Bernstein reaffirmed its Outperform rating and $36 price target on TeraWulf Inc. (NASDAQ:WULF), as the company is rapidly transforming from a Bitcoin (CRYPTO: BTC) miner into an AI infrastructure provider.

“Power Landlord Of AI“

Bernstein analysts led by Gautam Chhugani said TeraWulf’s latest earnings offered the clearest evidence yet that its AI infrastructure strategy is translating into financial results.

The company generated $32 million in high-performance computing (HPC) revenue during Q2, accounting for 71% of total revenue—a significant shift away from its legacy Bitcoin mining business.

The brokerage said the results validate its June investment thesis that TeraWulf is evolving into a "power landlord of AI" by monetizing large-scale power infrastructure for hyperscalers and AI companies, The Block reported.

Based on the company’s contracted backlog, TeraWulf expects to generate more than $1.8 billion in average annual revenue and over $1.5 billion in annual net operating income.

Bernstein also highlighted TeraWulf’s one-gigawatt Muskie campus in Kentucky as its next major growth driver. The first 500-megawatt phase is expected to come online in the fourth quarter of 2028, with full buildout targeted by 2030.

Cost Inflation Bears Watching

While remaining bullish, Bernstein cautioned that development costs are increasing.

The firm now estimates construction costs of $10 million to $12 million per IT megawatt, compared with its previous estimate of $8 million to $10 million.

TeraWulf’s flagship Lake Mariner campus also experienced rising development costs during the quarter.

Separately, the company amended its Fluidstack lease, increasing total contracted revenue by roughly $300 million to $7.2 billion over 10 years.

Bernstein values TeraWulf at 21x forward EV/EBITDA based on steady-state 2030 earnings and expects the company’s Bitcoin mining operations to be fully phased out by 2028.

The brokerage identified customer concentration risk as the primary downside to its investment thesis, given the company’s reliance on a limited number of hyperscale AI clients.

Image: Shutterstock