US Equities

Micron reported fiscal fourth-quarter revenue of 54.229 billion dollars for the quarter ended 3 September, against 41.456 billion in the prior quarter, 11.315 billion a year earlier and a reported market expectation of 51.5 billion. The gross margin was 86.8 percent and GAAP net income 37.701 billion dollars, or 32.87 dollars a diluted share. Full-year revenue reached 133.188 billion dollars against 37.378 billion, with net capital expenditure of 27.367 billion for the year. The company guided first-quarter revenue to 61.5 billion dollars plus or minus 1.5 billion with adjusted earnings of 38.15 dollars a share.

Why it matters here: This is the hardest observable evidence available that the artificial-intelligence capital cycle is accelerating rather than peaking: revenue has nearly quintupled in a year, the margin is 86.8 percent, and the company is guiding up again while committing 27.367 billion dollars of capital, which is itself a direct order flow into the semiconductor equipment, construction and power names elsewhere in the index.

Counterpoint: An 86.8 percent gross margin in a commodity memory business is a cycle-peak signature rather than a durable condition, and Michael Burry has publicly shorted memory names on exactly that reasoning. The consensus comparison comes from a weaker secondary account, and a company needing to beat by this much simply to be rewarded is a sign of how much is already priced.

Second-quarter growth was revised up to a 2.2 percent annual rate from 1.5 percent, an upward revision of 0.7 percentage point driven by investment, consumer spending and government spending, with real final sales to private domestic purchasers up 4.6 percent and real gross domestic income up 2.6 percent. First-quarter growth was revised to 2.5 percent from 2.1 percent. Corporate profits from current production rose 384.0 billion dollars against 63.4 billion in the prior quarter. Non-residential structures investment led by data centres was among the named contributors.

Why it matters here: Earnings estimates are built on a measured level of nominal demand, so a benchmark revision that lifts both the level and the composition towards investment feeds straight into the forward revenue line, and it does so most for the domestically geared and cyclical parts of the index rather than for the global megacaps.

Counterpoint: Stronger growth is precisely why the market still expects another increase in December and why long real yields are rising. For an index whose multiple is the dominant variable, a higher discount rate can take more away than better revenue adds, which is how the blue-chip average fell 0.86 percent on a day with good growth news.

The personal consumption expenditures price index rose 0.3 percent in August and 3.4 percent from a year earlier, while the measure excluding food and energy rose 0.2 percent on the month and 3.0 percent on the year, against surveyed forecasts of 3.7 percent headline and 3.3 percent core. Real consumer spending jumped 0.6 percent on the month and the saving rate fell to 4.1 percent, so the softer price data arrived alongside strong demand rather than weak demand. The broad index traded up as much as nearly 0.7 percent on the release.

Why it matters here: Equity multiples in this cycle are hostage to the tightening path, so evidence that inflation is decelerating towards something the central bank can live with lifts the valuation leg of the market directly, and it does so most where the duration of the earnings stream is longest — which is why the technology benchmark was the only major index to hold a gain.

Counterpoint: The market rejected the message. The broad index reversed its entire gain to close down 0.25 percent and the blue-chip average fell 0.86 percent, with new 52-week lows outnumbering new highs 262 to 23 on the New York Stock Exchange. Inflation relief that cannot hold an intraday gain is a weak tailwind, and one month does not settle a 3.0 percent core trend.

The implied probability of a quarter-point increase at the October meeting fell to about 35 percent from roughly 51 percent the previous day, and from above 80 percent at one point during the month, with traders moving the next expected increase to December. Goldman Sachs changed its forecast the same day, saying an October increase is now unlikely and seeing a strong chance the committee concludes no further increases are needed. The current target range upper bound is 4.00 percent and the committee has projected 4.1 percent by year end.

Why it matters here: The equity market's central problem this quarter has been a tightening cycle with no visible end. A credible argument from a major house that the cycle has only one move left changes the terminal rate against which every multiple is set, and it does so for the whole index rather than for one sector.

Counterpoint: Janus Henderson's reading of the same data is that strong labour and growth figures leave a year-end increase intact, and market pricing on the same day still implied three or four increases over the next twelve months. The index closed lower on the day, so equities plainly did not take the dovish interpretation, and the October meeting is close enough that Friday's payroll report can reverse the move.

Meta stock rose 28.8 percent as of noon on 30 September over the course of September, driven mostly by the reception of its first consumer artificial-intelligence agent, unveiled at a late-September event. A Deutsche Bank analyst wrote that agentic commerce has been touted as the next biggest evolution in digital commerce since the rise of electronic commerce, and that the product represents the first real large-scale consumer platform built around autonomous commerce. Bank of America told clients the fourth and first fiscal quarters have been the two best seasonal quarters to own chip stocks, with median outperformance of 300 to 500 basis points against the broad index from 2010 to 2025.

Why it matters here: Until now the artificial-intelligence trade in this index has been a capital-spending story priced through the supply chain. A consumer product that executes transactions rather than recommending them is the first evidence of revenue at the application layer, which is what justifies a multiple rather than just an order book, and it reaches the discretionary sector as a distribution-channel change as well as the platforms as a revenue line.

Counterpoint: Michael Burry is publicly short the complex and argues these companies will destroy trillions of dollars of invested capital; Trivariate Research has turned more negative on data-centre concerns; and the market's own breadth collapsed on the same day. A 28.8 percent monthly move in a megacap on a product launch is also precisely what a late-cycle narrative looks like, and the underlying claims are opinion rather than measured outcome.

Boeing secured a deal with the Pentagon and the Navy worth more than 20 billion dollars to develop the sixth-generation F/A-XX Strike Fighter over a long-term phase, replacing the service's F/A-18 Super Hornet fleet and advancing open mission systems architecture and operational reach. Boeing shares rose as much as 3.34 percent in premarket trading and were up about 2 percent at the open, while Northrop Grumman, reportedly also in contention, fell 3.5 percent.

Why it matters here: Development awards of this kind are the most predictable revenue in the index because they are funded by appropriation rather than by end demand, and replacing an entire carrier fighter fleet creates a programme running into the 2030s. It lands in a year when defence budgets are being driven by an active war, which is a distinctively defensive earnings stream inside an index otherwise geared to the cycle.

Counterpoint: The index effect is close to a wash in the short run, because the competitor that lost fell 3.5 percent while the winner gained about 2 percent at the open. Fixed-price development programmes of this scale also have a long history of cost overruns at this particular contractor, and the contracting authority's own announcement was not opened.

The announced South Korean programme covers 200 billion dollars of investment, including 54 billion dollars for an Alaska liquefied natural gas project with an 807-mile pipeline, a six-gigawatt power plant in Encinal, Texas, and nuclear power plants. It sits under a 350 billion dollar agreement signed last November that cut US tariffs on South Korean cars, auto parts and other products from 25 to 15 percent, with Seoul committing 150 billion dollars to US shipbuilding and 200 billion to other strategic sectors. A senator estimated 12,000 construction jobs on the Alaska project alone.

Why it matters here: Inward infrastructure investment is unusually high-quality earnings for the domestic industrial base because it is contracted, long-dated and largely insensitive to the business cycle, and project finance of this scale generates fee income for the domestic banking sector. It arrives alongside the 27.367 billion dollar semiconductor capital programme reported the same evening, which is the same order-book channel.

Counterpoint: Announced investment is not spent investment: coverage of the same event noted that the biggest deals are not yet concluded and Seoul has made commercial viability a condition. The timing immediately before midterm elections also invites scepticism about how much of the total is genuinely incremental.

The Chicago Business Barometer rebounded 11.7 points to 58.8 in September from 47.1, its highest reading since May and back in expansion after one month below 50, against an expected 51.0. Production rose 15.5 points, new orders 13.3 points and supplier deliveries 9.6 points for a twentieth consecutive month above 50, while employment softened 4.3 points back into contraction and prices paid eased 3.7 points. Respondents highlighted constraints in the availability and delivery of electronic components. The survey ran from 1 to 15 September.

Why it matters here: Factory surveys lead industrial revenue by roughly a quarter, and a double-digit rebound in both production and new orders is the kind of signal that forces order-book estimates higher. The component-shortage comments point specifically at the semiconductor supply chain, so the read-across runs to chip demand as well as to industrial order volumes.

Counterpoint: Employment within the same survey fell back into contraction and respondents attributed part of the order gain to seasonal improvement, so this may be a restocking bounce rather than a turn in end demand. It is one city's survey conducted in the first half of the month, and the series is volatile enough that a single swing rarely survives the next print.

The Nikkei 225 closed 1.94 percent higher at 66,753.72 on 30 September, its highest close since 19 August, with SoftBank Group surging 6.55 percent to 6,430 yen on reports of surging revenue at OpenAI and the memory producer Kioxia rising 1.20 percent to 18,095 yen. Overnight strength in US semiconductors, positive artificial-intelligence news and a decline in oil prices in the previous session were cited as lifting regional risk appetite.

Why it matters here: The artificial-intelligence capital cycle is a single global supply chain, so Japanese memory producers and investment-holding companies trade as leading indicators for the US semiconductor complex inside this index; a session in which those names re-rate hard is a same-day read on the demand the US supply chain is about to report.

Counterpoint: The read-across stopped at technology. The blue-chip average fell 0.86 percent and the broad index closed lower, with new 52-week lows outnumbering new highs 262 to 23 on the New York Stock Exchange. The stock-level moves also come from the weaker of the two publishers, and a narrow Asian lead did not translate into a US market gain.

ADP Research reported US private-sector employment rose by 90,000 in September against a surveyed consensus of 68,000 and a downwardly revised August gain of 36,000, with hiring accelerating for the first time since May. Education and health services added 55,000, leisure and hospitality 22,000, manufacturing 17,000 and construction 15,000, while financial activities shed 16,000 and professional and business services 11,000. Median base pay rose 3.2 percent from a year earlier and median gross pay 4.7 percent.

Why it matters here: Consumer earnings in this cycle rest on employment income rather than on credit, so a month in which both the job count and pay growth held up is direct support for the consumption-facing parts of the index. The 55,000 gain in education and health services speaks specifically to health-care services volumes, which is a different channel from discretionary demand.

Counterpoint: The same strength is what keeps a December increase in the market's base case, which works against the whole index through the discount rate. The composition was also narrow: two sectors shed jobs outright, financial activities most of all, and the gains came disproportionately from education, health and hospitality rather than from cyclical demand. This is a private survey with a known record of divergence from the official count.

Job openings and labour turnover data released on 29 September came in at 7.08 million against an estimate of 7.2 million. The shortfall had already cut the odds of an October rate increase from 70 percent a week earlier to roughly even by that Tuesday, before the following morning's inflation data pushed them below 35 percent.

Why it matters here: Vacancies lead wage growth, and wage growth is what has funded US consumption through this tightening cycle. A shortfall at this stage is an early signal that the income support under discretionary spending and under the most domestically geared parts of the index is fading.

Counterpoint: The private payroll report the next morning showed hiring accelerating for the first time since May with gross pay up 4.7 percent, which is the opposite read on the same labour market. The figure also appears only in a headline summary rather than in a release text, the official source page was not opened, and the official September count supersedes it within days.

The S&P 500 was up as much as nearly 0.7 percent during the session before reversing to close down 0.25 percent at 7,651.54, while the Nasdaq Composite advanced 0.24 percent to 26,861.06 and the Dow Jones Industrial Average fell 443.87 points, or 0.86 percent, to 50,906.05. In late trading new 52-week lows beat new highs 262 to 23 on the New York Stock Exchange with decliners exceeding advancers 1,478 to 1,154, and on the Nasdaq new lows topped new highs 294 to 84. Advancing volume was under 52 percent of the total on the New York exchange and under 60 percent on the Nasdaq, with composite volume at 52 percent and 69 percent of the thirty-day average.

Why it matters here: The index level is being held up by a handful of technology names while the median stock makes new lows. A market in which decliners outnumber advancers, volume runs at roughly half the recent average and new lows swamp new highs by more than ten to one is not in the condition its headline level suggests, and equal-weighted and small-cap exposures in this class register that directly.

Counterpoint: The quarter was still positive for both the broad and technology benchmarks, the S&P 500 gaining 2.0 percent and the Nasdaq 2.5 percent, and the day was a quarter-end rebalancing session with composite volume at 52 percent of the thirty-day average. The one index that rose did so on genuine earnings news; narrow leadership backed by an 86.8 percent gross margin is not the same as narrow leadership backed by nothing.

Morgan Stanley warns that a US export restriction would likely lower domestic diesel prices initially but with potentially adverse reactions downstream, including a feedback loop to US gasoline prices as refinery runs adjust, while the American Petroleum Institute argues restrictions would exacerbate refining challenges and hurt consumers. The Energy Secretary has described the options as restrictions rather than an outright ban, and the President is reported to be considering red-dyed diesel sales as an alternative. Average retail diesel stood at 6.50 dollars a gallon against a 6.53 dollar record.

Why it matters here: The policy is designed to transfer value from refiners and producers to households before an election, so its sector effects run in opposite directions inside the same index: freight and logistics margins and energy-sector earnings fall while discretionary spending power rises. The sign depends entirely on which link in the chain a company sits on, which is why it registers as net adverse for an index weighted towards the producing end.

Counterpoint: Morgan Stanley's own warning is that the consumer benefit may not survive the refinery response, in which case neither the discretionary tailwind nor the clean story about lower pump prices holds. The policy also remains undecided, publishers disagree about what form it would take, and European traders mostly doubt it will be imposed at all.

The President said he expects the war will not end until after the 3 November midterm elections, with the conflict now in its seventh month and the Strait of Hormuz still not reopened. US retail diesel sits at 6.50 dollars a gallon against a 6.53 dollar record, and the administration is weighing export restrictions in response. Iran's currency fell to a record low above 2.5 million rials to the dollar, and the US Treasury sanctioned ten further individuals and entities on 29 September.

Why it matters here: The conflict reaches the US consumer through the pump price and reaches the index through the inflation path and the policy response to it. A stated expectation that the fighting runs past the elections extends the horizon over which both apply, which matters most for freight-intensive industrials and for discretionary retail.

Counterpoint: Defence and aerospace within the same index are direct beneficiaries, with a development contract worth more than 20 billion dollars awarded the same week, and the technology complex that drives the index's earnings growth has negligible energy intensity. A war that leaves US supply intact is a cost shock for a minority of index earnings.

The 30-year fixed mortgage rate reached 7.58 percent on 29 September, the highest level since November 2023, and a daily tracker put it at 7.60 percent on 30 September. The Mortgage Bankers Association's weekly survey put the average contract rate for a 30-year conforming loan at 7.12 percent for the week ending 18 September, the highest since May 2024, with total applications down 1.5 percent, the refinance index down 3 percent and 62 percent lower than a year earlier, and adjustable-rate loans rising to 9.8 percent of volume from 8.4 percent.

Why it matters here: Housing transactions generate a long chain of earnings in lending fees, building products, furnishings and moving-related consumption, and that chain contracts with volumes rather than with prices. This is why the earnings effect arrives before any visible fall in house prices, and why it concentrates in financials, small-cap builders and suppliers rather than across the index.

Counterpoint: Housing-linked revenue is a modest slice of index earnings, and the same week's data showed the broader consumer still spending heavily, with real consumption up 0.6 percent in August. The daily rate figure is a lender-panel estimate rather than a transacted average, and the equity market's binding problem is the discount rate rather than mortgage volumes.

Brent closed the month above 103 dollars at 103.53 and US crude at 90.42 dollars, with Brent more than 14 percent higher over September. Average US retail diesel sits at 6.50 dollars a gallon just below its 6.53 dollar record, and US diesel futures rose 4.5 percent to 5.1175 dollars a gallon after the weekly inventory report showed distillate stocks down 2.3 million barrels.

Why it matters here: Fuel is a regressive tax on the consumer and a direct cost line for every goods-moving business, so a sustained energy shock shows up as margin pressure in industrials and volume pressure in discretionary retail before it shows up anywhere else in the index. Smaller companies have the least ability to hedge it or pass it through.

Counterpoint: Real consumer spending rose 0.6 percent in August despite these fuel prices, and the index's largest weights are technology businesses with negligible fuel intensity. The energy drag is real but concentrated in a small share of index earnings, and Gulf export flows are recovering towards their 2025 average.

The ten-year Treasury yield closed about four basis points higher at 5.298 percent after trading above 5.30 percent intraday, near its 2007 high, while the thirty-year rose to around 5.642 percent, its highest since 2002; the two-year was little changed at 4.895 percent. The ten-year's monthly rise of 0.49 percentage point was the steepest since September 2023, with inflation-protected yields adding 0.44 point to the ten-year real rate. The S&P 500 reversed an intraday gain of nearly 0.7 percent to close down 0.25 percent and the blue-chip average fell 0.86 percent, with new 52-week lows outnumbering new highs 262 to 23 on the New York Stock Exchange.

Why it matters here: An equity index is a long-duration cash flow, and the risk-free rate against which it is discounted has moved to a level most of the current market has never traded against. The breadth statistics show the damage spread far more widely than the headline index level admits, and it falls hardest on the equal-weighted, blue-chip and financial exposures that carry no growth duration to offset it.

Counterpoint: The technology benchmark still closed higher, and the underlying earnings cycle is accelerating hard enough that a memory company reported quarterly revenue of 54.229 billion dollars the same evening against 11.315 billion a year earlier. Where earnings growth is fast enough, it can outrun the discount rate for a long time.