Economic Update
Despite persistent inflation, higher borrowing costs, geopolitical uncertainty, tariffs and a slowing labor market, the US economy has demonstrated renewed momentum in recent months. The most recent evidence of strengthening business is the government's advance estimate of third-quarter GDP, which showed real economic growth at a 3.4% annualized rate. This followed growth of 2.5% in the first quarter and 2.2% in the second. While the quarterly progression was not uninterrupted, the third-quarter acceleration is encouraging. More importantly, the underlying components of domestic demand suggest that improvement is not simply a statistical consequence of inventory accumulation or foreign trade. Consumer spending overall remains resilient, and business investment has become an increasingly important source of economic growth. The latter is particularly noteworthy because it is occurring despite higher interest rates and includes an extraordinary commitment of capital to artificial intelligence (AI) infrastructure. There are, however, important qualifications. The labor market has lost some momentum, inflation remains above the Federal Reserve's objective, housing continues to struggle, and the economy's growing dependence on AI-related investment introduces risks that extend well beyond the technology sector.
A key question at this point is whether the current expansion has sufficient breadth and underlying strength to withstand these headwinds. The evidence increasingly suggests that it does, although the concentration of investment in AI deserves particular scrutiny.
One particularly useful measure of underlying economic activity is real final sales to private domestic purchasers, which combines consumer spending and private fixed investment. This measure increased 4.9% in Q3, following 4.6% in Q2. It is encouraging that private domestic demand has remained strong even as other components of GDP have fluctuated. Consumer spending increased at a 3.8% annual rate, while business investment, excluding residential construction, increased 12.7%. Investment in information-processing equipment surged 37.4%, reflecting the extraordinary spending on computers and related equipment associated in part with AI. Inventories also contributed to GDP growth, particularly in wholesale trade and manufacturing. Government spending increased, while exports declined and imports increased. Since imports are subtracted when calculating GDP, the increase in imports partially offset the contribution from domestic activity. The inventory contribution deserves some caution: inventory accumulation can temporarily boost GDP without representing an equivalent increase in final demand. Nevertheless, the 4.9% increase in private domestic final sales provides additional evidence that the expansion is supported by underlying consumer and business activity. The distinction is important. The economy is not simply benefiting from a temporary statistical boost; private demand is growing at a healthy pace.
One of the strongest confirmations of the economy's underlying health is the exceptional performance of corporate earnings. Third-quarter reporting season has been impressive, with approximately 60% of S&P 500 companies having reported by October 30. According to FactSet, the blended earnings growth rate for the index has reached 30.2% year over year, up from 29.1% when the quarter ended. Moreover, 82% of reporting companies have exceeded analysts' earnings expectations, while all eleven S&P 500 sectors are reporting year-over-year profit growth. Revenue growth is also strong, currently estimated at 13.4%. These results are particularly significant because they demonstrate that economic strength is translating into actual corporate profitability, rather than simply higher sales or increased investment spending.
Although a disproportionate share of earnings growth remains concentrated in technology, energy and AI-related businesses, the breadth of positive results is encouraging. Strong profits provide businesses with the internal resources to invest, hire, strengthen balance sheets and return capital to shareholders. They are an important underpinning of the expansion.
Consumers and businesses: Engines of Growth
Consumer spending, which accounts for over two thirds of GDP, continues to provide substantial support. September retail sales increased 0.8%, following a revised 1.3% increase in August. However, the September gain was heavily influenced by gasoline sales, which rose sharply as energy prices increased. Sales excluding gasoline stations were considerably less impressive. The September personal income and consumption report provides a more nuanced picture. Personal income increased 0.2%, while current-dollar personal consumption expenditures increased 0.5%. After adjusting for inflation, however, real consumption was essentially unchanged. Real disposable income declined 0.3%, and the personal saving rate fell to 3.8%. This is worth watching. Consumers are still spending, but the purchasing-power cushion is becoming thinner. Households with substantial financial assets and higher incomes remain better positioned to absorb rising prices than families dependent on wages and credit. This economic bifurcation helps explain why aggregate consumption can remain strong even when consumer sentiment is considerably less favorable.
Business investment, about 14% of GDP, is the more impressive part of the expansion narrative. September industrial production increased 0.7%, and manufacturing activity has continued to improve. The September ISM manufacturing index registered 54.5, its ninth consecutive month above the 50 threshold indicating expansion. The services index registered 54.9, its 27th consecutive month in expansion territory. The October flash S&P Global composite PMI registered 57.6, indicating continued expansion across manufacturing and services, although somewhat slower than September. The survey also identified renewed cost pressures, particularly in energy and other business inputs. While these surveys are not GDP measurements, their consistency with the official economic data adds credibility to the view that the economy entered the fourth quarter with considerable momentum.
Employment, Inflation and Housing
The labor market is not providing the same degree of reassurance of robust growth ahead as consumer spending and business investment. September payroll employment increased by only 29,000, while the unemployment rate remained at 4.2%. Employment across most major industries has changed only modestly. The preliminary benchmark revision to March payroll employment was also negative, indicating that earlier employment estimates had overstated job creation. An additional concern is real compensation. Median weekly earnings for full-time workers increased only 1.5% over the year in the third quarter, compared with a 3.5% increase in consumer prices. This suggests that the purchasing power of many working households has been under pressure. Initial unemployment claims remain relatively low, indicating that widespread layoffs which usually precede a broad business downturn have not materialized. The more important concern is that employers appear reluctant to add workers aggressively. This is consistent with an economy in which businesses are expanding output through capital investment and productivity improvements rather than substantial increases in employment.
Inflation remains the principal obstacle to a more favorable outlook. September CPI increased 0.6% for the month and 3.7% over the year. Gasoline prices accounted for more than half of the monthly increase. The September PCE price index increased 0.5% monthly and 3.6% over the year, while core PCE increased 0.3% monthly and 3.1% annually. Although the GDP report's quarterly annualized PCE inflation measures were lower, the monthly inflation data reinforces the conclusion that the Fed still has work to do.
Housing remains a weak spot. Higher mortgage rates have discouraged home purchases and residential construction. Housing investment declined in Q3, extending a period of weakness in an industry particularly sensitive to borrowing costs. The contrast is becoming clearer: the economy's strongest areas are consumer spending, services and business investment, while housing and employment are less dynamic.
Federal Reserve Policy
In its statement following last week’s FOMC meeting, where the Fed decided to leave its reference rate unchanged, officials described economic activity as expanding at a strong pace, characterized domestic spending as resilient, and identified productivity growth and capital investment as robust. At the same time, the rate-setting committee acknowledged that inflation remains elevated and reiterated its commitment to returning inflation to its 2% objective. Chair Kevin Warsh's press conference was particularly important for what it did not promise. He did not signal that rate increases were imminent, nor did he suggest that the Committee was prepared to begin reducing rates. Instead, he emphasized the need for additional economic information before determining the appropriate next step. Warsh observed that economic growth had strengthened, that productivity appeared to be improving and that capital investment was robust. He also cautioned against assuming that faster economic growth necessarily meant the economy was overheating or that higher interest rates were automatically required. His stated objective was not to throw sand in the gears of the economy or unnecessarily put people out of work.
The decision and chairman’s remarks reflect a difficult policy environment. Consumer spending remains resilient and business investment is expanding rapidly, yet inflation remains well above the Fed's target. The Committee must determine whether recent price increases, particularly those associated with energy and supply pressures, will prove temporary or spread into broader inflation. The timing of monetary policy's effects complicates that judgment. Financial markets respond almost immediately to changes in the expected path of interest rates. Mortgage rates, corporate borrowing costs and bank lending conditions adjust over days and months. Decisions involving housing, consumer purchases, capital expenditures and hiring generally respond over several quarters. The full effects on economic output and inflation can take a year or longer to become apparent. Consequently, the economy may not yet have absorbed the full impact of earlier tightening. If rates remain elevated or increase further, the additional restraint could become more evident during 2027, particularly in housing, commercial real estate and debt-financed business investment.
For investors, the message is that the direction of monetary policy remains uncertain. A pause should not be interpreted as a commitment to easier money. The Fed has retained the flexibility to tighten further if inflation persists, even at the risk of slowing an economy that is currently performing better than many anticipated.
Artificial Intelligence: Growing Financial Vulnerability
No assessment of October's economic and financial developments would be complete without a substantial discussion of artificial intelligence. The market reaction to the latest earnings announcements from Alphabet and Amazon has brought an increasingly important question into focus: how much capital can be committed to AI infrastructure before investors begin to question whether the returns will justify the expenditures? This is no longer simply a technology-sector issue. AI investment has become an important contributor to economic growth, corporate profits, employment, construction, electricity demand and financial-market valuations. Its continued expansion is therefore relevant to the durability of the broader economy.
In an extensive discussion of AI a few months ago, we identified in an Economic Update several risks: the extraordinary scale of capital commitments, the circularity of its financing, the possibility of overbuilding, the availability of electricity and supporting infrastructure, the uncertainty surrounding investment returns, and the potential consequences for financial markets if expectations prove excessive. While developments since then have made several of these concerns more pressing, we place particular emphasis on one risk that deserves increasing attention: the circularity of AI financing.
Circular Financing
Of all the risks associated with the AI investment boom, we believe the financing structure deserves the greatest scrutiny. The concern is not simply that companies are spending too much. It is that some financing arrangements may create an appearance of demand and revenue that is more dependent on the investment cycle itself than on independent end-user demand. Consider a simplified example. A large technology company invests billions of dollars in an AI developer or a smaller cloud-computing provider. That recipient then commits to purchasing computing capacity, chips or cloud services from the original investor or one of its partners. The technology company records revenue from those purchases. The resulting revenue and contractual commitments help justify further investment, higher valuations and additional borrowing. The capital has effectively moved through a circle. The investment helps finance the customer, and the customer uses that financing to purchase services from the investor or its affiliates. This does not mean that every such transaction is artificial or improper. The underlying services may be valuable, and the contracts may be commercially legitimate. The problem is that apparent demand may not be entirely independent of the financing supporting it.
The Bank for International Settlements (BIS) has identified these arrangements as a financial stability concern. Investments by hyperscalers and chipmakers in AI laboratories and neocloud providers can be accompanied by multi-year purchase commitments. In other arrangements, third-party developers finance data centers and lease them to hyperscalers. These structures can create complicated financial relationships, potentially involving overlapping collateral, concentrated exposures and obligations that are not always readily apparent to outside investors. The potential consequences are serious. If the ultimate demand for AI services proves insufficient to support the promised purchases, the AI developer may be unable to meet its commitments. The cloud provider loses expected revenue, the infrastructure owner may be unable to service its debt, and the original investor may suffer losses on both its equity investment and its commercial relationship. A problem that initially appears confined to a few technology companies can spread into private credit, corporate bonds, banks and other financial institutions.
The growing involvement of private-credit funds in financing AI infrastructure adds another dimension. These lenders frequently finance projects and businesses that have limited access to traditional public bond markets. Their exposures are generally less transparent, and the underlying loans may be difficult to value or sell during a period of financial stress. The danger is not that AI will necessarily fail. It is that a genuine technological revolution can still produce an excessive and potentially unstable financing boom.
Extraordinary Scale of Investment
The magnitude of the investment underway is difficult to overstate. The largest technology companies are committing hundreds of billions of dollars annually to data centers, advanced semiconductors, networking equipment and the electrical infrastructure required to support AI computing. AI spending extends well beyond the companies developing advanced models. Utilities are expanding generating capacity, construction companies are building specialized facilities, semiconductor manufacturers are expanding production, and equipment manufacturers are receiving substantial orders. These investments have created a powerful multiplier effect throughout the economy. It helps explain why business investment has become such an important contributor to GDP growth even as traditional interest-sensitive industries remain relatively weak. In conversations with clients, we have emphasized this point: AI infrastructure spending is generating a broader investment cycle involving manufacturers, utilities, data-center operators and suppliers. It is, in that respect, helping to support the manufacturing renaissance we have discussed previously. But there is a distinction between economic activity generated by an investment and the eventual economic return on that investment. Construction expenditures contribute to GDP today. The facilities must still generate sufficient revenue and profits in the years ahead to justify the capital committed. That distinction is becoming increasingly important.
The latest quarterly reports from Alphabet and Amazon illustrate both the promise and the financial demands of the AI revolution. Alphabet reported third-quarter revenue of $127.6 billion, an increase of 25% over the previous year. Google Cloud revenue increased an extraordinary 103%, reaching $30.7 billion. This is compelling evidence that demand for cloud computing and AI infrastructure is real. But capital expenditures reached approximately $51.4 billion during the quarter. The company's cash requirements are expanding rapidly as it builds the infrastructure needed to support future demand. Its substantial cash reserves provide considerable financial flexibility, but investors are increasingly interested in how quickly these expenditures will translate into sustainable free cash flow.
Meanwhile, Amazon reported third-quarter revenue of $201.9 billion, while AWS revenue increased 39% to $46 billion, its fastest growth in nearly five years. AWS operating income reached $17.8 billion, demonstrating that cloud computing remains a highly profitable business. Yet Amazon's trailing twelve-month free cash flow turned negative by $42.4 billion, primarily reflecting the substantial increase in property and equipment purchases associated with AI investment. Operating cash flow, meanwhile, increased 29% to $168.3 billion. Amazon expects 2026 capital expenditures of approximately $220 billion.
These results illustrate the central dilemma. Demand is growing rapidly, and the leading companies are generating substantial operating cash flow. But the cost of expanding capacity is increasing even faster in some cases. Investors are beginning to demand more evidence that the extraordinary capital commitments will generate commensurate returns.
Physical Constraints
A second major concern is physical rather than financial. AI computing requires extraordinary amounts of electricity. Data centers require reliable power around the clock, sophisticated cooling systems, transmission capacity and specialized equipment. The expansion of AI infrastructure is occurring faster than some parts of the electrical system can accommodate. The difficulties extend beyond generating electricity. They include delays in connecting new facilities to the grid, insufficient transmission capacity, shortages of transformers, permitting delays and the time required to construct new power facilities. Water availability and cooling requirements also matter. In addition, shortages of advanced memory chips, semiconductors and networking equipment can delay projects and increase costs. These constraints may ultimately limit the speed of AI expansion. They can also encourage companies to commit to long-term power and equipment contracts before they know precisely how much computing demand will materialize.
There is a broader economic consequence as well. If electricity demand grows faster than supply, higher energy costs could affect other businesses and households, adding another source of inflationary pressure. At the same time, investments in generation, transmission, electrical equipment and cooling infrastructure create opportunities for utilities, industrial companies and equipment manufacturers. The AI boom is therefore simultaneously a source of economic growth and a potential source of resource constraints.
Will AI Returns Justify the Investment?
There is a fundamental economic question underlying the entire AI investment cycle: will the revenues and productivity improvements generated by AI justify the capital being committed? There is major evidence that demand is real. Alphabet's cloud growth and Amazon's AWS results are impressive. Enterprise adoption is expanding, and businesses are increasingly incorporating AI into customer service, research, software development, manufacturing and administrative operations. But usage is not the same as profitability. Companies must demonstrate that AI applications can generate sufficient incremental revenue, reduce operating expenses or improve productivity enough to justify the costs of computing, software development and infrastructure. The economics of inference (i.e. the cost of actually operating models and responding to users' requests) will be particularly important as AI adoption expands.
There are other risks. Models may become less differentiated, competition may drive prices down, and technological advances may make existing chips and computing facilities obsolete sooner than anticipated. New applications may also disrupt established software businesses, changing the competitive position of companies that are currently among the largest beneficiaries of the boom. Regulation, cybersecurity, export restrictions and geopolitical tensions introduce further uncertainty.
Nevertheless, the productivity enhancement and new business development possibilities are substantial. If AI allows businesses to produce more with fewer resources, the resulting improvement in productivity could support stronger economic growth without a corresponding increase in inflation. It could also help offset slower labor-force growth and improve corporate profit margins. And it can speed the development of new products and services throughout the economy. These are the potential economic rewards that make the current investment cycle so consequential.
Market Reaction
The early investor reaction to the latest earnings reports suggests that investors are becoming less willing to accept capital spending as an end in itself. Here are two examples of this change. Alphabet's shares came under substantial pressure following its October 28 announcement that growth had slowed in its core Google Search business despite reporting exceptional cloud growth. Investors were concerned about the scale of capital expenditures and the timing of returns. Amazon's results, released October 29, also prompted scrutiny of spending and cash generation. Its impressive AWS growth did not eliminate concerns about the amount of capital being committed. On October 30, technology and semiconductor shares experienced substantial selling pressure, illustrating how quickly sentiment can change when investors question the relationship between spending, earnings and future returns. This is particularly important because, as we have warned in the past, the largest technology companies have become a substantial part of the S&P500 and of the other major cap-weighed equity indexes. Their investment programs have also supported equipment manufacturers, semiconductor companies, utilities, construction firms and other businesses. A sustained reversal in those expenditures could therefore have consequences well beyond the technology sector. There are estimates suggesting that as much as 1.9 percentage points of the third quarter's 3.4% annualized GDP growth were attributable directly or indirectly to AI information-processing equipment. These categories include AI-related activity but also encompass other expenditures as well. So there is little question that AI investment has become an important source of economic growth.
That said, a myriad of possible investment outcomes come to mind. At the extreme, in the favorable scenario, AI continues to improve productivity, create new products and services, reduce business costs and expand corporate profit margins. Investment in data centers, chips, power generation and networking continues, while the resulting productivity improvements gradually become visible throughout the economy. The expansion could become stronger and more durable. In the less favorable scenario, the investment cycle outruns the revenue it can support. Companies continue to borrow and spend, but customers do not generate sufficient independent demand to justify the capacity being constructed. Eventually, investment slows, specialized facilities become underutilized, and debt-financed projects become difficult to refinance. The resulting contraction could affect construction, manufacturing, utilities, semiconductor suppliers, private-credit lenders and equity markets. The consequences could be amplified by the circular financing arrangements described above. A decline in AI valuations could undermine the very financing structures supporting the investment boom.
It is important to distinguish a correction in AI-related stock valuations from a collapse in the technology itself or a broad-based economic contraction. AI may ultimately prove to be as transformative as the internet or electricity, while the companies financing its development may nevertheless have overestimated the speed and profitability of its adoption. That distinction is central to our assessment. Key for investors seeking to successfully navigate during this period of uncertainty will be broad equity portfolio diversification in high quality, well financed businesses both domestically and abroad.
The Outlook
The October economic data as well as Q3 earnings reports provide a stronger basis for confidence in the near-term expansion than we had several months ago. Consumer spending remains substantial, business investment is robust, manufacturing and services surveys are positive, and the third-quarter GDP report confirms that the economy has regained momentum. The expansion has demonstrated resilience in the face of higher interest rates, persistent inflation, geopolitical uncertainty and a weakening labor market. In that sense, the US economy has legs. But we should not confuse a strong current growth rate with an absence of vulnerabilities. Inflation could force the Federal Reserve to tighten further. Higher borrowing costs could gradually restrain housing, consumer spending and business investment. And the AI investment boom, while providing a significant stimulus today, has introduced a new and potentially consequential source of financial and economic risk.
In particular, the circularity of AI financing deserves close monitoring. The more revenues, purchase commitments and investment decisions become dependent on the same interconnected group of companies, the greater the possibility that a disappointment in one part of the system will spread to others. For investors, the appropriate conclusion is neither complacency nor pessimism. The economy is performing better than many expected, and the productivity potential of AI remains substantial. But the quality and sustainability of growth deserve as much attention as its reported rate. We enter the final months of 2026 with a stronger economy than we anticipated, but also with a financial-market environment in which the consequences of excessive expectations and capital commitments could become increasingly important. The challenge for the coming year will be to determine whether the investment boom is laying the foundation for a sustained productivity revolution or whether its financing and spending have moved too far ahead of the economic return.
Equity Investment Policy
Equity investment policy remains unchanged. Portfolios under our firm’s supervision remain fully invested within agreed guidelines reflecting our view that the US economy remains fundamentally sound and well financed. The US dollar has remained strong relative to other currencies despite our notable fiscal deficits. Corporate profits continue to provide a favorable backdrop for equity investment. Our large cap core equity portfolios remain well diversified, balanced between growth and value shares, providing exposure to a broad range of sources of earnings growth and market leadership over an economic cycle. That view is reinforced by readings of our firm’s Economic Model, updated last week, which remained above-trend, signaling expansion ahead. Though inflation remains elevated and interest rates have moved higher since the beginning of the year, the economy has shown resilience, expanding at an above-trend pace. Importantly, corporate fundamentals remain solid and profitability has continued to exceed expectations. Merger and acquisition activity remains healthy. The IPO pipeline still includes a number of high-profile technology offerings, though some recent postponements suggest investor appetites have become more selective as market sentiment toward technology shares in general has soured. Even so, there are few signs thus far of the broad deterioration in business fundamentals investors have periodically feared.
The stock market advances of September and early October, driven by mega-cap technology and semiconductor companies thought to benefit from a surge in spending for artificial intelligence, have given way to a wave of profit taking in these shares as doubts have arisen over whether their AI investments will be justified by their future returns. Reflecting this change in sentiment, the semiconductor index declined 17% last month, the tech-heavy NASDAQ corrected roughly 10% from its record high, while the S&P500 pulled back about 3.5% from its peak. The S&P500 equal weight index, less weighted with tech shares, declined by only 2.2% in October. In our view, this reversal does not signal that technology leadership is nearing an end. Rather, as the economic cycle evolves, we believe earnings growth and investor interest should continue to broaden to include industrials, financials, healthcare, energy and other areas of the economy as well.
As for stock market valuation, equities are neither uniformly cheap nor expensive. The cap weighted S&P500 remains above its longer-term average valuation, though less so since earlier this year, reflecting the premium attached to its largest and fastest-growing companies. By contrast, valuations across the equal-weighted S&P500, mid-cap and small-cap indices are considerably less demanding, supporting our view that opportunities remain attractive beneath the headline index. We, therefore, view the overall US equity market as fairly valued, with more attractive opportunities available away from large technology and AI-themed companies that have driven the remarkable stock market gains over the past couple of years.
While elevated valuations increase the likelihood of periodic sentiment-driven corrections of 10%+ in the overall market, and more severe drawdowns in more speculative, high-priced areas, absent deterioration in economic and earnings fundamentals, we retain our favorable long-term view toward equities.
Fixed Income Investment Policy
As we reported in our October Economic Update, we modestly increased the target duration for bond portfolios we manage to 3.0 years early last month from 2.75 years to take advantage of a sharp spike in yields since midyear. That rate increase was driven by a combination of inflation fears, uncertainties around Fed interest rate policies, the likelihood of huge sovereign and corporate bond issuance and the heightened global competition for capital. Bond yields have since retraced much of that rise with the benchmark 10-year US Treasury bond yield falling from a peak of over 5.56% to 4.94% at the end of last month. The over 50 basis point decline in rates, in our view, brings bond yields back toward the lower end of their expected trading range.
With yields at or above 5.0% on the 10-year benchmark US Treasury, the prospective income expectation from owning high-quality bonds is materially different from what investors faced when long-term yields were 2-3%. The critical question is whether today’s rates represent a new structural equilibrium or an overshoot that will eventually reverse as inflation and growth moderate. We believe the odds favor 10-year Treasury bond rates remaining in a 4.75-5.50% range, pointing to the current attractiveness of the investment grade corporate bonds and Treasuries we use to fill-out client’s laddered fixed income bond portfolios.
* * * *
MBF