Executive Summary:
After a volatile first quarter defined by the war in the Middle East and an 84% spike in oil prices, the second quarter delivered a sharp reversal as the conflict de-escalated and oil moderated from its March peak near $120 per barrel. As the ceasefire with Iran held, investors aggressively rotated back into the AI trade, propelling the S&P 500 higher. In this risk-on environment, the S&P 500 rallied 15.1% for the quarter, while the equal-weighted S&P 500 and the Russell 2000 small-cap indexes increased 11.3% and 21.4%, respectively. The technology sector exploded 43.3% higher, while the semiconductor index went parabolic — surging 87.8% in just three months, its best quarterly return since the index's 1994 inception. After this historic move, the top 10 companies in the S&P 500 now represent a record 38.6% of the index, and semiconductors alone account for nearly 20% of the S&P 500.
The U.S. economy is materially more vulnerable than the headline data suggests. Current growth is resting on three unsustainable pillars — historic deficit spending, the AI capex boom, and the wealth effect generated from the AI bubble and each is now showing signs of strain. Despite unemployment near record lows and the stock market at all-time highs, the U.S. is running an enormous budget deficit and spending nearly 4% of GDP servicing its debt. If short-term rates remain elevated, the interest burden will create a self-reinforcing spiral of wider deficits and higher interest rates. Most importantly, the AI capex cycle — with its deteriorating unit economics, Chinese open-source competition, FCF-negative hyperscalers, and growing dependence on debt issuance — cannot continue indefinitely. When the AI capex cycle slows, the U.S. economy will be hit from both directions — the direct contribution to GDP will reverse, and declining technology stocks will hurt consumption through a negative wealth effect.
We believe that we are in a very high-risk environment, and the risk-reward in U.S. equities is poor. Stocks are effectively priced for perfection, with investors crowding into AI-related names while largely ignoring the potential damage an oil shock or a bursting of the AI bubble can inflict on a fragile economy. Speculation and leverage drove the market higher in Q2. Retail investors borrowed at record levels to buy stocks, while corporate insiders and the largest AI-related companies themselves were selling. It is never a good sign when retail investors are gambling with leverage, while insiders are heading for the exit.
Today, we are underweight equities because their long-term risk-reward is unattractive, and the recent rally's narrow breadth reinforces our caution. In a regime of modest growth and rising inflation, we prefer assets that have historically performed well in similar inflationary environments: value stocks, international equities, gold, and short-duration fixed income. We will remain focused on preserving capital by maintaining a highly diversified and balanced portfolio until equities present a more favorable risk-reward opportunity
Second Quarter 2026 Market Review:
After a volatile first quarter defined by the war in the Middle East and an 84% spike in oil prices, the second quarter delivered a sharp reversal as the conflict de-escalated and oil moderated from its March peak near $120 per barrel. As the ceasefire with Iran held, investors aggressively rotated back into the AI trade, which propelled the S&P 500 market higher.
In this risk-on environment, the S&P 500 rallied 15.1% for the quarter, while the equal-weighted S&P 500 and the Russell 2000 small-cap indexes increased 11.3% and 21.4%, respectively. International stocks continued their strong performance. MSCI Emerging Markets increased by 21.1% during the quarter, propelled by commodity tailwinds and AI semiconductor exposure, particularly in South Korea and Taiwan. MSCI EAFE was up 8.6% in q2 as European equities benefited from attractive valuations and some easing of stagflation concerns.
Despite the Iran ceasefire and the reopening of the Strait of Hormuz — which drove oil prices 16.4% lower during the quarter — interest rates continued to march higher, extending the sharp move that began in Q1. The U.S. 2-year Treasury yield climbed another 39 basis points to 4.20%, while the 10-year yield rose 15 basis points to 4.47%, flattening the yield curve to 0.27% — well below its 50-year median of 0.80% and a clear signal of tighter financial conditions. Before the conflict, markets were pricing roughly two Fed rate cuts in 2026. As the oil shock lifted inflation expectations, the market has completely reversed its Fed outlook. Investors currently price in 0.41% of rate hikes in the second half of 2026 — a 57% probability of a 0.25% hike to 4.00% at the September meeting, and a 39% probability of a further increase to 4.25% in December.
Gold continued its poor performance in Q2. After spiking to $5,608/oz in January on geopolitical fear and central bank buying, gold has plunged nearly 30% over the past six months as real rates rose, the dollar surged, and the Fed pivoted from potential cuts to hikes — the same dynamic that drove gold lower in 2022, when the Fed's aggressive tightening pushed prices down even as inflation ran hot.
In our view, however, gold now offers an attractive long-term risk-reward at current levels. The structural bullish case remains fully intact — record peacetime budget deficits, an enormous federal debt burden, and persistent foreign central bank buying due to de-dollarization. And with long-term real rates at multi-decade highs, we expect real yields to fall from here, shifting from the primary headwind that drove Q2's decline to a meaningful tailwind.
The U.S. economy continued to show resilience. Q1 GDP grew at 2.7%, in line with its 50-year average, while the unemployment rate held near a historically low 4.2% in June. Inflation, however, is a significant problem. Headline CPI remained elevated at 3.46% in June, while core inflation remained stuck at 2.57% -- above the Fed's 2% target for an unacceptable five years running.
The AI Trade Narrows
We have long believed the S&P 500 offers a poor long-term risk-reward, given its extreme overvaluation and dangerous concentration in a handful of mega-cap technology names. Those concerns accelerated materially in the second quarter. The technology sector exploded 43.3% higher, while the semiconductor index went parabolic — surging 87.8% in just three months, its best quarterly return since the index's 1994 inception. After this historic move, the top 10 companies in the S&P 500 now represent a record 38.6% of the index, and semiconductors alone account for nearly 20% of the S&P 500.
Source: Citadel Securities
This concentration risk is particularly concerning because semiconductor stocks are exceptionally volatile. Since 1994, the SOX has experienced eleven separate drawdowns of 30% or more — averaging one every three years — with a median decline of roughly 40% and an average decline of 45%. Applying that historical median to today's index weight would translate to an 8% hit to the entire S&P 500 from semiconductors alone. And a decline back to the SOX's March low — a drawdown of over 50% from the June peak — would take approximately 10% off the S&P 500.
The Q2 strength in semiconductors alone accounts for 51% of the S&P 500's year-to-date performance, and together with the technology storage sector, these two industries have contributed nearly 75% of the index's return this year. In other words, three-quarters of the S&P 500's 2026 gain has come from a narrow slice of AI infrastructure names — not from broad economic strength, not from healthy earnings growth across the market, but from a single speculative AI theme.
Meanwhile, the market's former leadership has quietly rolled over. The Magnificent Seven — the group that carried the S&P 500 to record highs in 2024 and 2025 — has been range-bound for nine months, and its relative strength has broken down. Importantly, this is not the healthy, broadening rotation that typically accompanies a durable bull market. Leadership has not shifted from mega-cap tech into a wider set of economically sensitive stocks; it has simply layered a second concentrated bet on top of the first. Semiconductors and the Magnificent Seven now depend on the same underlying story — that the AI capex boom will continue indefinitely at its current pace — leaving the entire market exposed to the same single point of failure. History is clear on what this pattern signals --sharp leadership changes paired with narrow, speculative surges into a single industry are hallmarks of late-cycle bull markets, not healthy ones.
Source:Stockcharts.com
In summary:
The S&P 500 is near a record high, and by most long-term valuation measures the index has never been more expensive. The 15.2% surge in the second quarter masked a sharp deterioration in market health beneath the surface — the kind of decoupling between index performance and underlying breadth that historically precedes rather than follows major tops.
As an index, the S&P 500 is no longer diversified in any meaningful sense. Its returns are increasingly driven by a single, highly cyclical industry that is a direct beneficiary of the AI capex bubble. The S&P 500 now carries a level of concentration and speculative behavior that has historically preceded major market tops:
- Record concentration. The top ten companies represent 38.6% of the S&P 500, and semiconductors alone are nearly 20% of the index. By way of comparison, index concentration is now higher than it was at the peak of the 2000 dot-com bubble.
- Narrow leadership. Semiconductors alone account for 51% of the S&P 500's YTD gain, and together with technology storage stocks, the two industries account for nearly 75%. This is not a healthy, broad-based bull market; it is a single speculative mania.
- Broken former leadership. The Magnificent Seven — the group that carried the S&P 500 to record highs in 2024 and 2025 — has been range-bound for nine months, and its relative strength has broken down. Capital is rotating not into a wider set of economically sensitive stocks, but into an even narrower and more speculative semiconductor trade. Layering one concentrated AI bet on top of another is not a healthy rotation.
In our view, prudent investors should rebalance away from this concentrated bet and into parts of the market that offer both genuine diversification and materially more attractive valuations: value equities, mid-caps, and international stocks.
Economic Outlook
As value investors, our asset allocation is driven by long-term valuation measures and the risk-reward opportunities present in the market. Moreover, we analyze leading economic and market indicators to determine the likely paths of economic growth and inflation. This enables us to strategically position our portfolio to perform well in all economic environments.
Before the war with Iran and the ensuing oil shock, we were deeply concerned that the U.S. economic expansion was narrow, fragile, and resting on three unsustainable pillars: near-record budget deficits, a historic AI capital spending boom, and consumption increasingly concentrated among the wealthiest Americans who have benefited most from Fed-inflated asset bubbles. The events of the second quarter have done nothing to ease those concerns — and in several important respects, they have deepened them materially.
Pillar 1: The Budget Deficit
The federal budget deficit remains one of the most important risks facing the U.S. economy. Recently, the U.S. Department of the Treasury announced that the federal budget deficit totaled $1.4 trillion during the first nine months of fiscal year 2026, which was $29 billion higher than in FY 2025. The federal government continues to spend money at a pace normally associated with a deep recession — yet the unemployment rate is 4.2%, and the S&P 500 is at an all-time high. Large deficits are not a sign of a healthy, self-sustaining economy.
Source: U.S. Department of the Treasury
The core problem remains spending, not revenue. Through the end of Q1, government receipts stood at 18.2% of GDP, slightly above the 55-year average of 17.9%, while outlays reached roughly 24.1% of GDP — more than two percentage points above their historical norm of 22%. Compounding the problem, the U.S. Treasury has for the past several years financed our deficits by issuing short-term debt rather than locking in longer-term maturities, leaving the federal balance sheet acutely sensitive to any rise in short rates.
Since the war in the Middle East began, that sensitivity has become an urgent concern. The U.S. 2-year Treasury yield has jumped 85 basis points, from 3.50% to 4.35%, and if it holds at this level, the interest expense on the nation's $39 trillion debt is poised to rise by roughly $330 billion. That would push federal interest costs meaningfully above their current $1.2 trillion — already 3.8% of GDP and one of the largest expense items in the federal budget.
Despite unemployment near record lows and the stock market at all-time highs, the U.S. is already running an enormous budget deficit and spending nearly 4% of GDP servicing its debt. If short-term rates remain elevated, the deficit will widen further, pushing long-term rates higher and driving debt-service costs higher still — creating a self-reinforcing spiral. In our view, this dynamic will continue until the bond market, not the stock market, ultimately imposes the fiscal discipline that Washington has refused to impose on itself.
Since the Pandemic, government spending has surged, leading to a record peacetime budget deficit and rising interest expense. Since the Pandemic, interest expense increased from $514 billion to $1.2 trillion, which is a 136% increase to 3.8% of GDP.
Source: FRED
Pillar 2: The AI Capital Spending Boom
The second unsustainable growth driver is the AI investment boom. Morgan Stanley estimates that the hyperscalers (Google, Amazon, Microsoft, Meta, and SpaceX) are expected to spend $779 billion in 2026 and $1.23 trillion in 2027, equivalent to 2.4% and 3.86% of GDP, respectively. According to Paul Kedrosky, AI capex has now "exceeded rural electrification… interstate highway buildout… as a function of straight-up spending, as a percentage of the economy, as well as railroads, as well as canals."
Source: Morgan Stanley
Additionally, the AI capex boom now accounts for nearly 50% of U.S. GDP growth and an even larger share of corporate profits. In our view, both the stock market and the economy have become dangerously dependent on this historic level of capital spending, and any material setback will pop the stock market bubble and inflict severe economic consequences.
Source: Daily Chartbook, JPMorgan
In the second quarter, the four largest hyperscalers (Amazon, Alphabet, Meta, and Microsoft) raised their combined AI-related capital expenditure guidance to $700 billion for 2026 and $1 trillion for 2027, up dramatically from $410 billion in 2025. This surge in capex guidance drove the technology sector 43.3% higher in Q2, while the volatile semiconductor index went parabolic, surging 87.8% for the quarter.
While investors have focused on the headline capex forecast, we are deeply skeptical that spending on this scale will materialize. The hyperscalers currently sit on a backlog of roughly $2 trillion in customer commitments, and on the strength of that backlog they plan to invest an estimated $3.4 trillion over the next three years. The problem is who those commitments are with. Approximately $1 trillion — half of the entire backlog — comes from OpenAI (ChatGPT) and Anthropic (Claude), two private companies with dubious business models that are losing billions of dollars annually and ceding market share to cheaper, increasingly capable open-source models out of China.
In other words, the hyperscalers are committing trillions of dollars in real, debt-financed capital spending against demand promises from two cash-burning private companies whose competitive position seems to be deteriorating. That is not the foundation of a durable investment cycle.
Every investment ultimately comes down to return on invested capital relative to the cost of that capital. Because the economics of AI are still unknown, both operators and investors are forced to make assumptions — about total market size, revenue growth, expense trajectory, and eventual profitability. Wall Street's approach has been straightforward: take today's revenue, assume rapid growth and defensible market share, apply a "normal" profit margin on scale, and use that to justify the current spending. The core assumption is one of operating leverage — as revenue scales, costs should grow more slowly, and profitability should emerge.
The problem is that this assumption is not being validated by the actual results. There is no operating leverage yet among these private companies. Costs are ramping in line with revenue, and profitability continues to recede into the future. In our view, the reason is straightforward: demand is being inflated because OpenAI and Anthropic are effectively “selling $20 bills for $10”. Per SemiAnalysis: "A user who genuinely maxes their subscription every week for four weeks gives OpenAI a compute bill 70 times larger than the subscription revenue. Anthropic's equivalent is 40x — still catastrophic at scale, but significantly better insulated than OpenAI."
Stated plainly, the two private companies that represent half of the hyperscaler capex backlog are charging $200 per month for compute that costs them $8k to $14k. That is not a business model with a path to profitability. It is a subsidy program masquerading as demand — and once the subsidies stop, the actual, unsubsidized demand for AI compute is likely to be a fraction of what today's capex plans assume.
Source: semianalysis.com
OpenAI and Anthropic are almost certainly pricing their models below cost to capture market share and secure a first-mover advantage, on the assumption that once they achieve scale, they can raise prices and grow into profitability. This is a familiar strategy — but it only works if competitors cannot enter at similar quality with lower costs. That condition is not being met.
Chinese open-source models now deliver comparable performance at less than 20% of the cost, relying more heavily on software efficiency than on the expensive semiconductor infrastructure that dominates the U.S. approach. Estimates suggest these models already account for roughly 60% of global developer and API token consumption, and that share is likely to grow since the newly released Kimi model delivers similar performance at a small fraction of the cost of its U.S. peers. Chinese share of the consumer market is more modest at an estimated 20% to 30%, but the disparity between the two segments is itself telling: professional developers, who evaluate models on cost and performance rather than brand, are voting decisively for the Chinese architecture.
It seems that this is not a temporary competitive gap that Western hyperscalers can outspend — it is a structural cost advantage rooted in a fundamentally different architectural approach to the problem. And it may in fact be understated, since the degree to which the Chinese government is subsidizing the industry is unknown but almost certainly material.
The implication is significant. If Chinese open-source models continue to gain share at a much lower cost, OpenAI's and Anthropic's path to raising prices and reaching profitability becomes less likely. And if they cannot reach profitability, their ability to honor the more than $1 trillion in compute commitments they have made to the hyperscalers becomes unlikely as well.
Beyond the fragile unit economics at OpenAI and Anthropic, we are increasingly concerned about the AI industry's ability to fund its capex commitments through the capital markets. Because these companies are losing billions of dollars annually, they cannot self-finance their growth — every dollar of expansion has to be raised from equity or credit investors. The financing patterns of the past several weeks tell us that the market is beginning to demand a much higher price for that capital.
On June 8th, OpenAI filed an S-1 with the SEC in preparation for a Q3 or Q4 2026 IPO. On June 12th, SpaceX (a hyperscaler with a growing datacenter and AI business alongside its satellite and space operations) went public at $135 per share and raised $87.5 billion. Over the next three trading days, the stock surged 67% to $225.60 before collapsing nearly 50% and erasing roughly $1.5 trillion in market capitalization. On June 25th, the New York Times reported that OpenAI was leaning toward postponing its IPO to 2027. People familiar with the discussions cited management concern over the SpaceX debut and public-market investors demanding far more granular unit-economics disclosure than the company was prepared to provide.
At the same time, Alphabet (Google) raised $90 billion in a secondary equity offering — the largest in U.S. corporate history and Alphabet's first new share issuance since its 2005 IPO. This is a company with more than $100 billion in net cash going to the equity markets to fund AI capex. It chose equity over debt, and it chose it in size. In our view, that is not the action of a management team that believes the AI capex cycle will pay for itself in cash flow within a reasonable window. It is the action of a management team that wants permanent, non-callable capital on the balance sheet because it recognizes both the scale of the required spend and the risk that the returns may take much longer to materialize than the market currently assumes. When a hyperscaler with a fortress balance sheet issues equity at this scale, it is telling you something important about how it views the durability of the current cycle.
Finally, last September, Oracle’s stock surged 36% in a single day on the announcement of a $300 billion OpenAI compute contract, briefly making Larry Ellison the world's richest person as Oracle's market cap approached $1 trillion. Since then, the stock has plunged 67%, and its debt has been downgraded to BBB-, one notch above junk. S&P cited two reasons for the downgrade. First, the gap between capex and cash flow is widening, not narrowing. They expect Oracle's free operating cash flow deficit to reach $42 billion, nearly double its prior forecast. Second, customer concentration risk: approximately 50% of Oracle's $638 billion in remaining performance obligations comes from a single counterparty — OpenAI — whose own financial position is deeply uncertain.
For most of the past decade, the hyperscalers – Amazon, Google, Microsoft, Meta, and Oracle -- earned their premium valuations because of their asset-light business models, fortress balance sheets, industry-leading operating margins, and enormous free cash flow generation. The scale of AI capex has fundamentally transformed these companies into capital-intensive infrastructure businesses — massive data center buildouts, custom silicon programs, and multi-year power and land commitments that look far more like utility or industrial economics than like the software businesses these companies used to be. In our view, the market has not yet repriced this transition to asset-heavy businesses with large maintenance capex commitments and negative free cash flow. Wednesday's earnings reports from Alphabet and Tesla underscored just how far this transition has already progressed. Both companies reported negative free cash flow for the third quarter, and Bank of America expects the hyperscaler group as a whole to remain free-cash-flow negative through at least 2027.
Source: BofA Global Research
In summary: OpenAI delayed its IPO because public investors are growing skeptical of its unit economics; Alphabet raised unprecedented equity because internal cash flow is no longer sufficient to fund its AI ambitions; and Oracle's stock has collapsed, and its debt was downgraded because its cash burn is accelerating and a single unprofitable startup represents half of its backlog. In our view, the AI mania is faltering as investors are finally focusing on the poor fundamentals rather than the ebullient AI narrative that has driven markets for the past three years.
Pillar 3: The Wealth Effect and the “K-Shaped” Economy
While headline growth and unemployment continue to look solid, more than five years of elevated inflation have carved the U.S. economy into a distinctly K-shaped structure. The wealthy top 10% of households are doing well — the direct beneficiaries of record-high stock and home prices, which have inflated their net worth, reduced their perceived need to save, and encouraged them to spend more. Unfortunately, the majority of households have suffered from higher costs and a steady erosion of their standard of living.
The concentration of prosperity is now extreme. The top 10% of households account for nearly half of all U.S. consumer spending, own approximately 87% of the stock market, and 44.5% of homes. The wealthy have benefitted from the Fed's profligate post-COVID monetary policy: interest rates kept too low for too long, trillions of dollars printed to stimulate the economy, and the resulting asset bubbles in equities and residential real estate. Those same policies produced inflation that has now run above the Fed's 2% target for more than five years — an abject policy failure.
The wealth effect has been amplified by the AI capex bubble that began with OpenAI's launch of ChatGPT in November 2022. Since then, the mega-cap technology stocks that dominate the AI trade have driven the market to extreme valuations and now account for roughly 40% of the S&P 500.
For the bottom 50% of households, the asset bubbles have not helped. They own only 1% of the stock market and 13.2% of homes, which means they have received virtually no benefit from rising asset values, yet the full weight of the inflation produced by the Fed’s profligate policies.
Source: Citadel Securities
We believe that the economy is fragile because so much of it rests on the AI bubble. AI-related stocks and the historic capex cycle they have generated are driving the stock market, and the stock market is in turn producing the wealth effect that has kept aggregate consumption growing. With AI capex now contributing roughly 50% of GDP growth directly, and the AI-driven wealth effect powering an outsized share of what remains, the U.S. economy has become extraordinarily dependent on a single, speculative theme — one whose sustainability we have already argued is far from assured.
The wealth effect is not the only mechanism propping up consumption. Households have also materially reduced their savings rate because they must save less and spend more to maintain their standard of living.
The Fed's dual mandate is to achieve maximum employment and stable prices, with "stable prices" formally defined since 2012 as a 2% inflation target. The Fed has failed to provide stable prices for five consecutive years. Real wage growth — nominal wages minus inflation — has averaged 2.7% per year over the past five decades. Over the past eighteen months, real wage growth has collapsed from 2.8% to just 0.3% today. In other words, inflation continues to erode their purchasing power.
To maintain their living standard, households have reduced their savings. From April 2025 to today, the personal savings rate has declined from 5.5% to 3.0%. The impact of a low savings rate on future growth is significant because consumers have no cushion, and higher prices will lead to reduced spending. In simple terms, a low savings rate borrows demand from the future, and today's consumption comes out of tomorrow's spending power.
Source: MCM and FRED
Persistent inflation and the wealth effect from the AI bubble, along with the Fed’s easy monetary policies, have had a significant effect on households. While we doubt the AI bubble is sustainable, we believe that the new Fed Chairman Warsh understands the drivers of the K-shaped economy and will pivot to restore price stability and stop accepting higher inflation as the previous Chairman Powell did.
Despite headline inflation at 3.5% and the economy essentially at full employment, the Fed has been expanding its balance sheet by $60 billion per month since December and, according to the Taylor Rule (the widely used quantitative model for the appropriate Fed funds rate given inflation and unemployment), short-term rates are at least 100 basis points too low.
Unlike his predecessors, Chairman Warsh understands that the Fed's bloated balance sheet and the rapid post-pandemic expansion of the money supply were direct contributors to the inflationary surge — not the exogenous shocks the Fed dismissed as "transitory." Warsh has stated that "inflation is a choice made by the Fed" and that "inflation occurs when the government prints excessive money — meaning the central bank and the government spend beyond their means." He has further argued that the Fed's bloated balance sheet distorts asset prices and market pricing, which favors Wall Street over Main Street.
In our view, this represents the first genuine intellectual break from the post-2008 monetary regime — a recognition that the tools the Fed used to combat the financial crisis have, through overuse, become the source of the current problem rather than the solution. We expect Warsh to shrink the balance sheet meaningfully and abandon forward guidance, which for the last decade has effectively made the Fed the marginal price-setter of every asset class. Ending that practice will reduce the Fed's distortion of markets and, more importantly, allow market-based signals — the yield curve, credit spreads, dollar strength — to inform policy rather than the rearview-mirror data on which the Fed has relied.
Since the 2008 Financial crisis, the Fed has increased its balance sheet from $800 billion to $6.7 trillion. Chair Warsh, unlike his predecessors, believes the Fed’s bloated balance sheet distorted asset prices, created inflation, and favored Wall Street over Main Street.
Source: FRED
We are optimistic that Chairman Warsh understands the Fed's role in creating the K-shaped economy, sustaining asset bubbles, and enabling persistent inflation — and that his monetary framework will lead to lower inflation, more durable economic growth, and improved living standards over the intermediate to long term. Our optimism about the Fed Chair’s strategy to fight inflation is tempered by our concern about the situation in the Middle East and its implications for the global energy market. After Iran rejected what appeared to be a favorable memorandum of understanding, the fighting has escalated with no clear off-ramp in sight. A second front now appears to be emerging: the Houthis in Yemen have threatened to impose a naval blockade on Saudi Arabia, creating a direct and significant risk to Red Sea shipping lanes.
U.S. oil inventories have declined to a 45-year low since the war began and the Strait of Hormuz was effectively closed. Despite the severity of the supply shock, oil prices have not reached the extreme levels many analysts predicted — largely because the U.S. drained its Strategic Petroleum Reserve and China significantly cut its crude oil imports. The magnitude of China's demand response is striking: June crude imports collapsed 41.3% year-over-year to just 7.12 million barrels per day, down from 12.02 million bpd in June 2025 and the lowest monthly total since October 2016. Both of these buffers are finite. The U.S. SPR is now at levels not seen since the 1980s, and China is meeting the shortfall by drawing down onshore inventories that took years to accumulate. In our view, the energy shock is not over, and if the war doesn’t end soon, oil prices will surge if either the SPR reaches operational minimums or China imports more to rebuild inventories.
Source: BofA Global Research
Compounding the pressure on energy prices is a historic dislocation in the refined product market. The war in the Middle East and the ongoing Russia-Ukraine conflict have destroyed a meaningful share of global refining capacity, and crack spreads — the cost to convert crude oil into gasoline and diesel — have surged to record levels. Even if crude prices moderate from here, refined product prices — which are what consumers and businesses actually buy — can remain elevated indefinitely as long as the refining bottleneck persists. In our view, this is a meaningful headwind for both consumers and corporate margins, and it argues for higher-for-longer inflationary burden.
In summary
The U.S. economy is more vulnerable than the headline data suggests. Three unsustainable pillars have been supporting current growth — historic deficit spending, the AI capex boom, and the wealth effect flowing from a narrow, concentrated equity rally — and each of them is now showing signs of strain.
- Despite unemployment near record lows and the stock market at all-time highs, the U.S. is running an enormous budget deficit and spending 3.8% of GDP servicing its debt. Energy prices and interest rates have moved higher because of the war in the Middle East. If short-term rates remain elevated, the U.S. interest burden will increase, and the deficit will widen further, pushing long-term rates higher and driving debt-service costs higher still — creating a self-reinforcing spiral. In our view, this dynamic will continue until the bond market, not the stock market, ultimately imposes the fiscal discipline that Washington has refused to impose on itself.
- Energy prices and interest rates have recently moved higher as the Middle East conflict broadened. The oil market has been cushioned by SPR drawdowns and China's inventory drawdown, but both buffers are finite. Record crack spreads are already pressuring consumers. Additionally, the Taylor Rule suggests the Fed is currently 100 to 200 basis points too loose, and market-implied hike odds have shifted meaningfully hawkish under Chairman Warsh.
- The AI bubble is not sustainable. The combination of deteriorating unit economics, aggressive Chinese open-source competition, hyperscalers turning FCF-negative, and a financing structure that increasingly depends on debt and equity issuance rather than operating cash flow tells us that the current pace of AI capex cannot continue indefinitely. When the capex cycle rolls over, the direct contribution to GDP growth will reverse, and the equity market rally that has generated the wealth effect will reverse with it.
We believe that Chairman Warsh understands the Fed's role in creating the K-shaped economy, sustaining asset bubbles, and enabling persistent inflation. Over the intermediate to long term, we are optimistic his monetary framework will restore price stability and rebuild the economic foundation for households across the income distribution. Near-term, however, the only way to repair the K-shaped economy is to remove the monetary conditions that created it — and restoring a healthy economy for the bottom 90% likely requires letting the asset bubbles deflate. A Fed unwilling to tolerate above-target inflation is a Fed unwilling to bail out overextended asset markets, and that is precisely why the coming transition, however necessary, will not be smooth.
Stock Market Outlook:
The stock market, in our view, still offers a poor long-term risk-reward. The S&P 500 is extremely overvalued, and concentration risk has worsened materially over the course of the quarter. The seven largest AI-related stocks now account for 32.6% of the S&P 500, and the volatile semiconductor sector alone represents nearly 20% — meaning the fate of the broad U.S. equity market is dependent on the fate of the AI mania.
Beyond valuation and concentration, the market has taken on the character of a late-cycle speculative mania. Despite the war in the Middle East and its clear adverse implications for energy prices, interest rates and global growth, investors drove the technology sector up 43.3% in the second quarter, and the semiconductor index went parabolic — surging 87.8% in just three months, its best quarterly return since the index's 1994 inception.
The mechanics behind the surge are consistent with speculation and leverage. Leveraged ETF flows have exploded, and zero-days-to-expiration options — instruments that expire at the end of the trading day — now represent more than 50% of total options trading volume. Margin debt has climbed to an all-time record. Historically, surges in margin debt have corresponded closely with major market and speculative blowoffs.
Meanwhile, corporate insiders — the group with the most direct visibility into their own companies' fundamentals — are selling at the second-fastest pace in more than twenty years. While individual investors are borrowing at record levels to buy stock, corporate insiders are selling, and the largest AI-related companies themselves are using elevated equity prices to sell shares to the public — Alphabet with its historic $90 billion equity raise, its first since the 2005 IPO, and SpaceX with its $87.5 billion IPO. It is never a good sign when retail investors are gambling with leverage, while insiders are heading for the exits.
Historically, surges in margin debt have corresponded to major market peaks. It's interesting that individual investors are borrowing at record levels to buy more stock, corporate insiders are selling at record levels, and the largest AI-related companies themselves — Alphabet with its historic $90 billion equity raise and SpaceX with its $87.5 billion IPO — have been using elevated equity prices to sell shares to the public.
Source: Investech
While investors are borrowing at a record rate to invest in the AI mania, it is interesting that the market's former leadership has quietly rolled over. The Magnificent Seven — the group that carried the S&P 500 to record highs in 2024 and 2025 — has been range-bound for nine months, and its relative strength has broken down. Importantly, this is not the healthy, broadening rotation that typically accompanies a durable bull market. Leadership has not shifted from mega-cap tech into a wider set of economically sensitive stocks; it has simply layered a second concentrated bet on top of the first. Semiconductors and the Magnificent Seven now depend on the same underlying story — that the AI capex boom will continue indefinitely at its current pace — leaving the entire market exposed to the same single point of failure. History is clear on what this pattern signals. Sharp leadership changes paired with narrow, speculative urges into a single industry are hallmarks of late-cycle bull markets, not healthy ones.
Source:Stockcharts.com
Stocks remain incredibly overvalued. Two robust long-term valuation models—Market Value to GDP and Shiller’s CAPE—indicate that the S&P 500 is projected to deliver annual returns between -0.7% and 2.1% over the next decade. Given that 30-year Treasury Inflation-Protected Securities (TIPS) currently yield a real return of 3.0%, equities are substantially overvalued and present an unattractive risk premium.
Market Value to GDP – "Still, it is probably the best single measure of where valuations stand at any given moment." –Warren Buffett, December 10, 2001. Based on market value relative to GDP, stocks are more than 165% above their historical average and more expensive than they were during the 2000 technology bubble. Over the next ten years, the model forecasts an annual return of -0.7% for the S&P 500.
Source: MCM, FRED
Shiller's CAPE (a valuation measure that smooths out cyclical earnings fluctuations) indicates that stocks are more than 2-standard deviations above their long-term average, and the 10-year expected return is about 2.1% per annum. Since the 30-year Treasury Inflation-Protected Securities (TIPS) yield a real return of 3.0%, the S&P 500 offers an inadequate risk premium.
Source: MCM, FRED
In summary
- Valuation and concentration risk have reached historic extremes. The S&P 500 is extremely overvalued on the long-term valuation measures, and its fate has become inseparable from the AI speculative mania — the seven largest AI-related stocks now represent 32.6% of the index, and semiconductors alone account for nearly 20%. Meanwhile, the market's former leadership has quietly rolled over. The Magnificent Seven has been range-bound for nine months while capital has rotated into an even narrower and more speculative semiconductor trade — the layering of one concentrated AI bet on top of another, not the broadening rotation that characterizes healthy bull markets.
- Retail investors are borrowing at record levels to speculate, while corporate insiders and the large AI-related companies themselves are selling. It is never a good sign when retail investors are gambling with leverage, and insiders are heading for the exits.
Portfolio Review:
As value investors, our asset allocation is driven by long-term valuation measures and the risk-reward opportunities present in the market. Moreover, we analyze the leading economic and market-based indicators to determine the probable path of the rate of change for economic growth and inflation. This enables us to strategically position our portfolio to perform well in all economic environments.
Portfolio Performance and Positioning
Our diversified and balanced portfolio is underweight the tech-laden S&P 500 in favor of value stocks, international equities, gold, and short-duration fixed income. This positioning had been performing well, especially on a risk-adjusted basis, before the war with Iran. Once the war began, gold plunged sharply—primarily due to a rapid rise in real yields rather than any deterioration in gold's underlying fundamentals.
The ceasefire and MOU triggered a violent reversal in market leadership. Technology stocks exploded higher, semiconductors went parabolic, and the S&P 500 returned to all-time highs on the back of an increasingly narrow group of AI-related names. Since we believed the rally was driven by speculative excess and leverage rather than sustainable fundamentals, we remained underweight technology and this positioning—along with the continued weakness in gold—cost us relative to the benchmark in the second quarter.
Gold continued its poor performance in Q2. After peaking at $5,608/oz in January on geopolitical fear and record central bank buying, the metal has fallen nearly 30% over the past six months as real yields rose, the dollar surged, and the Fed pivoted from a bias toward cuts to a bias of hiking interest rates. This is the same dynamic that drove gold lower in 2022, when the Fed's aggressive tightening pushed prices down even as inflation continued to run hot.
In our view, gold now offers a very attractive long-term risk-reward. The structural bull case remains fully intact — record peacetime budget deficits, a historically large federal debt burden, and persistent foreign central bank buying driven by de-dollarization. With long-term real yields at multi-decade highs, we expect real yields to fall from current levels, converting what has been Q2's most significant headwind for gold into a meaningful tailwind in the second half.
Source:MCM, FRED
In recent weeks, the technology sector has come under significant selling pressure as the AI narrative has come under increased scrutiny. As investors reduce their AI exposure, capital is rotating into the diversified value, mid-cap, and international positions we have long held. Our positioning is designed to benefit from this rotation while preserving capital through what we increasingly believe will be a difficult period ahead. If the AI bubble pops or Middle East tensions keep energy prices elevated — and in our view, both are more likely than not — the U.S. economy will be materially more vulnerable, and capital preservation will matter far more than participation in the last leg of a narrow, leverage-driven rally.
In Summary
We believe that we are in a very high-risk environment, and the risk-reward in U.S. equities is poor. Stocks are effectively priced for perfection, with investors crowding into AI-related names while largely ignoring the potential damage an oil shock or a bursting of the AI bubble can inflict on an already fragile economy. Rather than attempting to predict what will happen, we rely on our value-driven investment framework that lets risk premiums and position sizing — not forecasts — govern our risk exposure.
Today, we are underweight equities because their long-term risk-reward is unattractive, and the recent rally's narrow breadth reinforces our caution. In a regime of modest growth and rising inflation, we prefer assets that have historically performed well in similar inflationary environments:
- International and emerging-market equities, which benefit from a weakening U.S. dollar and offer lower valuations than U.S. large-cap stocks.
- Value equities, which historically outperform in inflationary regimes and are trading at their widest discount to growth in more than two decades.
- Hard assets — gold and select commodities — as direct inflation and weak-dollar hedges
- Short-term fixed income, which mitigates interest-rate risk while capturing attractive real yields at the front end of the curve.
We plan to maintain a highly diversified and balanced asset allocation until economic uncertainty diminishes and equities present a more favorable risk-reward opportunity
