Guotou Securities: Comparing to the "Mao Index" in 2021, the historical peak of gold is basically clear.
```
In the past few years, gold and AI technology have almost simultaneously become the main lines for global capital. On the surface, one is a hedge and the other is growth; looking deeper, they represent two completely different worldviews: buying gold bets on continued loosening of order and ongoing doubt about the dollar’s credibility; buying AI bets that technology will once again elevate the efficiency and returns of the US economy.
Zou Zhuoqing, a strategy analyst at Guotou Securities, wrote in a strategy special on July 13: “Entering 2026, investors are facing the ultimate showdown between gold and AI technology. As AI accelerates and increasingly penetrates, from micro evidence to macro efficiency conversion, it is clear our view stands on the side of AI technology, and it’s evident the scales are already tilting towards AI technology.”
This does not mean gold will collapse immediately. More accurately, gold may have already reached a position similar to the “Mao Index” in 2021: the high point may have appeared, but true evidence for the market to completely abandon it will take time. After the Spring Festival in 2021, the "Mao Index" peaked, but it wasn’t until the second half of 2022, after the pandemic restrictions eased, that the logic behind consumption was disproven and became clearer. Gold is similar now; what first appears is “evidence of loosening,” not “evidence of collapse.”
The most critical loosening comes from the dollar. Much of the recent gold rise was built on the weak dollar, central bank buying, and de-dollarization narratives; now the weak dollar is turning into “the dollar is not weak.” If AI capital expenditure and productivity improvements are superimposed to restore dollar credibility, the monetary premium gold previously enjoyed will be systematically compressed. The so-called “M top,” the core is not the chart pattern, but the reversal of trading logic: the policy top comes first, then the data top.
Gold now resembles the 2021 “Mao Index”: Faith loosens first, evidence comes later
The “Mao Index” in 2021 didn’t peak only after everyone recognized fundamentals deterioration. On the contrary, at the peak only signs of loosening could be found: some liquor prices couldn’t rise, duty-free sales data dropped, and overseas rate hike expectations heated up. What actually led institutions to significantly reduce positions was later and more realistic verification.
The similarity with gold now is that the mid-to-short-term contradictions which have supported it—central bank buying, dollar credibility disputes, geopolitical conflicts, fiscal deficits—still exist. But these factors have been fully priced in, even exaggerated. The new variable is that the dollar has not continued to weaken.
The rising oil price center is one important reason the dollar isn’t weak. High oil prices boost inflation expectations, reduce room for rate cuts, and even strengthen expectations for hikes; after the shale oil revolution, as a major energy producer, high oil prices also support the US current account. For gold, this means its financial attribute is now constrained, while its commodity attribute comes to the fore.
This is the meaning of “clearly one top already”: not that gold can’t rebound, but that the previously one-sided narrative has started loosening. The real evidence determining whether gold enters a major bear market may await future verification of AI productivity and dollar credibility over 1–2 years.
M Top isn't just a pattern, but a timing gap between “policy top” and “data top”
Historically, gold bears mostly come from two forces: Fed rate hikes or the rise of technology. The former is gold’s most direct rival, but the capital siphoned off by technology often causes bigger pullbacks.
After the Bretton Woods system collapsed, gold went through several bull markets. Except for rare special stages, most bull-to-bear switches are sharp tops, only 1975–1981 and 2010–2013 are closer to M tops. Common point for M tops: policy shifts first, but market doesn’t fully believe; gold rises again, until economic and inflation data confirm policy effectiveness, the right-side top is then established.
In 1979–1980, after Volcker’s aggressive rate hikes, gold saw a first top, but the market still feared recurring inflation. Only when real rates turned positive, inflation confirmed its downtrend, and rising unemployment brought economic cooling, was the right top confirmed for gold.
Similar in 2009–2012. QE’s excess money worries pushed gold up, with the left top after QE3 failed and Operation Twist launched in 2011; gold surged again when QE3 rolled out in 2012. Then as US unemployment fell below 8% and consumer confidence recovered, the market confirmed the dollar’s credibility hadn’t collapsed, and gold’s bear market began.
Currently, gold is more like an M top than a sharp top. Policy side is under pressure: June FOMC showed a hawkish tilt, with half of voting members expecting at least one rate hike this year and the median policy rate revised upward; forward guidance has been weakened. Data side is still verifying: if inflation stickiness and job resilience persist, the right top could be confirmed sooner.
Dollar shifts from "weak" to "not weak", gold's most comfortable environment is gone
The dollar framework can be split into four states:
Rate cuts + long-term inflation not falling: weak dollar;No rate cuts + long-term inflation not falling: dollar not weak;Rate cuts + long-term inflation falling: dollar not strong;No rate cuts + long-term inflation falling: strong dollar.
Gold’s most comfortable scenario is the first: declining interest rates, inflation expectations still present, real rates subdued, dollar pressured. Over the period, gold benefited from this environment.
Now, it's closer to the second: no rate cuts, long-term inflation not falling, dollar not weak. Oil prices, geopolitical conflicts, and US economic resilience are all constraining space for rate cuts. Meanwhile, Europe faces weak growth, energy dependence, inflation pressure, and limited fiscal room; Japan faces low growth and high debt. Dollar strength doesn’t mean the US has no issues—other major economies have more.
A more distant risk lies in the fourth case: If AI boosts productivity, long-term inflation center falls, and US rates don’t rush downward, the dollar enters a stronger phase. That’s a deadly combo for gold, losing support from both a weak dollar and the narrative of dollar credit collapse.
The impact of tech rise on gold may exceed that of rate hikes
Gold’s biggest pullback may not come from rates themselves, but from tech assets continually siphoning capital.
The 1990s are an example. In the 1980s, computers flooded US companies, yet macro productivity failed to improve—this is the Solow Paradox: computers are everywhere but not in productivity statistics. The reason isn’t tech uselessness, but that companies hadn’t restructured workflows, roles, and organization.
Around 1993, US IT hit a critical watershed. Companies weren’t just buying computers, but re-writing business processes around them. After 1995, productivity surged, Nasdaq and the dollar rose together, gold was repeatedly sold off. Between 1995 and 2000’s internet boom, gold fell roughly 25%.
This history’s lesson is direct: If AI is just a chat tool or search substitute, it won’t shake gold; but if AI enters workflows for customer service, R&D, coding, sales, legal, and research, changing corporate organization, it turns from “tool” to “productivity variable.” Once capital believes AI will elevate US economic efficiency, gold's monetary attribute will be weakened.
AI is building a new capital cycle for the dollar
The dollar’s support from AI isn’t just “US tech companies are stronger.” More importantly, a new capital cycle centered on AI investment is forming.
US mega-tech companies are investing enormous capital expenditures, buying AI infrastructure, servers, storage, and network equipment from Asia, especially Korea and Taiwan. As Asia’s computing industry chain earns export surpluses, some funds circulate back to buy dollar assets, then indirectly support the next round of US tech capital spending.
Estimates show that in 2025, Microsoft, Amazon, Google, Meta, and Oracle together will invest about $449 billion in capital spending, with growth reaching 72%; in 2026 this may further rise to $805 billion. This scale isn’t just an internal tech cycle, but macro-level capital formation.
Korea and Taiwan’s AI-related exports as a proportion of GDP are also rising rapidly. The export surplus hasn't fully turned into local currency appreciation or domestic money supply expansion, but instead stays offshore in forms like overseas equity, deposits, and dollar assets. As long as US AI capital spending doesn’t collapse systematically, this cycle will buffer dollar weakness.
That doesn’t mean the dollar will only rise, but it dampens the unilateral “de-dollarization” narrative. Gold fears not a single rate hike, but the market’s renewed confidence in the dollar’s long-term asset returns.
The key in 2026 isn’t whether AI is useful, but whether productivity gains are realized
The real threat of AI to gold lies not in how much AI stock prices have risen, but whether productivity can translate from micro experience to macro data.
Already, there is some early evidence: Enterprise AI adoption rates rising, 65% of respondents say AI positively impacts their work efficiency; efficiency improvement is especially pronounced in customer service, writing, consulting, coding and other structured tasks. Related BEA research shows that since 2021, total factor productivity in high-AI-intensity industries grew 2.01% per year, while non-AI-intensive sectors fell 0.41%.
But this isn’t the endgame. AI agent deployment is still early, and company process restructuring has only just begun. In the short term, demand for data centers, electricity, copper, servers, chips and such may push up costs, causing inflation pressure. AI’s productivity dividends don’t materialize linearly; first comes capital spending, then process restructuring, finally macro stats.
In the next 1–2 years, the real focus isn't on model launches, but on a few key indicators: non-farm business labor productivity, unit labor costs, AI-intensive industry prices and TFP, degree of enterprise AI workflow integration, data center and cloud capital spending. If all these point to rising efficiency, the US could reproduce the mid-to-late 1990s model: growth resilience stronger than expected, inflation pressure weaker than traditional models, dollar credit strengthened.
A second top may still appear, but gold’s winning conditions have narrowed
An M-top does not exclude another surge for gold. The second top could come from future Fed rate cuts, or from setbacks in AI industry development. If AI capital spending is questioned, application fails expectations, or the dollar weakens again, gold would still have reasons to rebound.
But rebound is different from being the main line. Previously, gold rallied on central bank buying, de-dollarization, and hedge demand; now it must face dollar not weak, actual rates not low, AI’s capital siphon, and US productivity revaluation—all opposing forces.
The implication for equity markets is clear: In phases without rate cuts and with rate hike expectations swinging, gold is weak, high-prosperity assets perform better. Historically during policy wait-and-see periods, Nasdaq 100 often outperformed S&P 500, and growth styles beat value styles. The structural shift in 2026 may mean assets like gold—the “old Mao index”—peak, while tech and overseas-oriented “new Ning combination” keeps rising.
The boundaries must be clear: If overseas monetary policy changes unexpectedly or AI productivity fails to enter macro data, gold’s M-top validation will be prolonged. But if AI data keeps proving US economy is being reshaped by tech and dollar credibility strengthened, gold’s core monetary attribute pricing of recent years will be re-sealed.
Risk Disclosure and Exemption ClauseMarkets are risky, investment requires caution. This article does not constitute personal investment advice and does not consider the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their own situation. Invest accordingly at your own risk. ```