How Artificial Intelligence Is Bringing a Shift In the Chip Ecosystem: What Opportunities Does It Open For India's Semiconductor Aspirations Advanced Micro Devices (AMD) acquisition announcement of ZT Systems points at a globals shift towards chip designing and optimizing solutions
Opinions expressed by BIZ Experiences contributors are their own.
You're reading BIZ Experiences India, an international franchise of BIZ Experiences Media.
In the Artificial intelligence (AI) era, globally, there is a gradual shift towards chip designing and optimizing solutions. Advanced Micro Devices (AMD) acquisition announcement of ZT Systems in a cash and stock transaction valued at $4.9 billion, is a testament to this. The deal will help AMD compete against artificial intelligence chip rivals like NVIDIA: The need for custom designed, optimized and tightly integrated AI servers, racks or clusters is where ZT Systems brings in its expertise. "Our acquisition of ZT Systems is the next major step in our long-term AI strategy to deliver leadership training and inferencing solutions that can be rapidly deployed at scale across Cloud and enterprise customers," said AMD chair and CEO Dr. Lisa Su.
Counterpoint Research in its report – Move from AI to Gen AI – has predicted significant increase in 'semiconductors content' through 2030 from Cloud to Edge. "The growth in semiconductor capabilities has allowed us to compute and train the data-driven models in Cloud and inferencing at the edge efficiently. This is possible only with increase in semiconductor content both in the datacenters and the Cloud," the report quoted. The total semiconductor consumption across the three key AI segments – servers, PCs and smartphones will grow from $447 billion/year in 2024 to $1 trillion a year in 2030.
What does this deal mean for the global semiconductor ecosystem?
AMD is growing as a major challenger to NVIDIA with a series of acquisitions. "Having the right building blocks from chip to software to solution design will be key for AMD to create a highly valuable, sticky and differentiated ecosystem similar to NVIDIA. This deal aims not only to complete AMD's end-to-end AI portfolio to design a custom optimized DC solution for its biggest customers but also puts NVIDIA in a tough spot," said Neil Shah, partner & co-founder, Counterpoint Research.
"AMD's decision is wise to not compete fully with its partners and to stop at software stack, designing the solution and not venturing into manufacturing. But, whichever company acquires the ZT Systems manufacturing division would have great access to key hyperscalers and AMD-ZT Systems designs to win at major hyperscalers in the future," he added.
The specialized talent coming from ZT Systems will help accelerate future silicon and software roadmap for AMD. NVIDIA is still well ahead on ecosystem and hardware traction but AMD is catching up.
From selling chip to designing & solution: What this shift means for India
This global shift also marks an important chapter for India as it is the second or third largest market for most of the electronics end devices and equipment with growing amounts of semiconductor consumption. . Between 2021-2026 more than $300 billion worth semiconductor will be consumed in India alone."So, India offers an unique advantage not only in terms of being one of the largest semiconductor consumption markets but also as a future hub to churn out semiconductor solutions for global consumptions," Shah explained.
AI and (advanced Gen AI) will be penetrating smartphones, PCs, different software platforms, enterprise and public infrastructure powered by AI-centric data centers. "With billions of edge devices to become AI-centric by the end of this decade, India is up for an AI explosion, which will warrant Made in India designed and produced AI centric semiconductors to software stack. While producing advanced AI centric semiconductors will take time as it requires fabs with advanced process nodes. However, locally designing the software stack, AI data models training and inferencing beckons. The aspirations need to be set with building indigenous skills, capabilities and Intellectual Property (IP) to design, develop and deploy solutions across AI Stack," the partner added.
