Grayscale turns Bitcoin miners ETF into AI compute ETF

Grayscale Investments has followed the heavily beaten path of firms pivoting from Bitcoin mining to AI compute as the firm rebranded its Bitcoin Miners ETF (MNRS) into the Grayscale AI Compute ETF (GCPU), which is being marketed as a way for ordinary investors to cash in on the energy and compute demand driving the AI buildout.

The public mining sector is riding a big demand wave, with announced AI and HPC contracts totaling about $70 billion to $100 billion. Even then, available capacity is still well short of what the AI buildout needs to run at full strength.

That room to grow is what GCPU buyers are wagering on to deliver returns as more cash continues to flow in to finance the buildout.

How did Grayscale turn a Bitcoin miners index into a high-performance computing one?

Grayscale’s announcement states that the new fund now tracks the Indxx High Performance Computing Index, which pools companies buying data centers, securing power supply, and stacking the compute capacity that AI models run on, moving on from the previous one tied to the Indxx Bitcoin Miners Index, which tracked companies that solve blocks for Bitcoin.

Another difference that Grayscale explicitly stated is that the rebranded GCPU will not hold any digital assets directly and will not track any coin’s price.

Notably, the GPCU ETF will cover a mix of firms focusing on different parts of the AI buildout. For example, about 50% of the portfolio will track companies built around GPU cloud and AI-hosting capacity from the start.

The other half of the portfolio will be operators with roots in other parts of high-performance computing, including Bitcoin mining, who have announced publicly or are already in the process of pivoting into AI workloads.

The index rebalances every quarter.

Why Grayscale thinks compute is the scarce asset

Grayscale’s Head of Index, Steve Vanourny, is buying the hype around the scarcity of the compute needed to power the AI boom, writing in the September 22 announcement: “GCPU is a direct extension of that thesis, giving investors access to the companies closing the gap between AI’s demand for compute and the physical infrastructure needed to meet it.”

Grayscale’s release backed up the scarcity thesis with numbers highlighting the record-low of vacancies at North American data centers, with roughly six months of spare capacity in reserve, per CBRE data.

Building fresh capacity is not as straightforward as new facilities could take as long as five years to build. Goldman Sachs projects that AI-related capital spending, mostly directed at physical plant, will exceed $1 trillion annually.

AI compute providers were previously Bitcoin miners

Grayscale is packaging a trend that has been reshaping the mining sector for months. AI-integrated miners are up about 21% year to date while the broader Bitcoin market is taking a beating.

Firms that have already made the switch are riding a wave of momentum, including TeraWulf which posted gains near 73%. Riot Platforms is up roughly 94% over the same stretch.

Core Scientific booked $136.7 million in colocation revenue against just $27.5 million from mining in the second quarter of 2026.

The economics also make it a no-brainer for the firms contemplating making the switch. AI cloud services fetch an estimated median of $940 per megawatt-hour, versus $113 to $179 for Bitcoin mining.

Cryptopolitan has reported that miners carrying AI or HPC contracts trade at about 12.9 times forward sales, against 3.7 times for those without, as OpenAI’s roughly $280 billion compute push and a 2,600-gigawatt US interconnection backlog make an already-permitted, grid-connected site worth far more than the rigs sitting inside it.

The smartest crypto minds already read our newsletter. Want in? Join them.