding sharply after this year’s serious gold drawdown. As gold blasted higher on balance over the past five weeks, investors have been quick to return. This is certainly a bullish omen, as investment demand is the primary driver of major gold uplegs and bull markets. Investors’ capital inflows are self-reinforcing, as the higher they push gold the more investors want to chase its upside momentum.
Gold investment demand is challenging if not impossible to quantify. The global gold market is massive with countless buyers and sellers, and many prefer to keep their transactions private. Gold is the most-liquid asset in the world, easy to trade everywhere in any currency. The best-available global gold fundamental data is published once a quarter by the World Gold Council in its fantastic Gold Demand Trends reports.
Those divide global gold demand into four categories, jewelry, investment, central banks, and technology. Quarterly since Q1’20, those have averaged 43.3%, 32.0%, 17.3%, and 7.4% of overall gold demand. While investment is second behind jewelry, it is way more volatile. Gold investment demand soaring and collapsing is the dominant wildcard fueling gold’s big swings, major uplegs and corrections and bulls and bears.
The supreme importance of investment demand in driving gold price trends is readily evident in decades of GDT data. Consider a recent example, starting in 2024 where gold rallied a great 27.2%. Investment demand that year surged a similar 25.3% year-over-year to 1,205 metric tons. Then in 2025, gold blasted a spectacular 64.3% higher! The sole driver was global investment demand skyrocketing 83.3% YoY to 2,209t!
Ideally global gold investment data would be available daily, but gold’s world market is too dispersed and opaque to track. Even the venerable WGC’s quarterly numbers include many estimates. But there is an excellent way to approximate gold investment capital flows daily. Those GDTs further subdivide gold investment demand into traditional bars and coins and physically-backed gold exchange-traded funds.
These ETFs do report their gold bullion held in trust for their shareholders daily. When gold-ETF holdings are rising, that reveals capital inflows into gold. Conversely when they are falling, investors are selling and pulling money out. Every quarter those GDT reports list the world’s ten-largest gold ETFs. A trio of US ones always dominates it, GLD SPDR Gold Shares, iShares Gold Trust, and SPDR Gold MiniShares.
At the end of Q2’26, those respectively proved the world’s 1st-, 2nd-, and 4th-biggest physically-backed gold ETFs. Together they held a hefty 1,680t of gold bullion, 41.5% of the total across all the world’s gold ETFs! In some quarters over the past decade or so, changes in GLD+IAU+ holdings alone actually accounted for the entire swings in global gold investment demand and occasionally overall world gold demand.
Back to last year, total investment demand again soared 83% YoY driving gold 64% higher. Yet the bar-and-coin subcomponent merely grew 16.3% to 1,406t, while gold-ETF demand shot stratospheric from -3t in 2024 to +803t in 2025! The 372t build in GLD+IAU+GLDM holdings last year accounted for 46.3% of that world total. And that’s on the light side historically, with gold ETFs growing more popular around the world.
Two other factors make capital flows into and out of GLD+IAU+GLDM even more important for gold’s price trends. First, American stock investors control the largest pools of capital in the world by far. When they get interested in gold and increase their portfolio allocations, the capital inflows are vast. Second, again investment demand is the dominant driver of major gold uplegs and bulls and is usually led by gold ETFs.
All this makes following GLD+IAU+GLDM holdings an excellent proxy for overall global gold investment demand. Again published daily, they provide the highest-resolution view into gold investment capital flows ever witnessed.
This chart over the past few years or so superimposes GLD+IAU+GLDM holdings in metric tons in blue over gold in red along with its key technicals. Not only is gold investment rebounding dramatically which portends bigger gains to come, but the investment selling during gold’s recent serious drawdown proved surprisingly muted. American stock investors did sell gold ETFs forcing holdings draws, but sure didn’t panic.
From early October 2023 to late January 2026, gold skyrocketed 196.4% achieving its biggest cyclical bull ever by far in US-dollar terms! This chart carves that into uplegs and pullbacks, showing gold’s gains and losses as well as GLD+IAU+GLDM holdings’ parallel percent and tonnage changes. Gold’s biggest uplegs in that late monster record bull saw the biggest capital inflows into these dominant American gold ETFs.
That entire bull ran 27.8 months, remarkably without a single 10%+ correction! Over 4.8 months into late October 2024, gold surged 21.9% on a 4.9% or 65.0t build in GLD+IAU+GLDM. In 5.2 months into mid-April 2025, gold shot up another 33.5% on a 12.0% or 164.1t build. Then another 5.2-month span into mid-October 2025 saw gold soar 36.8% partially fueled by a massive 13.8% or 208.9t GLD+IAU+GLDM build!
That strong correlation between American gold-ETF capital inflows and gold price gains did weaken some into gold’s terminal surge. In just 2.8 months into late January 2026, gold rocketed up 37.1% into a near-parabolic popular speculative mania! The GLD+IAU+GLDM build during that was merely 4.7% or 79.4t, despite frenzied gold buying in extreme greed. Yet most of gold’s biggest bull gains were fueled by gold-ETF inflows.
Overall during that entire +196% tripling in gold over two-plus years, GLD+IAU+GLDM holdings enjoyed a large 29.3% or 402.3t build. Despite starting chasing gold later after that bull was already well-established, American stock investors eventually provided one of its primary fuel sources. But gold’s terminal gains accelerated so fast it stretched to epic overboughtness in late January, necessitating a serious reckoning.
At gold’s peak, it had rocketed a mind-boggling 43.4% above its baseline 200-day moving average! That was the most overbought it had been since March 1980, a whopping 45.9 years earlier! So less than a week after that late monster record bull climaxed, in early February I wrote an essay analyzing gold’s drawdowns after its next-twenty-four-largest cyclical bulls in US-dollar terms arguing a serious selloff looming.
I pointed out then as gold herd euphoria remained fevered, “…consider the reckonings after gold’s next-ten-biggest cyclical bulls after today’s. Those still averaged hefty 20.8% losses over just 2.1 months!” Yet because gold’s climax in late January 2026 was so crazy-extreme, I warned “Considering all this, I’d be shocked if gold doesn’t at least retreat 25% before this drawdown fully runs its course.” That proved prescient.
Indeed over the next 5.5 months into mid-July, gold plunged 26.3% in a massive drawdown snowballing into a new bear market! While that selling grew anomalous and excessive in June on irrational Fed-rate-hike fears, it served an important purpose. It rebalanced away the extreme sentiment and technicals in late January, hammering gold to its most-oversold levels in fully 9.6 years relative to its 200dma at just 88.8%!
Yet despite gold’s serious drawdown eradicating greed and rebuilding universal apathy threatening fear, American stock investors only sold enough GLD+IAU+GLDM shares to force a relatively-mild 5.7% or 101.1t holdings draw! Impressively that was just 25.1% of the late monster record gold bull’s entire build. For comparison, that 26.3% gold bear mauled away 39.8% of the preceding bull’s absolute price gains!
American stock investors’ resilience in not aggressively fleeing gold was remarkable. Especially since June’s 11.6% plummeting proved one of gold’s worst months ever on those Fed-rate-hike fears! That implied investors weren’t scared, wanting to keep most of their increased gold allocations added over the preceding couple years. And honestly given their meager gold allocations, there was sure no reason to panic.
With gold investment so murky, proxies must be used to estimate gold portfolio allocations. My favorite for American stock investors simply looks at the value of gold bullion held by GLD+IAU+GLDM compared to the total market capitalization of all the stocks. The day gold peaked in late January at an all-time-record $5,394, those dominant gold ETFs held $307.9b worth of bullion in trust for their shareholders.
While a big chunk of change, that was still merely 0.49% of the capital they had allocated in the giant S&P 500 stocks! With trivial implied portfolio allocations of one-half of one-percent, American stock investors had little reason to fear the looming gold drawdown. The day gold bottomed in mid-July, the value of the bullion held by those gold ETFs had collapsed 30.5% to $213.9b or the equivalent of 0.31% of the S&P 500!
Incidentally before we move on, I believe this is the single-most-bullish fundamental argument for gold. American stock investors’ portfolio allocations are around one-third of one-percent, effectively zero! So they have vast room to buy and chase gold higher, amplifying its gains. If these portfolio allocations just triple to a still-trivial 1%, the resulting enormous capital inflows would catapult gold way higher in coming years.
Sooner or later this crazy AI stock bubble will burst, ushering in a brutal bear market likely to at least cut the S&P 500 in half over a couple years. Weakening stock markets greatly boost gold investment, as it is the ultimate portfolio diversifier tending to rally during stock bears. Because of this, for long centuries the minimum-acceptable gold allocation universally recommended for investors was 10% and many advisors went 20%+.
Back to GLD+IAU+GLDM holdings, they peaked a month after gold in late February at 1,802.7t. Since American stock investors don’t closely follow gold, their gold-ETF capital flows often lag major reversals in gold. Investors usually don’t recognize tops and bottoms until subsequent trend changes are already well-underway. These gold ETFs’ bullion was worth $304.7b that day, challenging late January’s $307.9b record.
But interestingly in tonnage terms those GLD+IAU+GLDM holdings still hadn’t regained their previous best level. That was 1,861.9t way back in mid-October 2020. Then American stock investors’ bullion held in these ETFs was worth $113.8b, or 0.37% of the S&P 500’s collective market cap back then. Despite gold’s near-parabola earlier this year, American stock investors’ gold holdings sure weren’t extreme.
When gold bottomed in mid-July, surprisingly GLD+IAU+GLDM holdings followed just one day later at 1,671.3t. That implied American stock investors were somewhat paying attention to gold, nowhere near as apathetic as you’d expect after such a serious reckoning. And as gold blasted higher twice in August first in a falling-wedge technical breakout then later on Treasury bond buying, investors were quick to chase.
As I analyzed in depth in an essay on gold’s huge August a couple weeks ago, last month’s huge 10.0% gains were simply the mean reversion back out of June’s anomalous 11.6% plummeting. Gold exited August trading exactly where it had been on average at that point in all modern gold-bull summers! As of late August, gold had surged 17.4% at best in just 1.3 months. American stock investors played a big role.
During that span, GLD+IAU+GLDM holdings surged 3.2% or 53.3t! Extending gold’s latest young upleg into midweek, it had retreated sharply to +10.6%. Yet American stock investors continued buying some, growing these major US gold ETFs’ new build to 3.8% or 63.5t. That’s comparable with one of the earlier uplegs in gold’s late monster record bull, the metal up 21.9% into late October 2024 on a 4.9% or 65.0t build.
Gold investment is really rebounding, as evidenced by American stock investors’ sizable capital inflows into GLD+IAU+GLDM. As long as gold continues rallying on balance and doesn’t collapse, those should continue growing. Investor buying fuels a powerful virtuous circle in gold. The more they buy, the faster gold rallies. The faster gold rallies, the more they want to buy and the more those gains attract more investors.
Generally investors aren’t contrarians, they lack the motivation or discipline to buy low when assets are deeply out of favor then later sell high at big mean-reversion gains. Instead investors love chasing upside momentum, piling into big existing rallies that are growing. The bigger and longer-lived any gains, the more they attract investors further amplifying them. This crucial dynamic has been reestablishing itself in gold.
If gold’s major breakout surge over these past five weeks had been fueled entirely by gold-futures buying, it would look more fragile. The crazy 20x to 25x leverage those guys run forces them to maintain ultra-short-term trading time horizons, days or weeks on the outside. They have to reverse course on a dime when gold moves against them, or risk swift ruin. But investment demand is way more stable and durable.
Investors buy gold outright at 1x, so a 5% pullback is nothing more. They don’t care too much and can afford to weather any normal retreats. But that same 5% leveraged 20x by gold-futures speculators would obliterate 100% of their capital risked! So gold rallies are much more sustainable when investors are helping to drive them. And American stock investors’ GLD+IAU+GLDM holdings prove that’s the case recently.
The bottom line is gold investment is rebounding, per its leading high-resolution daily proxy. That’s the bullion holdings of the world-dominant US gold ETFs, reflecting American stock investors’ gold capital flows. After suffering a relatively-modest draw during gold’s serious reckoning this year, those holdings have surged sharply with gold since mid-July. That proves investors are already chasing gold’s new momentum.
Gold uplegs fueled by investment demand are much more durable than futures-driven ones, as investors own gold outright with far-longer time horizons. Investors’ impressive participation in gold’s young upleg portends bigger gains to come. Sizable gold investment buying usually becomes self-feeding, with gains attracting in investors amplifying that upside. Recent strong investment demand should herald a new bull.
