📰 Key Takeaways

Wall Street is re-evaluating the value of crypto companies—no longer just looking at Bitcoin holdings, trading volume, or mining revenue, but starting to focus on a brand new dimension: AI infrastructure. In June 2026, Galaxy Digital’s stock price surged significantly, but the driving force wasn’t the crypto market conditions—it was investors repositioning their large data center project “Helios Campus” in Texas. The campus is specifically designed for AI and high-performance computing (HPC), and Galaxy Digital’s management has hinted that Helios could account for a significant portion of the company’s total market value in the future, sparking immediate positive market response.

There’s a deeper structural reason behind this shift. Modern AI systems require massive amounts of GPUs and specialized networking equipment for training and inference. However, as demand has exploded, the bottleneck has evolved from “can we build more advanced models” to “can we secure enough power and data center space.” Many power companies report that waiting lists for grid connections for large AI data centers have stretched to several years in some regions. This makes access to power even more scarce than GPUs themselves in certain areas.

Crypto mining companies have long competed for cheap power, inadvertently accumulating land, power infrastructure, and data center assets—exactly what AI companies desperately need today. As a result, investors have started looking at some crypto firms through the lens of “AI infrastructure companies” rather than simple blockchain businesses. The Galaxy Digital case clearly shows that future valuation logic for crypto companies may depend in part on the physical computing resources they control, rather than purely digital asset exposure.

💬 JudyAI Lab Perspective

Galaxy Digital’s repositioning as an AI infrastructure player has driven stock prices up—this case is worth watching because it shows the capital market has started evaluating “old assets” through a completely new dimension.

The original text points out that the real bottleneck for modern AI systems training and inference has shifted from “can we build more advanced models” to “can we secure enough power and data center space.” Grid application waitlists are years long in some regions, and power access is even scarcer than GPUs themselves. In the process of competing for cheap power over the years, crypto mining companies have inadvertently accumulated land, power infrastructure, and data center facilities—exactly what AI companies need now. For AI builders, this reminds us: “compute accessibility” itself is already a prerequisite for whether a product roadmap can be executed, no longer an assumed background condition.

When evaluating AI project feasibility, consider asking a foundational question before diving into technical architecture: where does our computing power come from, and is it stable? The answer to this question may determine the project’s actual timeline earlier than any feature design.

📅 Source Information

🔗 Further Reading