Investing in AI & Cybersecurity
Key takeaways
Generative AI has the potential to be as revolutionary a technology as the printing press, democratizing information, idea communication and creativity.
The efficiencies that AI can bring to many basic business, legal and medical services will become readily apparent sooner than many expect.
Outside of technology, financials and fintech are likely to be among the most impacted economic sectors, with natural resources and climate tech the least impacted.
Generative AI also brings risks associated with IP infringement, misinformation, reinforcement of biases and tools that can increase the severity of cybercrime.
A new age of human progress has dawned. Generative artificial intelligence (AI) – systems that create content such as images, text and audio with human-like results – is a game changing technology. Indeed, we believe it could prove to be as revolutionary a technology as the printing press, allowing people everywhere to enhance their creativity and share knowledge and ideas more widely and cheaply.
The potential of generative AI to advance wellbeing and drive growth is tremendous. For example, new cures for afflictions such as cancer and Alzheimer’s disease may become more probable. Likewise, research to speed up the delivery of low-cost clean energy and the development of climate change–resistant crops may accelerate. For the global economy, this technology could transform the nature of work across industries and provide a much-required boost to long-sluggish productivity growth.
But generative AI also comes with risks. Skewed or inaccurate data can lead to results that reinforce gender, racial and many other biases. Gains may accrue disproportionately to select companies at the expense of those without access to similar capabilities. The production of incorrect materials and falsehoods could harm trust in generative AI itself, as can the issue of AI’s impact upon intellectual property rights.
Here, we examine generative AI’s potential disruptive impact across broad economic supersectors.
How generative AI may reshape technology
The first wave of potential opportunities centers on the technology value stack. This supersector can be divided into a “value stack” based on impacts to different technologies. The foundational layer of the stack is “silicon.”
Generative AI requires massive amounts of data processing power. It also needs significant memory (i.e., temporary storage as data is transferred). We therefore see generative AI driving significant demand for multiple kinds of specialized chips, such as graphic processing units (GPUs) and application-specific integrated circuits (ASICs), as well as for memory technology. Producing such technologies requires specialist equipment and materials such as silicon wafers.
A vast amount of data infrastructure will also be built to support generative AI. Hyperscalers, essentially large cloud service providers, and other digital infrastructure specialists will supply much of this. Examples include data center–oriented firms in the US and telecom operators in Asia. Models and machine learning operations cover critical activities such as the development and training of generative AI models. Smaller, more efficient models trained on proprietary data and on-device AI – where activity takes place on specific devices such as users’ handsets rather than on cloud computers – are key trends to watch here.
Generative AI is set to impact all types of software and applications. Front office software – including customer relationship management (CRM) and marketing automation and could become ever more insightful and efficient. The same is true of back-office software covering the likes of recruitment, talent management and other HR processes. And while cyber criminals are already seeking to harness generative AI for nefarious purposes, cybersecurity firms are deploying it in the fightback.
Lastly, we expect an AI-driven step forward in technology services. This includes the likes of consulting/advisory, third-party software development and business process optimization.
Generative AI and the rest of the economy
Dividing the economy into seven supersectors, Citi Research analysts have considered potential opportunities and challenges from generative AI. Excluding tech – the master sector – financials and fintech are seen as the most likely to be impacted, and natural resources and climate tech the least impacted.
Financials and fintech
Generative AI could potentially help democratize investing and markets, improve algorithmic trading, enable more robust data utilization and allow for better analysis and pricing of insurance risk. Other opportunities include 24/7 improved customer service via chatbots and digital assistants, better fraud detection and prevention and enhanced automated compliance monitoring. Risks to this super sector include generative AI enabling fraud; reputational and regulatory risks from mis-executed AI strategies; AI involvement in market manipulation; malfunctioning AI; and failure of investment managers to synthesize AI properly into their processes.
Consumer
Consumer staples companies – such as food, beverages, household and personal products makers – are exploring and deploying AI tools across their operations. This could mean faster innovation and more efficient data use in manufacturing, supply chain and marketing activities. Enhanced sustainability by reducing waste and energy consumption is a further avenue. Traditional luxury is adapting more slowly, by contrast. Jewelry firms see potential in design and modeling but not yet in cutting and polishing. Consumer retail is exploring the use of generative AI for handling customer inquiries, finding in-store items and supply chain management. The fight against counterfeiting could also receive a boost, particularly in high-end spirits and cosmetics.
Healthcare
AI is likely to have profound impacts in drug discovery and design. It will also create advances in patient selection and recruitment for clinical trials and optimization of sampling/sales calls. Generative AI can also help in communicating with patients, such as reminding them about healthy habits as well as streamlining physician tasks including note taking/data input into digital health records. Improvements could also come in patient risk coding, payment integrity and ensuring that treatments are cost efficient and necessary. Adoption of this technology involves the establishment of systems to protect patient, doctor and hospital privacy, among other data concerns.
Industrial tech and mobility
Industrial processes are increasingly digitized. Internet-of-things (IoT) platforms are allowing end-users to create industry-specific and process-specific apps. Generative AI could significantly broaden the ability to create code, massively opening up the market for analyzing data on industrial IoT platforms. In autos and mobility, AI has many uses in autonomous vehicles and may drive product development, manufacturing and customer-facing services. Smarter decision making in defense systems could also emerge.
Real estate
The real estate sector is generally less exposed to generative AI than other sectors. At least for the first few years of the adoption cycle, data center demand is set to strengthen. AI could have a role to play in smart buildings, assisting in the automation of temperature, lighting, alarms, leak sensors and security. Chatbots could enhance consumer-facing applications at hotel and residential properties, meanwhile. Office space could receive a potential boost from AI company demand in the near term but may suffer in the event of mass AI job displacement further out.
Natural resources and climate tech
Workforce productivity gains from AI are initially less likely for this supersector, given its heavy industry skew and relative lack of customer-facing roles. However, opportunities may arise alongside the likes of greater electrification and interconnectivity, along with more smart appliances and more distributed energy (e.g., rooftop solar). AI may be used to analyze the resulting increase in data produced to manage demand and plant usage, leading to more efficient energy use. Less wastage is a risk of sorts for the industry as it results in lower demand.
AI’s vast impact
There appears to be no area of the economy that will be untouched by AI or the applications it may enable. Productivity is likely to be enhanced across many industries and embedding AI capabilities in existing processes, like software, will make its value readily apparent to users that used to rely on others to complete tasks.
AI-propelled digitization in 2024: five areas of focus
The most obvious beneficiaries of the generative AI revolution have already seen an expanded growth in their market cap in 2023, but there are potential opportunities as we enter the next stages of the buildout.
It is, most certainly, not too late for investors to participate in the exponential growth of artificial intelligence (AI) technology. While mega-cap tech leaders will continue to provide reliable exposure to the AI trend, we see areas like semiconductor equipment, robotics, drug discovery and cybersecurity as clear beneficiaries from the coming integration of AI into everyday business and personal lives.
1. Mega-cap AI leaders
The basic theory behind AI models has been around for decades. The key to building human-like capabilities, for example writing code or even drafting this very article, is the development of a generative AI model that has access to lots of data and sufficient processing power. That’s how the mega-cap tech leaders have built their AI edge – they’ve spent decades aggregating data and designing superfast chips that stood to benefit first from the AI revolution. From this perspective, the big tech rally in 2023 is logical.
We see the rise of the internet as a useful parallel for what’s to come for AI. Initially, in the early and mid-1990s, telecommunications firms that already controlled phone lines and were laying the most fiber saw their shares surge as excitement around the internet’s promise attracted significant investment. But as internet access became ubiquitous, benefits grew more diffuse as nearly every individual and company saw productivity gains from the ability to use email or create a website to improve their business or quality of life.
2024 is likely to be a year of AI capacity buildout. This dynamic will play out in the aftermath of a period of restrictive financial conditions following the US Federal Reserve’s (Fed’s) two-year inflation fight. This means that scale and free cash flow will initially favor larger companies. That said, we see opportunities beyond the “Magnificent 7” 1 as ways to play AI over the long run. But we still think portfolios should have exposure to these companies in the year to come given their scale and incumbent status.
2. AI infrastructure (chipmakers, semiconductor equipment, data centers)
As demand for AI grows exponentially, we expect key suppliers along the global semiconductor value chain to be big 2024 beneficiaries.
Today, chips required for AI computing make up just a small fraction of overall semiconductor manufacturing (Figure 1). But the rise of generative AI applications has seen demand shift from general-purpose computing to high-performance chips.
Semiconductor equipment providers and fab facilities with the capacity and know-how to produce cutting-edge chips hold near-monopoly positions in their areas of expertise. Leading players in the US, Japan, Taiwan, South Korea and the Netherlands have dominated the market share for semiconductor fabrication, capital equipment and services that enable advanced chipmaking. For instance, one Dutch company is the sole provider of the lithography equipment needed to imprint silicon with the nanoparticle-sized circuitry. Taiwan’s largest fab facility is aiming to at least double its capacity for AI-related chipmaking processes in 2024.
Alternatives to this equipment and manufacturing know-how are hard to come by (though under pressure from Western import restrictions, China is trying), and there are significant barriers for potential new entrants. As the semi cycle turns upward in 2024, we see an opportunity for semiconductor manufacturer and equipment shares to catch up to the GPU design firms that led the market last year.
Figure 1: For all the promise, when it comes to chipmaking capacity, AI has a lot of catching up to do.
Source: Citi Research, Gartner, as of November 17, 2023. All forecasts are expressions of opinion and are subject to change without notice and are not intended to be a guarantee of future events. Indices are unmanaged. An investor cannot invest directly in an index. They are shown for illustrative purposes only and do not represent the performance of any specific investment. Index returns do not include any expenses, fees or sales charges, which would lower performance. For illustrative purposes only. Past performance is no guarantee of future results. Real results may vary.
3. AI-equipped robotics and automation
The same underlying technology used in generative AI – neural networks – is also the engine that will power more human-like robots and other forms of automation. When combined with long-term trends of evolving supply chains and automobile technology, we see the stars aligning for investment in this space.
As we discuss elsewhere in this report, 2023 saw a boom in US manufacturing construction as the US aims to shift production of high-end semiconductors and cutting-edge batteries closer to home. The buildout of warehouse capacity to support ever-growing e-commerce continues apace. While manufacturing square footage is rising, we doubt that every overseas factory job will be replaced one-for-one with American workers in an already-tight labor market. More likely, firms will continue to invest in automation and robotics to move packages and operate assembly lines, to the immediate benefit of the leading industrial automation and supply chain management companies.
Advanced driver assistance systems (ADAS) are another form of automation reliant on AI technology. These systems take in data from sensors, cameras and ultrasound devices positioned at every angle of a vehicle to gain 360-degree vision to assist and eventually replace human drivers. Creating human-like vision for a fully autonomous electric vehicle may have proved more elusive than originally anticipated (probably not available at scale now at least until 2035, according to the latest projections),2 but it will continue to arrive in stages. AI will be the key to powering each advancement.
4. AI-augmented drug discovery
Drug discovery and design is another field likely to benefit from significant AI disruption. In the same way that the latest chatbots can understand natural language, neural networks can also be tuned to interpret the language of the human body: DNA, RNA and the structure of proteins and drug molecules. Drug discovery today is often a game of trial and error. The world’s most highly sophisticated scientists pool their collective expertise and effectively make educated guesses on how certain combinations of proteins may affect the human body. Those tests then go through clinical trials, which can take over a decade to complete even before FDA review. According to Citi Research3, 92% of these candidate drugs fail.
We believe AI will be able to dramatically reduce the failure rate of clinical trials. Given the amount of time and resources required to get a drug tested, approved and marketed to the public, a higher success rate for new drugs will have a dramatic effect on both the cost of drugs and drugmaker profitability.
5. Cybersecurity – AI boosts demand due to threats from AI
Cybersecurity remains the No. 1 investment priority for chief technology officers at large firms. While AI will surely make us all more productive, it will inevitably become another tool for malicious actors to infiltrate networks and devices. Bad actors – even unsophisticated ones – can leverage the latest AI chatbots to help write the code needed to perform a cyberattack. Professional hackers can also leverage AI for pattern recognition and decryption.
But AI can also be deployed to improve cyber threat detection and prevention. Industry-leading cybersecurity firms are more likely to develop and deploy cutting-edge solutions relative to in-house security operations.
Download your copy
Discover our Wealth Outlook 2024 insights and themes.
