Dear Investor
General Market Overview
Global markets were broadly resilient over the quarter, despite an ongoing Middle East conflict and an inflationary backdrop. Much of the early strength was driven by artificial intelligence (AI) and semiconductor enthusiasm. That optimism was tempered, however, as semiconductor shares fell sharply from their highs, erasing over a trillion dollars in chip stock value in a matter of weeks, with the weakness spreading from Asian suppliers through to the major United States (US) names.
One recurring concern is that semiconductor stocks now make up close to 20% of the S&P 500 by weight, more than double their roughly 8% share at the peak of the dot-com bubble in 2000, which was followed by an 85% collapse in the sector. That comparison is being used as a warning by some and dismissed by others as a poor analogy; however, it captures why a retracement in a
handful of AI-exposed names was able to move the entire market.
Interestingly, the fall in share prices was not due to a lack of demand for their products. Second-quarter earnings were exceptionally strong, with 86% of all S&P 500 companies beating earnings estimates. The market’s concern is not that AI demand has disappeared, but that valuations have run ahead of it, and that the enormous capital expenditure underpinning the boom is starting to overwhelm the cash flow being generated in return.

The result was a significant rotation rather than a broad market decline, with energy markets rising amid continued Middle East tensions. While some Middle East news no longer dominates headlines, the Strait of Hormuz remains only partially open to commercial traffic. Iran has continued to strike ships attempting to bypass its traffic system, with at least two vessels hit on the Omani route since early August. Roughly 65 vessels that entered during the recent lull remain unable to exit, on top of more than 70 that have been stranded since the conflict began. Brent crude is below its initial high of $120 per barrel and is currently around $90–$95, but is significantly higher than $60 at the start of the year.
At his second meeting as Federal Reserve (Fed) Chair, Warsh’s committee held rates at 3.50–3.75%, but with a 9-3 dissent among voting and non-voting members. Warsh’s own language at that meeting left little doubt about where he intends to plant his flag on inflation: “There is no soft implicit target, not on this committee’s watch.” He also acknowledged that years of elevated inflation cannot be resolved in a matter of weeks. He signalled a preference for markets to react to incoming data in real time rather than to lean on Fed forward guidance, a deliberate change in communication style from his predecessor.
The Fed’s caution looks less unusual once set against its global peers. In the same week in June that the Fed held rates, the European Central Bank (ECB) raised its deposit rate to 2.25%, its first hike since 2023, as eurozone inflation reached 3% on the back of energy costs linked to the Middle East. The Bank of Japan (BoJ) went further still, raising its policy rate to 1% in mid-June, the highest level since 1995, explicitly to defend the yen and contain imported inflation. The Bank of England (BoE), like the Fed, also held its base rate at 3.75%, with Governor Bailey saying the institution was “in no rush to raise interest rates” amid weak growth. This period marks a change in the broadly synchronised global easing cycle among the major central banks.

Globally, inflation and long-term borrowing costs are making life difficult for central banks. US Consumer Price Index (CPI) was 3.4% year-over-year (y/y) in July, which is above the Fed’s 2% target. At the same time, the long end of the US yield curve rose to levels not seen in almost two decades, with the 30-year Treasury yield reaching 5.27%, its highest level since June 2007.

It is interesting, given the capex flows (largely through borrowing) that have poured into the AI/semiconductor space, and given increasingly higher long-term yields, to examine what fixed income markets are signalling about inflation and interest rate expectations. The US Treasury, when issuing bonds, can issue either “regular” instruments that pay a fixed coupon agreed at issuance or TIPS (Treasury Inflation-Protected Securities). A TIPS instrument pays a lower fixed rate but makes inflation adjustments based on CPI. If inflation rises, the coupon from a regular bond erodes in real terms while the TIPS payout keeps pace with inflation.
If the Treasury issues a 10-year regular bond and a 10-year TIPS simultaneously, then by examining the difference in the yield on those two instruments, it is possible to determine what “the market” expects inflation to do over the life of the bond. This is termed breakeven inflation because, at that rate, an investor would be indifferent between buying a TIPS and a non-inflation-adjusted bond.

Currently, the breakeven rate is about 2.3% over 10 years, implying the market is not anticipating runaway inflation and is broadly in line with the Fed target. The second fixed-income concept to examine is the term premium.
If an investor wants to invest in a 10-year fixed-income instrument, they can either buy one with a 10-year maturity, or, each year for 10 years, roll (reinvest) a one-year instrument into another one-year instrument. Buying a 10-year term instrument locks in a rate today but exposes the investor to a long-dated asset whose price will fall if rates rise, and whose real value will be eroded if inflation increases. Rolling for 10 years carries no view on long-term rates but exposes the investor to whatever short-term rates prevail in each future year.
The term premium is the extra yield the investor demands to accept long-dated risks rather than simply rolling short-term paper. For most of the decade following the financial crisis, that premium was negligible and at times negative, meaning investors were willing to accept a path of short-term rates below expectations to hold long bonds. Central bank bond-buying programmes and persistent safe-haven demand had made long-duration investing something investors were willing to pay for. Now, however, the term premium is climbing.

The reason this decomposition between breakeven rate and term premium matters is that it separates two very different explanations for why long-term borrowing costs have risen. If breakeven rates were widening, the market would be signalling it had lost confidence in the inflation outlook. That is not happening; breakeven rates are low. Therefore, the increase in long-term yields is occurring almost entirely in the real yield and in the term premium.
Increasing term premium means the market is demanding a materially higher real return for parting with capital over long periods, at a moment when artificial intelligence infrastructure, rearmament programmes, energy security investment, and a US federal debt now past $40 trillion are all competing for the same pool of global savings. The change in term premium, not breakeven rates, suggests a signal about the scarcity of capital rather than a loss of price stability.
General Conclusion
The rotation out of AI-exposed names over the quarter and into more value-oriented sectors shows that the market is no longer “just” looking at current earnings. The real test for the growth sectors of the economy will be the interpretation of December earnings and growth numbers from the big names and whether the continued heavy capex requirement to fuel that growth can be justified.
With a widening term premium in the fixed-income market, even with inflation expectations anchored, there is a suggestion that competition for capital to fund AI infrastructure, rearmament, and a federal debt load above $40 trillion will not be solvable by Federal Reserve decisions alone. The outcome of the Fed’s September meeting will be critical and a credibility test for Warsh’s new reign.
The Middle East also remains an unknown, and while the Strait is partially reopened, there is still no resolution. Further escalation remains a possibility that would move energy markets and inflation expectations quite rapidly.




Comments