In capital markets, some companies tell stories more dramatic than anything on screen.

In the past two days, U.S.-listed mining newcomer Cango announced a financing round totaling $75 million.

The structure is notable: $65 million comes from the company’s own top executives, while the remaining $10 million comes from Hong Kong capital.

If your impression of Cango is still the auto-finance platform that helped people buy cars, you may be behind the curve.

Today, it is plunging into the deep waters of AI and energy infrastructure.

The question is whether this large sum is paying for the future, or filling the hole left by past losses.

A Frantic Pivot

Cango’s transformation path reads almost like a survivalist evolution story.

At the end of 2024, the company abruptly announced its move into Bitcoin crypto mining. The result?

In 2025, it mined more than 6,500 coins.

That sounds impressive, and revenue surged to nearly $700 million. But the financial statements told a more awkward story: net losses reached as much as $450 million.

Whether it is reducing Bitcoin holdings, facing intensifying computing power competition, or investing in infrastructure, every path burns cash. Crypto mining is nowhere near as profitable as it looks from the outside.

Cango’s new direction for this financing is called EcoHash, whose core idea is to use its existing resources to pivot.

Cango previously built many mining farms specifically for Bitcoin crypto mining. Now that crypto mining is unprofitable and extremely capital-intensive, it plans to upgrade and retrofit those mining farms, moving away from Bitcoin and toward “distributed AI inference.” In practice, that means using existing sites, surplus power and legacy equipment inside the mining farms to run AI computing tasks for other companies.

The logic fits the moment: if crypto mining costs are too high, use surplus power and facilities to run AI instead, especially when the world is short of computing power.

A Real-Money Contest

The most telling part of this financing round is where the money comes from: $65 million is being put up by the chairman and directors themselves.

With the stock down 73% in the secondary market, insiders willing to put real money on the table are usually read in capital circles as a sign that they have not given up.

But there are two ways to read it.

On one hand, executive buying gives the market a shot of confidence and signals belief in the company’s AI pivot.

On the other, this kind of “internal circulation” also suggests Cango may not be finding it easy to raise money in public markets right now.

Everyone is watching the same question: can this AI computing power business actually make money, or is it just a more polished story repackaged to take investors’ money?

Not Every Pivot Makes It Ashore

The mining sector is now going through an unprecedented identity crisis.

Pure-play miners are finding it increasingly hard to survive, and everyone is rushing toward AI. The paths taken by peers also show where the mining sector is inevitably heading.

Marathon (MARA), a leading miner, saw fourth-quarter 2025 revenue fall 6% year on year and is under pressure from a $350 million Bitcoin-backed credit line. It is now deleveraging, has even authorized the sale of all its Bitcoin reserves, and is accelerating partnerships with investment institutions to build AI data centers.

Core Scientific, another leading miner, has moved even more decisively. It has sold large amounts of Bitcoin for cash and converted mining farms into “powered shell” data centers, offering AI companies bundled computing power and power services.

Cango’s EcoHash is essentially competing for the same market.

But is the AI business really that easy to digest? Mining rigs and AI servers are in completely different leagues when it comes to environmental conditions, stability and bandwidth requirements.

Turning a mining farm into an AI computing center is not as simple as swapping out a few plugs. It requires real money to rebuild infrastructure.

For example, transformer shortages have pushed delivery times to as long as 18 months, while liquid cooling retrofits can cost more than $3 million per megawatt. These are real barriers.

Cango’s distributed AI inference strategy avoids some of the heavy-asset pressure of large data centers and focuses on flexible GPU services for small and midsize companies, but it still faces plenty of challenges.

Its $75 million has to cover mining farm retrofits, technology upgrades and multiple other investments. Whether it works depends entirely on whether the model can run and whether it can sign stable orders.

More importantly, the core advantage for miners pivoting into AI is existing power and server-room infrastructure. Cango came from auto finance and has limited accumulated experience in mining and AI infrastructure, putting it behind peers such as Core Scientific and Marathon. With no clear large-scale partnership orders so far, the market remains skeptical about its transformation prospects.

Many companies facing difficulties assume they can turn around simply by switching tracks. In reality, it does not work that way.

Whether the sector is attractive matters less than whether the company can run the business well and stabilize its finances.

With this $75 million in hand, Cango has to manage Bitcoin’s volatility while also trying to ride the AI wave. Its success depends on whether it can find the right balance.

Plenty of people can tell a good story. Far fewer can turn that story into real cash and sustainable profits.