Micron, Samsung, SK Hynix Invest in Anthropic: AI Shift

According to Anthropic, the U.S.-based artificial intelligence (AI) company has successfully completed its latest Series H funding round. While the financing amount itself is significant, the most remarkable development is the unprecedented composition of its investor base: for the first time, the world's only three high-bandwidth memory (HBM) manufacturers—Micron Technology, Samsung Electronics, and SK Hynix—have simultaneously become shareholders in the same AI company.

Anthropic reportedly raised US$6.5 billion in the round, bringing its valuation to approximately US$96.5 billion. The new valuation surpasses OpenAI’s previous US$85.2 billion valuation, positioning Anthropic as one of the highest-valued AI companies globally.

A Historic Alignment Among HBM Rivals

For years, Micron, Samsung, and SK Hynix have been fierce competitors in the memory market, competing aggressively for supply positions with major AI chip customers including NVIDIA, AMD, and Google. Market share battles in every generation of HBM products have been intense, making their joint investment in Anthropic an extraordinary industry event.

Anthropic has described the three memory manufacturers as "strategic infrastructure partners" rather than traditional financial investors. Although the size of each investment has not been disclosed, the designation highlights their importance within Anthropic's long-term AI infrastructure strategy.

The company stated that these partnerships will help Anthropic "reliably scale compute capacity at the speed customers require."

HBM Emerges as a Critical AI Infrastructure Bottleneck

Industry analysts note that the statement reflects a growing reality in the AI ecosystem: HBM has become one of the most constrained and strategically important components in the global AI supply chain.

By 2026, HBM is no longer viewed as a standard semiconductor product. Instead, it has become a critical resource underpinning large-scale AI model training and inference. Industry estimates suggest that the annual production capacity of leading HBM manufacturers was effectively sold out during the first quarter of the year, with supply-demand gaps ranging between 20% and 50%. Market shortages are expected to persist through at least 2028.

Current market estimates place SK Hynix's HBM market share at approximately 50%, followed by Samsung at 28% and Micron at 22%.

Alongside the new financing round, Anthropic has reportedly secured major infrastructure commitments, including a US$5 billion investment from Amazon and additional commitments from hyperscale cloud operators totaling approximately US$15 billion. The company has also secured access to large-scale computing resources, including Amazon's planned 5GW AI compute capacity, next-generation TPU infrastructure from Google and Broadcom, and GPU cluster access through SpaceX's Colossus platform.

While compute availability is improving, memory remains a critical constraint. By bringing all three leading HBM manufacturers into its shareholder structure, Anthropic is effectively creating a strategic supply-chain alignment that strengthens its long-term access to advanced memory technologies while raising barriers for competitors.

Different Strategies, Shared Objectives

Although the three memory companies share a common interest in AI growth, each enters the partnership with distinct strategic goals.

SK Hynix currently holds a dominant position in NVIDIA's Rubin platform ecosystem and is expected to account for roughly 70% of HBM4 supply for the platform. HBM-related revenue has become a major profit driver for the company, with margins significantly exceeding those of conventional DRAM products. Deepening ties with Anthropic provides a pathway to secure long-term AI inference demand.

Samsung, meanwhile, experienced setbacks in the HBM market during 2024 and 2025 due to HBM3E yield challenges. The company returned to competitiveness with HBM4 production and has since secured key positions in AMD's next-generation AI accelerator platforms as well as a substantial share of Google's TPU-related memory demand. Supporting Anthropic helps Samsung diversify its AI customer portfolio beyond NVIDIA-centric opportunities.

Micron represents the smallest of the three participants by scale but arguably holds unique strategic value. As the only U.S.-based HBM manufacturer, Micron benefits from substantial support under the U.S. CHIPS Act and occupies an increasingly important position amid evolving geopolitical and semiconductor supply-chain considerations. The company's annualized HBM revenue run rate has expanded rapidly, reflecting strong demand from AI infrastructure deployments.

AI Competition Expands Beyond Models

The significance of this investment becomes even clearer when viewed within the broader transformation of the AI industry over the past 18 months.

Anthropic's annualized revenue reportedly increased from approximately US$3 billion in early April to US$4.7 billion within less than two months, highlighting the accelerating adoption of its AI services. Its Claude model family has become one of the few advanced AI platforms available across multiple major cloud ecosystems, including Amazon Web Services, Google Cloud, and Microsoft Azure. Meanwhile, the rapid growth of Claude Code is reshaping the market for enterprise AI development tools.

At the same time, the focus of AI investment is shifting. While earlier competition centered on model performance, the industry is increasingly recognizing that long-term leadership depends on control of underlying infrastructure—including compute, memory, networking, and energy resources.

Hyperscale cloud providers, semiconductor manufacturers, and energy companies are increasingly becoming direct stakeholders in AI companies, signaling a broader trend toward vertical integration across the AI value chain.

From Microsoft's deep partnership with OpenAI to Google's investment in proprietary TPU development, and now the simultaneous investment by the world's three leading HBM manufacturers in Anthropic, the competitive landscape is evolving from a race to build the best AI model into a race to secure the most comprehensive AI infrastructure ecosystem.

The decision by Micron, Samsung, and SK Hynix—companies that have traditionally competed intensely in the memory market—to become shareholders in the same AI company sends a powerful signal. It underscores how critical advanced hardware infrastructure has become to the future of artificial intelligence and highlights the growing convergence between semiconductor manufacturing and AI platform development.

This investment is not the culmination of the AI era. Rather, it marks the beginning of a new phase in which control over infrastructure, supply chains, and strategic technology partnerships will play an increasingly decisive role in shaping the future AI ecosystem.