Artificial Intelligence (AI) has made significant strides in recent years. Technology is redefining multiple industries including creation. The launch of AI tools such as ChatGPT, OwlyWriter AI, Dall-E, and more is redefining content creation. However, the processing of these AI algorithms requires advanced chips for efficient computing. AI algorithms are primarily based on Machine Learning (ML) and Deep Learning (DL) capabilities, which require a huge number of cores and Graphic Processing Units (GPUs) to work efficiently.
In easy terms, the processing of complex AI algorithms requires the computing power of a supercomputer, and they are exorbitant and out of the reach of the masses. Thus, cloud computing and parallel processing have emerged as the most viable alternatives catering to the huge demand. However, this storage tool is also expensive, making storage a serious challenge for the increasing inflow of unprecedented amounts of data.
Thus, efforts are initiated for the development of quantum computers, which will be able to process a bulk of possibilities simultaneously. These would potentially speed up the processing of the AI algorithms and large datasets, leading to the creation of more powerful AI models. Quantum computing is still in its infancy resulting in immature quantum hardware. Thus, the industry has redirected its focus towards the development of more advanced chips, namely GPUs. These provide a high-octane system for training and running complex AI algorithms. Their key role is supporting Generative AI (Gen AI). Thus, many market leaders including Google, Amazon, Nvidia, and more are making significant investments in the manufacturing of such chips.
New AI Chip Venture by OpenAI CEO Sam Altman: Case Synopsis
Sam Altman, chief executive officer of OpenAI, during a panel session on day three of the World Economic Forum (WEF) in Davos, Switzerland, presented his interest in establishing a new chip venture. He is targeting potential investors for setting up a network of factories for manufacturing semiconductors. He presented the proposal at WEF, on 18th January 2024, to some prospective investors including UAE's Sheikh Tahnoon bin Zayed al-Nahyan (brother of UAE President Sheikh Mohammed bin Zayed al-Nahyan, the National Security Advisor of UAE, and in charge of the country’s most powerful investment fund), Taiwan Semiconductor Manufacturing Co (TSMC) and Softbank.
It is still not clear whether the new company will be a subsidiary of OpenAI or a separate entity. However, OpenAI will be the first client of the new company. The new company would come as a solution to the prevailing issue of AI chip shortage. Further, OpenAI is also looking forward to diversifying its current chip supplier portfolio and networking with some chip manufacturers beyond Nvidia. The new venture is aimed at meeting the massive demand for AI chips, post the launch of Gen AI capabilities such as ChatGPT.
Overview of the Quantum Chip Market
The Quantum Computing Chip market is anticipated to grow at a CAGR of 45.1%, by 2030. The market is segmented by type (into Photonic Chip, Semiconductor spin qubits, Superconducting Chip, and Trapped ion) and application (into below 30 qubit quantum computers, 30-50 qubit quantum computers, 50-60 qubit quantum computers, and above 60 qubit quantum computer).
Industry growth is driven by the rising adoption of quantum computing technologies. Apart from the Gen AI endeavors, quantum chips are used in the management of data in sectors including fintech, simulation, defense, space research, and more. Additionally, government initiatives and major market players are also driving the market's growth.
For instance, in accordance with the CHIPS and Science Act passed in August 2022 by the 117th Congress, under the Presidential leadership of Joe Biden, the US Department of Defense (DoD) launched the Rapid Assured Microelectronics Prototypes (RAMP) using advanced commercial capabilities program. This was aimed at developing sustainable microelectronics for defense technologies. In line with the program, Microsoft, within the Azure Government cloud environment, developed three novel state-of-the-art chips, to ensure compliance with DoD supply chain requirements. It is also extendable to the commercial design environment in Azure, with a motive to accelerate the goals of the CHIPS Act and ensure the development of a sustainable domestic supply chain for semiconductors.
The market has huge potential across the globe with Europe anticipated to hold a prominent share. This is attributed to the rising adoption of new technologies by medium and large-scale businesses in the industrial sector. This would be further supported by the rising number of startups and efforts by governments to boost the usage of cloud-based technology. However, the largest market share would be held by the Asia Pacific region, primarily from emerging economies such as China, Japan, and South Korea. This would be attributed to the adoption of quantum computing across industries such as space and defense, healthcare and pharmaceutical, energy and power, and more.
Further, the adoption of quantum computing would be more evident in cloud infrastructure as against on-premises infrastructures. The users are anticipated to invest in Noisy Intermediate-Scale Quantum (NISQ) systems to solve real-world problems. The enhanced flexibility offered by these systems is estimated to contribute towards the large-scale adoption.
Thus, the successful integration of AI and quantum computing requires innovations in the semiconductor designs to meet the data demands of AI applications, however, the development of such semiconductor chips suffers from such potential challenges. The primary issue is the hefty investment required for the production of such chips. Another major challenge faced by the industry is the unstructured supply chain. The huge chip production processing time largely affects the semiconductor supply chain. The time between initial processing and the final product takes weeks. Also, up to 30.0% of production costs, is lost during this time to testing and yields.
Encapsulating, the effective designing and development of semiconductor chips would revolutionize the computing, processing, and storage of complex AI algorithms. The chips should be designed to reduce the power consumption of AI silicon while still maintaining performance robustness. Also, with the expansion of the AI market, the need for semiconductor chips would accelerate, making it vital for manufacturers to optimize the production capacity to meet the escalating demand. The manufacturers are also required to keep a check on the manufacturing cost of the AI chips. Thus, it can be rightly said that the future of AI is highly dependent on the roadmap of the semiconductor industry, representing the huge market potential.