Dead Stocks Walking: 10 Overhyped AI Names to Sell Today

Every great technology boom comes with a cast of impostors. Not frauds, exactly. They can be real companies with real equipment. Some even have real revenues. But their stocks trade at prices that have nothing to do with what the business actually earns – or is likely to earn anytime soon.

It happened with the dot-com crash and the cloud computing shakeout. And now I’m seeing it with the AI boom.

The script is always the same. A genuine technological shift creates enormous wealth for a handful of companies that have built something irreplaceable. Meanwhile, a far larger group of companies borrows the narrative, rides the excitement – then crashes when the market finally realizes the truth.

The 10 names in this report are in that second group. And they fall into four camps.

Four are Bitcoin miners who rebranded as AI infrastructure plays. Two are quantum computing startups burning cash on technology that is years from commercial viability. Two are crypto platforms whose costs have outrun their revenues, and one of them has turned its balance sheet into a leveraged Bitcoin bet. And there are two outliers – including one company that has almost nothing to do with AI at all.

That last one tells you something about how far the narrative has drifted.

In this report I’ll walk through what each company does, what it actually earns, what the numbers say, and why I think it’s best to steer clear.

The Bitcoin Miners Wearing AI Clothing

Let’s start with the four companies that have done the most creative rebranding in recent memory.

They all started as Bitcoin mining operations. They built large data centers, filled them with power-hungry hardware, and pointed that hardware at earning Bitcoin. The business is simple and cyclical, driven by two variables: the price of Bitcoin and the cost of electricity.

Then the AI boom arrived.

All four pivoted their marketing. And while their underlying businesses still focus on Bitcoin mining – pulling in only modest AI revenue – the stock market rewarded them anyway.

But the party won’t last forever.

IREN Limited – A $18 Billion Promise Built on $7 Million in AI Revenue

IREN Limited (Nasdaq: IREN), formerly Iris Energy Limited, has one of the better stories in this group. The company has secured a $9.7 billion AI cloud contract with Microsoft and is targeting $3.4 billion in AI annualized run-rate revenue by the end of 2026. Those are real numbers. I don’t doubt the ambition.

But here is what IREN actually earned in AI Cloud Services in its most recent full quarter: $7.3 million. Total revenue for that quarter was $240 million. Bitcoin mining accounted for $232.9 million of it.

The market cap sits at roughly $18 billion.

Run the math. IREN’s AI revenue, annualized from its latest quarter, runs about $29 million. The stock is priced at more than 600 times that figure. Its business must execute flawlessly on a multi-year timeline for that valuation to make sense. A Needham analyst recently slapped a “Hold” rating on the stock specifically because of “delayed AI revenue.” The stock jumped 10% anyway.

When a Hold rating sends a stock up double digits, the market is not thinking about fundamentals.

TeraWulf Inc. – Nuclear Power, Post-Halving Squeeze

TeraWulf Inc. (Nasdaq: WULF) has built its identity around nuclear-powered Bitcoin mining. The Nautilus Cryptomine facility draws power from a Pennsylvania nuclear plant. That is a real differentiator. Sustainable energy and low power costs matter in this business.

The April 2024 Bitcoin halving, however, did not care about your power source.

During a Bitcoin halving, the reward miners receive per block drops 50% overnight, squeezing margins across the entire industry.

The impact on TeraWulf’s results was immediate. In Q1 2024 – before the halving – Adjusted EBITDA was $32 million. In Q3 2024, it fell to $6 million, a drop of more than 80%, as the company’s own filings acknowledged “an approximate doubling in network difficulty and the bitcoin reward halving.” The power cost per Bitcoin nearly tripled, from $9,322 in Q3 2023 to $30,448 in Q3 2024.

By Q1 2025, the squeeze had deepened. Revenue fell 19% year over year to $34.4 million. Power cost per Bitcoin hit $66,084 – more than four times what it cost the year before. Adjusted EBITDA swung to negative $4.7 million. The company carries $500 million in convertible debt against $90 million in cash.

Now, TeraWulf has real high-performance computing (HPC) hosting ambitions. The Core42 deal is genuine. But the company is not yet earning HPC revenue at any material scale. It is a miner in a post-halving world, carrying a significant debt load, valued as though the transition is already complete.

Cipher Digital Inc. – The Name Change That Changed Nothing

Cipher Digital Inc. (Nasdaq: CIFR) was Cipher Mining Inc. until it decided a rebrand was in order.

The underlying business is still Bitcoin mining. The Q1 2025 numbers make that clear – and note that by this point the post-halving squeeze had fully worked its way through Cipher’s results, just with a quarter’s lag compared to miners that felt it immediately. Cipher reported $49 million in revenue for Q1 2025 but posted a GAAP net loss of $39 million. That was a stark reversal from the $18 million profit it earned in Q4 2024. Adjusted earnings fell from $51 million to $6 million in a single quarter. The stock, which had traded as high as $7.99 in the prior twelve months, sat below $3 at the time of the report.

The company has since signed a 15-year data center campus lease with Amazon Web Services for AI workloads. That is real progress. But for most of 2024 and into 2025, Cipher’s profitability swung entirely with Bitcoin’s price and the post-halving math. The AWS deal has yet to produce the kind of recurring revenue that would justify the premium the stock has commanded.

Plans are not profits.

Core Scientific, Inc. – Bankruptcy Graduate, Still Paying Tuition

Core Scientific, Inc. (Nasdaq: CORZ) has one of the more remarkable stories in this group, and not entirely in a bad way. The company filed for Chapter 11 bankruptcy in December 2022. It emerged in early 2024, restructured and relisted.

What happened next was predictable, if you’ve seen this before. The relisted shares surged as AI enthusiasm swept through any company with data center infrastructure. Core Scientific signed an HPC hosting agreement with CoreWeave. Investors immediately began pricing in a full transformation from miner to AI infrastructure provider.

The actual numbers tell a more complicated story. In Q3 2025, total revenue was $81.1 million – down from $95.4 million a year earlier. Bitcoin self-mining revenue fell from $68.1 million to $57.4 million. High-density colocation – the renamed HPC hosting business, the AI pivot – produced $15 million. That’s real growth. But the company is spending $121 million in capex per quarter to build toward the CoreWeave contract. And its GAAP net loss in Q3 was $146.7 million.

The stock has priced in a destination the company has not reached.

The Quantum Computing Bets

Quantum computing will matter someday. The underlying physics are real. The potential applications – drug discovery, cryptography, logistics, materials science – are genuinely enormous. The research is serious. I am not dismissing any of it.

But the two quantum computing stocks below are priced as though the revolution is already in full swing.

Rigetti Computing, Inc. – $1.9 Million in Revenue, $20 Million in Expenses

Rigetti Computing, Inc. (Nasdaq: RGTI) is pursuing gate-based quantum computing – the architecture most likely to produce general-purpose quantum advantage if the engineering challenges can be solved.

Gate-based systems work more like traditional computers. They process information step by step, applying a sequence of programmed operations to quantum bits to work through a calculation. Because those operations can be arranged in any order, these machines are designed to run any quantum algorithm – including the code-breaking and drug-discovery applications that dominate the headlines. The technical ambition is real.

The financial reality is stark. In Q2 2025, Rigetti posted total revenues of $1.9 million against operating expenses of $20.4 million. For the full year 2025, revenue came in at $7.1 million – a 34% decline from the prior year. Total net losses for 2025 were $216 million.

Most of Rigetti’s revenue comes from government contracts and research grants. Commercial revenue from private enterprises is negligible. The company burns tens of millions of dollars per quarter while producing revenues that a mid-sized restaurant would match in a year.

Rigetti routinely moves 10% to 15% on no news. It has had drawdowns of 50% or more, repeatedly. That is what happens when a stock is driven by narrative rather than earnings. At some point, the market demands something to show for it. Rigetti has not yet found that answer.

D-Wave Quantum Inc. – Two Decades, Still Not Profitable

D-Wave Quantum Inc. (NYSE: QBTS) takes a different approach: quantum annealing, a physics technique suited to optimization problems rather than general-purpose computation. Unlike gate-based machines, D-Wave’s systems don’t execute step-by-step programs. They are good at certain narrow problem types – route optimization, scheduling, resource allocation – and not much else.

D-Wave has been generating revenue for years, which puts it ahead of most quantum peers. In Q1 2025, the company posted record quarterly revenue of $15 million – a figure that looks impressive until you recall that D-Wave has been operating commercially for over two decades. Full-year 2024 bookings (future revenue under contract), which the company described as a record, came in at $23 million. Annual revenue for 2024 was $24 million. The Q4 2024 net loss alone was roughly $42 million.

The market has periodically priced D-Wave as though commercial quantum is around the corner. The stock has moved 200% or more on no news, then given it all back.

This one is pure speculation at this point.

The Crypto Platforms

The next two companies on this list have both posted impressive revenue numbers. I want to acknowledge that upfront, because the bear case here is not about sales. Both companies have benefited from higher Bitcoin prices and expanded mining capacity.

The bear case is simpler: neither controls its own destiny, and both are valued as though they do.

Riot Platforms, Inc. – Record Revenue, Then a $663 Million Loss

Riot Platforms, Inc. (Nasdaq: RIOT) is one of the largest Bitcoin miners in the United States by hashrate – the measure of raw computing power indicating how many calculations a miner’s machines can perform per second. The company posted record annual revenue of $647 million for full-year 2025. Not bad, right?

Then came the net loss: $663 million for 2025. That is a swing of more than $770 million from the $109 million in net income the company earned in 2024.

How does a company post record revenue and then lose $663 million? Several ways. Levered free cash flow ran to negative $992 million in 2025. Capital expenditures for data center buildout are massive.

And Bitcoin mining economics work in one direction at a time. When Bitcoin is up, margins look fine. When Bitcoin corrects, every miner’s income statement deteriorates at once, because nobody controls the price, the block reward, or the network difficulty.

Riot has signed data center deals with AMD, which is real progress. But the core business is still a commodity. And burning nearly $1 billion in free cash flow per year is not a setup for a premium valuation.

Marathon Digital Holdings, Inc. – Doubling Down on the Bet

Marathon Digital Holdings, Inc. (Nasdaq: MARA) is the other heavyweight in U.S. Bitcoin mining. Like Riot, it has expanded aggressively: acquiring hardware, developing sites, pursuing international energy deals.

Revenue increased 92% year over year in Q3 2025, reaching $252 million. Net income came in at $123 million for that quarter.

Those are good numbers. Here is what concerns me.

MARA has built a large Bitcoin treasury – 52,850 Bitcoin as of the end of Q3 2025, up 98% year over year. The strategy mirrors what MicroStrategy has done: hold Bitcoin directly and let it appreciate alongside the mining business.

The problem is that a Bitcoin price decline hits both the mining margins and the treasury at the same time. There is no hedge. There is no offset. The whole structure is a single concentrated bet.

When Bitcoin rises, MARA looks like a genius. When Bitcoin falls, the income statement and the balance sheet deteriorate together. A $123 million profit quarter can become a deep loss in a single bad Bitcoin cycle.

The Outliers

Strategy (MicroStrategy) – Paying a Premium to Own Something You Could Buy Cheaper

MicroStrategy Inc. (Nasdaq: MSTR) – now operating under the brand name Strategy – is something close to unique in the public markets. It began as a business intelligence software company. That business still technically exists. Under CEO Michael Saylor, however, the company has become the world’s largest corporate Bitcoin holder.

As of late June 2026, Strategy holds approximately 847,000 Bitcoin, funded through relentless equity and convertible debt issuances. The strategy is a leveraged bet on Bitcoin appreciation.

Here is the problem.

Strategy trades at a persistent premium to the net asset value (NAV) of its Bitcoin holdings. That premium has ranged from nearly 2.8 times NAV at peak enthusiasm in late 2024 to below 1.0 times NAV in November 2025, when the stock briefly traded at a discount to the Bitcoin it owns.

Bitcoin ETFs now exist. You can buy Bitcoin exposure directly, at zero premium, through iShares or Fidelity’s funds. There is no rational basis for paying a sustained premium on top of that.

The premium itself is the fuel for Saylor’s strategy: when MSTR trades above NAV, issuing equity and buying Bitcoin is accretive to per-share Bitcoin holdings. When the premium collapses – as it did in 2025 – that engine breaks down. And with approximately $6 billion in convertible debt tied to the Bitcoin position, the downside in MSTR when Bitcoin corrects tends to exceed Bitcoin’s own decline. The stock has fallen 50% or more in past Bitcoin drawdowns.

The software business is not growing. The Bitcoin position is leveraged. You can own the underlying asset for less.

Reservoir Media, Inc. – The One That Has No Business Being Here

Reservoir Media, Inc. (Nasdaq: RSVR) owns music publishing rights and recorded music catalogs. It acquires rights to songs and recordings, then collects licensing fees as those works are streamed and performed. It’s a real business with real cash flows. The company generated roughly $175 million in annual revenue and $114 million in gross profit in the most recently reported fiscal year.

So what is it doing on this list?

Reservoir has been swept into the AI conversation. And it’s pushed the stock up 25% over the past year.

Now, could Music rights companies benefit from generative AI? Sure. The theory is that AI music generators might be required to pay licensing royalties for training on copyrighted catalogs.

But that is untested, unresolved, and years away from gaining any legal clarity.

Does AI-generated music compete with catalog content? Do AI companies owe licensing fees for training on copyrighted works? Can rights holders monetize those relationships at all? Nobody knows. When that narrative moves on – and it will – Reservoir will reprice to reflect what it actually is: a slow-and-steady rights aggregator, not an AI powerhouse.

The Pattern Behind All 10

All of these companies have legitimate businesses. Several have genuine technology and capable teams. A few have started generating real revenues.

But a capable team running a real business is not the same as a stock worth owning at any price. The question I always come back to: what am I paying, and what am I getting for it?

These stocks are all priced to perfection. And perfection is rarely what you get.

Every boom has a larger group of companies that borrows the narrative, rides the excitement, and gives it back when the market asks for proof.

Phase 2 of the AI boom is real. The companies that define it are building something durable: irreplaceable software, mission-critical infrastructure, services customers cannot function without.

These 10 are not in the same league.