By Peter Reagan
Your News to Know rounds up the most important stories about precious metals and the overall economy. This week, we’ll cover:
- The massive and not-so-quiet AI boom
- What the Fed’s bank-capital proposal really says about gold
- Why “cheapflation” confirms what families already know
The American economy has bet an awful lot on AI
A strange new contact recently appeared in one of the chat applications on my phone.
It was an “artificial intelligence assistant,” prominently displayed and accompanied by the sort of verified badge usually reserved for a celebrity or public official. The application was not shy about promoting it, either.
Is it just me, or is AI appearing everywhere now?
It summarizes web searches. It joins meetings. It wants to draft your emails. I see AI in word processors, phones, browsers and customer-service chats. Sometimes it’s helpful, and other times it’s like a vacuum cleaner salesman with his foot in the door who “Just wants a moment of your time” and isn’t taking no for an answer…
Why the sudden, aggressive push?
Well, the optimistic explanation is that every provider of software on the planet has simultaneously developed an extraordinary technology and want everyone to use to use it.
Now, that may be true in some cases.
There’s another possibility we must consider, though: Technology companies have spent so much money on AI that they need widespread adoption to arrive quickly. Reuters recently tried to measure the scale of this buildout and came up with $800 billion this year, on top of $800 billion last year after $260 billion in 2024. That’s nearly $2 trillion in just three years.
But the spending isn’t slowing down! In fact, Goldman Sachs estimates a total of $4-$8 trillion in capital investment over the next five years.
These numbers are mind-boggling. AI is no longer merely a new product category. It is an economic engine.
Think about it: The construction of data centers supports electricians, equipment manufacturers, utilities, construction companies and local governments. It creates economic activity before the AI systems inside those buildings have produced a single dollar of additional productivity.
That spending is helping support economic growth right now. (As much as 74% of total first-quarter GDP came from AI datacenter construction and related activity.)
Here’s my concern… These so-called “hyperscalers” are spending money today. The actual payoff, the profits required to turn AI into a thriving industry, remains somewhere in the future.
The AI buildout is increasingly funded with borrowed money.
All this leads to my major concern: As a nation, we’ve put down a very large bet before all the cards are on the table.
Now, that doesn’t mean AI is useless. Our nation has a rich history of speculative technological manias that, eventually, paid off. Railroads transformed America. So did automobiles, electricity and the Internet. Every one of those technologies attracted enormous spending because businesses correctly believed they would change the world.
Every one also produced periods of overbuilding, wasted money and broken promises.
See, a genuinely useful invention can still become a speculative asset participating in an economic bubble.
That is the concern raised by David Woo, who spent two decades at Bank of America and eventually served as its head of global rates and currency research.
Woo recently told Kitco that the dollar’s strength has become increasingly dependent on global demand for American AI:
“If this AI bubble were to burst, the dollar’s toast.”
For clarity, Woo did not call AI “America’s biggest defeat since the Vietnam War.” (That remark referred to his interpretation of the U.S.-Iran agreement, and frankly I think it’s an obnoxious overstatement.)
His AI warning was different – and, frankly, more interesting.
Woo argues that businesses around the world need dollars to gain access to American-designed computer chips, data-center capacity and leading AI systems. That international demand helps support the dollar as global reserve currency.
If spending on AI slows sharply, he believes a significant portion of that support would disappear.
It’s a fair point. But I don’t know whether Woo is right. After all, the dollar is influenced by far more than one industry. Interest rates, trade, economic growth, political stability and its role in international commerce all matter.
But the role of AI in the U.S. economy concerns me. Our current economic story leans heavily on the assumption that AI will soon generate enough productivity and revenue to justify not only $2 trillion already spent, but hundreds of billions in annual revenue.
What if it does not? What if the disappointing Bain & Co. report on AI return on investment (reported by Bloomberg) turn out to be normal?
AI does not have to disappear. It does not even have to fail. It merely has to disappoint the businesses who are spending hundreds of millions to use the latest generation of AI tools.
Maybe businesses discover that they need fewer AI tools than they paid for. Maybe the technology improves productivity, but much less than promised. Maybe less expensive competitors make today’s enormous data centers too expensive…
Any one of those outcomes would turn a this spending boom into an expensive hangover. And the effects would not remain confined to Silicon Valley.
The data-center boom now reaches construction, manufacturing, commercial real estate and electricity production. It also competes with households for power and equipment, potentially raising utility and consumer costs.
That is what happens when an entire economy starts leaning too heavily on one support beam.
The beam may be strong.
But we should not build our entire financial futures on it! That’s the opposite of diversification.
By the way, this is not a prediction that the AI sector will collapse. It is simply an acknowledgment that none of us should depend completely on one technology, one sector or one economic story unfolding exactly as we hope.
What the Fed’s proposal actually says about gold
Gold fell below $4,000 an ounce on June 24, reaching its lowest price in more than seven months.
Whenever gold declines sharply, explanations multiply like dandelions after a spring rain.
Some people blamed Gulf oil producers. Others blamed private banks. Still others claimed the Federal Reserve had quietly instructed banks to sell gold by dismantling Basel III regulations.
That last claim does not survive a close reading of the actual proposal.
Reuters attributed the June 24 decline primarily to a stronger dollar and growing expectations that the Federal Reserve could raise interest rates.
Both pressures are familiar.
A stronger dollar makes dollar-priced gold more expensive for buyers using other currencies. Higher interest rates can also reduce short-term demand for gold because the metal itself does not pay interest.
That does not mean either pressure will last forever. It does mean we do not need to invent a secret order from the Fed to explain the price decline.
Here is what actually happened.
On March 19, the Federal Reserve, Federal Deposit Insurance Corporation and Office of the Comptroller of the Currency jointly requested comments on proposed changes to bank-capital regulations.
The proposal was published in the Federal Register.
The public-comment period ended June 18.
That deadline did not automatically convert the proposal into law. Banks did not become free to interpret it as they pleased the following morning. Regulators must still consider comments, potentially revise the proposal and adopt a final rule before new requirements take effect.
More importantly, the proposal does not direct banks to sell gold.
It continues to assign a zero-percent risk weight to qualifying gold bullion held in a bank’s own vault or held on an allocated basis in another institution’s vault, provided those gold assets are offset by corresponding gold liabilities.
In plain English, the proposal continues to recognize properly allocated physical gold as carrying very little credit risk under those specific circumstances.
That does not mean the proposal is harmless or beyond criticism.
The agencies estimate that some of the combined changes would reduce capital requirements for several categories of banks. Regulators argue the changes would better match capital requirements to actual risks and reduce unnecessary regulatory burdens.
Banks have an obvious incentive to support that argument.
Money held as a capital cushion cannot be used as aggressively for lending and other potentially profitable activities. Banks naturally prefer greater flexibility. Regulators naturally face pressure to make credit more available and less expensive.
The benefit is easier lending.
The risk is that thinner cushions can leave institutions more vulnerable when loans fail or economic conditions change suddenly.
That is an important debate. It is also a more credible criticism than claiming the Fed ordered banks to dump gold.
Bank-capital regulations are a lot like building codes.
During calm weather, stronger requirements can look costly and excessive. Why use thicker beams or deeper foundations when the building is standing perfectly well?
Then the earthquake comes.
Suddenly, yesterday’s unnecessary expense becomes today’s only line of defense.
We should scrutinize any attempt to weaken the financial system’s shock absorbers. But we should criticize what regulators actually proposed – not what we imagine they secretly intended.
Woo offered another possible explanation for the spring gold decline. He suggested Gulf oil producers may have sold gold because disruptions around the Strait of Hormuz reduced their access to cash.
That theory is plausible, but it remains Woo’s interpretation. There is not enough public evidence to identify those producers as the principal sellers.
The honest conclusion is less satisfying than a conspiracy:
We do not know precisely who sold every ounce.
We do know the dollar strengthened. We know expectations for interest rates changed. We know gold prices can move violently over short periods even when longer-term demand remains intact.
A correction does not mean physical gold has stopped serving its traditional purpose.
Gold is not a promise that its price will rise every week. It is a tangible form of savings that does not depend on a bank remaining solvent, an AI company meeting its projections or the Federal Reserve making the right decision at the right time.
That distinction matters more than any one week’s price.
“Cheapflation” is not an excuse – it is an admission
“Cheapflation” sounds like the sort of word economists invent when the old vocabulary is no longer confusing enough.
But the idea behind it is both simple and useful.
The Bureau of Labor Statistics reported that consumer prices rose 4.2% during the 12 months ending in May.
That was the highest annual reading in roughly three years.
An average inflation rate, however, does not mean every price rose 4.2%. Nor does it mean every household experienced the same increase in its cost of living.
A retiree buying prescription medicine has a different personal inflation rate than a young couple paying for child care. A commuter who drives 60 miles each day experiences energy inflation differently from someone who works at home.
New research by Northwestern University economist Kunal Sangani adds another layer.
Sangani found that during periods of high inflation, lower-priced versions of products can rise faster in percentage terms than premium alternatives.
Consider a simple example.
Suppose higher input costs force two coffee companies to add 20 cents to the price of a package.
Adding 20 cents to a $5 package is a 4% increase.
Adding the same 20 cents to a $10 package is only a 2% increase.
The dollar increase is identical. The inflation rate is not.
Sangani’s data showed that between 2020 and 2023, prices for less expensive coffee brands rose 36.4%, compared with 9.7% for premium brands. Across groceries more broadly, prices for the cheapest products rose 51.5%, compared with 16.3% for higher-priced alternatives.
Those findings do not explain away what families experienced.
They confirm it.
And there is an important point that deserves emphasis: Lower-income customers are not the only people who buy less expensive products.
Plenty of middle-class and affluent families buy store brands, compare unit prices and choose the less expensive cereal because they see no reason to pay extra for a cartoon mascot on the box.
That is ordinary prudence.
The distinction is not that wealthier households always buy premium products.
It is that they usually have more room to adjust.
A household with significant disposable income can absorb a higher grocery bill, switch products or cut back somewhere else. A family already buying the least expensive version has nowhere left to trade down.
That is what makes cheapflation so punishing.
Official inflation statistics combine thousands of prices into broad categories. That approach is necessary. No single number could perfectly reproduce the experience of every American household.
But averages can conceal as much as they reveal.
Imagine ten people sitting in a diner. Nine are struggling to pay for breakfast. Then a billionaire walks through the door.
On average, everyone in the room is now extremely wealthy.
The calculation is correct.
The conclusion is absurd.
Inflation measurements face a milder version of the same problem. An average can be statistically valid while failing to capture how rising prices are distributed.
There is no need to claim government statisticians are deliberately falsifying the data. The more grounded criticism is that national averages cannot fully describe the pressure families experience at the grocery store, pharmacy or gas station.
Cheapflation helps explain why Americans can hear that inflation is 4.2% while watching individual necessities rise much faster.
It also explains why inflation damages more than purchasing power.
It damages trust.
When official reports seem disconnected from everyday life, people naturally begin questioning whether the institutions producing those reports understand their circumstances at all.
Physical gold does not rise in perfect step with grocery prices. It should not be presented as a guaranteed short-term hedge or a cure for every increase in the cost of living.
Its relevance is broader.
Inflation steadily reduces what each dollar of savings can buy. Physical precious metals offer a way to diversify a portion of those savings outside the currency system responsible for that loss of purchasing power.
The common thread beneath all three stories
AI spending, bank-capital rules and grocery prices may appear unrelated.
They share one important characteristic:
Each depends on assumptions that are easy to overlook during good times.
The AI boom assumes future productivity will justify today’s enormous spending.
The banking system assumes institutions can operate with greater flexibility without becoming more fragile.
Official inflation figures assume a national average provides a useful picture of household experience.
Maybe all three assumptions hold.
AI may deliver extraordinary benefits. Banks may remain stable. Inflation may retreat quickly.
I hope so.
But hope is not the same thing as resilience.
Families cannot control how much technology companies borrow, how regulators measure bank risk or how economists calculate inflation. They can control whether all their savings remain dependent on those systems working as intended.
That is the enduring argument for diversification.
Physical precious metals do not require an AI forecast to come true. They do not require a bank to honor a promise or a government agency to measure your personal cost of living correctly.
Gold and silver simply exist – tangible savings held outside those layers of expectation.
Learn more about the potential role of physical precious metals in diversified savings, and request your free Precious Metals Information Kit today.