2026 MID-YEAR OUTLOOK

Executive Summary

Our January outlook argued that 2026 returns would be driven by earnings delivery rather than multiple expansion. That call was correct — but the market has arrived there by a route few could forecast, and the composition of returns has been far more important than their level.

The S&P 500 closed at 7,674 on 21 August, roughly 12% higher year to date and already above the upper bound of our 7,400–7,600 year-end target. Yet the index is cheaper than when we published: the forward twelve-month multiple has compressed from approximately 22x to 20.0x, because earnings have grown faster than prices. Since the end of the second quarter alone, the index has risen 2.8% while forward earnings estimates have risen 4.7%. This is an unusual and, in our assessment, constructive combination.

Three developments dominated the first half. First, the conflict involving the United States, Israel and Iran disrupted transit through the Strait of Hormuz, drove Brent crude from the high $60s to a peak above $110 and reset the inflation path for the year; Brent trades near $94 today, roughly 39% higher than a year ago. Second, the Federal Reserve has not cut once. The target range has been held at 3.50–3.75% at every meeting this year, and in July three regional presidents dissented in favour of a hike. Third, the long end of the Treasury curve has repriced hard, with the 10-year near 4.74% and the 30-year above 5.20% — the highest since 2007 — as federal debt passed $40 trillion and AI-related corporate issuance surged.

Corporate earnings absorbed all of it. Second-quarter blended earnings growth for the index is 50.4%, or 32% excluding two exceptionally large non-operating gains, with revenue growth of 15.0% and a record net margin. Our revised base case is a year-end index level of 7,800–8,100, implying a total return of roughly 15–19% for the calendar year. The distribution around that base case is wider than in January, and it is skewed by the long end of the curve rather than by equity valuation.

Scorecard: The January Outlook Reviewed

We publish our own scorecard because a research process that never marks its forecasts to market is not a research process. Six months on, the record is largely positive in an instructive way: the equity framework held, the rates framework did not, and the risks we identified were expressed.

What Held

Our central thesis — that returns would come from earnings delivery rather than multiple expansion — has been vindicated in a stronger form than we anticipated. We projected index earnings of $306–312 on growth of 12–15%. Bottom-up estimates for the calendar year now sit near $355 on a headline basis, or approximately $335–340 excluding two exceptionally large non-operating gains discussed below. Our index target of 7,400–7,600 with a total return of 8–12% was consequently too conservative, and we raise it in this document.

Our warning on earnings concentration was the most useful call in the January outlook, although it expressed itself differently than we described. We anticipated that a narrow leadership cohort created index-level fragility. What occurred was a rotation rather than a decline: large-cap value has returned 20.7% year to date against 0.3% for large-cap growth. Investors who owned the index as a proxy for the artificial intelligence trade have had a difficult six months. Investors who owned earnings have not.

We also assessed that expansion would continue without recession. That has held, though the labour market support beneath it has deteriorated materially and now represents the principal threat to the forecast rather than a supporting assumption.

What Did Not

Our monetary policy framework had anticipated more balanced FED views, with three regional presidents dissenting in favour of an increase in July. We expected one further 25 basis point cut and a terminal rate of 3.25–3.50%. The target range has been held at 3.50–3.75% at every meeting this year. Medium-duration, high-quality paper with the 10-year near 4.0% was the view; the 10-year now yields approximately 4.74% and the 30-year sits above 5.20%, levels last seen in 2007.

Finally, we identified inflation persistence or policy error. No index-level decline of such magnitude occurred, although the Nasdaq fell 10.1% into 29 July and semiconductors fell 29% during that month. We regard that risk as deferred rather than eliminated, and the following sections explain why we now believe its most likely trigger is the long end of the Treasury curve rather than corporate fundamentals.

Macroeconomic Framework

Labour: The Cushion Has Gone

The labour market is the clearest deterioration of the first half. Nonfarm payrolls fell by 23,000 in July, and revisions removed a combined 103,000 jobs from the May and June reports, bringing average monthly job creation over the past year to roughly 34,000. The unemployment rate ticked down to 4.1%, but for the wrong reason: the participation rate fell to 61.4%, its lowest in more than five years. Average hourly earnings grew 3.2% over the year, below headline inflation of 3.4%, so real wages are contracting.

This matters more than the monthly noise suggests. In January we described a bifurcated consumer supported by asset prices at the top and pressured by price levels at the bottom. That divergence has widened, and the labour market no longer has the slack to absorb a policy error in either direction. It is the single variable most likely to convert an inflation problem into an earnings problem.

Inflation: A Headline Problem, Not Yet a Core Problem

Headline CPI rose 0.1% in July and stands at 3.4% year on year; core CPI rose 0.2% and is at 2.5%, close to target. The gap between the two is almost entirely energy: energy prices are up 14.7% over twelve months, with gasoline up 24.6%. The June and July prints suggest the energy-led spike earlier in the year is losing momentum, but the risk distribution remains skewed upward while the Hormuz question is unresolved. Our concern is second-round effects — the durability of the pass-through into services and expectations — rather than the level of the headline itself.

Monetary Policy: The Asymmetry Has Flipped

In January the risk was that the market expected more cuts than it would receive. That risk has now fully inverted. The Committee has held rates for five consecutive meetings; the July vote was 9–3, with three regional presidents preferring an immediate 25 basis point increase, and the minutes indicated that many participants viewed tightening as likely if inflation did not decline. The June projections implied one increase by year-end, and market pricing has at points carried two.

Our expectation is that the Committee remains on hold through year-end, with a hike a live rather than theoretical risk should the September and October core prints firm. The practical implication for portfolios is that the front end is no longer a waiting room — it is a competitively priced asset — and that any equity thesis relying on rate relief in 2026 should be retired.

Fiscal Policy and the Long End

The most under-discussed development of the first half is the behaviour of the long end. Federal debt passed $40 trillion, the Treasury has at least doubled its long-maturity buybacks in an effort to contain financing costs, and a recent institutional survey found two-thirds of respondents expecting the 10-year yield above 5% before year-end. Long-dated government supply is now competing directly with an unprecedented volume of AI-related corporate issuance for the same duration buyers. That competition, not the policy rate, is what sets the discount rate applied to equities from here.

Corporate Earnings: Delivery, and the Quality of It

Second-quarter results were exceptional on any measure. With 88% of the index reported, blended earnings growth is 50.4% and revenue growth is 15.0% — the strongest revenue quarter since late 2021 — with 86% of companies beating estimates and a blended net margin of 16.9%, a record in the data series.

The headline requires an important qualification. Alphabet reported a quarter including a $98 billion gain in other income from unrealised gains on equity securities, and Amazon a $53.4 billion gain largely tied to its investment position. Excluding those two companies, index earnings growth is 32%, the aggregate earnings surprise falls from 29.2% to 10.9%, and the net margin falls to 15.0%. A material share of reported index earnings this quarter is a mark-to-market on holdings rather than operating cash generation, and it will not repeat. We assess underlying operating growth at roughly 30%, which is still exceptional.

The sector composition also shifted decisively. Energy earnings rose 147% on oil averaging $92.55 in the quarter, 45% above the prior year. Semiconductor revenues grew 77%. Health care was the only sector to report a year-on-year earnings decline. The index is no longer a single-theme story, which is precisely the change we argued would matter.

Valuation: Cheaper, But the Cushion Is in the Wrong Place

The forward twelve-month P/E of 20.0x sits modestly above the five-year average of 19.9x and the ten-year average of 19.0x. The trailing multiple of 28.2x illustrates how much of the current valuation depends on earnings that have not yet been delivered. On its face this is a better entry point than January.

The complication is the discount rate. A forward earnings yield of approximately 5.0% against a 10-year Treasury yield of 4.74% leaves an equity risk premium of roughly a quarter of a percentage point. This is the most important number in this document. It means equities are not being compensated for their risk relative to government bonds, that the entire return case rests on earnings growth continuing, and that a sustained move in the 10-year above 5% mechanically compresses the multiple regardless of how good the earnings are. Valuation risk in the second half is a rates story, not an equity story.

The AI Infrastructure Cycle: From Capital Expenditure to Capital Structure

Our January estimate of hyperscaler capital expenditure approaching $600 billion is now too low. Following second-quarter guidance, sell-side estimates for the largest five have been revised toward $700–750 billion, with capital expenditure reaching approximately 86% of sales at Oracle, 54% at Meta, 47% at Microsoft, 46% at Alphabet and 25% at Amazon. Demand for compute has not been the constraint.

What has changed is how the buildout is funded, and this is where we were too gentle in January. Incremental annual debt has risen from roughly 9% of capital expenditure in fiscal 2024 to approximately 32% by mid-2026. Alphabet priced an equity raise of nearly $85 billion in June. AI-related debt issuance globally reached approximately $236 billion by the end of May, running at roughly four times the prior-year pace, with full-year forecasts near $570 billion. A further $662 billion of hyperscaler data centre lease commitments sits off balance sheet under current accounting. The buildout has moved from self-funded to externally funded, and that shifts the risk from equity holders to the credit and duration markets — which is exactly where the strain has shown up.

The July semiconductor drawdown of 29% was not a demand event. It followed indications that surplus AI capacity might be sold externally, which challenged the scarcity premium the market had been capitalising into the supply chain. The relevant questions for the second half are no longer whether AI demand exists. They are who carries the depreciation, who carries the debt, and what the assets are worth on a five-year rather than a fifteen-year schedule.

Market Leadership: The Concentration Risk Resolved as Rotation

July is the month worth studying. The index finished essentially flat, down 0.1%, while beneath it energy rose 12.6%, financials rose 6.2%, technology fell 3.4% and the Nasdaq 100 lost 6.6%. The equal-weighted S&P 500 set a new high in the same month. Year to date, large-cap value has returned 20.7% against 0.3% for large-cap growth.

This is the concentration risk from our January outlook expressing itself in the most benign available form. Investors who owned the index as a proxy for the AI trade have had a difficult six months; investors who owned earnings have not. Dispersion has risen materially, and the opportunity set outside the ten largest constituents is the widest it has been in several years. For a concentrated, valuation-disciplined manager this is a materially better environment than the one we described in January.

Fixed Income: A Correction to Our Prior View

We were wrong on duration. In January we described medium-duration, high-quality fixed income as offering an attractive balance of income and interest rate risk with the 10-year near 4.0%. Yields have risen approximately 70 basis points at the 10-year point and more at the 30-year, and that position lost money.

Our revised stance is to take yield at the front and intermediate part of the curve and to decline the term premium risk beyond it. The two-year yields approximately 4.2% against a policy rate of 3.50–3.75%, which is adequate compensation without exposure to the fiscal financing question. We would not extend duration until the long end has repriced further or the fiscal trajectory becomes clearer. Credit spreads remain compressed and continue to offer thin compensation; within that, we would scrutinise issuers whose leverage is tied to the AI capital expenditure cycle, where the debt is new, the assets are short-lived and the revenue is not yet contracted.

Revised Base Case and Scenarios

We raise our year-end target to 7,800–8,100, implying a calendar-year total return of roughly 15–19%. The probability weighting below reflects a wider distribution than in January, driven principally by rates and energy rather than by corporate fundamentals.

Our base case, to which we assign a probability of approximately 55%, is a year-end range of 7,800–8,100. It assumes the Committee remains on hold through December, the 10-year yield is contained within a 4.50–5.00% range, Brent trades between $85 and $100 with the Hormuz risk premium intact but without further escalation, third and fourth quarter earnings deliver growth above 20% on an operating basis, and the forward multiple is sustained near 20x.

Our downside case, at approximately 25%, is a range of 6,600–7,000, representing a decline of 9–14% from current levels. It requires only one of several triggers: a rate increase in September or October, a 10-year yield sustained above 5.25%, an energy re-escalation carrying Brent beyond $120, or labour market deterioration converting into consumption weakness and negative earnings revisions. In each pathway the mechanism is the same — the forward multiple compresses toward 18x — and in three of the four the proximate cause is the discount rate rather than the earnings stream.

Our upside case, at approximately 20%, is a range of 8,300–8,600. It requires the Strait of Hormuz to reopen with Brent falling into the $70s, headline inflation converging toward core in the high 2s, a cut delivered in December, and a multiple re-rating toward 21.5x on credible evidence of artificial intelligence monetisation. We note that this scenario depends more heavily on geopolitical resolution than on anything within the control of corporate management.

Key Risk Factors

We regard long-end rates and fiscal financing as the highest-probability route to a double-digit equity drawdown. With federal debt above $40 trillion and the Treasury intervening through an expanded buyback programme, a disorderly repricing of the term premium would compress equity multiples irrespective of earnings delivery.

Energy remains the second-order risk with first-order consequences. The memorandum signed in June has not produced a reopening of the Strait of Hormuz, and Brent above $120 would push headline inflation back toward 5% and exhaust what remains of the Committee's patience.

Labour market deterioration is the variable we watch most closely. Average monthly job creation of 34,000 and falling participation leave no buffer should the Federal Reserve be obliged to tighten into a slowdown, and the transmission from employment to consumption to earnings is direct.

Within the artificial intelligence complex, the risk is one of capital structure rather than demand. Off-balance-sheet lease commitments, growing private credit exposure and shortening asset lives are the channels through which a monetisation disappointment would reach the broader market, and they sit largely outside the equity market where the enthusiasm is priced.

Earnings quality warrants specific attention. Non-operating gains flattered the second quarter materially, and comparisons in 2027 will be demanding against a very high base, with consensus already assuming growth of 13.6%. A shortfall against that assumption would arrive with the multiple offering little protection.

Finally, political and institutional risk is elevated in a way that resists modelling. November midterm elections, continuing questions regarding central bank independence and an unsettled tariff regime each carry the capacity to reprice assets independently of fundamentals.

KSE Investment Approach

KSE Capital Management employs a research-intensive approach focused on businesses with durable competitive advantages, strong free cash flow generation and management teams with demonstrated capital allocation discipline. We seek situations where intrinsic value can be estimated with reasonable confidence and where current prices offer an adequate margin of safety against our assessment of downside scenarios.

The first half has improved the environment for that process. Rising dispersion, a rotation of leadership away from a narrow cohort and a market that is now discriminating between companies that spend on artificial intelligence and companies that earn from it all favour security selection over index exposure. Our portfolio construction continues to consider correlation risk explicitly, with particular attention to positions whose apparent diversification conceals a common dependence on the same capital expenditure cycle. Scenario-weighted drawdown potential is assessed alongside expected return for every position.

Important Disclosures

This document is provided for informational purposes only and does not constitute investment advice, an offer to sell, or a solicitation of an offer to purchase any securities or interests in any fund or investment vehicle managed by KSE Capital Management LLC ("KSE"). Any such offer or solicitation may only be made pursuant to the Confidential Private Placement Memorandum and related subscription documents (collectively, the "Offering Documents"), which contain important information concerning investment risks, fees, and other material terms.

Interests in KSE Secure LP are offered solely to investors who qualify as "accredited investors" as defined in Rule 501(a) of Regulation D under the Securities Act of 1933. Prospective investors should carefully review the Offering Documents in their entirety and consult with their own legal, tax, and financial advisors prior to making any investment decision.

Interests in the fund have not been registered under the Securities Act of 1933 or any state securities laws and are subject to significant restrictions on transfer. There is no public market for interests in the fund, and none is expected to develop. Investors should be prepared to bear the economic risk of their investment for an indefinite period and be able to withstand a total loss of their investment.

The views and projections expressed herein are based on publicly available information and reflect analysis as of August 2026. Market data is stated as of 21 August 2026 unless otherwise indicated. Such views are subject to change without notice and should not be construed as a guarantee of future performance. All investments involve risk, including the potential loss of principal. Past performance is not indicative of future results. Forward-looking statements are inherently uncertain, and actual outcomes may differ materially from those expressed or implied. References to prior projections published by KSE are included for transparency of process and do not imply any particular level of forecasting accuracy.

KSE Capital Management LLC is registered as an Exempt Reporting Adviser with the U.S. Securities and Exchange Commission pursuant to the Investment Advisers Act of 1940 (CRD No. 339485; SEC No. 802-134911) and is registered with the UK Financial Conduct Authority under the NPPR regime (FRN No. 1061419).