Fluid and Dynamic

Last week, we highlighted the more favorable Consumer Price Index (CPI) and Producer Price Index (PPI) data reported by the Bureau of Labor Statistics, noting that CPI fell the most in one month since COVID’s peak and that lower energy prices accounted for more than half of the improvement in specific PPI components.^1,2 Readers may recall that while we were pleased to see some progress in both measures, Middle East instability may prove June’s inflation data ephemeral. The bond market, which we will cover in more detail below, seemed to agree with our skepticism, with yields extending to their highest levels in the past year.

Energy markets demonstrate a few key properties. First, Brent Crude, the global oil benchmark, peaked at $126.41/barrel in April but fell to just over $70/barrel in early July. On Thursday, it touched $102/barrel on news that Houthi rebels claimed they closed the Bab el-Mandeb Strait, a Red Sea passageway that sees 12% of global trade and 25% of global container traffic pass through its waters.³

Second, more regional measures, including Dated Brent Crude (which tracks North Sea oil prices) and UK and German natural gas, continue to display local market energy risks. UK and German natural gas proxies are now at their highest levels since the Iran War began, and Dated Brent remains 67% above where it stood at the year’s start. Markets continue to express concerns about both global trade tensions and the lingering tensions between the U.S. and Iran, which were exacerbated even before nuclear talks began in earnest. Markets affix a 33% chance that the U.S. Federal Reserve will raise interest rates next week, but pressures may mount should energy cost persistence endure.

Our working hypothesis on AI remains that we have migrated away from markets asking the binary question of whether a company had an AI strategy to investors increasing scrutiny on returns on invested capital and free cash flow generation. The AI ecosystem is extremely complicated across layers, including infrastructure, data, model development, applications and governance, and each of these layers interacts with the others to shape outcomes.⁴ Chief financial and chief technology officers are weighing cost and return considerations, and change is a daily phenomenon. Within the model development layer, as we highlighted last week, new entrants including Moonshot’s open-source Kimi K3 can instantaneously change dynamics around model preferences and potential spend.

We are still early in the technology earnings season, but we did have two significant technology companies, Alphabet (formerly Google) and Intel, report their second quarter results this week. Capital expenditure guidance was a common thread for both companies.  In discussing forward guidance, Intel CFO Dave Zinsner noted “Due to strong customer demand signals, we are raising our outlook for 2026 and now expect our cap ex to be more than $20 billion, which is up significantly versus our expectations entering the year.”⁵

Alphabet’s capital expenditure guidance also increased when they reported earnings on Wednesday. They guided analysts to a midpoint of $200 billion for 2026 from their prior $185 billion for the year, and CFO Anat Ashkenazi highlighted that “we continue to expect our cap ex to increase significantly in 2027, and we'll provide more details at a later date.”⁶ Further, the company noted they produced negative free cash flow in the quarter due to their technical infrastructure capital expenditures to support AI.⁷

Both stocks traded down following their earnings releases. Investors acknowledge the current demand construct, but open questions on total spend for next year coupled with cash flows dipping into negative territory can raise eyebrows. Our interpretation is that overbuild risks are growing, and while we view AI as a transformative technology, capital markets will not write checks unsupported by returns. Applying the adage “nothing cures high prices like high prices,” seeing newer entrants and competing technology may increase AI adoption, but it may also place return constraints on certain layers of the ecosystem.

First, we note that bond market volatility has not been contained within the U.S. With a global lens, 10-year government bonds in the U.S., Mexico, the UK, most of continental Europe, Japan, Australia and South Korea all sit at the upper end of their one-year range. A common refrain from central banks across those regions is balancing policy responses to energy price movements. Further, corporate bond issuance has been high, and increased supply tends to coincide with higher yields. U.S. investment grade, high yield and convertible bond issuance through June totaled $1.5 trillion, up 28% year-over-year.¹⁰ M&A activity, AI infrastructure spending and general refinancing activity have all provided issuance tailwinds.

The level we continue to focus on for U.S. government bonds is 5% for 10-year maturities. 10-year Treasury notes have not exceeded 5% in 18 calendar years, and that level has significance for everything from consumer mortgages to credit card rates to commercial real estate to how foreign lenders set their interest baselines. It is worth highlighting that both 2-year and 30-year U.S. yields have exceeded 5% in recent years. When the Fed became more aggressive with interest rate policy, 2-year rates touched 5% in 2022, 2023 and 2024, but have been below that threshold for the past two years. 30-year bonds traded above 5% for 29 days this year through Friday, representing the longest streak since 2007. Breaching 5% for the 10-year yield would be more than symbolic, and how government bond auctions perform as well as how commercial real estate liquidity responds will be important variables in our assessment.

So far, earnings season remains extremely strong. With 25% of S&P 500 companies reporting earnings through Friday morning, sales growth is up 13% and earnings growth up a whopping 70%, though the results are skewed by a few companies and sectors. We are furthest along among financial services companies with half of the index constituents reporting, but technology, energy, consumer staples and materials are all still very early days. Sales surprises have been more subdued with only 3% growth outside of expectations, but earnings surprises within the early communications reporting companies skew that figure.¹¹

As we have shared in prior publications, earnings expectations continue to ramp higher. Two sectors with strong tailwinds include technology and energy, but they offer idiosyncratic storylines. Energy companies had been punished for over-investing in infrastructure and chasing high prices. This was a near-Pavlovian relationship between higher spot oil prices that often drew the industry into pricey fixed-cost discovery projects, only for more supply to come online and drive prices lower, impacting returns on investments. As highlighted in the earlier section on Intel and Alphabet, one can see a potential analogous situation emerging: strong demand for AI ecosystem build draws in more suppliers, and the potential for lower returns to emerge increases. Further, demand may also soften as companies manage their tech and model spend in unique ways. We still see AI as the dominant earnings theme, but we do not want to lose sight of consumer health, which we will continue to monitor through retail, credit card and airline earnings reports, along with the usual smattering of macro data from around the globe. 

¹ U.S. Bureau of Labor Statistics. “Consumer Price Index Summary.” Economic News Release. July 14, 2026. https://www.bls.gov/news.release/cpi.nr0.htm. Accessed 24 July 2026.

² U.S. Bureau of Labor Statistics. “Producer Price Index Summary.” Economic News Release. July 15, 2026. https://www.bls.gov/news.release/ppi.nr0.htm. Accessed 24 July 2026.

³ Khaled, Fatma and Al-Haj, Ahmed. “A New Threat by Yemen’s Houthis Could Widen the Iran War and Put Another Trade Chokepoint at Risk.” The Associated Press. July 23, 2026. https://apnews.com/article/yemen-saudi-houthis-attack-shipping-red-sea-iran-6ee98d611669dc84953d2f2e15958bf9. Accessed 24 July 2026.

⁴ Stryker, Cole. “What Is an AI Stack?” IBM Think. https://www.ibm.com/think/topics/ai-stack. Accessed 24 July 2026.

⁵ Intel Corporation. “Comments from CEO Lip-Bu Tan and CFO Dave Zinsner.” CEO/CFO Earnings Call Comments. July 23, 2026. https://www.intc.com/news-events/press-releases/detail/1776/intel-reports-second-quarter-2026-financial-results. Accessed 24 July 2026.

⁶ Alphabet Investor Relations. “2026 Q2 Earnings Call.” July 22, 2026. https://abc.xyz/investor/events/event-details/2026/2026-Q2-Earnings-Call-2026-GgTAq7Is0z/default.aspx. Accessed 24 July 2026.

⁷ Ibid

⁸ Northern Trust Wealth Management Research, Bloomberg data.

⁹ Ibid

¹⁰ SIFMA. “US Corporate Bond Statistics.” SIFMA Research and Statistics, July 17, 2026. https://www.sifma.org/research/statistics/us-corporate-bonds-statistics. Accessed 24 July 2026. 

¹¹ Northern Trust Wealth Management Research, Bloomberg. Bloomberg data accessed on terminal, July 24, 2026.