KEEL

Keel does not have a true classic network effect. Adding one tenant to a data-center site does not materially improve the product for another tenant in the way a marketplace or social platform would. There is some ecosystem reinforcement: more committed capacity can improve relationships with utilities, equipment vendors, financiers, and hyperscale customers that want to colocate near other large users. But those benefits are indirect and mostly reflect better site economics rather than self-reinforcing user value. Customers can multi-home across providers with little loss of utility, so the network value created by Keel remains limited, one-sided, and highly location-specific rather than a durable companywide moat.

Switching costs are real, but only moderately strong. Keel’s facilities are purpose-built for high-density compute, with custom power, cooling, and interconnection requirements that make relocation expensive and operationally disruptive. Moving workloads can require re-racking specialized hardware, reworking power density assumptions, renegotiating site-specific service terms, and accepting downtime risk. Those frictions create lock-in for existing customers, especially when contracts are tied to a specific site and power profile. Still, large hyperscalers and enterprise AI users routinely design for redundancy and multi-sourcing, so they are not permanently trapped. The result is meaningful migration friction, but not the kind of deep, system-wide lock-in that would support a strong, enduring moat.

Keel’s intangible assets are weak relative to stronger infrastructure franchises. The company does not appear to have a dominant consumer or enterprise brand, and its reputation is still being shaped by a transition from crypto-mining roots into AI-focused infrastructure. Any proprietary know-how around site selection, power procurement, or infrastructure design is valuable, but it is largely execution-based and difficult to defend through legal exclusivity. There is no clear evidence of a patent wall, exclusive license portfolio, or brand premium that would let Keel charge meaningfully more than similarly situated competitors. Customers care far more about available power, interconnection speed, and economics than about brand. That makes intangible assets a minor, not decisive, source of advantage.

Keel has some cost advantages, mainly from scale and procurement leverage, but they are not overwhelming. A 2.2 GW pipeline in constrained markets can support lower unit costs over time through bulk purchasing, standardized development practices, and better bargaining power with equipment vendors. The Vertiv agreement also suggests some supply-chain certainty and potentially favorable economics for power and cooling components. In addition, operating in power-rich or demand-response-eligible regions can offset energy costs. However, these benefits are partly cyclical and depend on power pricing, supplier availability, and project timing. Well-capitalized rivals can pursue similar scale strategies, so Keel’s cost position is more a moving target than a permanently superior structure.

Keel benefits from a market with real capacity constraints, which raises barriers to entry. High-density HPC and AI infrastructure requires scarce grid interconnections, large capital commitments, utility approvals, and specialized build capabilities, all of which limit the number of viable participants. That creates an oligopolistic feel in some geographies. Still, the company is not a natural monopoly and does not occupy a protected, unavoidable chokepoint. Several peers compete for the same scarce power markets, financing sources, and hardware suppliers, and new entrants can still arrive if they secure capital and permits. Efficient scale exists at the project level, but it is not exclusive enough to provide a broad, durable corporate moat.