AFMC - AI: A Mega Force Driving Returns

2023-11-14 06:10:00 ET

Summary

• We see artificial intelligence (AI) as a mega force shaping the new regime. We are overweight the AI theme in developed market stocks.
• U.S. stocks and 10-year Treasury yields both rose last week. We see investors demanding more compensation for the risk of holding long-term bonds.
• U.S. CPI data is in focus this week. The presidents of the U.S. and China will meet for a summit at a time of heightened strategic competition and tensions.

Transcript

As tech shares continue to outperform and top tech companies prepare to roll out new AI tools, we expect more widespread use of artificial intelligence.

Digital disruption and AI is one of five mega forces, or structural shifts, that we think can drive returns now and in the future.

1) An intelligence revolution

We think we are at the dawn of an intelligence revolution with exponential advances. Just consider the exponential growth of OpenAI’s ChatGPT.

We see a shift in the tech industry towards AI-centric business models, igniting a competitive race among a handful of top companies.

2) Stacking tech layers

We view this evolution in terms of a technology “stack.” The first layer covers cloud infrastructure and chipmakers that have already begun to reap the benefits of AI advancements. The second layer covers models, data and data infrastructure.

At the pinnacle of the stack are the applications that leverage these innovations.

3) Potential to boost productivity

AI-powered automation has the potential to boost worker productivity, and may offer a competitive advantage to companies, sectors and economies that more adeptly adopt this technology.

AI’s interaction with other technologies and mega forces is likely to yield the biggest investment opportunities, in our view.

We are tactically overweight the AI theme in developed market stocks.

Exponential advances

Notes: The chart shows the increase in parameters in selected iterations of Open AI’s GPT models between 2018-2023. Data available on page 8 of this paper. Open AI has not disclosed details on the training data and architecture for GPT 4. The estimate is based on independent analysis, industry watchers and media reports, as well as cited in sell-side research including JPMorgan.

Advances in computing hardware and deep learning innovations led to an inflection point for AI in late 2022. We think we are at the dawn of an intelligence revolution - with exponential advances. For example, the number of “parameters” between OpenAI’s ChatGPT-1 and GPT-4 models has likely surged exponentially in five years. See the chart. Parameters are elements of a model learned from historical data that allow it to generate text or multimedia content based on a prompt. The higher the number of parameters, the more sophisticated a model’s understanding of patterns and the more nuanced its output. We expect exponential growth to persist. The tech sector is largely benefiting so far: Mega-cap tech names account for most S&P 500 returns this year, yet when stripping out their impact broader stocks are down, LSEG data show. Mega-cap tech stocks have outperformed by delivering on earnings, even with the yield jump that can hurt long-term valuations.

We observe a fundamental shift in the tech industry towards AI-centric business models, igniting a competitive race among a handful of mega-cap companies. We view this evolution in terms of a technology “stack,” with increasingly foundational technologies appearing closer to the bottom. The first layer covers cloud infrastructure and chipmakers - the building blocks of computing power that have already begun to reap the benefits of AI advancements. The second layer covers models, data and data infrastructure. This layer may be underappreciated as demand for digital infrastructure like hardware, data and server farm locations is likely to outstrip supply. At the pinnacle of the stack are the applications that leverage these innovations. We think we are currently somewhere between the first and second layers, with the top layer likely coming next.

Far-reaching impact

AI-powered automation has the potential to boost worker productivity, and may offer a competitive advantage to companies, sectors and economies that more adeptly adopt this technology. The implications may impact economies and markets faster than generally expected. AI’s impact is set to span multiple domains, in our view, intersecting with mega forces like aging populations and geopolitical competition. Companies will likely succeed by attracting top talent and being able to invest in scaling up computational power to harness their data. The flood of AI patents may help identify some pioneers and laggards. Our research suggests there is a positive correlation between an uptick in AI patents and earnings growth in the one-two years after the patent registration. Our work also finds that private companies primarily produce AI patents, making private markets another way to tap into the AI theme. While not every patent leads to commercial success, the rising market value of AI patents indicates investor enthusiasm for the potential outsized returns for companies who are working to incorporate AI.

Like any technology, AI has adoption limits. Cybersecurity risks abound yet may spawn opportunities for consultancies or start-ups that specialize in setting up secure AI environments for other firms. Generative AI, which learns from massive data sets to create new content, still faces reliability issues. The tech can be prone to “hallucinations,” generating outputs that aren’t grounded in the input data. We think future versions will likely gradually improve. Global governments are trying to address the risks - potentially slowing adoption in some sectors - and shape AI business conduct, spurring opportunities.

Bottom line

AI’s interaction with other technologies and mega forces is likely to yield the biggest investment opportunities, in our view. We see a multi-country and multi-sector AI-centered investment cycle unfolding that we think will support revenues and margins. We’re overweight the AI theme in DM stocks on a six- to 12-month horizon, as we see it set to keep unfolding.

Market backdrop

U.S. stocks gained and are up 7% from their October lows. U.S. 10-year yields edged up but are still below their 16-year highs hit last month. We think the yield volatility is one reason why investors are starting to demand more compensation for the risk of holding long-term bonds. U.S. Q3 corporate earnings results have showed a handful of mega-cap stocks are still propping up earnings - and analysts are downgrading expected earnings due to the cautious tone from companies.

We’re watching U.S. inflation data this week to see how much inflation falls as pandemic-driven economic mismatches keep unwinding. Yet, we think tight labor markets and a shrinking workforce will keep inflation on a rollercoaster. U.S.-China tensions will be in focus with Presidents Joe Biden and Xi Jinping set to meet at a summit in San Francisco.

This post originally appeared on the iShares Market Insights.

For further details see: