Jonathan Kitchen/DigitalVision via Getty Images
Summary
- Global equity markets extended their rally in Q3 2025 as enthusiasm for generative AI and strong capital flows into semiconductors remained the dominant market themes, while investors largely looked past ongoing trade and valuation concerns.
- Regional performance was mixed. U.S. equities outperformed following robust economic growth and earnings, while Japanese stocks maintained momentum on the back of corporate reforms and reflation. Emerging markets, especially China, saw significant gains due to AI optimism and supportive government policies.
- Our emphasis on quality growth investing was challenged by the market’s preference for high-beta growth stocks, contributing to underperformance relative to broader indexes during the quarter.
- Top relative contributors to the Portfolio’s performance included Oracle (ORCL), Shopify (SHOP), and Icon. The top absolute contributors were Oracle, Shopify, and Alphabet ((GOOG) (GOOGL)). The largest relative and absolute detractors in the quarter were Paycom (PAYC), SAP (SAP) and Adyen (OTCPK:ADYEY).
- During the quarter, we established new positions in NVIDIA, Broadcom (AVGO), TSMC, Boston Scientific (BSX), and Uber (UBER), and added to several existing holdings to capture emerging opportunities and evolving company fundamentals.
- To fund these investments and optimize portfolio positioning, we eliminated positions in ADP, Airbnb (ABNB), Accenture (ACN), and Thermo Fisher Scientific (TMO), and trimmed selected exposures across the Portfolio.
Seeks Growth & Capital Preservation (Performance (%) as of 9-30-2025)
Commentary
Global equity markets continued to rally across Q3 2025 as investors looked past lingering trade policy uncertainty and valuation concerns, driving the MSCI (MSCI) All Country World Index to record highs during the quarter, returning 7.7%. Similar to the prior quarter, hype around generative AI (Gen AI) was again the dominant theme and semiconductors were one of the primary drivers of returns as the continued capital flow into AI-related infrastructure projects shows no signs of abating.
From a regional perspective, non-US equities lagged their US counterparts, which was a reversal from the prior two quarters of 2025, as the big rebound in European equities stalled out producing only modest gains on the quarter and where the US Dollar stabilized following a ~10% depreciation during the first half of the year (as measured by the Bloomberg DXY US Dollar Index (DXY)). However, Japanese equities continued their stellar run so far this year thanks to corporate governance reforms and the nascent economic reflation that has boosted investor confidence and foreign investor appetite in the region. Emerging markets produced the best returns for the quarter driven in large part from a Chinese equity market propelled by Gen AI hype and both fiscal and monetary support from the Chinese government. Trade policy moved to the periphery early in the quarter as the Trump administration was able to reach interim agreements with most major trading partners, while a surprise improvement in US economic growth along with potential fiscal tailwinds from the ‘One Big Beautiful Bill’ and the strong US earnings reported during the quarter provided further impetus to the market rally.
Although the US economy surprised to the upside, growing at a revised annual rate of 3.8% in Q2, it came with an artificial boost thanks to a sharp decline in imports which was essentially an unwinding of the huge surge in Q1 as businesses stockpiled inventory ahead of anticipated tariffs. However, what was most notable in the data was the contribution from technology capex, and AI-related spending in particular, outpacing the contribution from consumer spending despite its much smaller share of the total economy, reflecting the sheer magnitude of capital currently flowing into AI-related projects.
A notable example of this continued investment was Oracle’s quarterly earnings report, which highlighted their remaining performance obligations (essentially contracted future revenues mainly related to their cloud infrastructure business) increasing 359% to $455 billion—in just one quarter—with the stock rallying over 30% following the news. The size and scale of those numbers is hard to fathom, however it speaks to the extraordinary demand for cloud computing and AI infrastructure currently.
Fortunately, Oracle was one of Global Growth’s largest holdings prior to these results and so provided an important boost to performance during the quarter (and the year so far), and combined with our positions in Microsoft (MSFT), Alphabet and Amazon (AMZN) reflects our continued conviction that the majority of the long-term value from Gen AI will accrue to the cloud infrastructure providers, along with semiconductor companies and select software businesses.
While headline US economic growth numbers had been positive prior to this quarter, further under the surface there have been areas of concern, most notably the continued softening in the labor market and where the burden of consumption is increasingly being shouldered by higher income cohorts—reflective of the so called ‘K-shaped’ economy. Despite the ‘last mile’ of inflation remaining stubbornly above the Federal Reserve’s target of 2%, the Fed’s focus shifted more to the employment side of their dual mandate and they lowered interest rates by 25bps for the first time this year, guiding for additional rate cuts across the remainder of the year and into 2026.
The other noteworthy observation during the quarter was guidance from the hyperscaler management teams around future capital expenditures (CAPEX) intentions—not just for the remainder of 2025 but for future years as well. Historically, these firms have tended to be conservative in their guidance around future spending, however we have recently witnessed prodigious increases to AI-related spending, flagging to us the spigots are wide open as the hyperscalers struggle to keep up with the voracious demand.
With investors maintaining their enthusiasm towards anything AI-related, the traditionally defensive sectors like healthcare, consumer staples and real estate lagged meaningfully on the quarter. The market has been quick to bifurcate sectors, industries and companies as ‘AI winners or losers’, and while there are specific headwinds that relate to each of these areas and companies within, it would appear many have been categorized into the latter until proving otherwise.
The persistent ‘risk-on’ behavior of the market has meant ‘high-beta growth’ companies have dominated the contribution to returns while ‘quality’ and ‘low-volatility’ stocks have lagged meaningfully.
While our heavy emphasis on ‘quality’ characteristics has been rewarded over the long-term, in narrowly driven and ebullient markets like the one we’re in today, quality can get left behind. Despite these continued headwinds to our philosophy and approach, we retain conviction that our emphasis on quality will be rewarded across the full market cycle.
Portfolio Performance & Attribution
In Q3 2025, the Polen Global Growth Composite Portfolio (the “Portfolio”) returned 3.0% gross of fees and 2.7% net of fees compared to 7.7% for the MSCI All Country World Index (the “Index”). Top relative contributors to the Portfolio’s performance included Oracle, Shopify, and Icon. The top absolute contributors were Oracle, Shopify, and Alphabet.
The largest relative detractors in the quarter were Paycom, SAP and Adyen. The largest absolute detractors were SAP, Paycom, and Adyen.
Following on from Q2’s substantial rally, Oracle was again our top-owned relative contributor following their impressive quarterly results showing a huge acceleration in demand for their cloud infrastructure services, with the stock up almost 30% in Q3. When reacquiring Oracle in September last year, our investment thesis predicted an anticipated acceleration in growth across their application, databases and cloud infrastructure businesses (and the expectation that cloud infrastructure would be growing the fastest), and clearly that thesis is playing out, with Oracle in the midst of a substantial acceleration in revenues.
The primary drag on relative performance was our positioning within information technology. In a similar vein to Q2, with Gen AI as the dominant theme and the high-beta growth areas benefitting most, our underweight exposure to semiconductors (especially early in the quarter) proved to be a headwind, while several of our software holdings detracted as the market continues to categorize large parts of the industry as ‘AI losers’. While we acknowledge this market perception, we believe it is an oversimplification that overlooks the high recurring revenues, strong margins, significant switching costs, and durable competitive advantages many of these businesses possess. Additionally, we see Gen AI as a potential catalyst to further enhance productivity and innovation for select companies in this space. Elsewhere, our exposure to the historically consistent and stable areas of digital payments and parts of health care also dragged.
While the risk-on sentiment that has propelled markets these past 6-months (and most of the past 3-years) has been focused on a very narrow set of drivers, our purpose is to build a high-quality portfolio that can compound earnings at a mid-teens rate, through all manner of economic regimes and market backdrops in the expectation that the Portfolio’s returns will converge with the underlying earnings profile. Striving to have balance across the spectrum of growth opportunities is a critical element of our purpose. While recent drivers of market performance have been increasingly concentrated to the detriment of our Portfolio’s ‘balance’, our focus is on the long-term and maintain our conviction in the approach that has stood the test of time for multiple decades.
Portfolio Activity
In Q3 2025, we initiated new positions in NVIDIA, Broadcom, Taiwan Semiconductor (TSM), Boston Scientific and Uber, and eliminated our positions in ADP, Airbnb, Accenture and Thermo Fisher Scientific . We also added to our positions in Microsoft, Eli Lilly (LLY), ServiceNow (NOW), and trimmed our positions in Amazon, Shopify, Oracle, Workday (WDAY) and Abbott Laboratories (ABT) .
During the quarter, we initiated positions in NVIDIA, Broadcom and Taiwan Semiconductor (TSMC), after having avoided semiconductor holdings over the past 2½ years following the initial wave of enthusiasm around Gen AI. While we have long admired these companies, their highly cyclical earnings have made it extremely difficult to forecast future earnings growth with any degree of conviction. Given our approach of seeking durable and persistent earnings growth that compounds over long holding periods, our concern in holding these names was that we would be forced to endure a punishing downcycle within our typical holding period – there is very little room for that in a concentrated portfolio of 25-35 companies. Notably, pre ChatGPT, NVIDIA had two punishing downcycles over the preceding five years.
Our team is constantly reevaluating each company that meets our quality guardrails, seeking to be intellectually honest and being willing to evolve our opinions when the facts change or when more information is available—that is specifically what has occurred for NVIDIA, Broadcom and TSMC.
While the sheer magnitude of demand for AI chips, servers and networking equipment was something that we clearly underappreciated, new incremental data points over the past few months lead us to conclude the current boom in AI chips and related hardware will likely continue for the foreseeable future giving us greater conviction over the trajectory of future earnings for all three companies. These new data points included: Oracle’s $30 billion annual revenue cloud contract (6/30), U.S. tax incentives reinstating 100% bonus depreciation for short-lived assets such as servers and networking equipment, boosting data center investment (7/4), Meta (META)’s pledge to invest “hundreds of billions” in multi-gigawatt data center clusters (7/14), Google (GOOG, GOOGL)’s capex hike to $85B for 2025, with more planned for 2026 (7/23) and the Trump administration’s executive order to accelerate the federal permitting of data center infrastructure (7/23). These data points combined with our existing fundamental research led us to believe that the growth potential for these companies remains strong even as growth moderates from the extremes of the past few years—Oracle’s recent earnings results (highlighted earlier), along with Microsoft and NVIDIA’s (NVDA) recently announced $45 billion investment in data center and AI-related projects in the United Kingdom reinforce that belief and our decision to acquire these three companies during the quarter.
NVIDIA produces the fast chips (GPUS) that are able to process compute intensive tasks like Gen AI training models extremely efficiently, are very flexible so can be used for any various workloads, and as a result are in high demand as the hyperscalers build out their Gen AI infrastructure. We believe NVIDIA has a very strong competitive moat, which is partly about the speed of their chips, but also the entire ecosystem they have built around them (programming language, training models and networking equipment).
We anticipate they will be able to generate earnings growth at ~20% per year for the near-term future and feel their mid-thirties multiple is reasonable for that level of expected growth.
Broadcom is the other major player in the AI chip market and the leading provider of custom chips. The company captures a significant share of enterprise spending in this area. As Gen AI use cases mature, and as inference workloads become a bigger piece of the compute pie, we expect that custom chips (and Broadcom’s in particular) will account for a larger share of the total market. We also expect they will be able to grow their earnings by ~20% for the next 3-5 years and believe their multiple in mid-to-high thirties is fair.
TSMC dominates global semiconductor manufacturing with over 60% market share and possesses some of the strongest competitive advantages of any business we’ve studied outside the US—significant scale, high switching costs and close customer relationships—which allow TSMC to easily meet our guardrail criteria. Roughly 60% of revenues come from high-performance computing (including ~33% from AI datacenter) and another ~25% from smartphones, where Apple (AAPL) and NVIDIA are their biggest customers. While China-Taiwan tensions remain a key risk—which we believe is reflected in the company’s sub-20x P/E multiple—we expect mid-teens revenue growth, modest margin expansion, and high-teens EPS growth through the cycle. Given these expectations, we view this risk as acceptable while remaining vigilant for any signs of escalation.
We also acquired a 2% position in Boston Scientific, a global leader in medical products that treat various cardiovascular and other conditions. Over the last two years they have witnessed a meaningful acceleration in their growth profile based on two primary catalysts: their Farapulse platform for pulse field ablation (PFA) and their Watchman platform for atrial appendage. PFA is a newer medical procedure used to treat atrial fibrillation that is less invasive, more precise and faster than more traditional ablation procedures and with fewer risks than medication. We believe PFA is likely to become the standard of care in this treatment paradigm and Farapulse is poised to be the market leader. The Watchman segment which already dominates market share, is a permanent implant designed to reduce the risk of stroke in patients with atrial fibrillation, represents a meaningful and accelerating percentage of their revenues. With an overall revenue growth rate in the low double-digit range combined with modest margin expansion, we think Boston Scientific has the potential to grow their earnings in the mid-teens over the next 3-5 years.
Additionally, we acquired a new position in Uber during the quarter. We have followed Uber for many years and believe their scale, network effects, growth opportunities and market position, combined with their current valuation make a compelling investment thesis. They have become one of the most recognizable consumer brands in the world and anticipate nearly $200bn in booking transactions for 2025. Over the past three years, they have compounded revenues at 36%. Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) at 69% and Free Cash Flow (FCF) margins have gone from negative to mid-teens. While the threat of autonomous vehicles looms and is likely weighing on the valuation, we believe that threat is many years away and so view that risk as low, and we expect Uber to compound earnings at ~20% p.a. over the next five years. In fact, we think the fastest way for autonomous vehicle companies to scale is to partner with a large and highly utilized platform like Uber who has dominant market position where it competes.
To help fund the NVIDIA acquisition, we sold our small remaining position in ADP. While ADP remains reasonably valued and continues to deliver steady performance, we also hold a position in Paycom, which provides a similar exposure with greater long-term growth potential in our opinion.
We also exited our Airbnb position to help fund our acquisition of Broadcom. Airbnb’s growth has slowed in recent quarters, and the consumer environment appears weaker today compared to recent history and so we view the risk/reward in Broadcom as more favorable.
In addition, we also sold our remaining position in Accenture to help fund our acquisition of TSMC. Growth has slowed across the IT services industry, and while AI-related projects should benefit Accenture (and the industry) longer-term, for now that remains a small component of overall IT services spending and so we feel TSMC offers a more attractive growth opportunity.
We also eliminated our remaining position in Thermo Fisher Scientific to help fund the purchase of Boston Scientific, which we feel represents a better opportunity in a similar industry. Thermo Fisher Scientific is currently navigating a combination of macro, policy and funding headwinds that while cyclical, are likely to persist for the foreseeable future and therefore we felt the Portfolio would be better served by a reallocation of proceeds to what we believe to be a superior portfolio candidate.
Outlook
Over the course of our 36-year history, we have never sought to be at the vanguard of nascent, emerging trends in markets, instead, waiting for evidence of durability and persistence in revenues and earnings before embarking on our intended long-term journey—and this approach has generally served us well for over three decades. We recognize that the timing of our recent acquisitions of NVIDIA, Broadcom and TSMC are after the prices for all three have appreciated materially, however it is our belief that we’re still in the early innings of this infrastructure buildout and so expect these new additions will deliver important contributions to the earnings growth and total return of the Portfolio. Along with the other highly competitively advantaged businesses we own, we anticipate mid-teens or better earnings growth over the long-term and expect that the portfolio’s returns has significant potential to converge with the underlying earnings profile in time.
While the ‘Magnificent 7’ companies—a homogeneous grouping we believe oversimplifies the differences among those businesses—continue to capture the headlines, we have also seen increasing divergence in performance among the constituents so far this year which might suggest that their heterogeneity is becoming more prominent. Combined with the notable performance of businesses outside of this group (e.g. Oracle, Broadcom) it may be that we’re finally seeing a broadening of market performance away from this grouping, which we believe bodes well for our portfolio of attractive growth businesses that we believe span the spectrum of durable growth opportunities across sectors and industries.
Thank you for your interest in Polen Capital and the Global Growth strategy. Please feel free to contact us with any questions or comments.
Sincerely,
Damon Ficklin and Steve Atkins
Global Growth Composite—GIPS Composite Report
Performance % as of 12-31-2024: (Annualized returns are presented for periods greater than one year)
Editor’s Note: The summary bullets for this article were chosen by Seeking Alpha editors.