Artificial intelligence remained one of the most powerful forces shaping equity markets in the first half of 2026. The scale of investment continues to be significant, spanning semiconductors, memory, data centers, cloud infrastructure and the power systems required to support them. Hyperscalers and major AI labs have raised their combined 2026 capital expenditure guidance to approximately $700 billion, illustrating the exceptional scale of the investment cycle currently underway.

But as the AI investment cycle has accelerated, so has investor enthusiasm. During the second quarter, AI-related stocks became a dominant driver of market performance, contributing to an increasingly narrow and momentum-driven rally. In some areas, share prices moved considerably faster than the underlying fundamentals, raising an important question for long-term investors: when a compelling structural theme becomes a crowded trade, how do you separate durable opportunity from market enthusiasm?

For us, the answer remains the same as it has through previous market cycles: focus on business fundamentals, valuation and the durability of earnings.

From structural opportunity to market momentum

Earlier this year, we explored how the AI infrastructure buildout was creating opportunities beyond the market’s most obvious beneficiaries. We argued that the scale of investment in data centers, power and enabling infrastructure could create meaningful opportunities for established businesses across the broader value chain.

That view has not changed. What has changed is the market environment surrounding it. By the second quarter of 2026, the AI theme had expanded well beyond a fundamental discussion about long-term infrastructure demand. Momentum increasingly became a major driver of stock prices, particularly in areas such as semiconductors, memory and other AI-related infrastructure exposures.

This acceleration was particularly visible in semiconductors. The Philadelphia Semiconductor Index rose 88% in the second quarter, while semiconductor stocks more than doubled over a period of just a few months, something the industry had not experienced since 2001.

The distinction matters. A strong secular growth opportunity does not automatically translate into an attractive investment at any price. As expectations rise, investors must increasingly assess how much future growth is already reflected in valuations, how sustainable current earnings trends may be and whether returns are being driven by underlying business performance or simply by continued multiple expansion.

That is where fundamental discipline becomes particularly important.

Not all growth is created equal

One of the risks we see in highly sought-after areas of the AI ecosystem is that earnings expectations can increasingly become dependent on shortages, pricing increases or continued speculative demand.

These conditions can be powerful in the short term, but they are not necessarily durable.

The memory market provides a particularly striking example. In the case of SanDisk, strong demand and supply constraints associated with the current cycle helped gross margins rise from 23% to 78% in a single year. Despite that extraordinary improvement, memory remains a historically cyclical and commoditized industry, where periods of elevated profitability can eventually attract new capacity.

Periods of constrained supply can support pricing and profitability, but elevated prices also encourage new capacity, substitution and changing customer behavior. Over time, those forces can normalize the very conditions that initially drove exceptional earnings growth.

For that reason, we distinguish between businesses whose growth depends heavily on maintaining unusually favorable industry conditions and those where the opportunity is supported by more durable drivers.

We generally prefer companies whose growth can be driven by increasing volumes, market-share gains and recurring capital requirements. These businesses may still participate meaningfully in the AI investment cycle, but their long-term investment case does not depend entirely on continued scarcity or ever-higher expectations.

In our view, that creates a more resilient path to compounding through different stages of the cycle.

A narrow market can obscure broader opportunities

Another defining feature of the recent market has been concentration. Despite strong headline index returns, performance beneath the surface has been highly uneven. A relatively small group of AI-related companies captured a disproportionate amount of investor attention, while many high-quality businesses in other parts of the market were left behind.

This creates an interesting dynamic for active investors. When momentum becomes the dominant factor driving returns, company-specific fundamentals can temporarily lose influence over stock prices. Businesses can continue to execute well, grow earnings and strengthen their competitive positions without necessarily being rewarded by the market.

Conversely, companies associated with the strongest prevailing themes can experience substantial valuation expansion even when the improvement in their underlying fundamentals is more modest.

These periods can be uncomfortable for fundamental investors, but they can also create opportunity. The wider the gap between stock prices and business fundamentals becomes, the greater the potential for differentiated outcomes when markets eventually begin to discriminate more carefully between companies.

Toward the end of the second quarter, we began to observe early signs that the AI trade could be broadening, as some crowded exposures moderated and investors began to revisit quality companies elsewhere in the market.

Whether that shift persists remains uncertain. But importantly, our investment approach does not depend on predicting exactly when market leadership will change.

Participating without chasing

Artificial intelligence continues to represent a significant investment cycle, and we remain exposed to businesses that can benefit from its expansion.

The key distinction is how we choose to participate. We are not seeking AI exposure simply for the sake of having it. Nor are we willing to relax our standards for profitability, balance-sheet strength, competitive advantage or valuation because a company operates within a popular theme.

Instead, we continue to look for businesses where AI can enhance an already compelling long-term investment case.

DigitalOcean provides one example within our portfolios. The company has benefited from accelerating demand related to AI and cloud infrastructure, particularly in inference, production applications and core cloud usage.

The stock rose 86% during the second quarter. In results reported during the period, the company delivered revenue and EBITDA growth in the low 20% range while beating earnings expectations by 59%. Management also increased its 2026 revenue growth outlook to 26%, from 21% previously, and outlined a framework for more than 50% revenue growth in 2027, alongside a 40% EBITDA margin.

Those figures illustrate that the share-price appreciation was accompanied by a meaningful improvement in fundamentals. Even so, as the valuation increased, we reduced our position in accordance with our valuation discipline.

That combination is important. Recognizing a structural opportunity and maintaining valuation discipline are not contradictory. In fact, we believe both are essential to managing long-term investment risk. A great company can become an unattractive investment if expectations move too far ahead of fundamentals. Likewise, a strong secular theme does not eliminate the need to continuously reassess expected returns.

Fundamentals still matter

The long-term potential of artificial intelligence is significant. The investment required to build the infrastructure supporting it may continue to create opportunities across a broad range of industries for years to come.

But transformative technologies do not eliminate investment cycles, competitive dynamics or valuation risk. If anything, periods of exceptional enthusiasm make those considerations more important.

We believe the strongest long-term opportunities will ultimately be determined not by which companies are most closely associated with AI today, but by which businesses can translate structural demand into sustainable earnings, attractive returns on capital and durable free cash flow.

Market leadership may remain concentrated for some time, or it may continue to broaden. We do not believe successful investing requires predicting that turning point. It requires maintaining discipline through it. For us, that means continuing to focus on high-quality businesses, durable competitive advantages, sustainable growth and disciplined valuations, whether they are benefiting from artificial intelligence or any other powerful investment theme.