Source: AFP
The race for AI supremacy is on
The breakneck speed of AI development has captured both the attention and imagination of investors globally. From an obscure piece of technology, AI is making significant inroads into businesses and the everyday life of consumers.
The level of investment dedicated to AI infrastructure at this early stage is significant and is only set to increase exponentially in the near term.
The IDC forecasts GenAI spending to increase at a 73% CAGR over the next four years. This would bring global GenAI core IT spending from USD16b in 2023 to an estimated USD143b by 2027.
This is unsurprisingly as hyperscalers are spending significant CAPEX on silicon and hardware. 3rd party estimates put cloud AI compute revenue (including Graphics Processing Unit (GPU)/AI accelerators, Central Processing Units (CPU), High Bandwidth Memory (HBM) and Dynamic Random-Access Memory (DRAM) for AI servers) at ~28% of 2024E global semis revenue, and for this figure to grow to ~35% of total semis revenue by 2026E. This figure could be even higher if we account for derivative demand from Networking, Storage (SSD) and semis content from Edge AI devices.
Thus, we do not see growth in AI semiconductors plateauing just yet. In fact, on the back of relentless computational demand, technology giants have unveiled their GPU roadmaps for the next few years, which currently point to a very speedy product development calendar ahead. This would also have broader implications for the wider semiconductor and hardware complex, such as the accompanying demand for more complex HBM and cooling solutions.
We also see notable demand for more AI semiconductor alternatives in the form of custom solutions such as application-specific integrated circuit (ASIC) chips.
In fact, AI ASICs are projected to grow at a rate that is faster than that of the total AI semiconductor market over the next few years, thus increasing its share over time. This is likely due to its ability to position itself as a strong alternative for compute accelerators in comparison to off-the-shelf merchant solutions as hyperscalers strive for greater differentiation, better performance, lower power consumption and lower overall silicon costs.
While direct AI ASICs players could benefit from the continued demand from Cloud Service Providers, valuations appear rather stretched at this moment given the run-up in share prices. However, we see other ways to play the growing ASICs market, which will also broadly benefit from the existing GPU buildout.
First, chip design activity in AI/high performance compute should remain elevated due to the efforts of cloud/hyperscalers, who are actively deploying their custom ASICs solutions. This should be beneficial for electronic design automation (EDA) software/intellectual property firms.
Second, leading foundries should benefit regardless of whether ASICs or GPUs turn out to be more popular, and competition will likely further drive AI chip cadence and enhance specifications, thus leading to more demand for leading edge production.
Third, the growing ASICs market should see greater demand for HBM.
Fourth, testing solutions are likely to become increasingly critical.
At the same time, larger CAPEX levels to support AI should also be constructive for the longer-term opportunities for the semiconductor capital equipment industry as well.
Consumer adoption rates rising
Enterprises increasingly see GenAI and LLMs as having a direct impact to their investment priorities, which is certainly encouraging. But apart from enterprise use cases, what has been striking is the adoption of chatbots by consumers. Unsurprisingly, younger cohorts are driving adoption rates.
Use cases include price comparisons and recipe research, with the latter indicating potential runway to further monetise online grocery shopping.
GenAI chatbots have significant network effects that can help them to scale. However, we will remain watchful for further innovation to exponentially increase their utility for users so as to increase retention and build further traction.
AI industry collaborations and regulations
We have seen recent agreements between the US AI Safety Institute (AISI) and leading AI companies mark a notable development in the constantly evolving landscape of federal AI model regulation and oversight. This also marks a significant step in the US government’s efforts to engage with the AI industry on issues around safety and ethics. These agreements empower the AISI to access major new AI models before/after their public release.
In our view, this unprecedented collaboration between government agencies and private AI developers could have wide-ranging implications for the AI industry, regulatory frameworks and market dynamics. At the heart of these agreements is joint research aimed at assessing the capabilities and safety risks associated with new AI models.
From an investor’s perspective, companies that participate in such collaborations may be deemed favourably as they showcase their dedication to safety and transparency, which in turn, could help engender investor confidence.
Important information
This product may only be offered: (i) in Hong Kong, to qualified Private Banking Customers and Professional Investors (as defined under the Securities and Futures Ordinance); and (ii) in Singapore, to Accredited Investors (as defined under the Securities and Futures Act) and (iii) in the Dubai International Financial Center to Professional Clients (as defined under the Dubai Financial Services Authority rules) only. No other person should act on the contents of this document.This product may involve derivatives. Do NOT invest in it unless you fully understand and are willing to assume the risks associated with it. If you have any doubt, you should seek independent professional financial, tax and/or legal advice as you deem necessary.
Please carefully read and make sure that you understand all Risk Disclosures, Selling Restrictions, and Disclaimers. This document must be read together with the relevant Prospectus & Offering Documents &/or Key Fact Statement.
Disclaimer
The Bank, its Affiliates and their respective employees are not in the business of providing, and do not provide, tax, accounting or legal advice to any clients. The material contained herein is prepared for informational purposes and is not intended or written to be used, and cannot be used or relied upon for tax, accounting or legal advice. Any such client is responsible for consulting his/her own independent advisor as to the tax, accounting and legal consequences associated with his/her investments/transactions based on the client’s particular circumstances.
This document and other related documents have not been reviewed by, registered or lodged as a prospectus, information memorandum or profile statement with the Monetary Authority of Singapore nor any regulator in Hong Kong or elsewhere.
This document may not be published, circulated, reproduced or distributed in whole or in part to any other person without the Bank’s prior written consent. This document is not intended for distribution to, publication or use by any person in any jurisdiction outside Singapore, Hong Kong, or such other jurisdiction as the Bank may determine in its absolute discretion, where such distribution, publication or use would be contrary to applicable law or would subject the Bank and its related corporations, connected persons, associated persons and/or affiliates (collectively, “Affiliates”) to any registration, licensing or other requirements within such jurisdiction.
