Commentary: Appreciation Strategy

• Benchmark performance remains highly concentrated: only three of 11 GICS sectors outperformed the index in the quarter, with technology — AI in particular — the principal driver.
• Extreme valuation disparities are beginning to appear between AI-related and unrelated companies just as they did between technology-related and unrelated companies in the late 1990s.
• We keep a close eye on balance sheet and cash flow durability across our holdings to ensure our exposure centers on companies with the financial wherewithal to weather a potential cooling in today’s white-hot environment.

The S&P 500 Index continued its upward grind, rising 8% in the third quarter and advancing 15% year to date. The S&P 500’s strength is even more impressive given the index is up 34% since the April 8 post-Liberation Day low, at which point it was down 15% for the year. Clearly, tariff turmoil is in the rearview mirror for investors, while the AI economy powers ahead.

Benchmark performance remains highly concentrated: only three of 11 GICS sectors outperformed the index in the quarter, with technology — AI in particular — the principal driver. Leadership mostly mirrored the second quarter’s. Information technology (IT) led all sectors, up 13%, followed by communication services at 12% (these two led second-quarter performance as well). Within IT, tech hardware gained 23% as AI-driven storage demand lifted memory players while Apple shares performed well on expectations for an iPhone 17 upgrade cycle. The AI-themed semiconductor industry advanced 18%, building on a 43% return in the second quarter; the semiconductor industry is now up 39% year to date after declining 18% in the first 90 days of the year.

Consumer staples was the worst performer overall and the only sector to decline in the quarter. Weakness was broad based amid soft demand, worries about an influx of cheaper goods from China and limited pricing power to offset rising costs. Underperformance in materials, financials and industrials appears tied to persistent malaise in the traditional economy, as investors grow impatient with the still-elusive reshoring-driven boost to output.

After a bumpy start to 2025, U.S. equities are poised for a third straight year of outsize returns. The S&P 500 is within reach of delivering 20%+ gains in three consecutive calendar years (2023 up 26%, 2024 up 25%, year-to-date 2025 up 15%) — a feat achieved only once in the past century (1995-98 delivered four straight years above 20%, with 1999 close at +19.5%). Given the euphoria around AI and scorching stock prices, it’s all too easy to draw parallels to the late 1990s froth.

As long-term investors, we’re at a crossroads. On one hand, the backdrop for stocks remains constructive: the Fed has begun easing rates; the yield curve is positively sloped with short rates anchored; credit spreads are tight; and capital markets are wide open. On the other hand, speculative risks are building alongside untenable valuations. The current AI land grab carries an “if you build it, they will come” mindset despite a lack of profitable AI applications today. Until recently the data centers were funded by the large tech companies from ongoing cash flows. This shifted in September 2025 when several large vendor-supported AI megaprojects were announced in partnership with cash burning, profitless enterprises. That circularity — where vendors or capital markets are needed to fund buildouts — poses risk, especially given the blue-sky forecasts used to justify the spend.

"Given the euphoria around AI and scorching stock prices, it’s all too easy to draw parallels to the late 1990s froth."

It is too early to call a bubble, in our view, but the current environment is reminiscent of the Internet craze in 1998 or 1999. Extreme valuation disparities are beginning to appear between AI-related and unrelated companies just as they did between technology-related and unrelated companies in the late 1990s. The consumer staples sector appears to offer buying opportunities in terrific franchises at reasonable valuations.

Valuation alone is rarely a reason to sell, but by several measures equities have never been more expensive. For example, the Buffett Ratio, which measures total equity market cap to GDP, stands at 219%, its highest reading ever (dating back to 1990) and over two standard deviations from the long-term average (Exhibit 1). To us, this implies an equity market already discounting a sizable share of AI’s hoped-for GDP and corporate profit gains, with little cushion for delays or disappointments. We would also note that the Fed Model — which compares the 10-year Treasury yield relative to the S&P 500 earnings yield (earnings yield is simply the inverse of the P/E ratio) — is in equilibrium, a setup rarely seen since 2002 (Exhibit 2). Put simply, today’s risk-free rate is a far more competitive alternative relative to risk assets than at almost any time over the past 25 years.

Exhibit 1: Buffet Ratio at Extremes

Exhibit 2: Risk-Free Rate Competitive with Risk Assets

As in the second quarter, policy remains supportive of near-term growth, and capital markets conditions suggest risky assets such as stocks can continue to do well. That said, the market appears to be entering a more speculative phase, raising the risk of correction. In our view, capital markets will determine whether equities can sustain their advance. If investors keep rewarding AI megaprojects with higher share prices the cycle can persist. But if markets tire of new multiyear infrastructure commitments or funding AI firms at ever-richer valuations, there is little valuation support to fall back on.

We are long-term investors focused on through-the-cycle outperformance via downside protection, and we increasingly believe risks may be starting to outweigh the potential rewards. Accordingly, we keep a close eye on balance sheet and cash flow durability across our holdings to ensure our exposure centers on companies with the financial wherewithal to weather a potential cooling in today’s white-hot environment.

The ClearBridge Appreciation Strategy underperformed the benchmark S&P 500 Index in the third quarter. On an absolute basis, the Strategy had positive contributions from eight of 11 sectors. The IT and communication services sectors were the main positive contributors, while consumer staples, industrials and real estate were the main detractors.

In relative terms, overall stock selection and sector allocation detracted. Stock selection in the industrials, communication services and IT sectors and an IT underweight detracted from relative results, while stock selection in and an underweight to the health care sector and a communication services overweight proved beneficial.

On an individual stock basis, the biggest relative contributors during the quarter were ASML, TJX, Vulcan Materials, Alphabet and not owning Costco. The biggest detractors were Synopsys, Honeywell International and Netflix and underweights to Apple and Tesla.

During the quarter, we initiated new positions in Marsh & McLennan in the financials sector, Lennar in the consumer discretionary sector and International Paper in the materials sector. We exited Starbucks in the consumer discretionary sector, Old Dominion Freight Line in the industrials sector, Progressive in the financials sector, Crown Holdings in the materials sector and Synopsys in the IT sector.