Deal activity tied to artificial intelligence infrastructure remains active as competition for power intensive data center capacity continues to shape valuations and capital flows. According to Joe Nardini, head of investment banking at B. Riley Securities, demand for megawatts has not weakened despite broader concerns that enthusiasm around artificial intelligence has cooled. Buyers linked to AI, high performance computing, and bitcoin mining continue to compete for limited power resources, keeping transactions moving even late in the year. The central constraint is not appetite but access to energy and suitable locations. GPU ready facilities are drawing multiple tenants with strong credit profiles, supporting pricing and deal momentum. This dynamic has sustained merger and acquisition discussions across Wall Street, reinforcing the view that AI infrastructure is anchored by real operational demand rather than short term market sentiment.
Bitcoin miners remain a significant part of this landscape, particularly after the latest block reward halving tightened margins across the sector. Many operators have shifted strategy by repurposing existing sites to host AI and high performance computing workloads, allowing them to monetize power capacity more efficiently. This repositioning has supported higher equity valuations and improved access to capital for miners that successfully attract long term tenants. In several transactions reviewed by bankers, assets with high quality power and favorable locations have commanded valuations exceeding four hundred thousand dollars per megawatt, with some negotiations approaching higher levels depending on contract structure. Less attractive sites still receive interest, though bids are more conservative. The divergence highlights how power quality, location, and tenant demand now play a decisive role in determining asset value.
The buyer universe for these assets extends beyond crypto native firms and includes hyperscale technology companies, AI focused operators, and traditional industrial players seeking to reposition underused facilities. Older industrial and office sites with sufficient power are being evaluated as candidates for modular data center development, often drawing interest from dozens of prospective buyers. In some cases, tenants have shown willingness to commit capital before projects are completed, reflecting the scarcity of suitable capacity. While some high profile AI related stocks have retraced earlier gains, the underlying economics of data center infrastructure remain intact. As long as facilities secure tenants at viable rates, deal activity is expected to persist, reinforcing the role of energy and infrastructure as foundational elements of the AI economy.
