H1’26 Recap / H2 Setup — Stronger for Longer, But Rougher

The first-half scorecard is stronger than it may feel. The S&P 500 has generated a +10% total return at the halfway mark, which looks somewhat anticlimactic versus the convexity of recent years and lags both the equal-weight S&P at +12% and small caps at +23%, but is still a very solid outcome relative to market history. It is also stronger than the first half of last year. The bigger historical framing is that if the S&P closes the year above roughly 7,530, it would mark a fourth consecutive year of double-digit returns. In the 68-year history of the index, that has only happened once before: the late-1990s run from 1995 through 1999.

The surface-level return also understates the scale of the right-tail opportunity set beneath the index. While several high-profile stocks have suffered meaningful drawdowns from their highs, the internal distribution has been heavily skewed toward winners. Within the S&P 500, 44 stocks are up more than 50% year-to-date, versus only 6 down more than 50%. Even more striking, 29 stocks are up more than 75%, with none down that much, and 22 stocks are up more than 100%, again with none down comparably. This is a market defined by large upside tails, high dispersion, and a meaningful reward for being in the right parts of the tape.

The most important driver remains the accelerating AI buildout. The capital cycle around AI infrastructure has moved from large to extraordinary, and the market has rewarded the direct beneficiaries accordingly. Goldman’s custom baskets tell the story: memory +250% YTD, data centers +115%, and AI semis +101%. This theme has run through almost every major market leadership discussion this year. That said, the narrative has not traveled in a straight line. The latest tension between US frontier models and Chinese open-weight models is a reminder that even within a secular boom, the market will constantly reprice who captures the value and for how long.

That distinction is critical because a rich opportunity set has not meant an easy trading environment. The year has featured sharp downside gaps in March, upside gaps in April, and significant swings across rates, FX, commodities, and factor leadership. Front-end rates have moved from pricing roughly 50bps of cuts in 2026 to around 37bps of hikes. The dollar has gone from extreme bearishness in March to broad bullishness today. Oil rallied through a period in which the Strait of Hormuz was closed for nearly four months, only to give back most of the move. Point-to-point returns look clean; the path has been anything but.

Globally, North Asia has been the defining equity story of the first half. In local-currency total-return terms, the top of the leaderboard is dominated by Korea +102%, Taiwan +62%, and Japan +40%. Each market has its own domestic story, but the throughline is obvious: all sit close to the center of the AI infrastructure supply chain. Japan deserves special mention because the Nikkei rallied 37% in Q2, the largest quarterly rally in its history going back to 1970. Korea deserves equal attention for a different reason: despite the extraordinary headline return, it has already experienced five intraday trading halts this year, nearly half of the total halts seen this entire century. That is “spot up / vol up” in its purest form.

The Asian story is not uniformly positive, however. The bottom of the global leaderboard is also Asia-heavy, with India -7%, Hong Kong -10%, and Indonesia -33%. This is less a generic “buy Asia” market than a highly specific “own the AI infrastructure coalface” market. The expected second-half message from the region is therefore probably “stronger for longer, but with higher volatility.” That phrase also applies well beyond Asia.

Small caps have been one of the biggest positive surprises. The Russell/small-cap complex is tracking for its largest annual outperformance versus the S&P 500 since 2003. Ex post, the drivers make sense: a better cyclical growth backdrop, visible in names like CAT and industrial cyclicals; some exposure to biotech and select AI infrastructure beneficiaries, particularly power and data-center plays; and positioning, with RTY futures carrying a record short base as recently as one month ago. Small caps have benefited from both fundamental broadening and a violent positioning reversal.

High dispersion and low correlation have also created a fertile environment for momentum. Goldman’s flagship momentum pair trade is up 57% YTD, following gains of 48% in 2024 and 30% in 2025. The AI winner/loser framework has been highly stimulative for the factor, as the market increasingly separates capex recipients, capex spenders, bottleneck beneficiaries, and structurally impaired business models. But momentum’s realized volatility has increased meaningfully over the past few months, and the fundamental community has substantial leverage to it. The factor remains powerful, but it is becoming more violent.

Hedge funds have generally handled the environment well. Coming off their best year since 2009, the community continues to generate positive returns across all nine tracked cohorts, with particular strength in fundamental long/short and macro. The GS Hedge Fund VIP basket is up 22% YTD. In a world where the technological, political, and geopolitical landscape changes rapidly, the ability to move quickly and adjust risk has real value. This is an environment that rewards active managers who can adapt rather than passively extrapolate.

A few first-half anomalies stand out. The US tech sector has done extremely well in aggregate, with SOX up 102% YTD and having just posted its best quarter in history, yet the Magnificent Seven are flat YTD while the S&P 493 is up 16%. That is one of the clearest signs that this is not simply a repeat of the 2023–2024 megacap leadership trade. The AI theme has broadened and migrated away from the platform leaders toward semis, memory, power, industrials, and infrastructure.

The tech-plus-energy barbell has also remained a remarkably powerful construction. Since the end of 2020, a daily rebalanced mix of those two sectors has cumulatively returned +204%, with zero down years, versus +114% for the S&P 500. That pairing captures two major features of the current cycle: technological capex intensity and commodity/geopolitical scarcity. It is a reminder that the most durable portfolios in this environment have often combined secular growth with hard-asset or energy exposure.

Europe has also done better than many expected, with the Stoxx 600 returning 11%. As often happens, Europe has found a way to muddle through better than the consensus feared. Meanwhile, despite all the volatility, issuance, and macro noise, IG credit spreads have barely moved from where they began the year. That stability is important. Equity dispersion has been intense, but credit has not yet sent a broad stress signal.

Precious metals tell a different story. Gold and silver were making a forceful run to new all-time highs before peaking around the news that Warsh would take over the Fed. The first leg lower reflected a washout of speculative enthusiasm, the second reflected the Iran war dynamic, and the more fundamental third leg may reflect a changed perception of the Fed put. If investors believe the days of responding to every major problem with money printing are over, that changes the opportunity cost and narrative support for gold.

The politics of AI are likely to become a bigger second-half storyline. The quote that “the AI industry today is the only industry in America that has less regulations than sandwich shops” captures the risk. For now, the equity market is focused on capex, productivity, and earnings. But as the labor-market and regulatory debates intensify, especially into the political calendar, AI’s social license may become more relevant for valuations. The market has created and impaired enormous amounts of market cap on the expectation of a productivity miracle. That expectation is not yet implausible, but it increasingly needs confirmation.

This is the most dynamic market environment in decades because two forces are colliding: immense technological change and significant political/geopolitical fragmentation. Both the public and private sectors are pressing the accelerator. Record deficits, industrial policy, AI capex, energy security, supply-chain reorientation, and geopolitical competition are all pushing capital into motion. The rate of change and the scale of deal flow have been highly stimulative for markets, but also destabilizing for traditional valuation anchors.

The second-half setup therefore turns on the sustainability of the AI investment boom. The macro strategy framing is the right one: to square the value added in AI-related names with potential broad-economy gains from AI requires increasingly optimistic, though not yet implausible, assumptions. The growing risk is that the market is overestimating the persistence of above-average earnings, particularly earnings generated by the investment boom itself. As long as the capex boom looks secure, near-term earnings should dominate valuation concerns. But pricing has made the market more vulnerable to any challenge to that optimistic view.

The base case remains that global equities trend higher in H2’26. That view rests on the expectation that double-digit S&P earnings growth persists through 2027. The macro backdrop is still pro-cyclical, AI capex remains enormous, small caps and industrials are broadening, and earnings rather than multiples have driven much of the S&P’s gains. However, the trading environment is likely to become even more “live by the sword, die by the sword.” The ante has been raised across the system: record G3 deficit spending, hyperscalers’ unrelenting AI capex commitments, and significant leverage across the trading community, including roughly $570bn of global exposure in levered ETFs.

That argues for a long delta / long vol posture into the second half. Directionally, the market can continue higher because earnings are strong and the investment cycle is intact. But the path is unlikely to be smooth because positioning is crowded, factor volatility is rising, corporate buybacks are in blackout, CTAs have downside convexity, and the market is increasingly dependent on a narrow set of AI-related earnings streams. Long exposure still makes sense, but so does owning convexity against the possibility that the market begins to question the persistence, ROI, or political durability of the AI boom.

The practical implication is to stay invested, but not complacent. Favor AI infrastructure, memory, power, data centers, industrial broadening, and quality cyclicals, while remaining selective around hyperscalers until the 2027 capex trajectory becomes clearer. Respect the fact that the Magnificent Seven being flat while the S&P 493 is up 16% marks a real leadership transition. Use volatility tactically rather than viewing every spike as a reason to de-risk. And watch earnings commentary closely, because the second half will likely be determined less by whether AI is important and more by whether the market continues to believe the investment boom can translate into persistent earnings power.

The Buffett quotes are a useful closing frame. Markets are dealing with immense technological change, geopolitical fragmentation, political uncertainty, and valuation questions, but the long arc of American capitalism has repeatedly absorbed major shocks and compounded through them. That does not eliminate drawdowns or policy mistakes, but it argues against confusing volatility with terminal impairment. The US market’s ability to finance, scale, and monetize new industries remains extraordinary. The second half may be volatile, but the larger story is still one of capital formation, technological acceleration, and earnings growth.