CoreWeave shows how crypto-era infrastructure quietly became AI’s backbone

CoreWeave’s transformation from a crypto-mining operator to a large-scale AI infrastructure provider highlights a broader shift in how computing resources are reused across technology cycles.

In its latest newsletter, The Miner Mag outlined how Ethereum’s move away from proof-of-work reduced demand for GPU-based mining, pushing companies like CoreWeave to redeploy hardware toward AI training and other high-performance computing workloads as demand for compute began to rise.

As Cointelegraph previously reported, CoreWeave began moving away from crypto mining as early as 2019, shifting first into cloud and high-performance computing before fully repositioning itself as a GPU infrastructure provider for AI workloads.

That pivot has since gained momentum. Chipmaker Nvidia recently agreed to a $2 billion equity investment in CoreWeave, a move Miner Mag said reinforced the company’s position as one of the largest independent GPU infrastructure operators outside the major cloud providers.

CoreWeave’s growth has also translated into significant liquidity for company executives, who have generated roughly $1.6 billion in proceeds from stock sales since the company’s initial public offering in March last year, Miner Mag said.

CoreWeave (CRWV) stock. Source: Google Finance

Related: Bitcoin mining’s 2026 reckoning: AI pivots, margin pressure and a fight to survive

From crypto mining to AI data centers

The shift toward AI workloads has proven profitable for several crypto miners, including HIVE Digital, TeraWulf, Hut 8 and MARA Holdings.

Like CoreWeave, these companies have repurposed energy infrastructure and computing capacity originally built for mining into data centers that support AI and high-performance computing.

However, AI data centers are beginning to face some of the same challenges that Bitcoin (BTC) miners encountered in their early years. As Cointelegraph recently reported, local opposition tied to power consumption, grid strain and land use is emerging in several regions hosting large AI facilities.

Even so, the market remains in flux. Data cited by Bloomberg, based on research from DC Byte, shows thousands of new entrants entering the data center business. By 2032, Big Tech companies could see their share of global computing capacity fall below 18%, suggesting a more fragmented and competitive market.

If that trend holds, AI data centers, much like crypto mining before them, may increasingly operate outside the direct control of large technology companies.

AI data centers may become less concentrated among Big Tech companies as new operators enter the market. Source: Bloomberg

Related: What role is left for decentralized GPU networks in AI?