Gold's $4,000 Line: The Pause That Decides the Bull Market

Gold reached $5,595 per ounce on January 29, 2026, the culmination of a run that included a 60% surge in 2025 alone—the best annual performance since 1979. By mid-June, gold had fallen 20% from that peak.

Today, the SPDR Gold Shares ETF (GLD) sits at $396.36, down 1.73% on the day, roughly 30% below its 52-week high of $509.70. The RSI rests at 45.7. The 200-day moving average sits at $415.81.

Everything runs through $4,000 per ounce.

The record that turned into a problem

Gold's January high was real. The pullback was real, too. By late June, the metal had dipped below $4,000 for the first time since November 2025, at one point touching $3,972. It bounced, dipped again in July, and spent the summer compressed in a roughly $3,970 to $4,600 range—a 15% band after a 20% decline from peak.

Two forces drove the correction.

Federal Reserve Chair Kevin Warsh took over with a hawkish mandate. His July 29 press conference and August 28 Jackson Hole speech signaled that inflation would not be tolerated, but left the timing of rate action deliberately ambiguous. Money markets, which priced in less than a 30% chance of a September rate hike in mid-August, swung to approximately 67% by late August. The Fed currently holds its target at 3.50%–3.75%.

Simultaneously, the US-Iran conflict closed the Strait of Hormuz and disrupted energy, agriculture, and metals markets. Goldman Sachs noted the conflict disrupted reserve accumulation in energy-exporting regions—money that historically recycles into precious metals has been diverted.

The chart shows what headlines don't: gold has been coiled since late June. An elongated pause, not a trend reversal.

The floor institutions defend

Goldman Sachs, in separate September and late-August research notes, maintains $4,000 as the level where sovereign buying and institutional sponsorship create a durable bid. Their year-end forecast calls for gold to reach $4,900 per ounce.

The call is anchored to data that most short-term traders ignore.

Central banks have more than tripled their gold purchases. Before 2022, they bought roughly 400–500 metric tons annually—about 17 tons per month. In 2026, that figure has risen to approximately 50 tons per month on average, spiking to 100 tons in June on a seasonally adjusted basis. China's central bank was the largest identifiable buyer in June and added 19.9 tons in July, marking 21 consecutive months of accumulation and the largest monthly increase since October 2023.

Annual mine production sits at roughly 3,500 metric tons. Central banks now absorb about one-third of global supply, narrowing the funnel available for ETF flows, jewelry, and physical investment. Less investment capital is required to move prices higher.

Goldman Sachs describes this as a structural floor: central banks diversifying reserves away from currencies that can be frozen, buying at scale regardless of short-term price action. $4,000 is where that accumulation has historically shown up as price support.

What the GLD chart says

The SPDR Gold Shares ETF is the most liquid way to read the market's price action. The current technical picture is neither broken nor bullish—it's loaded.

GLD is at $396.36, below both its 50-day moving average ($390.65) and its 200-day moving average ($415.81). A close below the 200-day MA places the ETF in bearish territory. The RSI at 45.7 sits below the 50 midpoint—sellers still have the edge. The 14-day ATR of $7.83 means GLD normally moves roughly $7.80 a day.

Year-to-date returns are essentially flat at 0.01%, but the rolling one-year return of 17.86% shows gold is still materially higher than a year ago. The 52-week range runs from $333.81 to $509.70; at $396, GLD sits roughly in the middle of its annual range.

Capital flow in GLD today is split. Block and large-order traders—typically institutional—are net sellers, with outflows exceeding inflows by roughly $7 million at the block level and $10.5 million at the large-order level. Retail activity is approximately even, at $233.5 million in inflows versus $230.7 million in outflows. Medium orders show slight buying. This is not a one-directional market. It's contested.

The September decision

The upcoming September FOMC meeting is the event that can resolve—or complicate—this setup.

July CPI came in at 0.1% month-over-month, 3.4% annually, with core at 2.5%. That looked tame enough to lower rate-hike expectations. But Warsh's language at Jackson Hole pushed the probability of a 25 basis point hike back toward two-thirds. Bank of America analysts described the market's tepid response to his July speech as a "central bank inflation credibility shock".

Gold's relationship with rates is the engine of the trade. Higher real yields have historically pressured gold. The twist is that gold rallied in August even in a "higher for longer" environment, suggesting the traditional correlation may be weakening.

Warsh's Jackson Hole speech included a line that should not be overlooked: investors should not look "primarily to the Fed" for trade guidance. That is the language of an official who expects fiscal forces to outpace monetary ones. With US debt concerns mounting and the Treasury doubling long-term bond buyback operations to at least $4 billion per operation, gold could rise regardless of where the Fed sets its target if fiscal sustainability fears dominate.

The line that matters

$4,000 per ounce. GLD at approximately $390.

Gold broke below $4,000 on June 24, bounced, broke it again in July, and has tested it multiple times since. That repetition is what gives the level memory—and what makes a break below it dangerous or a hold above it meaningful.

Hold $4,000 through the September FOMC and the elongated pause turns into a base. The Goldman Sachs $4,900 year-end forecast becomes a working target, and the 200-day moving average at roughly $416 becomes the first ceiling to test.

Fail $4,000 and the chart opens toward $3,700–$3,800, the zone that supported the post-January decline. That would represent a further 5–7% decline from current levels, a deeper correction that would shake out weaker holders and extend the bearish structure.

The setup has until the FOMC to prove that the floor holds. Warsh's reaction function—whether he signals a hike, a hold, or continued ambiguity—will determine whether central bank demand at $4,000 is enough to absorb selling pressure.

Hold $4,000 and the long-term bull case stays alive. Lose $4,000 and the pause becomes a trend.