AI — Beyond the Investor Buzz
Our five mega forces —
digital disruption and AI
geopolitical fragmentation
economic competition
demographic divergence
the future of finance
transition to a low-carbon economy — is affecting investing now and in the future. The key to harnessing mega forces and their potential is to first identify the catalysts that can supercharge them and how they interact with each other. And then take a view of what markets are pricing in.
Our early research suggests a positive correlation between an uptick in patents and broad earnings growth in the short term. Moreover, investors are increasingly attributing economic value to these patents, indicating a market attempt to assess the potential of innovation.
However, the significance extends beyond patents. Success in this realm will likely depend on attracting top AI talent, substantial investments to support their aspirations, and acquiring the necessary computing power to drive their innovations. Identifying companies excelling in these areas may reveal leaders and laggards.
Revenue exposure and anticipation of product placement in the tech stack will also play pivotal roles. The digitalization surge across sectors is generating massive data, necessitating preparation not only for utilizing it with AI but also for its secure storage and processing.
The most significant investment impact may stem from AI’s integration with other technologies and mega-forces—a focus of future research. We anticipate substantial investment in developing the infrastructure required to leverage AI fully.
Despite the AI revolution underway, there are limits to its adoption, and not all potential outcomes are positive.
Cybersecurity risks loom large, especially with generative AI algorithms facing reliability issues. Governments worldwide are grappling with these risks while attempting to shape AI business conduct, potentially influencing adoption rates in certain sectors.
We perceive digital disruption and artificial intelligence (AI) as key components of five mega-forces shaping the landscape with increased macroeconomic and market volatility. Tracking these risks and opportunities will be crucial for investment decisions.
To leverage these mega-forces effectively, it’s imperative to identify the catalysts that can propel them and understand their interplay.
The subsequent step involves uncovering potential beneficiaries not yet priced in by markets.
AI is already reshaping markets, with companies rushing to adopt the technology and investors deciphering between hype and reality. While large language models (LLMs) have proliferated, progress into other applications necessitates time, new skills, and vast data sets. The overarching race to develop the smartest machines resembles a revolution akin to past industrial and information revolutions.
As innovation accelerates, advances are expected to be exponential, as evident with each LLM iteration. The technology “stack,” comprising layers of technology facilitating further innovation, may provide a roadmap for assessing investment opportunities. The base includes computing essentials like processing chips and semiconductors, while powerful applications enabling AI utilization sit atop.
We observe the entire tech industry, led by a few major firms, pivoting their focus to AI.
Assessing the innovation underway poses a challenge. While AI patents offer some insight, they don’t capture the entire story, given significant AI development occurs in open-source environments where code inspection, modification, and enhancement are possible.
Generative AI: A Platform Perspective
Assessing the investment implications of generative AI demands viewing the technology through a platform lens, akin to the transformative impact of technologies like 4G. Just as 4G laid the groundwork for app-based economies, we foresee a stack of generative AI technologies emerging, poised to reshape various industries.
The tech industry as a whole is undergoing a seismic shift towards AI-driven business models, triggering an arms race led by major players. Chipmakers are positioned as immediate beneficiaries, while software advancements in translation, summarization, and text analysis hint at automation’s potential. Yet, the true innovation lies in the evolution of AI capabilities, particularly in processing visual data.
Although AI promises productivity gains, its synergy with other technologies and mega-forces presents the most compelling investment opportunities. While outcomes remain uncertain, we anticipate a global, multi-sector AI investment cycle supporting revenues and margins. In the short term, we are bullish on the AI theme in developed market equities.
Our analysis of five mega-forces, including digital disruption and AI, underscores the need to identify catalysts propelling them and their interplay. Understanding what markets price in is crucial to capitalizing on these forces.
AI isn’t novel but has undergone a resurgence, catalyzed by recent computing advancements and deep learning innovations, epitomized by OpenAI’s ChatGPT. This sophisticated language model represents a paradigm shift, as evidenced by the exponential surge in parameters from GPT-1 to GPT-4 over five years. We anticipate this intelligence revolution to yield profound, unpredictable implications beyond near-term productivity gains.
Advancing Through the Layers: AI Stack
It’s crucial to delineate between two primary LLM data applications. First, leveraging vast datasets to train new LLMs, exemplified by OpenAI’s ChatGPT. Second, utilizing existing LLMs to extract value from existing text resources. Companies adept at monetizing proprietary data may offer enticing investment prospects. Adjacent businesses involved in storing and managing data-feeding AI models may also present enduring, high-quality opportunities.
AI, like any technology, faces adoption limits. Cybersecurity risks are prevalent, paving the way for consultancies and startups specializing in secure AI environments. Generative AI confronts reliability issues, yet future iterations are expected to improve progressively. Governments worldwide are striving to address risks and shape AI business conduct, fostering opportunities despite existing limitations.
While automation can enhance worker productivity, the interplay between AI and other key tech innovations, spanning connectivity, security, and physical automation, extends its impact. Companies securing top talent and investing in scaling computational power stand poised to lead the AI race.
Navigating investment opportunities across geographies and sectors entails high uncertainty. Our technology stack roadmap outlines potential pathways, with cloud infrastructure and chips serving as foundational building blocks, followed by models, data, and data infrastructure. The evolution from the first to the second layer is underway, with the application layer likely to follow suit.
AI’s influence spans multiple domains, intersecting with mega forces like aging populations and geopolitical competition. Generative AI, proficient in creating new content from vast data, could accelerate this trend. Market enthusiasm surrounding AI has primarily focused on chipmakers, while the significance of harnessing data remains underappreciated. Demand for digital infrastructure, including hardware and server farm locations, is projected to outstrip supply, offering investment potential.
The AI Tech Stack
Each layer builds upon the preceding one, facilitating further innovation. As the AI ecosystem evolves, some categories may give way to newer ones:
Apps Software
Data & Data Infrastructure
Foundation Models
Cloud Infrastructure
Semis and Hardware
Sitting at the top of the stack are the apps, where AI applications could be open to thousands of companies. The data infrastructure software layer involves the storage, management and manipulation of the huge data sets being used. The data used to train and feed AI applications — both public and proprietary. The proprietary or open-source large language models are being built—think Open AI’s ChatGPT for example.
The infrastructure needed to train AI models on large clusters of high-performance machines, such as cloud computing providers and Hardware manufacturers that make the chips and semiconductors — are the building blocks of computational power.
Quantifying Innovation Impact
Our research underscores the positive correlation between an uptick in AI patents and near-term earnings growth in the broader equity market, factoring out other variables like GDP. An analysis spanning 2000–2020 reveals that a one standard deviation surge in patents above their trend corresponds to a 0.3 standard deviation increase in earnings growth over the subsequent two years, as indicated by the yellow bar in the chart. The IT sector experienced the most substantial impact, followed by healthcare and financials, showcasing AI’s intersection with sectors propelled by other macro forces, such as ageing populations and financial evolution.
Yet, patent volume alone doesn’t illuminate their economic worth. Not all patents translate into profitable ventures, and the future viability of a new concept remains highly uncertain. Hence, we rely on a proxy — the market’s response to patent grants. A significant uptick in the market’s valuation of AI patents over the past decade indicates investors’ keenness to seize nascent ideas with potential for wider adoption.
Identifying frontrunners and stragglers proves challenging, mainly due to the scarcity of directly measurable innovation metrics. AI patents provide valuable insights, although with caveats. Given the open-source nature of much AI development, collaboration and idea expansion thrive, unlike in closed systems. Nonetheless, the surge in AI-related patents suggests a potential — albeit imperfect — proxy capturing innovative entities.
The U.S. Trademark and Patents Office (USPTO) classifies a patent as “AI” if it pertains to machine learning, natural language processing, vision, and other components. Tech dominates AI patents, comprising around 65% of total grants over the past decade, with communications, industrials, and consumer discretionary sectors following suit.
But the number of patents on its own says little about their economic value. Not all patents lead to profitable enterprises, and the future value of a new idea is highly uncertain. Therefore, we turn to a proxy. In our view, the public market reaction to the grant of patents is one useful gauge. The rising value markets ascribe to AI patents hints at investors’ eagerness to capitalize on nascent ideas that may potentially see broader adoption in the future.
