A new record predicts a skyrocketing demand for AI-related products and services, which manufacturers may struggle to satisfy, leading to yet another global chip shortage.
According to consulting Bain and Company, AI workloads could increase by between 25 % and 35 % annually up to 2027. However, a 20 % increase in demand has a high chance of destabilizing the equilibrium and causing the world to experience yet another chip shortage.
According to the creators of the Global Technology Report 2024,” The AI blast across the intersection of the big end industry could easily surpass that level, creating resilient straits throughout the supply chain.”
Additionally, larger data centers with more than a gigawatt of capacity may be necessary due to our thirst for AI. Existing information areas tend to be between 50 and 200 mw.
The market for AI software and hardware is anticipated to grow between 40 % and 55 % annually over the next three years as a result of the combination of demand for AI-enabled products and AI-enabled infrastructure.
If large data centres currently cost between$ 1 billion and$ 4 billion, in five years they could reach between$ 10 billion and$ 25 billion, the report states. This results in a total AI market prediction of between$ 780 billion and$ 990 billion ( £584 billion and £741 billion ) for 2027.
Notice: Gartner Estimates Worldwide AI Chip Revenue Will Increase by 33 % in 2024.
The offer spider’s online, and the stress it ’s under
The supply chain for AI aspects may be able to scale up at the same rate in order to maintain this growing need. But, in fact, the string is more like a complicated spider’s internet, with the device raw materials at the centre.
The fabs and facilities needed to expand device manufacturing are one thing, and the data centers needed for the operation of AI products are another. Each has a direct period of between three-and-a-half times to over five years, according to Bain, posing a major blocker for keeping up with desire.
According to the report, bleeding-edge fabs that produce the most cutting-edge cards are the most vulnerable website. They will need to raise their output by between 25 % and 35 % between 2023 and 2026 to keep up with the predicted 31 % and 15 % sales growth in PCs and smartphones respectively.
Up to five more bleeding-edge fabs would need to be constructed to keep up, costing an estimated$ 40 billion to$ 75 billion.
Additionally, there is the supply chain involved in turning bits into smartphones and Devices with on-device AI capabilities, such as Apple Intelligence products, which are becoming more popular as more people are looking for information security.
Notice: Gartner: AI-enabled Computers to Dominate Laptop Options for Companies
In order to accommodate the on-device neural processing engines, the silicon surface area on the average notebook core processing unit and smartphone processor has already increased by 5 % and 16 %, respectively. By 2026, Bain anticipates that these products will increase the demand for upstream components by 30 % or more.
Another area of the internet is presentation, and manufacturers would need to triple their production capability if demand for GPUs doubles by 2026. Additionally, different requirements for power and cooling will connect utility companies, who will also need to scale up to meet demand.
The next worldwide chip shortage
Chipmakers have prospered since the onset of the latest conceptual AI boom. NVIDIA, a leader in the sales of graphics processing units, announced record revenues of$ 30 billion ( £24 ). 7 billion ) in the second quarter of 2024, and has a stock market value of over$ 3 trillion ( £2. 2 trillion ). Similar achievement has been reported by transfer manufacturers Broadcom and SK Hynix, both of which have had similar success.
Notice: Nearly 1 in 10 Organizations to Spend Over$ 25 Million on AI Initiatives in 2024, Searce Report Finds
Only a few main businesses, which have significant influence over the supply chain, have been able to realize these record profits. NVIDIA, an British firm, designs the majority of GPUs that are used to teach AI models. But, they are manufactured by Taiwan’s TSMC. The only two organizations that can produce the most cutting-edge bits on a massive scale are TSMC and Samsung Electronics, respectively.
However, the industry has n’t always been straightforward. A global chip shortage was sparked in first 2020 due to the COVID-19 crisis. Over the course of more than three times, supply issues between this relatively small number of businesses have had an impact on sectors like consumer electronics and AI.
The semiconductor supply chain was weak ground even before the pandemic, as a result of a number of activities, including trade war between the U.S. and China. S. and China, and Japan and Korea, impacting product sales and distribution. In addition, natural tragedies, such as a drought in Taiwan and three grow fires in Japan between 2019 and 2021, contributed to fresh stuff shortages.
“Extreme weather, natural disasters, geopolitical strife, a pandemic, and other major disruptions over the past decade have made abundantly clear how supply shocks can severely limit the industry ’s ability to meet demand, ” the Bain and Company report states.
A lack of AI independence may lead to an even worse device lack.
Not just a lack of production capacity could cause a second world chip shortage.
“Geopolitical conflicts, trade restrictions, and foreign tech firms ’ decoupling of their supply chains from China continue to pose serious risks to silicon offer. Delays in factory design, materials scarcity, and various unpredictable factors may also create squeeze items, ” the report says.
The U. S. , for instance, has applied chip-related trade settings on the price of transistors to China, as well as the Netherlands and Japan. The U. K. In 2023, companies that wanted to trade transistor technology to China also ceased to submit the majority of certificate applications.
In an effort to protect national security and pursuits, China’s Ministry of Commerce announced it would place export controls on chromium and germanium-related goods. ” These rare metals are essential in chip production, and China produces 98 % and 54 % of the world’s supply of gallium and of germanium respectively.
Governments around the world are also investing billions of dollars to increase their individual capacity for producing semiconductors, with the goal being to lower their rely on other nations as a main source of funding. But, data security also plays a piece; Officials can better defend against spy and cyberattacks by keeping the supply network within their territories.
In 2022, the U. S. passed the CHIPS Act, to give needed silicon research opportunities and developing opportunities as well as reinforce America’s business, national safety, and supply stores. In addition to funding the proof-of-concept for a shared regional AI research system, the White House has also released a blueprint for an AI Bill of Rights to aid in domestic AI regulation.
Intel, TSMC, Texas Instruments, and Samsung — the world’s largest storage intel — have all announced plans to build fabs in the U. S.
It was made known that the U.S. would be in the U.S. in August 2023. K. government will devote £100 million ($ 126 million ) to fostering AI hardware development and shoring up possible computer chip shortages. Amazon Web Services only this month made the announcement that it would invest £8 billion in data centers in the nation over the next five times.
Observe: UK Government Announces £32m for AI Projects After Scrapping Funding for Supercomputers
With the adoption of the European Chips Act in July 2023, the European Union provided €43 billion ($ 46 billion ) in subsidies to grow its semiconductor industry. The group also has the lofty goal of producing 20 % of electronics by 2030.
But Anne Hoecker, mind of Bain’s Global Technology exercise, said that the adventures for data sovereignty does become “time-consuming and very expensive. ”
Although less challenging in some ways than building silicon factories, these tasks demand more than just getting local grants, she stated in a media release. HPSCALERS and other major software companies perhaps continue to invest in regional AI operations that will give them significant competitive advantages. ”
The Bain record adds that little speech models with algorithms that use RAG, or retrieval-augmented technology, and vector embeddings, may stand to benefit from data sovereignty, as they handle a lot of the computing, networking, and storage tasks close to where Artificial data is stored.
Advice for executives in the supply chain for surviving a chip shortage
The Bain report makes some suggestions for businesses that use semiconductors to survive another global chip shortage:
- Forge a deep understanding of and track the entire AI supply chain, including data centre components, PCs and smartphones, and peripheral devices like routers and network equipment.
- Sign long-term purchase agreements to ensure that chips can be accessed despite potential disruption.
- To maximize compatibility with various suppliers and flexibility in sourcing, design products to use industry-standard semiconductors rather than application-specific chips.
- By diversifying suppliers and sourcing components from multiple regions, strengthen the supply chain against geopolitical uncertainties like tariffs or regulations.
According to the report’s authors,” Executives may still be weary of the semiconductor supply disruptions that were brought on by the pandemic, but there is no time to stop because the upcoming major supply shock is looming.” This time, however, the signs are clear, and the industry has a chance to prepare.
“The path forward demands vigilance, strategic foresight, and swift action to reinforce supply chains. Business leaders can ensure their resilience and success in an increasingly AI-enabled world by taking proactive measures. ”
