Bitcoin Miners Become AI's "Power Landlords": Bernstein

Wall Street analysts are betting Bitcoin miners are quietly becoming indispensable to the artificial intelligence industry as "power landlords," per a Wednesday research note from Bernstein, the investment arm of Société Générale. The report initiates coverage on two miners—TeraWulf and Cipher Digital, both rated "Outperform"—arguing that former crypto-centric operators are unusually well-placed to solve one of AI's more stubborn bottlenecks: large-scale, ready-to-use power. The numbers show how fast the pivot has happened. Over the past two years, miners have signed 17 deals worth more than $110 billion, contracting roughly 6 gigawatts to Google, Amazon, Microsoft, Nvidia, and CoreWeave—about 10% of all AI data centers currently under construction in the U.S. "Bitcoin miners remain best positioned to solve 'time to compute,'" the analysts wrote, citing the industry's planned 30-gigawatt power portfolio and its experience delivering "warm powered shells"—facilities with electricity already flowing, ready for someone to plug in the GPUs. After years of being mocked for "wasted" energy, the miners are now the ones holding the keys.

Bernstein projects that the aggregate AI revenue of the companies it covers will climb from $1.2 billion this year to $10.7 billion by 2030. TeraWulf, anchored by a partnership with Fluidstack and Google, is expected to reach $1.7 billion in AI revenue by that point, with EBITDA margins approaching 84%. Cipher Digital, whose client base is majority hyperscaler, is projected to hit $1.2 billion in AI revenue with margins near 93%.

The colocation model the companies favor—leasing powered facilities under long-term, take-or-pay contracts—has drawn particular investor attention for its stability. Project financing markets now cover 75–85% of construction costs for these facilities, Bernstein noted, at interest rates well below the returns the underlying contracts generate. Cheap debt, expensive leases—a combination that tends to make bankers smile.

Bitcoin continued sliding lower on Thursday, dragging the rest of the crypto market with it and offering no clear sign of a bottom. The leading crypto has dropped roughly 17% over four days, from nearly $74,000 on Monday to Thursday's intraday low of $61,556, per CoinGecko data. In under four days, total crypto market liquidations hit $4.47 billion, with bullish bets contributing $3.82 billion—roughly 93% of all positions wiped. At time of publication, BTC remains underwater, trading at...

The report reflects a broader shift in how investors and technology executives think about power infrastructure. As demand for AI computing accelerates, securing reliable electricity at scale has become as strategically important as the chips themselves—and the miners who spent years chasing cheap power now find themselves sitting on a valuable asset.

Shares in both TeraWulf (WULF) and Cipher Digital (CIFR) are down on the day as of this writing, but both have put up towering gains so far in 2026, with WULF up nearly 122% and CIFR rising about 69% during that same span.

As people increasingly turn to AI chatbots for advice, companionship, and emotional support, a new study suggests that even the most advanced models still struggle to maintain healthy boundaries with users. The study by researchers at the University of Southern California introduced EUDAIMONIA, a benchmark designed to measure what they call undesirable dynamics in human-AI conversations. "Large language models are increasingly used as conversational partners for companionship, emotional disclos...

Law professors preferred answers generated by artificial intelligence over answers written by fellow professors, according to a recent study led by Stanford University that examined how large language models perform on legal reasoning tasks. In the study, 16 professors from 14 U.S. law schools—including Stanford, Yale, New York University, the University of Chicago, Georgetown, UCLA, and the Univers