ChatGPT on the wall, should 5G be abandoned at all?
When asked to answer the question “if 5G can be abandoned” in the form of a poem, ChatGPT responded like this:
Not only impressing with its poetry and recognition of the importance of 5G, the artificial intelligence (AI) chatbot also managed to pass some of the most highly regarded exams, including the U.S. medical licensing exam and a Wharton Business School exam for an MBA program. No wonder it quickly attracted 100 million users just two months after launch, making it the fastest-growing consumer application in history.
The emergence of conversational AI chatbots is revolutionary and represents a tipping point toward a future powered by AI. As a result, competition is heating up, with major tech companies around the world vying to create their own versions of Microsoft-backed ChatGPT. Google revealed its plan to roll out “Bard” based on its existing large language model Lamda. Meanwhile, China’s Baidu announced that its “Ernie bot” will likely complete internal testing in March before being made public. Major e-commerce players such as Alibaba are also joining the global race to develop ChatGPT-like dialogue robots.
As companies strive to create more powerful AI models, demand for high performance computing (HPC) solutions is increasing. AI chatbots need massive training on huge amounts of data to constantly make improvements and provide more accurate responses. This is where AI processing chips such as Graphics Processing Units (GPU) and network processors play a crucial role. GPUs are optimized for training deep learning models as they can process multiple parallel tasks much faster and more efficiently, while network processing chips enable large dataset communication with low latency. As the success of AI-enabled applications continues to grow, we believe leading semiconductor company AMD is a key enabler and is well positioned to benefit given its strength in cutting-edge AI computing chips.
Advanced Micro Devices (AMD)
AMD’s Growing CPU Market Share
Source: PassMark CPU Benchmarks, AMD vs. Intel Market Share.
Xilinx Acquisition Expands Addressable Market
Source: AMD.
From modest beginnings, AMD’s early products were mainly general microprocessors for PCs. Thanks to CEO Lisa Su’s foresight when she took over the reins in 2014, AMD has become an industry leader in high-performance computing. Known for its CPU processors and GPU graphic cards, its products power next-generation computing.
AMD has enjoyed strong market share gains in the server business, growing from low single digits to over 20%. A large part is due to Lisa’s decision to bet on high-performance computing architecture, including powerful computer processors and graphics chips for gaming, AI and supercomputing. Strong partnerships with leading foundries such as TSMC have also enabled AMD to deliver flawless product execution. AMD’s latest transformational move was the acquisition of the world’s largest FPGA company, Xilinx, in 2022. FPGA chips enable flexible and reconfigurable computing, which is critical to programmability for AI accelerators and new technology developments such as 5G base stations in the early-mid stage. The Xilinx acquisition further bolstered AMD’s position in computing and broadened its reach to end markets (industrials, healthcare, auto, aerospace), thereby expanding the addressable market to an estimated US$ 300bn in 2027, a 30% CAGR from 2020.
AMD is leaning into AI. It has named AI as one of its strategic pillars and launched a PervasiveAI initiative, recognition of the importance and growing demand for adaptive computing. To drive its next phase of growth, AMD has a strong line-up of products based on 5nm processes and leading-edge architecture, including optimizations for AI and machine learning. Apart from hardware, AMD has a unified AI software stack to empower developers.
As big as AI is today, it is still in the early stages. We believe AMD should continue its leadership in traditional computing environments (data center, gaming), but also be a significant player in broad embedded markets. With solid management, strong product portfolio and an expanded addressable market, we believe AMD is well positioned to proliferate AI technology.
