Why strong US growth could not stop the bond rally and AI selloff

The US economy grew at a 3.4% annual rate, yet Treasury yields fell and the Nasdaq ended October with its worst monthly performance since March 2025. The market is no longer rewarding capital expenditure simply because it supports growth. Investors are testing whether that expenditure can produce cash flow and returns above the cost of capital.

The market event

The US economy expanded at a 3.4% annualised rate in the third quarter of 2026, up from 2.2% in the second quarter and above the 3.2% consensus forecast. Real final sales to private domestic purchasers, which combine consumer spending and private fixed investment, increased 4.9%. The Bureau of Economic Analysis attributed the acceleration mainly to stronger investment and an upturn in government spending, while consumer spending also increased. Within equipment investment, information-processing equipment was the largest contributor.

The Federal Reserve had already acknowledged the same strength. On 28 October, it held the federal funds target at 3.75% to 4.00% by a 10 to 2 vote, describing economic activity as strong, domestic spending as resilient and capital investment as robust. Governors Lisa Cook and Beth Hammack preferred a 25-basis-point increase. The hold therefore reflected a decision to gather more information, not a declaration that inflation or growth risks had disappeared.

The market response was unusual for a positive growth surprise. On 30 October, the Nasdaq fell 2.86%, the S&P 500 declined 1.82% and the semiconductor index lost 7.2%. The Nasdaq ended October with its worst monthly performance since March 2025. The US Treasury's official par-yield curve placed the ten-year yield at 4.94%, down 35 basis points from 5.29% on 30 September, while the two-year yield ended at 4.56%. The bond rally therefore survived both the strong GDP release and the Federal Reserve's hawkish dissents.

Corporate results explained the change in emphasis. Alphabet reported 25% revenue growth to $127.6 billion and a 30% increase in operating income, while Google Cloud revenue more than doubled to $30.7 billion. Yet third-quarter capital expenditure reached $51.4 billion and free cash flow was negative $9.1 billion. Alphabet shares fell 12.6% on 29 October, erasing about $500 billion of market value. Amazon reported 12% sales growth and 39% growth at AWS, but trailing twelve-month free cash flow moved to a $42.4 billion outflow from a $14.8 billion inflow a year earlier. Its expected 2026 cash capital expenditure remains approximately $220 billion.

The market did not reject the GDP data. It questioned the quality and price of the investment supporting part of that growth. GDP records expenditure when equipment and infrastructure are built. Equity valuation depends on the cash flows that follow, their timing and the return earned on the capital deployed. A data centre can add to current output before investors know whether future revenue will justify the servers, power, inventories, depreciation and financing required to operate it.

The shift from financing capacity to return on capital

The largest technology groups still have substantial revenues, operating profits and access to capital. The issue is not an immediate inability to fund investment. Alphabet generated $42.3 billion of operating cash flow during the quarter, but spent $51.4 billion on property and equipment. It also issued $28.7 billion of senior unsecured notes during the quarter. Amazon generated $168.3 billion of operating cash flow over twelve months, but its investment in property and equipment was large enough to turn free cash flow deeply negative.

The expenditure is producing measurable activity. Google Cloud revenue increased 103%, its operating income reached $11.1 billion and the company reported a $561 billion cloud backlog. AWS sales rose 39% and its quarterly operating income increased to $17.8 billion. These figures prevent a simple conclusion that the investment is being wasted. They also explain why management teams continue to spend. The unresolved question is whether revenue conversion, margins and asset utilisation can rise fast enough to restore free cash flow.

This is a change in the market’s test. Earlier in the investment cycle, access to chips, power and financing signalled competitive strength. The latest results make cash conversion and return on invested capital more important. Investors now need to distinguish expenditure that expands a profitable service from expenditure required merely to defend an existing franchise.

Negative free cash flow during an investment cycle is not proof of value destruction. Capital expenditure occurs before the associated revenue, and inventory can temporarily absorb cash. The warning signal becomes stronger when spending forecasts keep rising while revenue expectations, margins or backlog conversion fail to improve. The relevant comparison is therefore the prospective return on new capital against the company’s cost of capital, not the absolute size of the investment.

Implications across markets

For currencies, strong US growth would normally support the dollar through higher expected interest rates. That channel weakened when Treasury yields fell and investors reduced rate-hike expectations. The yen and Swiss franc initially gained as the technology selloff increased demand for liquidity and portfolio protection. The new week adds a further question: if the shock originates in US technology and investment expectations, the dollar may not receive the same safe-haven support as it would during an external crisis. Dollar traders should therefore separate the growth signal from the yield and risk signals.

For interest rates and government bonds, October showed that duration can still hedge an equity shock even when reported growth remains strong. The rally also implies that investors see a risk of slower future investment or expect the Federal Reserve to respond cautiously to tighter market conditions. The bullish bond case would weaken if productivity and corporate revenue accelerate enough to validate the capital spending and revive expectations of further tightening.

For equities, the relevant divide is no longer simply technology against the rest of the market. Suppliers of scarce computing capacity can benefit while platform owners absorb the cost, but that advantage depends on customer spending remaining durable. Among the platforms, positive free cash flow, improving margins and visible revenue conversion deserve a premium. A broad equity decline would signal a macroeconomic problem; concentrated weakness in capital-intensive technology points first to valuation and return-on-capital risk.

For commodities, data-centre construction supports demand for copper, electricity, natural gas and specialised equipment. Continued investment can keep those markets tight even if technology shares fall, while a material reduction in capital budgets would affect industrial metals before it appears in national GDP. Oil remains driven primarily by geopolitics rather than the AI investment cycle: prices fell sharply on 2 November as renewed diplomatic hopes reduced part of the Middle East supply premium. Traders should keep the infrastructure-demand thesis separate from fast-changing energy-supply risk.

Applying the DCRADO Principle

The D.C.R.A.D.O.™ Principle converts the gap between reported growth and market confidence into a conditional trading process. The objective is to identify whether the market is repricing one company’s execution, the return on a sector’s capital or the outlook for the wider economy.

Define: Specify the decision. Is the position based on continued technology investment, improving monetisation, lower bond yields or a broader risk-off move? A long-duration bond trade and a short technology-equity trade may react similarly for a few sessions, but they depend on different evidence.

Contextualize: Place the move inside strong GDP, the Federal Reserve's hold, volatile energy prices and a week containing the US midterm elections, further mega-cap earnings and the October employment report. Falling yields after strong data carry more information when they coincide with negative cash-flow surprises and weaker technology breadth.

Retrieve: Track capital-expenditure guidance, operating and free cash flow, cloud backlog and its conversion, margin trends, earnings revisions, credit spreads, Treasury yields, equity breadth, the dollar, yen and Swiss franc, and prices for copper, power and natural gas. This week's remaining earnings and employment data provide immediate tests of whether the repricing is company-specific or becoming macroeconomic.

Analyze: Use three scenarios. In successful monetisation, revenue and margins catch up with spending, free cash flow improves and technology leadership resumes. In a return-on-capital repricing, investment remains high but platform valuations fall and bonds outperform. In a macro spillover, capital budgets are cut, credit spreads widen, commodity demand weakens and safe-haven currencies strengthen further.

Decide: For a bullish technology position, require evidence that cash conversion or margins are improving, not only that revenue is growing. For a duration trade, require weaker rate expectations and continued equity stress. Avoid extrapolating one earnings shock across the entire economy. Use smaller size while price gaps remain large and invalidate the trade when the confirming cash-flow, yield or breadth indicators reverse.

Oversee: Review quarterly capital budgets, free-cash-flow revisions, backlog conversion, debt issuance, the Treasury curve and market breadth. Reduce exposure when the original thesis becomes dependent on one company’s guidance. Oversight should test whether a sector repricing is remaining contained or becoming a broader contraction in investment and demand.

Conclusion

The third-quarter GDP report confirmed that the US economy remains strong. Friday's technology selloff and the completed October bond rally did not contradict that evidence. They showed that markets assess a different question: whether today's exceptional investment can generate tomorrow's acceptable return. Strong revenue and large backlogs support the investment case, while negative free cash flow and rapidly rising capital budgets increase the burden of proof.

The modest improvement in US equity futures and the flat European session on 2 November indicate stabilisation, not resolution. For traders, the next signal will not come from GDP alone. It will come from the relationship between capital expenditure, cash conversion, margins and the cost of capital. If those variables improve together, the selloff can become an opportunity. If spending continues to rise while cash generation and earnings expectations deteriorate, lower yields may be warning that the investment cycle will slow before its promised productivity gains arrive.