Samsung Electronics and Broadcom have signed a major partnership worth more than $200 billion through 2030 to jointly develop memory chips, manufacturing processes, and advanced packaging technologies for artificial intelligence infrastructure. The memorandum of understanding, announced at an AI Summit in San Francisco, represents a significant shift in how semiconductor companies are organizing their supply chains to meet the explosive demand for AI chips.

Why Are Samsung and Broadcom Joining Forces on AI Chips?

The partnership reflects a fundamental change in how AI infrastructure is being built. As AI systems grow more powerful and demanding, they require close coordination between multiple specialized semiconductor technologies working in harmony. Neither company alone can deliver everything needed; Samsung brings memory and manufacturing expertise, while Broadcom brings AI accelerator design and networking prowess.

The collaboration covers several critical areas that together form the backbone of modern AI systems. Samsung will supply Broadcom with High Bandwidth Memory (HBM), which is specialized memory designed to move data extremely quickly between processors and storage. Broadcom will also use Samsung's advanced 2-nanometer manufacturing process for its own chips, and the companies will jointly develop advanced packaging techniques that allow different chip components to work together more efficiently.

"As AI infrastructure continues to scale, close collaboration across the semiconductor ecosystem becomes increasingly important. By combining Samsung's memory and foundry expertise with Broadcom's AI and connectivity leadership, we aim to continue to deliver technologies that power the next generation of AI infrastructure," said Charlie Kawwas, President, Semiconductor Solutions Group at Broadcom.

Charlie Kawwas, President, Semiconductor Solutions Group at Broadcom

What Technologies Are Samsung and Broadcom Developing Together?

The partnership spans multiple technology domains that are essential for building competitive AI systems. Understanding these areas helps explain why this collaboration matters so much for the future of AI infrastructure.

• High Bandwidth Memory (HBM): Samsung will supply specialized memory chips that allow data to move between processors at extremely high speeds, critical for training and running large AI models efficiently.
• Advanced Manufacturing at 2 Nanometers: Broadcom will use Samsung's cutting-edge 2-nanometer process technology to manufacture its own semiconductor products, including wireless broadband communications devices.
• Advanced Packaging Technologies: The companies will collaborate on 2.3D and 2.5D heterogeneous integration, which allows different types of chips to be stacked and connected more tightly, improving performance and reducing power consumption.

These technologies are not optional extras; they are becoming mandatory for companies trying to build competitive AI systems. The combination of fast memory, efficient manufacturing, and clever packaging allows AI chips to deliver better performance while using less electricity, a critical advantage as data centers consume more and more power.

How Does This Partnership Change the Semiconductor Industry?

The Samsung-Broadcom deal signals a broader industry trend toward vertical integration in semiconductor supply chains. Rather than relying on dozens of independent suppliers, major AI chip companies are increasingly locking in long-term partnerships with foundries, memory makers, and packaging specialists. This approach allows companies to optimize their entire supply chain for AI workloads rather than settling for generic off-the-shelf components.

The $200 billion value of the agreement over five years underscores how much money is flowing into AI infrastructure. For context, this single partnership represents a massive commitment of resources, reflecting the intense competition among chipmakers to supply the hyperscale cloud providers and AI companies building the next generation of AI systems.

Samsung's position is particularly interesting because it operates across multiple parts of the semiconductor ecosystem. The company manufactures memory chips, runs a foundry business that manufactures chips for other companies, and develops advanced packaging techniques. This breadth allows Samsung to offer what executives call "end-to-end" solutions, meaning a customer can source memory, manufacturing, and packaging all from one vendor. Broadcom, by contrast, specializes in AI accelerators, networking chips, and custom processors for cloud providers, making it a natural partner for Samsung's manufacturing capabilities.

Steps to Understanding Semiconductor Supply Chain Integration

• Recognize the Bottleneck: AI systems require multiple specialized components working together seamlessly; no single company can excel at all of them, creating pressure for partnerships.
• Understand Vertical Integration: Companies are moving away from buying components on the open market and instead locking in long-term deals with suppliers to ensure they get the latest technology and priority access.
• See the Competitive Advantage: Companies that can combine memory, manufacturing, and packaging from a single source can optimize their designs more effectively than competitors using multiple vendors.

The partnership also highlights how AI infrastructure is becoming increasingly specialized. Broadcom's portfolio includes custom accelerators for specific AI workloads, Ethernet networking silicon for connecting AI systems together, optical interconnects for high-speed data transmission, and switching platforms for routing data efficiently. Each of these components requires cutting-edge manufacturing and integration with high-performance memory, which is exactly what Samsung offers.

Looking ahead, this type of strategic partnership is likely to become more common as AI infrastructure becomes more complex and competitive. Companies that can offer integrated solutions combining memory, manufacturing, and packaging will have significant advantages over those relying on fragmented supply chains. The Samsung-Broadcom agreement, valued at more than $200 billion through 2030, represents a bet that this integrated approach is the future of AI chip development.