Sources: LBMA Gold Price (PM London auction, USD/oz) for the price, U.S. Bureau of Labor Statistics, CPI-U (All Urban Consumers, SA) for the index, live spot from the live spot market. Real figures are stated in July 2026 dollars throughout

The live number and what moved it this week belong on our gold price today page, which tracks records as they are set. This page is the one that says what a record is worth, because a nominal high reached in a year of 3.3% inflation is a different event from the same high reached in a flat year.

The nominal answer and the real answer disagree by more than 18 years, and neither is a trick. Measured on the highest monthly close of the era, gold recovered in 26 years and 7 months, which is the figure on our gold price history page and the one you will see on most nominal charts. Measured in constant dollars it took 44 years and 9 months, because the $666.75 of September 1980 bought roughly 4.0 times what the same number of dollars buys today. Both pages are reading the same LBMA series and asking different questions of it.

The two records are not even set in the same month. Gold reached its nominal high of $666.75 in September 1980, 8 months after the real price peaked, because CPI rose 7.6% over that stretch while gold managed 2.1%. The real price fell 5.1% across the exact eight months in which the nominal price was still climbing to its record. That is what 13.9% inflation does to a chart, and it is why a site plotting only the nominal series has the wrong date for the peak before anyone gets to the argument about the recovery.

Two moments in that 536-month stretch are worth knowing about, because both looked at the time like the recovery had arrived. August 2011 carried the real price to $2,669, which was 96% of the 1980 level and the highest close since February 1980. Four years of falling prices followed rather than a breakout. September 2024 then came within $5.20 of the peak before clearing it the following month, and the actual crossing happened quietly enough that nobody marketed it.

None of this says gold was a bad thing to own. It says that the most-cited number in the category took 44 years and 9 months to reach again once you measure it in the money people spend. Only 21 of the 699 months on record have ever closed above that level, which is 3% of the record and about one month above the line for every 32 below it. A page that plots the 1980 high without adjusting it is showing you a chart of the dollar as much as a chart of gold.

Stretching the holding period helps less than the sales material implies. Going from ten years to twenty cuts the failure rate from 41% to 34%, which still leaves 155 of the 459 twenty-year windows underwater in real terms. The worst of them ran from September 1980 to September 2000 and destroyed 80.2% of the purchasing power it started with, which is the span of a whole working career. A reader who is 62 and expects to draw the account down at 72 is buying the ten-year distribution rather than the twenty-year one.

The starting decade decides almost everything, and the 1980s are the extreme case. All 120 ten-year holds begun in that decade lost purchasing power, and the best of them still came out at -0.8% a year. Holds begun in the 1960s and the 2000s never lost, which are the two windows every gold advertisement chooses to plot. That spread is not a reason to avoid the metal. It does mean a single long-run average return tells a reader almost nothing about the decade they are buying into.

Deflating by CPI is one of two corrections that change the answer rather than decorate it, and the other one runs the same way against the asset gold is usually compared with. Our gold vs S&P 500 page finds that counting reinvested dividends flips 57 of 577 rolling ten-year windows from a gold win to a stocks win, which is a correction of the same size and the same character as the one on this page. Neither adjustment is a rhetorical trick. Both are the difference between the number a chart shows you and the number you would have banked.

The hedge claim comes in two forms, and a dealer will usually make both in the same breath. One is that gold holds its value when consumer prices rise, which is what the rest of this section tests against the record. The other is that gold rises when central banks expand the money supply. Our page on how gold has tracked the money supply takes the second one apart, and on the same series the first is the better measured of the two. Gold's trailing 12-month returns correlate 0.35 with CPI and 0.09 with M2 across 687 windows, and that M2 figure falls to 0.07 once CPI is held constant.

The honest complication is that the mean points the other way. Average real returns in high-inflation months were 8.6% against 6.0% elsewhere. That gap exists because 1979 and 1980 produced gains large enough to lift an average on their own, while leaving the typical month untouched. A median of -1.0% next to a mean of 8.6% is the signature of a distribution carried by a handful of observations. Both figures are on this page for that reason, and what they show together is that the inflation-hedge case rests on one episode in the late 1970s rather than on a pattern that repeats.

Gold lost purchasing power in three of those four stretches, and the exception is the one everybody cites. Over the 113 months from April 1973 to August 1982, CPI rose 123.6% while gold rose 353.5%, which is a real gain of 102.9% and easily the best inflation defence anywhere in the record. The most recent test went the other way. From June 2021 to February 2023 CPI rose 11.4% while gold rose only 3.5%, which left a real loss of 7.1% over exactly the scenario the product is advertised for.

1. April 1968

   The record starts at $39 an ounce

   The London Gold Pool had collapsed the month before, and a two-tier market let gold trade freely for the first time since 1934. That is why our series starts at $39.10 rather than at any of the fixed prices before it. Restated in July 2026 dollars that close is worth $378, so the whole 58-year record begins about a twelfth of the way to where the price sits now.

   On the day $39.10 In Jul 2026 dollars $378 Against the 1980 peak -86.4%
2. January 1980

   The January 1980 peak, and the level everything since is measured against

   Gold closed the month at $653 after a decade of double-digit inflation, a second oil shock and the Soviet invasion of Afghanistan. That monthly close is worth $2,786 in July 2026 dollars. It is the number behind every advertisement that plots a gold chart starting in 1970, because the nominal version of this month looks like an entry point rather than a top.

   On the day $653.00 In Jul 2026 dollars $2,786 Against the 1980 peak +0.0%
3. June 1990

   The worst 10-year hold in the record ends here

   Anyone who bought in June 1980 and sold ten years later had lost 65.8% of their purchasing power, which works out at 10.2% a year going backwards. Gold was still $352 in nominal terms and looked flat on a price chart. CPI had risen 57.5% underneath it and taken most of the value with it.

   On the day $352.20 In Jul 2026 dollars $902 Against the 1980 peak -67.6%
4. March 2001

   The real low, 82% below the 1980 peak

   Twenty-one years after the top, gold closed at $257.70. That is $487 in July 2026 dollars, and the lowest real month since May 1972. Central banks were selling under the first Washington Agreement, while the dot-com boom had made a metal that pays no income look like a museum piece.

   On the day $257.70 In Jul 2026 dollars $487 Against the 1980 peak -82.5%
5. August 2011

   Gold gets within 4% of the 1980 peak and turns back

   The post-crisis rally carried the real price to $2,669, which was 96% of the January 1980 level and the highest close since February 1980. It then fell for more than four years rather than breaking out. A reader who bought at that top on the argument that gold was still under its old high turned out to be right about the argument, though the trade took another decade to work.

   On the day $1,813.50 In Jul 2026 dollars $2,669 Against the 1980 peak -4.2%
6. December 2015

   The bottom of the post-2011 slide

   Gold closed 2015 at $1,060, or $1,484 in July 2026 dollars. It had given back 44.4% of its real value in four years while CPI kept rising underneath it. Nothing about the inflation story changed over that stretch, which is the clearest evidence here that gold and CPI move independently for years at a time.

   On the day $1,060.00 In Jul 2026 dollars $1,484 Against the 1980 peak -46.7%
7. February 2023

   The 2021 inflation surge ends with gold behind CPI

   Trailing inflation first entered the top quarter of the record in June 2021. Over the 21 months that followed, CPI rose 11.4% while gold rose 3.5%, leaving a real change of -7.1%. This is the most recent test of the hedge argument on live data. A reader who bought for exactly this scenario came out with less purchasing power than they went in with.

   On the day $1,824.60 In Jul 2026 dollars $2,014 Against the 1980 peak -27.7%
8. October 2024

   The 1980 peak is regained, measured on the monthly close in constant dollars

   The real price closed at $2,883 and cleared the January 1980 level for the first time, having come within $5.20 of it the month before. Central bank buying had run above 1,000 tonnes a year since 2022, according to the World Gold Council. Most of the analysis at the time credited that demand rather than inflation.

   On the day $2,734.15 In Jul 2026 dollars $2,883 Against the 1980 peak +3.5%
9. February 2026

   A real all-time high

   February 2026 is the first month in the record where gold was worth more in real terms than at any point before it, at $5,308 in July 2026 dollars. The record now holds one month above the 1980 peak for every 32 months below it. A chart of the last two years gives you none of that context.

   On the day $5,222.30 In Jul 2026 dollars $5,308 Against the 1980 peak +90.5%

What is the inflation adjusted gold price today?

Gold closed July 2026 at $4,026.60 an ounce in July 2026 dollars. That is the most recent month where both the LBMA price and the CPI-U index exist. The nominal close for August 2026 was $4,562.75, and it carries no real figure yet because the inflation reading for that month has not been published.

Is gold a good hedge against inflation?

No, not on the horizons most buyers actually hold for. Across the 687 twelve-month windows in this record, gold's median real return in the top quarter by inflation was -1.0%, against +3.5% in every other window. It beat CPI in 49% of high-inflation years and 57% of the rest. The average points the other way because the 1979 and 1980 move was enormous, so the case for gold as a hedge rests on one episode rather than on a pattern.

What was the gold price in 1980 adjusted for inflation?

Gold closed January 1980 at $653 an ounce, and that monthly close is worth $2,786 in July 2026 dollars. The real price did not close above it again until October 2024, a wait of 44 years and 9 months, and 678 of the 699 months in this record closed below it, or 97%. The answer depends entirely on which 1980 peak you mean. The $850 intraday spike of 21 January took 28 years to regain in nominal dollars, and the highest monthly close of the era took 26 years and 7 months. We use monthly closes throughout, so every figure comes from one consistent series.

Has gold beaten inflation since 1968?

Yes. An ounce bought at the April 1968 close of $39.10 is worth 10.6 times as much in purchasing power today. That works out at 4.1% a year above CPI over 58 years. The figure depends heavily on the start date, because 1968 sits at the bottom of a fixed-price era. Every one of the 120 ten-year holds begun in the 1980s lost purchasing power instead.

Why does the inflation adjusted gold price change when the gold price does not?

The deflator moves as well as the price. A real figure is the nominal price multiplied by the ratio of CPI now to CPI in the month being restated. A hot inflation print therefore lowers every historical figure on this page without gold trading a dollar differently. The Bureau of Labor Statistics also revises its seasonally adjusted series each February for the previous five years, which moves recent real figures after the fact.

Which inflation measure does this page use?

CPI-U for all urban consumers, seasonally adjusted, on the 1982 to 1984 base, published by the Bureau of Labor Statistics as series CPIAUCSL. It is one measure among several, and reasonable people prefer others, including chained CPI and the PCE deflator, each of which would produce a slightly different real series. We use CPI-U because it is the most widely published and the easiest for a reader to check against the source.