“The case for digital transformation has never been more urgent or clearer. Digital technology is a deflationary force in an inflationary economy.”
Satya Nadella
Key takeaways
- How can investors profit from the AI hype right now? We believe that investors should split their rationale into two parties; how to invest in and how to profit from that AI concept.
- With AI all about computer power, data, and storage which areas currently offer profitable exposure to AI right now?
- We have identified three areas where we believe investors can profit from AI today.
- Find out why we think ChatGPT's launch at the end of 2022 could be an "iPhone moment" for AI.
Throughout history, global economies have experienced hype cycles and bubbles: from the Dutch Tulip mania in the 17th century and the dot-com boom in the late 90’s to more contemporary topics like 3D printing and fake meats. Thus, investors making and losing paper fortunes on the back of a shiny new “game changing” idea which captures the market's imagination, is nothing new. In light of the fact that Artificial Intelligence ("AI") is currently on investors' minds, we argue that investors should think critically and objectively while attempting to profit from AI. In order to help investors navigate hype cycles, First Trust has devised a framework that poses the crucial question: are you aiming to invest in or profit from the current hype?
“The most advanced AIs require vast data, tremendous computer power and skilled technicians. Unsurprisingly, organizations with access to such resources, both commercial and governmental, drive much of the innovation in this new field.”
- Henry Kissinger, Eric Schmidt & Daniel Huttenlocher1
AI has arguably become the most significant and disruptive general-purpose technology since the iPhone, but AI dates back decades to Alan Turing’s “Turing Test” in the 1950’s (which is a test devised to determine whether a computer exhibits behaviour indistinguishable from that of a human). The release of ChatGPT at the end of 2022 might be an "iPhone moment" for the technology as suddenly, we're asking ChatGPT to write our reports and code our software; this technology reached over 100 million users in one week (it took Netflix 3.5 years)2, propelling it to the top of investors' minds. And, as this excitement surrounding AI grows, investors are ever eager to profit from this technology's transformative potential. However, to navigate and potentially profit from this AI hype cycle, we think investors should focus on two key and separate thought processes: identifying potential concepts worth investing in or figuring out how to profit from the current hype.
We explore three areas that we believe investors could potentially “profit from” AI, today.
Semiconductors (the backbone of AI)
AI is heavily reliant on these powerful processors to handle the massive volumes of data it requires to operate. At the heart of AI’s advancement are semiconductors, but only a select few designers and producers will profit from the rising demand for AI applications. The semiconductor value chain can be divided into four groups: foundry operators (those who manufacture the chips, such as Taiwan Semiconductor Manufacturing Company), integrated device manufacturers (run all processes in producing the chips from planning to producing the final product, such as Intel), chip designers (design and develop semiconductors, such as Nvidia), and equipment manufacturers (create the necessary machinery and equipment, such as ASML).
The cost curve for these chips has enabled machines to be built at a substantially lower price, and as AI evolves and permeates numerous industries, demand for these chips is expected to soar, presenting substantial growth potential. We believe semiconductors are the picks-and-shovels play for the AI theme, representing a potentially compelling way to capitalise on the AI “gold rush”, particularly when access to these sought-after chips is limited for many. Demand continues to outstrip supply as semiconductors and AI fuel each other's growth, ushering in an era of technological transformation, leading to some lofty valuations within this space, warranting caution. For those companies unable to access these chips they are forced to turn to the cloud.
Infrastructure (IaaS) & Platform as a Service (PaaS) Cloud Computing
In our opinion, there’s little doubt as to why cloud companies appear under every rock in the AI universe. The Cloud has democratised the computer, allowing computing power and storage to effectively become a digital utility. The cloud model eliminates traditional infrastructure costs by enabling any organisation (from start-ups to governments) to buy on an as-needed basis and easily scale up or down depending on consumption, user requirements, and growth. AI feeds off data, and the more it can access and the faster it can process it, the smarter it gets. Thus, due to the prohibitive cost of the required infrastructure, we would expect that almost every company utilising AI will turn, in some part, to the cloud. Cloud providers are involved at practically every level of development, profiting from the investments made in AI, whether for data storage, training, hosting, or deployment at scale, here’s why:
- Firstly, as referenced above, the Cloud solves the problem of access to computing power posed by the lack of available chips, like Nvidia’s H100. Herein lies the first surprising discovery; Nvidia themselves partner with cloud companies to use their own chips. Oracle just revealed that Nvidia use Oracle’s clusters for Nvidia AI infrastructure3. In fact, Nvidia has revealed multiple key AI partnerships with cloud names like data giant Snowflake, ServiceNow and Adobe.
- Secondly, platform and infrastructure as a service solution offer the computational tools and resources required for the development and deployment of AI, giving investors a potentially already profitable investment. For example, Microsoft provides full stack cloud infrastructure purpose-built for AI. Azure AI’s infrastructure offers the performance, scalability, and built-in security required to create, train, and deploy AI workloads. While not identifying AI as a separate business segment, revenue third-quarter fiscal 2023 in Microsoft’s Intelligent Cloud business, which includes Azure, grew 27% year-over-year4. Google meanwhile have investing heavily in chatbot systems Bard and DeepMind Sparrow. Google’s cloud business, for the quarter ended March 2023, grew by an impressive 28% year over year4. Further, there are a litany of cloud companies that offer specific services tailor-made for AI, like Arista Networks' data-driven networks and platform solutions and MongoDB's Atlas Platform, built to optimize the scaling of client's databases.
- Thirdly, perhaps the most exhilarating segment of the AI hype: Large-Language Models (LLMs) and Generative AI. Surprisingly, numerous companies have emerged and are capitalising on the tremendous potential of these two fields. Through countless research papers and articles, we discovered that companies like Adobe, Twilio, MongoDB, and even Amazon are reaping the rewards of the significant investments flooding into Generative AI and LLMs. While the spotlight has often shone on Google and Microsoft, Amazon’s AWS, is also at the forefront of LLM development and has made substantial investments in Generative AI. Notably, during Oracle's Q4 earnings call, Larry Ellison, CTO and Founder, revealed that “Oracle Gen 2 Cloud has become the leading choice for running Generative AI Workloads, citing its collaboration with Nvidia and its investment in the Generative AI Startup Cohere”. Ellison further disclosed that due to these developments, "cutting-edge companies" engaged in LLM development, such as Mosaic ML, Adept AI, Cohere, and over 30 other AI development firms, have recently agreed to purchase more than $2 billion worth of capacity in Oracle's Gen2 Cloud”.5 This is why we are spotlighting that a wide range of cloud companies are profiting now, already, from the hype.
Data Centres (the unsung heroes)
As the demand for digital IT infrastructure increases, hyperscalers and edge computing are poised to play an increasingly important role in the deployment of AI. Generative AI, the latest catalyst for this technology, is expected to accelerate demand for computing power in data centres, with an estimated $343.6 billion market size by 20306. As we highlight above, even the most basic levels of storage and computing power can be leased via data centres. Perhaps that’s why Nasdaq call them the “unsung hero of the AI boom”7.
“The data-centre leasing companies like Equinix represent a different facet of the AI market - the physical housing of AI systems. Moreover, Equinix's new GPU "as a service" offering has further diversified its portfolio, positioning it as a potential rival to cloud computing behemoths such as Amazon, Alphabet, and Microsoft.”
- Nasdaq article
Holding these "unsung heroes" has the added advantage of allowing investors to potentially profit from the AI wave because many data centres (such as Equinix and Digital Realty Trust) are set up as REITS. The latter pays a substantial dividend of about 4.5%⁸.
How can investors potentially profit today?
The explosive rate at which generative AI and Large Language Models are evolving and becoming ingrained in people's lives may truly represent an "iPhone moment" for artificial intelligence, in our opinion. Computer power, data, and storage are all crucial to AI, so we have highlighted three areas where we believe investors can find companies that are profiting from AI today, in contrast to investing in some expected future outcome like self-made television shows, video games etc, which may or may not yield solid investment returns. Hopefully, this will help cut through the hype.
With the wider benefits of AI likely to be felt in years rather than months, we believe that thematic ETFs offer access to a wider array of companies connected by a theme, rather than a country or sector while also doing away with the risk of selecting the wrong individual stocks. We believe the First Trust Cloud Computing UCITS ETF (“FSKY”) offers a way for investors to gain exposure to this theme, benefitting from the immediate demand for computer power and storage while continuing to provide exposure to several other themes (such as broader digitalisation) and potentially benefitting from AI growth in the years ahead. For similar reasons, we also believe the First Trust Alerian Disruptive Technology Real Estate UCITS ETF ("DTRE") offers a compelling option for those looking to capture the increasing pivotal role datacentres are playing in the AI ecosystem.
We will be following this article up with a review of industries that have already been using AI for core products and services for years. Spoiler alert: Energy efficiency and Cybersecurity companies to name but a few.
Share