Matthew Sigel, head of digital-asset research at VanEck and portfolio manager of the VanEck Onchain Economy ETF (NODE), says the strongest opportunity in the crypto-adjacent equity market has shifted from bitcoin production toward the physical infrastructure needed for artificial intelligence data centers.
NODE has outperformed bitcoin by nearly 100 percentage points over the past 15 months, Sigel said, largely because of its exposure to bitcoin mining companies that are repurposing power capacity and land for AI computing facilities. The trade reflects a growing divide between companies able to monetize scarce electricity connections and crypto networks facing weaker institutional demand for their native tokens.
Sigel’s argument places mining companies’ energy infrastructure at the center of the AI buildout. In his view, application-specific integrated circuit, or ASIC, machines used to mine bitcoin are replaceable equipment. Grid access, land, transmission capacity and permits are harder assets to replicate, particularly in regions where new large-scale electricity demand can take years to connect.
For miners with suitable sites, an AI data-center lease could create a revenue stream that is less directly tied to bitcoin’s price and mining difficulty. Sigel said that economic calculation can be stark: CleanSpark would need bitcoin to reach $360,000 before it made sense to cancel existing AI leasing arrangements and return fully to a pure bitcoin-mining strategy.
AI leases reshape the mining investment case
The conversion of mining sites into AI facilities has become a defining part of the NODE portfolio’s positioning, according to Sigel. Mining operators already control sites designed for energy-intensive computing, often with substations, cooling systems and industrial-scale connections that can support expansion.
AI customers need much of the same underlying infrastructure, although their hardware and commercial models differ from bitcoin miners. AI data centers require high-performance chips, fiber connectivity and dependable power, while their income can come from multi-year agreements with cloud providers and enterprise customers rather than block rewards.
Sigel contrasted that model with earlier infrastructure booms, particularly the U.S. railroad expansion of the 19th century. Rail construction absorbed roughly 3% of U.S. gross domestic product for nearly two decades, he said. AI spending, by comparison, has reached approximately that level for only one year since GPT emerged five years ago, while many forecasts point to a spending peak around 2030.
The comparison also highlights a financing difference. Sigel said the 1862 federal railroad policy promised rail companies hundreds of millions of acres of land, but transferred title only after construction requirements were met. The U.S. Treasury also issued development bonds that ranked behind private capital. Railroad companies then sold bonds internationally as safe investments, even though repayment depended partly on future land sales tied to property rights that had not yet been fully secured.
AI data centers can begin earning revenue earlier in their construction cycle, Sigel argued. Once a facility has power, fiber and chips, it can process workloads without waiting for a nationwide system to be completed. That structure gives current AI infrastructure projects a more immediate path to cash generation than rail networks that depended on completing long stretches of track.
Demand visibility is another part of the case. Sigel said the four largest cloud providers have more than $2 trillion of contracted backlog, with Microsoft and Oracle accounting for roughly half of that total. He added that many contracts extend beyond five years and may include customer prepayments or arrangements in which customers provide their own graphics processing units.
Those terms can make AI leases especially attractive to mining companies seeking to finance expensive site conversions. Long-duration contracts can support construction spending and reduce reliance on the volatile economics of bitcoin mining, though execution risks remain substantial for firms attempting to build and operate facilities at cloud-computing standards.
Capital-spending leaders lost momentum after June
Sigel described June 1 as the turning point in a market rotation that changed the rewards for heavy capital expenditure. During the first five months of the year, companies spending aggressively on infrastructure performed well, he said. Since June, the market has punished the biggest spenders most severely.
That reversal coincided with selling pressure across software-related assets. Sigel said bitcoin and crypto tokens were traded as part of that broader “software” basket, even though bitcoin miners with AI data-center exposure offered a more direct connection to physical infrastructure demand.
The distinction complicates the idea that crypto assets should move together. A mining company with a power-constrained AI lease may be valued increasingly on contracted data-center cash flows, while a layer-1 token remains dependent on network activity, token issuance, regulatory clarity and the willingness of institutions to hold it.
VanEck cuts exposure to major layer-1 tokens
Sigel said VanEck reduced its exposure to major layer-1 assets, including ether and Solana, in the period following the U.S. election. His explanation was that many tokens doubled in price without a similar acceleration in real-world adoption.
Attention has instead moved toward enterprise-focused blockchain systems, he said, pointing to efforts involving Circle, Stripe, Robinhood and Wells Fargo to develop or use customized chains. The trend could limit the winner-take-all narrative often applied to public blockchain networks.
Regulated financial firms may be reluctant to put funds directly onto open-source networks, Sigel said. When they do engage with public chains, he said they often support three to five networks simultaneously. That approach spreads activity across competing systems and reduces the likelihood that one token captures all institutional demand.
The result is a more demanding environment for large layer-1 valuations. Public networks need to show that transaction activity, stablecoin use, application revenue or other forms of adoption can justify their token economics, rather than relying mainly on expectations of institutional allocation.
Disclosure rules remain a potential catalyst
Sigel identified the proposed CLARITY Act as a potential turning point for token markets. He said a disclosure regime could clarify beneficial ownership and token holdings associated with blockchain labs and foundations, potentially reducing uncertainty for institutions considering exposure to digital assets.
He expects passage could trigger a relief rally in affected tokens, but said the bill’s odds of advancing were at their lowest point of the year. That political uncertainty has helped keep institutional allocations restrained, according to Sigel.
Token supply is also central to his cautious view. Sigel said proposals to reduce validator inflation on Ethereum, Solana and Near reflect concern over dilution from ongoing issuance. Validator inflation refers to newly created tokens distributed as rewards to participants who help secure a network.
Lower issuance could improve the supply outlook for token holders, but it would also affect companies whose models depend on staking revenue. Sigel cited Bitmine as an example of a business that markets staking yields and could face second-order effects if network reward rates fall.
The split in Sigel’s outlook is therefore increasingly defined by cash flows and structure: mining companies with power assets and long-term AI contracts offer exposure to a buildout backed by customer commitments, while major layer-1 tokens await clearer disclosure rules, stronger adoption evidence and a more favorable balance between staking rewards and token dilution.
Explore how AI and crypto intersect in infrastructure and trading with our guide on AI–blockchain synergy today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
