Opportunity in innovation: H2 2023 Technology Outlook

Opportunity in innovation

As we approach the second half of 2023, we believe three themes are likely to drive the technology sector forward: the ascent of generative artificial intelligence, a shift from growth-at-all-costs to operational efficiency, and the increasing global emphasis on localization & reshoring of pivotal technologies.

First half 2023 review

Looking back on the first half of 2023, we see that the year-to-date performance in technology equities has been a tale of concentration, with investors overwhelmingly favouring mega-cap technology stocks over their smaller peers.

The technology equity sector as a whole has delivered strong performance year-to-date, with the tech-heavy Nasdaq composite up 25%¹. However, the rally has been extremely narrow, with mega-cap tech clearly outperforming the rest of the market. The top 3 constituents by market cap (Apple, Microsoft, and Nvidia*) in the technology sector, as represented by the MSCI ACWI Information Technology Index, have contributed over 59% of year-to-date index performance¹. When considering mega-cap technology names outside of the GICS Sector (Meta, Alphabet, Amazon, and Tesla), the dispersion is even wider.

YTD return of technology companies¹
by market cap (median, USD)

Why have mega-cap tech so clearly outperformed in the first half of 2023?

We believe there are two core reasons:

1. Investors continue to favour size and quality in an uncertain economic environment. The largest and most profitable companies in the world are mega-cap technology stocks².
2. Generative artificial intelligence (AI) is emerging as a transformational theme. Most mega-cap tech stocks stand to be initial beneficiaries of the AI arms race.

Generative AI: a watershed moment

The technology sector has been at the forefront of profound innovations that have reshaped modern society. In the 1980s, personal computers began to make their way into households. People started to use these devices to connect to the internet in the 1990s. Internet usage went mobile in the 2000s with the advent of smartphones. By the 2010s, cloud computing began to magnify the capabilities of each of these prior innovations and fueled an exponential growth in the amount of data that exists in the world. These transformations have not only revolutionized the world but also unlocked trillions in value, catapulting the technology sector to become the largest in the economy.

Today, we believe that we are on the precipice of the next transformative platform in technology: generative AI. Advancements in compute power, big data, and new methods in machine learning have led to an AI breakthrough and an inflection point in technological evolution.

AI isn't a new notion. It has long been a fixture in the realm of science fiction, conjuring visions of autonomous systems equipped with human-like intelligence.

As a scientific domain, the origins of AI date to 1950 when computer science pioneer Alan Turing published the paper ‘Computer Machinery and Intelligence’ in which he explored the question of whether machines can exhibit intelligent behavior³. In the decades that followed, the pursuit of AI faced significant challenges due to limitations in computer technology.

In 2017, a paradigm shift arrived in the form of a landmark publication by Google researchers titled “Attention is All You Need.” This paper introduced the world to the Transformer model, whose distinguishing feature, the attention mechanism, allows an AI system to selectively focus on various portions of the input data sequentially⁴. This revolution accelerated the AI research community’s learning curve and amplified the ability for AI models to execute complex tasks.

Soon after, the emergence of Transformer-based large language models (LLMs) capable of learning from extensive datasets and producing novel output, sparked a major shift in the AI community. These generative AI models showcased superior efficacy over their statistical counterparts in a variety of applications, but most notably in comprehending, summarizing, and generating human-like text. Rather than just predicting a user’s next word or datapoint, generative AI can create new responses in the form of text, code, images, or other modalities. In the past year, the world has witnessed remarkable advancements in the size and capabilities of generative AI models.

The capabilities of 2023’s most advanced generative AI models are pushing boundaries unimaginable even less than a year ago. OpenAI's GPT-4 (released in March 2023), not only passed the Uniform Bar Exam, but scored in the 90th percentile⁵. Given such, 44% of professionals in the legal industry are exposed to AI automation, according to a Goldman Sachs analysis⁶. These pioneering strides in generative AI carry the potential to significantly reshape not only the technology industry but also the broader global economy, with knowledge workers facing a greater risk of disruption than workers in physically-intensive professions.

Exam Results⁵
estimated percentile lower bound
(amongst test takers)

Generative AI has clearly become a focal point of fascination across the world. There is little doubt that it will reshape the global economy in the years ahead. According to a 2023 survey by KPMG, 77% of global executives believe that generative AI will have a bigger impact on broader society in the next three to five years than any other emerging technology⁷. Yet, we are still in the early days of this theme. While 58% of US adults had heard of ChatGPT by May 2023, only 14% had tried to use it⁸.

We believe generative AI is set to follow in the footsteps of PCs, the internet, and smartphones, becoming a deeply ingrained part of daily life in the coming years. The biggest initial beneficiaries of today’s AI eruption have been mega-cap firms that offer the silicon and supercomputers required to train and run generative AI models. However, we believe we’re in the early stages of a multi-year secular theme with wide potential for value creation beyond just the mega-cap tech companies providing AI’s infrastructure layers. Among companies exposed to AI, we expect to see a broadening of performance going forward. It is as critical as ever for management teams to have sound AI strategies. Winners and loser will emerge as they navigate this evolving landscape. Companies that are able to effectively incorporate AI into their operational and strategic frameworks stand to gain significant advantages over those who fail to adapt in this rapidly changing environment.

Deconstructing the Generative AI technology stack

In technology, a set of software and hardware components layered together that work in unison to execute specific tasks or functions is known as a stack. Generative AI requires a new stack. To fully harness the power of generative AI technology, the world will need a profound overhaul of computing infrastructure, innovative methods for data handling, and a new suite of applications. 

At the base, semiconductors and supercomputers make up the physical infrastructure required for generative AI training and inference. Training is how models learn from massive datasets. Inference is when models are used to generate output. Currently, the focus of spending is on building the infrastructure layer that enables training, which requires a significant amount of graphic processing units (GPUs) run, at scale, in data centres equipped with supercomputers. As such, the early beneficiaries have been a handful of mega-cap tech stocks that dominate the GPU and cloud services markets. As the use of generative AI grows and enterprises begin to implement the technology into their workflows, we believe more focus will be placed on the infrastructure that enables inference. Relative to the market for the chips and hardware that enable training, the inference market is more complicated and more competitive. On the whole, the growth in spending on the physical infrastructure required by generative AI models is substantial. The proportion of data centre server spending on dedicated AI servers will grow from 20% today to 45% over the next five years, predicts market research provider Dell’Oro⁹.

On top of the physical infrastructure are the foundation models themselves. There is currently a race to build the largest, best-performing, and most cost-efficient models. Developers of closed-sourced models are competing against each other and against the open-sourced community. There also vertical generative AI models, trained on datasets specific to an industry. The pace of development in the first half of 2023 was remarkable. Going forward, we will likely see further advancements and new custom-tailored AI models trained on unique datasets. 

To manage the vast and intricate data necessary for generative AI, the need for tools such as platforms for hosting and sharing open-source models, as well as software for data labeling and fine-tuning, will be indispensable. Proprietary data access will be a key factor for competitiveness in model development and corporate AI application.

Finally, at the top of the stack is the application level, where a multitude of start-ups are endeavoring to develop the next breakthrough application. These future applications will be more sophisticated than current iterations, offering valuable tools for professionals across all industries of the economy.

The Generative AI Technology Stack¹⁰

Investing in the AI landscape: opportunities and risks

The dawn of Generative AI introduces a new set of opportunities and risks for investors. As we navigate this transformative phase, our strategic approach involves investing in companies that are well-positioned to benefit from the evolving AI landscape, while assessing those at risk of disruption. We've observed a positive trend already in 2023, with several firms noting incremental revenue growth tied to generative AI — a pattern we anticipate will continue in 2H 2023 and accelerate in the coming years.

A ripple effect is surging through the tech industry, driven by escalating demand for resources critical to AI development. From GPU and memory chip designers to server manufacturers, the need for infrastructure essential to AI implementation is beginning to expand.

This demand extends beyond hardware, foretelling growth in services and solutions offered by software companies.

Firms possessing unique data sets are also uniquely positioned, enabling them to train models on patterns unique to specific industries, thereby bolstering their competitive edge.

However, we are mindful of the inherent risks accompanying these prospects; this is not a distant concern but a reality we face today. As we explore new opportunities and mitigate potential downsides, our commitment to understanding this rapidly evolving landscape and adapting our investment strategy remains unwavering, as does our mindfulness of valuations.

Overall, we believe the tech sector will continue to drive the economy in the AI era, much like it did during previous technological breakthroughs.

Operational efficiency: Profitability vs. Growth in the Tech Sector

Over the past decade, many tech companies adopted undisciplined spending habits, fueled by historically low interest rates and investor preference for growth over profitability. This led to the rise of unprofitable initiatives, unaligned with core competencies. 

This year, with higher rates, poor stock returns, and pressure from activist investors, new spending habits are being formed. Companies are increasingly focused on operational efficiency, finding opportunities to cut costs and expand profit margins. 

The refocus on profitability and operational efficiency is driving a deeper analysis by investors. Analysts are examining profitability metrics at varying levels of revenue. For many CFOs, it's no longer about growth at all costs; instead, the value lies in sustainable, profitable growth. 

Tech companies are finding they can make significant cuts without adversely impacting their operations or sales growth. In the first five months of the year, tech firms have laid off over 200k employees, surpassing the 165k tech employees let go in all of 2022¹¹. Over the long-term, we believe the cost reductions will ultimately prove beneficial to the tech sector.

One area where cost-cutting measures are not being made is in AI spending. The potential return on investment that AI offers is too significant to ignore, and companies are accordingly channeling their resources into this emerging field.

As many firms emerge leaner and more profitable than in the previous era, we believe their stocks will be re-rated. The push towards operational efficiency and enhanced profitability, while simultaneously increasing investment in AI, illustrates that there is still considerable room for growth in the tech sector. This evolution is reshaping the landscape, redefining the rules of success, and offering a promising outlook for long-term investors.

Global sector margins
12m forward net profit margin

Localization: reshoring critical technologies

The viability of globalization has been called into question following various challenges, including the COVID-19 pandemic, the war between Russia and Ukraine, and geopolitical tension between the US and China. In response, countries are adopting a strategic mantra focused on reshoring critical technologies. This shift, far from being merely strategic, is increasingly vital for the future of sovereign economies in the 21st century.

The reshoring trend is centered on two fundamental principles: geopolitics and supply chain security. The escalating US-China tech war emphasizes national security and sovereignty concerns. Accordingly, US manufacturing is shifting towards technology, with a focus on semiconductors, batteries, and solar energy. 

The second principle is about strengthening supply chains to mitigate vulnerabilities. The reshoring of supply chains for booming industries like semiconductors, solar energy, and EV infrastructure is not just strategic but necessary, creating a multi-year tailwind of growth and innovation.

Unlike the layoffs impacting the broader tech sector, industries tied to the reshoring trend are adding jobs. According to the Reshoring Initiative, over 100k manufacturing jobs were added in Q1 2023 due to reshoring efforts and foreign direct investment (FDI), mainly in the electrical equipment industry as a result of the increase of investment in EV battery manufacturing¹².

This strategic realignment, driven by geopolitics and supply chain security, is set to play a fundamental role in the sustainability and resilience of sovereign economies over the next decade.

CEO Survey: Have you considered reshoring manufacturing operations to the US?¹³

Solar Energy

Evolving consumer preferences, supportive policy, and the trend towards de-globalization will likely support the buildout of solar energy in developed economies. In the US, the infrastructure spending bill, the Inflation Reduction Act (IRA), includes incentives for the domestic production and manufacturing of renewable energy, leading to positive price reactions among public companies that appear well positioned to shift some or all of their supply chain and production onshore. Some of the first beneficiaries of the IRA may potentially include those involved in the production of solar inverters, which convert the direct current (DC) created by solar panels into alternating current (AC) which can be used for electricity, as well as solar trackers, which rotate solar panels to follow the arc of the sun to maximize the DC generated.

IRA: $369bn of support for decarbonization technologies¹⁵

Electric Vehicles

We believe we are still in the early stages of the transition from Internal Combustion Engine (ICE) cars to EVs, with an estimated 13% penetration of the global market at present¹⁶. As is the case with solar energy, the auto market is localizing as a result of policy incentives from the IRA which encourage domestic production and supply chains. Auto companies with the margin structure to make low cost EVs will benefit in the long-term, likely those with full vertical integration of their supply chain. We believe the current opportunity set is in companies creating necessary EV components such as silicon carbide wafers which can operate at higher voltages and temperatures, with better energy efficiency for increased mileage. Elsewhere, we see opportunity in EV batteries and auto manufacturers that stand to benefit from evolving policy in the US and upcoming policy in the EU.

EV Market Penetration (%)¹⁶

Second half of 2023: opportunity for long-term growth

It appears evident that the transformative power of generative AI, the focus on operational efficiency, and the strategic reshoring of pivotal technologies are redefining the tech sector and presenting new investment opportunities in the second half of 2023 and beyond. As fundamental research-driven technology equity investors, we aim to translate our evolving insights on the emerging set of opportunities and risks into investment action, ensuring we remain adeptly positioned to navigate and capitalize on this era of rapid technological transformation and economic realignment.