If Kevin O'Leary had to pick one investment theme for the next three years, he said he would choose energy.
That was the clearest takeaway from his appearance on The Deep End, a podcast under Stock Sharks, where the Canadian investor laid out how he is rethinking AI, crypto, quantum computing, and risk management. His argument was not that AI has lost momentum. It was that the next layer of value may sit beneath the model race, in the infrastructure that keeps the industry running.
For O'Leary, chips and model capability still matter. So do the large technology companies that led the last wave of AI investing. But as data centers scale up, power supply, land, grid access, and long-term electricity contracts start to matter more. In that setting, he said, investors may need to spend less time trying to identify the single winning model company and more time asking which resources the whole industry cannot operate without.
Why O'Leary is shifting from AI models to energy infrastructure
O'Leary said he is not bearish on AI. He acknowledged that the last cycle produced strong returns for investors who owned major technology stocks or related indexes. The difference now is where he sees the next opportunity.
He described his approach as a classic picks-and-shovels trade. In a gold rush, the better business may be selling tools rather than trying to guess who finds the gold. He sees a similar setup in AI. Model companies need chips. Chips need data centers. Data centers need large amounts of stable electricity.
That logic leads him to energy. O'Leary said his focus over the next 36 months is on energy companies and on areas tied to power generation and delivery, including natural gas generation, turbines, power infrastructure, and uranium. He is not buying those assets because they carry an AI label. He is looking at them because they may become required inputs for the entire sector.
He repeatedly pointed to Canada and the Nordics, especially Finland, Norway, and Alberta in Canada, as places with an energy advantage. In the interview, he said some long-term power contracts in those regions can come in below $0.06 per kilowatt-hour. For large data centers that need to run for years, electricity costs can have a direct effect on margins.
Even if the competitive order among model companies changes, businesses that control land, grid access, and long-term power contracts may still be able to serve whichever customers show up. That is the part of the AI buildout he wants exposure to.
Bitzero as an example of the thesis
O'Leary used Bitzero to show how that thesis works in practice.
According to his description, Bitzero has generated cash flow from Bitcoin mining, but that is not the main reason he finds the company interesting. What stands out to him is the long-term power access and infrastructure the company has secured in Norway and Finland.
In his view, mining is one current line of business. The longer-term asset may be the company’s ability to source energy. If cloud computing or AI companies decide to build data centers in those markets, Bitzero could use its existing contracts and infrastructure to participate. The attraction of that model is simple: an infrastructure provider does not need to know in advance which AI model will win.
He also made clear that access to power does not automatically mean durable profits. A project still has to solve for grid connection, equipment procurement, customer contracts, and capital spending. Cheap electricity only matters if the full build and operating costs still leave room for returns. Even so, he sees energy infrastructure as another way to participate in AI growth without making a concentrated bet on one model developer.
From electricity to uranium and SMRs
O'Leary’s energy view extends beyond current power supply. He said he is also looking at the energy mix over the next seven to eight years, with particular interest in small modular reactors, or SMRs.
SMRs are a nuclear technology path built around smaller reactor modules. Some designs aim to improve flexibility through modular manufacturing and deployment. O'Leary’s view is that if AI data centers keep driving long-term power demand higher, nuclear energy could become an important source of stable supply, and uranium would sit upstream in that chain.
That is why he has been paying more attention to uranium investments.
He did not present that as a certainty. Whether SMRs can reach competitive generation costs on the timeline investors expect still depends on engineering, regulatory approvals, financing, and fuel supply. But the logic is consistent with the rest of his framework: rather than trying to call the final technology winner, he wants to study the base resources that large-scale expansion depends on.
Why Canada matters in his allocation
Canada is a major part of that framework. O'Leary said the country has natural gas, uranium, potash, aluminum, and other strategic resources, while also sitting next to the US, a large energy consumer. If North America keeps adding computing capacity and power infrastructure, he believes Canada could benefit.
He said his team built roughly a 10% allocation to Canada through the large-cap ETF XIU. He added that the trade did not have unanimous support inside the team at first because trade friction between the US and Canada had increased concern about the Canadian economy.
O'Leary said he viewed that pessimism as a contrarian opening. As policy conditions change and resource projects move ahead, he believes the market could reassess Canadian assets.
In the interview, he said that position at one point outperformed the S&P 500 by 158 basis points, or 1.58 percentage points. The interview did not provide the full comparison period or pricing basis, so the figure stands only as his description of the trade’s performance.
Whether Canada can keep outperforming will depend on project execution, commodity prices, trade relations, and policy follow-through. His point was narrower: global technology expansion may raise the strategic value of energy and natural resources again.
Crypto has to move beyond trading flows
O'Leary said he is applying a similar reassessment to crypto.
In the earlier institutional crypto framework, one common assumption was that more institutional money would lift the whole market. He said that may not happen evenly. Recalling an industry conference with heavy institutional participation, O'Leary said some analysts argued that Bitcoin and Ethereum alone could provide about 97% of the price-volatility exposure to the crypto market they were studying.
He was relaying a research view, not citing BTC and ETH market share. Still, the portfolio implication is clear. If a small number of large assets already satisfy most institutional allocation needs, then the case for buying a wide range of smaller tokens becomes weaker. Capital concentration inside crypto could become more pronounced.
O'Leary said his portfolio once held 27 crypto-related positions. He later sold some of them, and some of the remaining investments suffered large losses. The main assets he kept were Bitcoin and Ethereum.
That did not lead him to conclude that every other blockchain had lost investment value. Instead, it changed the question. If the next phase of crypto is not driven by broad market liquidity, how do other networks prove that they have durable economic value?
His answer was enterprise adoption, but with an important caveat.
Why he is revisiting Avalanche
O'Leary said he has been studying Avalanche again, not because of short-term token action, but because different industries may need different blockchain infrastructure.
He argued that finance, logistics, real estate, contract management, and sports assets do not necessarily need the same type of network. Financial institutions may care more about stablecoin payments and asset transfers. Logistics companies may focus on supply chain management and contract execution. Sports clubs may want blockchain systems for collectibles and digital assets.
That is why Avalanche caught his attention. He specifically mentioned its ability to support customized blockchain deployments for enterprises and said that model could fit sports assets, collectibles, and other use cases that need separate business networks.
In his view, the market could develop in one of two directions. One path is that a small number of general-purpose networks dominate. The other is that multiple networks serve different industries and connect through technical interfaces.
He did not say which model will win. He stressed something else: enterprise adoption does not automatically create token demand. A company can use Avalanche technology without guaranteeing that AVAX captures the same economic value. Investors still need to examine how network fees are paid, how revenue is generated, and whether that value accrues to token holders.
For that reason, O'Leary did not say he was preparing to buy altcoins at scale again. He said he is considering smaller public-chain positions after studying specific applications. The shift is important. Instead of relying on market liquidity to lift the whole sector, he is looking for evidence that businesses actually need these networks and that the demand can produce recurring revenue.
Quantum computing as a long-term crypto risk variable
While revisiting the commercial case for blockchains, O'Leary also raised another issue that could affect digital assets over time: quantum computing.
He said Bitcoin risk has usually been framed around price volatility, liquidity, regulation, and macro conditions. As quantum computing develops, the durability of current cryptographic systems becomes another question worth studying.
O'Leary referred to the point at which quantum computing might break some existing cryptographic algorithms as Q-Day.
He was not saying Bitcoin is under quantum attack today. His point was that market expectations can move before the technology does. If institutional investors begin to worry that future quantum machines could attack current digital signature systems, that concern alone could reduce willingness to allocate capital, even before any real-world attack appears.
He said he personally suspects some institutions may avoid Bitcoin over the long term because of quantum risk, though he did not cite public data to support a specific share. He presented that as his own judgment.
He also drew a technical distinction. A potential threat to some public-key cryptography algorithms does not mean every part of Bitcoin’s security model fails at once. Networks could also adopt post-quantum cryptography through protocol upgrades.
For O'Leary, that risk also points to a possible investment theme. If financial institutions need to upgrade security systems, companies that provide quantum computing technology or quantum-resistant security solutions could see new demand. He said he is watching IBM and Google in that context.
At the same time, he cautioned that research capability, commercial revenue, and post-quantum security products are not the same thing. Investors cannot assume that technical progress will automatically show up as profit growth.
From picking winners to managing uncertainty
Even with positive views on energy, blockchains, and quantum computing, O'Leary said investors should not concentrate too much capital in any single technology trend. Diversification remains one of his core rules.
He said no single stock should account for more than 5% of a portfolio, and no single sector should exceed 20%.
That discipline comes from experience. During the dot-com era, O'Leary shorted Yahoo. As the stock kept rising, he faced repeated margin pressure. By his account, the trade at one point caused a drawdown of roughly 40% to 50% in his net worth. Yahoo later fell, but the experience of carrying losses and meeting margin calls led him to stop shorting stocks. It also changed how he manages portfolios.
Today, he prefers to build core exposure through index funds and then add to companies he likes while keeping position sizes under control.
The Tesla example and his sell discipline
He used Tesla as one example.
O'Leary said he initially rejected Tesla because he thought the valuation was too high. His son changed his mind by arguing that Tesla was not just a car company but also a technology company collecting large amounts of data through its vehicles. O'Leary eventually built a position and then trimmed it several times as the stock rose so that it would not become too large inside the portfolio.
That approach may have left some upside on the table, but it reduced the effect of one stock on total performance. He said that for some individual stock trades, he typically uses about a 17% return as a minimum target and considers selling more than half the position once that threshold is reached.
He did not present that as a universal rule for everyone. The point was that he does not want the outcome of the entire portfolio to depend on a single company call.
What he looks for in private companies
O'Leary said he applies the same discipline to private-company investing. When he evaluates a business, he focuses heavily on customer acquisition cost, or CAC, and churn rate.
CAC shows how much a company has to spend to win a customer. Churn shows whether those customers stay. If a business has to keep raising marketing spend just to maintain revenue growth while customers continue to leave, then headline growth may not be economically attractive.
He also looks at cash flow, debt levels, and operating data across multiple quarters. That is one reason he remains constructive on enterprise AI applications. In private companies he has invested in, AI tools are already being used for financial analysis, customer acquisition, and day-to-day operations. He said one of AI’s real commercial benefits is the ability to lower operating costs and improve margins.
He also said his own business team chose Anthropic’s product as an internal tool. That, he said, showed him that model choice inside companies does not depend only on public rankings. Real procurement decisions are made on performance in actual business use.
Still, whichever vendor a company chooses, the final test is operational improvement. A technology can be advanced without producing profit for the buyer. That is why O'Leary can stay positive on AI over the long run while spending so much time on energy, infrastructure, and cash flow.
The common thread in his framework
Across energy, crypto, quantum computing, and AI applications, O'Leary keeps returning to the same question: how does growth turn into cash flow?
For energy companies, that means asking whether low-cost power can support profitable long-term contracts. For blockchains, it means checking whether enterprise adoption becomes real network revenue. For quantum computing, it means separating research progress from commercial products. For AI applications, it means measuring whether businesses can keep cutting costs and improving margins.
If those conditions do not materialize, long-term industry growth may still fail to produce the returns investors expect.
His broader conclusion was straightforward. Rather than trying to predict the winner of every technology race, investors may be better served by identifying the companies that can supply essential resources, earn commercial revenue, and convert expansion into cash flow as the industry grows.