SAINTS spotlight: still a ‘wicked learning environment’

As with any investment, your or your clients’ capital is at risk. Any income is not guaranteed and can fall as well as rise.

When Demis Hassabis, co-founder of DeepMind, set out to test his early machine-learning models, he sought a “kind learning environment”: one with predictable inputs and quick, clean feedback. 

He chose video games, starting with Pong. After 100,000 hours of practice, the DeepMind agent mastered the game, consistently beating the computer 21–0. Later generations scaled to far more complex challenges, culminating in the 2016 defeat of Lee Sedol, champion of Go. Given the complexity of the East Asian board game, it was an achievement many believed decades away.

Hassabis opted for games because, as he put it, the real world is a “wicked learning environment”. Inputs are noisy, outcomes are unpredictable, and feedback loops can be long. That truth has been front of mind after a disappointing quarter for clients, despite robust company fundamentals. Many holdings reported sound first-half results and healthy dividend progression, yet share prices did not reflect this progress. 

Such disconnects are unsurprising over short periods. While operating results drive share prices over the long run, the relationship can break down dramatically in the short run. The outputs eventually reflect the inputs but the feedback loop is long.

As a growth manager, being underweight the so-called Magnificent Seven has been unhelpful for relative returns. The picture elsewhere is similar. Europe is our largest overweight, and there the best-performing segments have been traditional value sectors. Banks have benefited from higher net interest margins, and defence names have rallied as European nations accelerate spending to reduce reliance on the US.

Our positioning has therefore been out of step with near-term trends, but it reflects our commitment to building a dependable portfolio for the long run. Defence and banking stocks have done well recently, yet they are cyclical, capital-intensive and beholden to external forces, qualities that don’t typically provide steady, durable returns. By avoiding such areas, we believe we are safeguarding the portfolio’s resilience, even if that discipline has been uncomfortable in the short term.

We are also mindful of potential overbuild as big-tech firms commit astonishing sums to datacentre infrastructure with limited evidence, so far, of returns. The five largest AI hyperscalers are expected to spend more than $380bn in 2025 alone – larger than Ireland’s annual GDP – echoing the land grab of earlier technology rollouts, such as the dotcom era.

Further down the AI value chain we find businesses with exposure to the same structural tailwinds but with greater diversification, and therefore more durability. History suggests this is where value often accrues over time. NVIDIA, AMD and Broadcom, today’s anointed AI winners, must all turn to TSMC to manufacture cutting-edge designs. This advanced segment accounts for only about one-third of TSMC’s revenues, with smartphones, automotive and Internet-of-Things providing valuable diversification. The blend of growth and resilience makes TSMC better aligned with the objectives we pursue on clients’ behalf.

Our largest absolute position, Microsoft, sits a step further down the stack in software. Its three core businesses are productivity (Office 365), cloud computing (Azure) and personal computing (Windows). Each contributes materially to its roughly $250bn in annual revenues and each stands to benefit from AI. Microsoft’s prescient 2019 investment in OpenAI positions it uniquely to deliver productivity gains, whether via GitHub Copilot for code or summarisation and workflow tools embedded across Office.

Beyond hardware and platforms are firms embedding AI to augment existing products and processes. Machine-vision specialist Cognex, a top performer this quarter, is a good example. Historically focused on barcode readers and quality-control cameras, Cognex is broadening end-markets with deep-learning software layered on to its VisionPro devices.

Variability in packaging: shapes, textures and materials across food, pharmaceuticals and beyond, has historically required human inspection. By integrating AI, Cognex helps manufacturers remove humans from the loop where appropriate, countering fatigue, improving quality and lowering labour costs. 

In total, approximately 15 per cent of the portfolio by weight is directly exposed to the AI revolution and should benefit from strong earnings growth in the years ahead. While this is ‘underweight’ versus the benchmark’s concentration, we believe our approach is more durable. Importantly, our exposure is distributed among firms that should benefit from long-term deployment of the technology while remaining resilient to inevitable bumps in the road. In a domain defined by disruption and a learning environment likely to remain wicked, prudence matters.

Looking through short-term headwinds

While US and European positioning explains much of this quarter’s underperformance, there were stock-specific disappointments. Most notable was Novo Nordisk, pioneer of the obesity-drug market.

Novo underperformed due to reduced guidance and intensifying competition. The company cut 2025 sales-growth guidance from 17 per cent to 11 per cent, prompting a sharp share-price reaction despite the market already braced for weaker numbers. Challenges include the persistent presence of copycat GLP-1 treatments, which still account for nearly a third of the US market, and ongoing competitive pressure from Eli Lilly. Sentiment also softened after Novo named internal candidate Mike Doustdar as CEO. Some saw this as a conservative choice for a period demanding significant adjustment. Our view is different: internal leadership can better optimise for the long term, and we welcomed the appointment.

Near-term performance is being weighed down by workforce reductions and execution challenges, yet the company is simultaneously ramping efforts to address copycat producers through litigation and regulatory engagement, while investing in R&D, capacity expansion and manufacturing efficiency.

Also, we’ve seen this movie before. In 2016, Novo faced pressure in the US insulin market. By focusing on execution, broadening access and investing in its pipeline, it restored growth and ultimately created the GLP-1 category now transforming obesity care. Today’s reset looks similar: an acknowledgement that sharper execution and leaner operations are needed to capture the next phase of growth. With the current valuation below 2016 levels, we believe the odds from here are skewed in favour of outperformance.

Controlling the controllables

Across the portfolio, the vast majority of holdings by weight are performing in line with, or ahead of, the expectations in our original investment cases. Where they are not, we commission further research, often asking a team member other than the original proponent to provide fresh eyes.

We are applying this process at Wolters Kluwer, the professional-services publisher, to assess the likelihood and potential magnitude of AI-driven disruption to its business model. Our investigative researcher has also completed work on Nestlé, where we are examining what she calls “confident rigidity” in the corporate culture and how that may affect the company’s ability to adapt in a consumer-staples market that may be changing faster than at any point in recent memory. This kind of deep research is our best mechanism for weeding out poor performers and improving future returns.

Beyond a handful of big winners, many high-quality US companies have lagged in share-price terms. We have prioritised these in our research pipeline. In September, after meetings with management in New York and London, we initiated a position in MSCI, the leading provider of global equity indices and risk analytics. About three-quarters of earnings come from the core index business, where long-term contracts and assets under management-linked fees drive resilient, high-quality revenue with gross margins above 80 per cent. MSCI calculates close to 300,000 indices daily, licenses benchmarks to passive giants such as BlackRock, and has seen assets in MSCI-linked funds grow to more than $1.3tn over the past decade.

We see multiple growth drivers: further proliferation of indices across active and passive strategies:

• increasing adoption by banks and other financial institutions
• expansion of risk-analytics offerings
• new opportunities in private-market benchmarking. 

MSCI’s shares have historically traded at a premium, reflecting strong growth, sticky revenues and robust cash generation, but that premium has recently narrowed on what we see as short-term concerns. We have taken the opportunity to initiate a position.

More broadly, many high-quality compounders in the portfolio continue to trade on attractive valuations. The portfolio’s forward price-to-earnings ratio sits only one point above the market average, a five-year low. We view this as a compelling entry point for a portfolio with superior growth, quality and resilience characteristics. Dividends are rising faster, gross margins are higher and balance sheets are stronger than the market average. On fundamentals – stock by stock and for the portfolio overall – we see a collection of businesses far stronger than recent returns imply.

The prize for taking the long view

In 2024, fifteen years after founding DeepMind, Hassabis was awarded the Nobel Prize in Chemistry. His team’s work solved the protein-folding problem and published more than 200 million previously unknown protein structures, a contribution to open-source science that could help tackle antibiotic resistance and even the proliferation of ocean plastics.

Investing is humbling. Depending on the quarter, one can look like an idiot or a genius as prices ebb and flow with sentiment. The antidote is to focus on inputs and control what can be controlled: deep company research, measured portfolio construction and patience.

Above all, it pays to take the long view. Our organisational structure equips us to make that wicked learning environment a little kinder, extending our time horizon and supporting us as we stick to a tried-and-tested process. While dividend progression has been strong and the path of returns smooth, with downside protection evident during the turmoil around ‘liberation day’, capital growth has lagged increasingly exuberant markets this year. 

These facts are linked. We run a diversified, resilient portfolio. That prudence means many of the index’s biggest drivers, such as tech stocks geared to ever-greater AI capex or deep cyclicals like banks, are a poor fit for us.

We appreciate this period has been disappointing, particularly when low-cost passive funds are the alternative. Yet resilience and discipline, qualities that may seem dull today, are exactly what should shine if expectations for today’s darlings prove too ambitious. In such an environment, we expect the portfolio’s strong fundamentals and steady dividend growth to be rewarded.

Past performance is not a guide to future returns.

Source: London Stock Exchange Group plc and its group undertakings (collectively, the “LSE Group”). © LSE Group 2025. FTSE Russell is a trading name of certain of the LSE Group companies. “FTSE®” “Russell®”, “FTSE Russell ®, is/are a trade mark(s) of the relevant LSE Group companies and is/are used by any other LSE Group company under license. All rights in the FTSE Russell indexes or data vest in the relevant LSE Group company which owns the index or the data. Neither LSE Group nor its licensors accept any liability for any errors or omissions in the indexes or data and no party may rely on any indexes or data contained in this communication. No further distribution of data from the LSE Group is permitted without the relevant LSE Group company’s express written consent. The LSE Group does not promote, sponsor or endorse the content of this communication.”

The value of the trust's shares and any income from them can fall as well as rise. Past performance is not a guide to future returns.

This communication was produced and approved in September 2025 and has not been updated subsequently. It represents views held at the time of recording and may not reflect current thinking.

This communication should not be considered as advice or a recommendation to buy, sell or hold a particular investment. This communication contains information on investments which does not constitute independent investment research. Accordingly, it is not subject to the protections afforded to independent research and Baillie Gifford and its staff may have dealt in the investments concerned.

The investment trusts managed by Baillie Gifford & Co Limited are listed UK companies and are not authorised or regulated by the Financial Conduct Authority. The value of their shares, and any income from them, can fall as well as rise and investors may not get back the amount invested.

Baillie Gifford & Co and Baillie Gifford & Co Limited is authorised and regulated by the Financial Conduct Authority (FCA).

The specific risks associated with the Trust include:

• SAINTS invests in overseas securities. Changes in the rates of exchange may also cause the value of your investment (and any income it may pay) to go down or up.
• The Trust invests in emerging markets where difficulties in dealing, settlement and custody could arise, resulting in a negative impact on the value of your investment.
• Market values for securities which have become difficult to trade may not be readily available and there can be no assurance that any value assigned to such securities will accurately reflect the price the Trust might receive upon their sale.
• The Trust can make use of derivatives which may impact on its performance.
• Share prices may either be below (at a discount) or above (at a premium) the net asset value (NAV). The Company may issue new shares when the price is at a premium which may reduce the share price. Shares bought at a premium may have a greater risk of loss than those bought at a discount.

Further details of the risks associated with investing in the Trust, including a Key Information Document and how charges are applied, can be found in the Trust specific pages at www.bailliegifford.com, or by calling Baillie Gifford on 0800 917 2113.

178408 10058580