When Marc van der Chijs began shifting a large portion of his bitcoin holdings into artificial intelligence, it felt like the natural next step for a market that had already embraced speculative technology. AI was no longer just a software story; it was becoming a capital-intensive, infrastructure-heavy, and potentially economy-defining sector. For a crypto-native investor, the move made sense. The same appetite for high-risk, high-reward innovation that drove early bitcoin adoption was now pointing toward compute, data centers, chips, and enterprise AI platforms.
But conviction in a technology is not the same as confidence in its stability. As the AI buildout has accelerated, so have concerns about how concentrated, interconnected, and financially leveraged the ecosystem has become. In a recent warning, van der Chijs argued that the risks surrounding AI have grown serious enough to trigger systemic shocks across banking and critical infrastructure. His message is blunt: the system may be moving faster than its safeguards can keep up with.
From Bitcoin to AI: A High-Convexity Bet
Van der Chijs did not enter the AI trade from a place of caution. He sold a meaningful amount of bitcoin to invest in AI, a move that reflects how the two asset classes have increasingly shared the same investor psychology. Both are associated with technological disruption, network effects, and the possibility of outsized returns. Both also attract intense speculation, meaning momentum can move markets quickly in both directions.
For a while, that seemed like a reasonable trade. Bitcoin had already captured much of the speculative attention, while AI offered a broader, more institutionalized growth narrative. Hyperscalers, chipmakers, energy providers, and software companies all appeared to benefit from the expansion of the AI stack. The investment thesis was less about a single asset and more about a full technology transition.
Now, according to van der Chijs, the picture has become more complicated. The same factors that made AI attractive have also made it riskier. The question is no longer whether AI will be transformative, but whether the financial and physical systems supporting it are prepared for the volatility, concentration, and interdependence that come with it.
Why AI Could Create Systemic Banking Risk
The phrase “systemic risk” is often used loosely, but in this context it carries real weight. AI is not just a software layer running on top of existing banks and markets. It is increasingly embedded in the infrastructure that supports the modern economy. That includes cloud computing, power generation, data transmission, enterprise operations, risk modeling, and financial trading systems.
Several channels make this concern plausible. First, the AI buildout is capital intensive. Data centers, advanced semiconductors, cooling systems, and grid upgrades require enormous spending. Much of that spending is financed through bank credit, private capital, bonds, and corporate debt. If AI valuations were to correct sharply, the stress would not stay isolated to a single group of tech stocks. It could ripple through lenders, suppliers, and infrastructure providers.
Second, the sector is highly concentrated. A relatively small number of companies and suppliers dominate key parts of the AI stack. That concentration creates a kind of single point of failure. If one major player faces a disruption, the knock-on effects can spread quickly through the supply chain and the broader financial system.
Third, AI is being integrated into banking and trading environments themselves. Algorithms are used to monitor risk, price assets, detect fraud, manage liquidity, and execute trades. That can improve efficiency, but it can also amplify instability. If multiple systems are making rapid, correlated decisions based on similar data or models, a small shock can turn into a much larger one. In fast-moving markets, that is where fragility becomes dangerous.
Infrastructure Is the Hidden Weak Point
One of the less discussed risks of AI is physical. The technology is often presented as digital, but it is powered by very tangible infrastructure. Data centers consume enormous amounts of electricity. They require advanced cooling. They depend on reliable fiber networks, hardware supply chains, and skilled labor. In many regions, grid capacity is already under pressure, and the pace of AI deployment is straining utilities, real estate, and local planning systems.
That physical dependence matters because infrastructure is slow to build and hard to replace in a crisis. You cannot quickly scale up power capacity during a demand surge. You cannot instantly retool a semiconductor supply chain. You cannot instantly reroute traffic through a network that is already saturated. If a major disruption hits one part of the system, the delay in recovery can cascade through markets, businesses, and public services.
That is why van der Chijs’s warning is less about AI being “bad” and more about AI being too advanced, too integrated, and too exposed to failure. The concern is not that the technology will stop working. The concern is that the systems built around it may not be resilient enough to absorb a serious shock.
Why Some Investors Are Moving Back Into Crypto
In this context, van der Chijs’s decision to move some AI profits back into crypto becomes easier to understand. It is not necessarily a statement that crypto is a traditional safe haven. It is more a reflection of a broader search for systems that are less dependent on centralized intermediaries and more transparent in how they operate.
Bitcoin, in particular, offers a different structure. It is permissionless, censorship-resistant, and transparent in a way that many centralized financial systems are not. On-chain activity can be inspected. Settlement is global. The protocol does not rely on a single corporate balance sheet, a single data center, or a single regulatory decision to keep functioning. In a world where AI is making centralized systems more powerful but also more vulnerable, that structural independence can be appealing.
That said, crypto is not risk-free. It remains volatile, and in times of market stress it can move sharply alongside other risk assets. But for some investors, the appeal lies in the fact that the system itself does not collapse because one bank fails, one model misprices risk, or one supplier stumbles. The decentralization is imperfect, but it changes the failure modes.
The Core Warning: We May Be Moving Too Fast
The most striking part of van der Chijs’s message is not the asset allocation shift. It is the sense that the pace of change has outrun the ability to manage risk. AI is being adopted across banking, trading, energy, logistics, and public infrastructure at a speed that leaves little room for correction. The incentives are powerful, the capital is flowing, and the competition is intense. But the guardrails are still catching up.
That is where the phrase “we have lost control” becomes meaningful. It does not necessarily mean that machines are making autonomous decisions without human oversight. It means that the complexity of the system has grown so large that no single institution, regulator, or market participant fully understands the second- and third-order effects of a disruption. In a highly connected economy, that is a dangerous position.
If AI continues to expand as the dominant engine of economic growth, the risks will not remain confined to tech stocks or AI vendors. They will show up in credit markets, in energy bills, in supply chains, in cyber defense, and in the stability of financial institutions. The question for investors and policymakers alike is whether the system is being built to survive the next shock, or whether it is being built only to capture the next growth cycle.
Van der Chijs’s move back into crypto may look like a contrarian trade, but it is also a signal. It suggests that even someone who made a major bet on AI now sees enough fragility in the system to seek diversification outside the most exposed parts of it. In a market that has already priced in years of AI growth, that kind of caution may be the most important data point of all.
Related read: ECB President Lagarde Intervenes to Block Binance’s EU MiCA License, WSJ Reports
