A gold town’s bust offers a warning for the AI boom

This month, I embarked on an adventure for the ages in Canada’s larger-than-life Yukon Territory. From climbing subarctic peaks in Kluane National Park to chasing auroras after midnight sunsets, I immersed myself in the story of the Klondike Gold Rush of the 1890s.

I travelled to Dawson City, the gateway to the rivers of gold that drew more than 100,000 prospectors between 1896 and 1899. I wandered past the hulking dredges that once chewed through riverbeds for placer gold, tried my hand at panning, and walked streets that still feel like a wild west set - all in an effort to understand what brought this feverish boom to life.

One of my favourite days was spent in the Yukon Archives, where shelves of books and records are devoted to this extraordinary moment in the global economy. There, I came across a line from historian Pierre Berton, author of The Klondike Quest, that struck me with its eerie relevance to today’s AI frenzy - and to stocks like Palantir and Nvidia.

“…the Klondike stampede began, not quietly or gradually, but instantaneously and with explosive force."

That single sentence captures the essence of the Klondike. Berton argued that what made it unique  - unlike even the great Victorian discoveries of the 1850s and ’60s - was its sheer immediacy: a sudden bang that ended almost as soon as it began.

Déjà vu in today’s AI mania

As I stood in Dawson, surrounded by relics of a frenzy that erupted overnight and collapsed just as quickly, I couldn’t help but think of today’s AI stampede.

Companies are pouring billions of dollars into AI projects, but most have yet to see any measurable returns.

A recent MIT report, "The GenAI Divide: State of AI in Business 2025", revealed that while U.S. businesses have collectively invested between $35 billion and $40 billion in AI initiatives, 95% of them are seeing zero return or no measurable impact on profits. Only 5% are seeing any genuine “value” from AI.

It’s not unlike the Klondike, where miners spent heavily on dredge machines to capture placer gold - many of which failed to pay off and were shut down quickly. Today, AI’s capital intensity carries the same risk.

Even the insiders are sounding cautious. OpenAI CEO Sam Altman recently admitted:

“When bubbles happen, smart people get overexcited about a kernel of truth. Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes," he told The Verge this month.

The comparison to Dawson is uncanny. The gold was real in the Yukon - but only a fraction of participants struck it rich. AI’s potential is real too, but right now, most are staking claims on promises, not proven profits.

Who really wins in a frenzy?

Dawson wasn’t just a mining camp - it was dubbed the “Paris of the North.” 

In just a few years, it grew from a sleepy riverbank to a bustling city. Ornate banks, hotels and dance halls sprang up alongside muddy tent encampments. Fortune-seekers arrived from as far as Australia, South Africa, and Europe, each hoping to find their share of riches.

But the goldfields were finite. By the time many arrived, the best claims were already taken. By 1900, the easy nuggets were gone and the human tide began to ebb. What was once a roaring hub of global speculation dwindled into a shadow of itself. Today, Dawson counts just over 2,000 permanent residents, a far cry from the 30,000 at the height of the rush.

Of those who struck it rich, only a handful were miners. The real winners were the service providers: ships ferrying dreamers from San Francisco, outfitters selling prospecting gear, and even Mrs. G.I. Lowe’s laundry - which doubled as a fortune-telling parlour for extra income! They didn’t gamble on gold; they built steady businesses serving those who did.

Lessons for investors

The winners of the AI boom may not be the headline-grabbing start-ups, but the picks-and-shovels providers: chips, cloud, power, and infrastructure. Even old-world industrials could benefit more from using AI than from trying to build it.

But history rhymes: Klondike prospectors once paid outrageous sums for pans and passage; today, investors are paying sky-high multiples for AI exposure (see the chart below). Capital is being deployed at unprecedented scale, yet returns remain uncertain.

History shows us that asset frenzies often feel unstoppable in the moment. Dawson City had electricity before many American cities; champagne flowing in its saloons, and more newspapers than towns five times its size. Few in 1898 would have guessed that within a decade, the party would be over.

That’s the sobering echo I took from the Yukon: hype and capital can run far ahead of reality, and when they do, the collapse tends to be brutal.

As investors, we need to ask hard questions:

• Are we chasing shiny promises or backing companies with durable business models?
• Are we buying into hype at the peak, or finding the picks and shovels that thrive no matter what?
• Most importantly, are we prepared if today’s bang ends as abruptly as Dawson’s?

Berton described the end of the Klondike stampede bluntly:

“By the summer of 1898, the easy gold was gone, and the romantic allure of striking it rich had faded for the majority of gold-seekers who suffered hardship, disease, and financial ruin.”

--

With thanks to the Government of Yukon Archives for access to research materials.