The move underscores a growing trend among top AI labs to vertically integrate, with major implications for chip suppliers and crypto mining infrastructure alike.
SK Group Chairman Chey Tae-won has confirmed that Anthropic, the AI company behind Claude, has opted to develop its own computing power rather than rely entirely on third-party providers. The statement, made during events surrounding SK Hynix’s Nasdaq listing in July 2026, highlights a seismic shift in how the world’s most advanced AI companies think about infrastructure.
Anthropic’s $50 billion bet on self-sufficiency #
Anthropic committed $50 billion to building custom AI data centers across the US, an investment first announced in November 2025. The facilities, located in Texas and New York, are being designed specifically for AI workloads through a partnership with Fluidstack.
By April 2026, reports surfaced that Anthropic was also exploring the development of custom AI chips, both ASICs and GPUs. The goal is straightforward: reduce dependency on external chip suppliers during a period of persistent compute shortages.
This push for proprietary silicon mirrors moves by other tech giants. Google has its TPUs, Amazon has Trainium, and Meta has been investing heavily in custom hardware.
SK Hynix plays both sides of the table #
Chey Tae-won isn’t just an observer in this story. SK Hynix, the semiconductor subsidiary he oversees, participated in Anthropic’s May 2026 Series H funding round alongside Samsung and Micron. The company is one of the world’s leading producers of high-bandwidth memory, or HBM, the specialized chips that make large-scale AI training possible.
SK Hynix completed its Nasdaq listing on July 10, 2026, valued at $26.5 billion.
Chey has led SK Group since 1998, and his strategic pivot toward AI infrastructure and semiconductor research over the past several years has positioned the conglomerate to benefit from exactly this kind of convergence.
Why crypto investors should care #
The competition for compute power directly affects the economics of crypto mining. When AI companies like Anthropic commit $50 billion to data center construction, they’re competing for the same energy resources, the same cooling infrastructure, and in some cases the same GPU supply chains that Bitcoin miners depend on. Several publicly traded mining companies have already begun pivoting to offer AI hosting services precisely because the margins are better.
There’s also the tokenized compute angle. Projects like Render, Akash Network, and io.net have built decentralized GPU marketplaces that attempt to address the very supply constraints Anthropic is trying to solve through vertical integration. If major AI labs increasingly build their own infrastructure, the addressable market for decentralized compute networks could narrow, at least at the high end. Conversely, smaller AI developers who can’t afford to build custom data centers might become more reliant on decentralized alternatives.
SK Hynix’s positioning as both an investor in and supplier to Anthropic offers a template for how traditional semiconductor firms might navigate this transition. The $26.5 billion Nasdaq valuation suggests the market is already pricing in a future where memory suppliers sit at a critical chokepoint in the AI supply chain.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our