Earnings Scoreboard - Scarcity on contract

Investor Lens

Earnings remain solid but fall short of expectations in a demanding market that rewards conversion and higher guidance.

• According to FactSet, with 24.1% of the S&P 500 reported, the blended Q3 earnings growth rate stands at 29.5% from 29.2% a week earlier, on revenue growth of 12.7%. If that holds, it would be the third straight quarter above 25% and the eighth straight quarter of double-digit growth.
• The beat rate remains strong at 82.6%, but the magnitude is modest. Companies have exceeded estimates by an aggregate 2.1%, well below the five-year average of 7.5%.
• This week the focus is on potential returns on capital across the investment chain, from electricity, copper and optical connectivity to cooling, storage, cloud infrastructure and the applications expected to monetise them. It distinguishes selling scarce capacity from earning adequate returns on owning it. The two can generate markedly different cash-flow profiles despite sharing the same growth story.
• Technology intensity should be assessed differently from economic breadth. More storage, connectivity and power per installation can support selected suppliers without establishing a general recovery in devices, industrial demand or enterprise budgets. Pricing, volumes, product mix and utilisation must tell compatible stories.
• Backlog quality matters more than its headline size. A larger order book is more valuable when execution requires proportionately less working capital and customers have credible capacity to pay, enhancing conversion.

This is the season's heaviest week. There are 186 S&P 500 companies scheduled to report, 11 of them Dow components. The docket covers three of the four largest cloud platforms, the most valuable consumer-technology franchise in the world, the leading incretin franchise, constituents of the Memory & Storage and Site-readiness and Physical Infrastructure layers of our seven-layer AI framework, copper, and two of the most closely watched aerospace and defence programmes in the US.

The Week Ahead

Questions Earnings Will Answer

Is technology intensity being mistaken for economic breadth?

Much of this season’s growth is coming from more content per installation rather than more installations. AI clusters require far more fibre per rack, far more cooling per megawatt, far more storage per workload and far more power per site. That intensity is lifting Corning, Vertiv, Seagate, Sandisk and GE Vernova at rates that look like a broad industrial recovery, but are not. Corning’s Display business, Apple’s unit volumes and the industrial end-markets served by the rest of the docket tell a more nuanced story.

A broad-based recovery spans customers, geographies and budgets. An intensity-led recovery depends on a few buyers with correlated spending decision. This means one change to hyperscaler capital plans can affect all suppliers at once, as GE Vernova and Vertiv experienced after Meta’s results last week.

Intensity also generates inflation. The Dynamic Random Access Memory (DRAM) and NAND shortage enriching Sandisk and Seagate pushed Amazon to raise its CapEx outlook to $220 billion, explicitly citing memory costs. It forced Apple into broad hardware price increases in June, and consensus still expects Products gross margin to fall 338 bps q/o/q. At Freeport-McMoRan, copper scarcity lifts realised prices while lower by-product credits push unit costs above guidance. Scarcity shifts margins along the chain without raising them uniformly.

Eli Lilly offers the instructive counter-case. Its growth is coming from breadth, with volumes expanding through direct-to-consumer, employer and Medicare channels even as US realised prices fall 9.0%. This contrasts with the AI supply chain, where price growth outpaces volume growth. Investors should seek results supported by consistent trends in pricing, volumes, product mix and utilisation. Growth driven solely by pricing is real, but lasts only as long as scarcity persists.

Who earns the return on AI capital?

This week’s results span the investment chain, distinguishing sellers of scarce capacity from companies that must own it. Upstream suppliers Sandisk, Seagate, Vertiv, GE Vernova and Corning benefit through pricing, allocation and customer prepayments. Sandisk’s long-term agreements total at least $93.9 billion, Seagate’s nearline output is allocated through 2028 and GE Vernova’s gas-turbine slot reservations include down payments. For these suppliers, the AI cycle benefits working capital, with cash arriving before products leave the factory.

Hyperscalers face the opposite cash-flow profile. Consensus forecasts Alphabet’s Q3 CapEx at $55.195 billion, with FCF remaining negative. Microsoft’s FCF is expected to fall 68.1% y/o/y despite Azure growth in the mid-forties. Amazon’s trailing twelve-month FCF outflow is projected to widen to $9.941 billion. Suppliers’ revenue becomes hyperscalers’ depreciation, which continues to accumulate for years after spending peaks.

This asymmetry concerns who captures the economic rent, not demand, which remains contracted and visible in Google Cloud’s $514 billion backlog and AWS’s $496 billion backlog. If cloud operating margins withstand rising depreciation, as Google Cloud’s 35.6% and AWS’s 39.4% in Q2 suggest is possible, capacity owners retain a fair share of the value they fund. If margins narrow as new capacity reaches the income statement, that rent shifts upstream to component makers and downstream to applications such as Palantir and Copilot, which monetise compute without owning it.

Investors should therefore focus less on the size of this week’s hyperscaler earnings beats and more on incremental margin relative to incremental capital. That ratio determines if owning AI capacity offers franchise economics or utility-like returns.

Major Earnings on the Docket

Alphabet
Operating leverage at the steep end of the CapEx curve

Alphabet reports Wednesday after the close, with the call at 4:30 pm EDT. Consensus centres on revenue of $127.911 billion, an increase of 24.9% y/o/y, with EPS estimates of $3.03. Q2 set the bar high, with revenue growth of 24.9% y/o/y to $119.796 billion. Google Search & Other grew 16.8% to $63.271 billion and Google Cloud grew 81.8% to $24.768 billion at an operating margin of 35.6%.

Google Cloud is now the centre of the investment case. The segment’s annual revenue growth has accelerated from 33.5% in Q3 2025 to 81.8% over four quarters, and the backlog reached over $514 billion. Margin expansion is also part of the Cloud story, with the segment’s operating margin expanding by 14.8 percentage points from 20.7% to 35.6% over the same period. Alphabet has begun selling complete Tensor Processing Units (TPU) systems to third parties, which makes its custom silicon a product line as well as an internal cost advantage. OpenAI models are now also available on Google Cloud, a sign of how far the industry has moved from exclusive partnerships towards multi-cloud distribution.

Google Cloud reflects Alphabet’s AI ambitions across a broader range of the ecosystem than other hyperscalers’ CapEx strategies. For Q3, the Street anticipates Cloud revenue to increase 89.5% y/o/y to $28.723 billion, alongside an operating margin of 33.9%, which would imply a sequential contraction of 165 bps from Q2.

Search anchors the case's defensive side. With AI Overviews and AI Mode fully integrated, query growth, paid clicks and cost-per-click will indicate if generative answers strengthen or cannibalise commercial intent. This remains a key investor concern as competing LLMs reshape online search behaviour. For Q3, Search revenue is anticipated to reach $64.957 billion, reflecting 14.8% y/o/y growth.

YouTube advertising and subscriptions, forecast to grow 11.5% y/o/y to $11.439 billion, form Alphabet's third revenue pillar.

Should the top line fail to exceed expectations, and perhaps more consequentially, should margins erode, then the most sensitive signals will be CapEx and FCF. The 2026 CapEx range guidance is $195 to $205 billion, against $80.598 billion spent H1, implies a sharp step-up in H2. Q2 FCF was an outflow of $5.855 billion. For Q3, consensus estimate CapEx of $55.195 billion, up 130.4% y/o/y and FCF to be an outflow of $2.271 billion.

An earnings beat, higher cloud margins, sustained high-teens Search growth and 2027 CapEx growth below cloud backlog growth would reinforce Alphabet's claim to the most complete AI stack across models, chips, cloud and distribution. Increased spending without operating leverage would echo Meta's position last week, leaving investors to fund distant optionality.

Microsoft
New segments, higher scrutiny

Microsoft reports Wednesday after the close, with the call at 5:30 pm EDT. The Street calls for revenue of $90.616 billion and EPS of $4.71, increasing 16.7% and 14.1% y/o/y, respectively. Q3 revenue guidance is $89.850 to $ 90.950 billion. This is the first quarter reported under Microsoft's new two-segment structure, ‘Agents and Infra’ and ‘Devices and Consumer’, and Azure revenue is now disclosed in dollars rather than as a growth rate alone.

Azure is the print. Management guided Azure y/o/y growth of 44% to 45% in constant currency, against 43.0% in the prior quarter on the restated basis. For Q3, the consensus estimates anticipate 44.3% y/o/y growth in constant currency revenue for Azure and revenue of $32.280 billion. Capacity is the constraint. Azure is reported under the segment Agents and Infra. For this segment, revenue is projected to reach $75.541 billion with operating income at $38.895 billion, reflecting an operating margin of 51.5%.

Devices and Consumer’s revenue for Q3 is anticipated to be $15.013 billion, and its operating margin is pencilled in at 32.9%. Quarterly CapEx is expected at $46.536 billion, while FCF is anticipated to decline 68.1% to $8.189 billion from $25.663 billion.

The OpenAI relationship is the strategic variable. OpenAI models are now also available through Google Cloud. This ends Microsoft's de facto exclusivity at the distribution level, even as the restructured agreement preserves Microsoft's economic interest and its access to the models. Investors will look for evidence that Azure's growth comes from a broad enterprise base rather than from OpenAI's own compute consumption.

Copilot monetisation, through Microsoft 365 seat expansion and the average revenue per user uplift from AI tiers, is the other half of the return-on-capital argument. Agentic products need to show adoption that can be measured in revenue. Engagement alone is insufficient if the associated compute and support costs rise without corresponding monetisation.

A beat with Azure at or above guidance, a disciplined CapEx trajectory and visible Copilot revenue would affirm Microsoft as the enterprise default for AI deployment. It would further strengthen its position in the Applications Layer of our AI architecture framework.

Amazon
Can Amazon Prime afford to subscribe to AWS AI ambitions?

Amazon reports Thursday after the close at 4:00 pm EDT, with the call at 5:00 pm EDT. Management guided Q3 net sales to $197 to $202 billion and operating income to $22.5 to $26.5 billion, including a roughly four-point headwind from moving Prime Day into Q2. Consensus calls for revenue of $202.222 billion, up 12.2% y/o/y, operating income of $26.214 billion, up 50.5% y/o/y and EPS of $1.97. The EPS year-ago comparison included mark-to-market gains on the Anthropic stake, as did Q2, when a $53.4 billion pre-tax gain lifted EPS to $5.75. Operating margin is the cleaner gauge.

Like Microsoft, and Alphabet before it, the cloud segment, AWS, is the driver. Revenue reached $42.232 billion in Q2, up 36.8% y/o/y, its fastest growth in eighteen quarters, with operating income of $16.621 billion, reflecting an operating margin of 39.4%. The backlog stands at $496 billion. Management said both the AI business and the custom-chip business had passed a $25 billion annualised run rate (ARR). Broader Trainium adoption beyond Anthropic will be the clearest measure of Amazon's ability to rely less on third-party accelerators. For Q3, the Street has pencilled in AWS y/o/y revenue growth of 36.9% to $45.195 billion and an operating margin of 37.6%. This is a sequential contraction of 180 bps from Q2’s 39.4%, but an expansion of 291 bps from the year-ago figure.

CapEx remains the offset. In Q2, Amazon lifted its 2026 CapEx outlook from about $200 billion to $220 billion, citing higher memory costs. Trailing twelve-month FCF recorded a $7.689 billion outflow. For Q3, investors anticipate Q3 CapEx to be $58.659 billion, up 67.1% y/o/y, and trailing twelve-month FCF to be an outflow of $9.941 billion, from an inflow of $1.297 billion a year-ago.

Online Stores, Advertising, Subscription Services and Physical Stores fund much of its AI infrastructure. Regionalised fulfilment, robotics, advertising and other initiatives are expected to improve Online Stores margins. Revenue is forecast to rise 3.4% y/o/y to $69.680 billion, while Advertising is expected to grow faster, up 17.6% y/o/y to $20.815 billion. Amazon Leo, the satellite broadband network, adds a further investment line.

An earnings beat, AWS growth matching or exceeding Q2's pace, improving retail margins despite tariffs and a disciplined 2027 CapEx framework would position Amazon best among hyperscalers to justify its capital intensity. Strong custom silicon adoption, digital advertising growth and resilient Prime consumer trends could bolster confidence in long-term earnings growth.

General Dynamics
Long-cycle demand across segments, short-cycle execution challenges

General Dynamics reports Friday before the open at 7:00 am EDT, with the call at 9:00 am EDT. For Q3, the Street has pencilled in EPS of $4.12, up 6.3% from $3.88 a year earlier, and revenue of $13.718 billion, also up 6.3% y/o/y. Q2 set a strong base, with revenue of $14.094 billion reflecting y/o/y growth of 8.1%, EPS reached $4.24, up 13.4% y/o/y from $3.74 a year earlier, and backlog rose 31.7% y/o/y to $136.498 billion on a book-to-bill ratio of 1.4x. Then, management raised full-year EPS guidance to $16.80 to $16.90 on revenue of $55.7 billion, at a 10.6% operating margin.

Aerospace is the margin driver. Gulfstream delivered 41 aircraft in Q2, and segment revenue rose 15.1% to $3.525 billion at an operating margin of 14.4%. The G700 and G800 production ramp, and the G300 entering the line-up, will determine delivery cadence and pricing at a time when demand for large-cabin business jets remains firm among corporate and high-net-worth buyers. For Q3, the segment’s revenue is anticipated to increase 4.5% y/o/y to $3.378 billion, with an operating margin of 14.4%, relatively stable sequentially, but reflecting a 108bps y/o/y expansion.

Marine Systems is the largest and most strategically important segment. Q2 revenue rose 10.4% to $4.660 billion, driven by the Columbia-class ballistic missile submarine and Virginia-class attack submarine programmes. Labour, supplier capacity and throughput at Electric Boat and its partner yard at Newport News constrain production. Margins remain thin as the yards invest in higher output. Q3 revenue is forecast to grow 6.3% y/o/y to $4.354 billion, with an operating margin of 7.5%, up 43 bps y/o/y.

Combat Systems, where European rearmament supports munitions and armoured vehicle demand. Technologies, which provides IT and mission systems to the US government, complete the portfolio. A beat with Gulfstream deliveries on schedule and improving Marine margins would confirm that General Dynamics can turn both of its super cycles, business aviation and naval shipbuilding, into earnings. A strong backlog is more valuable when production becomes more predictable and requires less extraordinary intervention.

Apple
The replacement cycle tests the price ladder

Apple reports Monday after the close at 4:30 pm EDT, with the call at 5:00 pm EDT. This is the first earnings call under Chief Executive John Ternus, who succeeded Tim Cook. Management guided fiscal Q4 revenue growth of 9.0% to 11.0%, or roughly $111.690 to $113.740 billion, including about 2.5 points of currency headwind, and a gross margin of 47.0% to 48.0%. That compares with 50.1% in the prior quarter, which included a roughly two-point benefit from a tariff refund. Consensus calls for Q3 revenue of $113.389 billion, up 10.7% y/o/y, gross margin of 47.5% with gross income y/o/y growth of 11.4% and EPS of $1.99, up 7.6% from $1.85 a year earlier.

The iPhone is the key driver, with management guiding growth in the low teens. The iPhone 18 Pro and 18 Pro Max launched on 18 September, contributing only about two weeks of sales to the fiscal quarter, making holiday-quarter guidance more important than reported results. Their higher prices follow broad hardware price increases in late June, largely driven by memory costs. The DRAM and NAND shortage benefiting Sandisk and Seagate is also raising Apple's input costs. Q3 iPhone revenue is forecast to rise 15.5% y/o/y to $56.616 billion, with 59.38 million units sold. Products gross margin is expected at 36.7%, down 338 bps q/o/q from 40.1% but up 52 bps y/o/y from 36.2%.

Services support margins but carries regulatory risk, with the App Store facing legal rulings and legislation across several jurisdictions. Management guided growth in line with Q2. The Q3 consensus forecasts revenue of $31.604 billion, up 9.9% y/o/y and 2.8% q/o/q. Gross margin is expected at 75.4%, down 17 bps q/o/q from 75.6%, but up 13 bps y/o/y from 75.3%.

At its Surprise and Shine event on 9 September, Apple unveiled refreshed iPhone 18 Pro and 18 Pro Max models, Watch Series 12, Watch Ultra 4 and AirPods 5, alongside the new iPhone Duo.

Analyst reactions were mixed. Some expect broad price increases and the delay of lower-end models until spring 2027 to weigh on unit sales. The iPhone Duo is viewed as an innovation milestone under CEO John Ternus, but its high price, ergonomics and weak category demand raise concerns. Analysts also question if Apple's privacy-first AI strategy can compete with less restricted, open-source models. More bullish analysts argue that the $100 Pro price increases, split launch strategy, progress on the custom C2 modem and aggressive carrier promotions will lift Average Selling Prices (ASPs) and revenue enough to offset margin pressure.

China’s demand and supply concentration remain persistent narratives. A beat with a gross margin at the upper end of the range and a holiday guide that implies demand is absorbing higher prices would give Ternus a strong start.

Also on our radar

Tuesday

Freeport-McMoRan reports on Tuesday before the open and much of the operating picture is already public. The company has indicated Q3 copper sales of approximately 830 million pounds and gold production of around 230,000 ounces. Gold sales are closer to 100,000 ounces, with roughly 60,000 ounces deferred into Q4 as shipments from Indonesia normalise.

Price is doing much of the work. Realised copper above the $6.50 per pound guide reflects tight concentrate markets, the US tariff premium on refined metal and the supply disruptions that have kept the market in deficit. Higher costs temper the benefit, with consensus forecasting unit net cash costs 6.0% above the $2.00 per pound guidance as lower Grasberg gold by-product credits increase costs per pound. Grasberg recovery is the swing factor. After the mud-rush incident, the operation is running at about two-thirds of pre-incident rates.

In the Americas, Arizona’s Bagdad expansion awaits an investment decision by year-end and, if approved, would be one of few additions to US copper supply comparable in scale to a greenfield project.

Corning’s print will be driven by the Optical Communications segment, given that it sells fibre, cable and connectivity into hyperscale data centres. AI clusters need far more fibre per rack than conventional cloud architecture, because GPUs must be linked to one another at very high bandwidth inside the data centre as well as across campuses.

Investors should assess if long-term Corning’s customer commitments accelerate growth or merely extend visibility, if future platforms incorporate glass substrates and co-packaged optics, and if Display and Specialty Materials segments remain stable enough to avoid weakening the optical investment case.

For Seagate, the fundamental picture has rarely been stronger, as evidenced by both guidance and the Street’s estimates. Mass-capacity nearline drives, the high-capacity hard disks that store the data AI models train on and generate, are fully allocated through 2028 under long-term agreements with cloud customers. The Mozaic 4 platform, based on heat-assisted magnetic recording (HAMR) and offering up to 44 terabytes per drive, raises capacity per platter. That increases exabytes shipped without a matching increase in units or capital spending, and it is the mechanism behind Seagate's gross-margin expansion.

Supply is investors’ main concern. Reports of Toshiba expanding capacity renewed oversupply fears in an industry with just three manufacturers and a history of damaging pricing cycles. Investors will seek confirmation that capacity growth still comes from areal-density improvements rather than new factories, and that long-term contract prices remain fixed rather than linked to spot prices.

In the case of Boeing, a year ago, a $4.9 billion 777X charge produced a core loss of $7.47 per share, so the y/o/y comparison flatters the Q3 EPS estimate. Investors are not buying Boeing’s income statement, they are buying the production rate.

Commercial Airplanes remains loss-making, but is improving. For Q3, the Street anticipates an operating loss of $342 million, following Q2’s loss of $322 million, despite projected revenue growth of 2.9% y/o/y to $11.414 billion. The 737 programme is producing at 47 aircraft a month, and moving to 52 depends on opening the North Line in Everett. The 787 is at eight a month with a target of ten.

For Defence, Space & Security, the key question is if fixed-price development programmes, including the KC-46A tanker, T-7A trainer and MQ-25, have stopped incurring losses after years of charges. Global Services is the relatively steady profit centre that funds the recovery.

Management expects positive Q3 FCF despite an approximately $700 million Department of Justice settlement payment and full-year FCF of $1 to $3 billion. Outside the factory, risks include ongoing Society of Professional Engineering Employees in Aerospace (SPEEA) engineering union negotiations and supplier capacity at the now-integrated Spirit AeroSystems.

Wednesday

Power is the core investment case of GE Vernova. GE Vernova has 116 gigawatts of gas equipment under contract or slot-reservation agreements and targets at least 125 gigawatts. That queue now extends deliveries well into the decade's second half. It gives the company pricing power over utilities and developers that need firm generation. Backlog stood at $176 billion after Q2, with management expecting it to exceed $200 billion in early 2027.

Electrification, which covers grid equipment, switchgear, transformers and high-voltage systems, is the second growth engine and more sensitive to data-centre sentiment.

FCF guidance of $11.5 to $12.5 billion for the full year is supported by customer down payments on slot reservations, which raises the question of whether cash conversion relies on advances.

A print with sustained gas pricing, firm Electrification orders from utilities as well as hyperscalers, and narrowing wind losses would separate GE Vernova from the AI CapEx narrative and anchor it in electrification more broadly.

Vertiv’s top and bottom line estimates also demonstrate sound position in Site-readiness and Physical Infrastructure, the fifth layer of our seven-layer architecture of AI. Vertiv sells the power distribution, thermal management and liquid cooling that allow dense AI racks to operate. Nvidia's DSX reference architecture for AI factories, presented on 21 October, reinforced Vertiv's position as a co-designer of the data-centre stack rather than a component supplier.

For positioning ahead of Vertiv’s print, investors should consider three drivers. First, orders and book-to-bill, which show whether hyperscaler and colocation demand is holding after the summer. Second, the share of revenue from liquid cooling and prefabricated modular systems, which carry higher content per megawatt. Third, incremental margin, because tariffs on components and the pace of capacity expansion can absorb operating leverage.

An earnings beat, a book-to-bill well above one and higher full-year guidance would suggest Vertiv’s recent share decline priced in a slowdown absent from its order book.

Thursday

Eli Lilly reports Thursday before the open, and is expected to be heavily driven by its incretin franchise: the class of metabolic hormones underlying its blockbuster diabetes and weight-loss therapies.

For Q3, Mounjaro sales are anticipated to be $10.002 billion, up 53.5% y/o/y, and Zepbound’s are pencilled in at $5.049 billion, up 40.7% y/o/y. Foundayo, the oral GLP-1, contributed $98 million in its launch quarter. Oral delivery could widen the treatable population by removing the injection barrier and easing manufacturing constraints. Its early prescription trajectory will be the most closely watched launch metric in the sector.

However, price is the counterweight. US realised prices fell 9.0% as access expanded through direct-to-consumer channels, employer coverage and the Medicare GLP-1 Bridge programme. Volume is growing much faster than price is falling, but the trade-off is what defines the franchise's long-term margin.

The pipeline adds further potential. A filing for Retatrutide, a triple agonist that could set a new obesity efficacy benchmark, is planned for Q1 2027. An earnings beat, raised guidance and early signs of Foundayo adoption would confirm Lilly's incretin lead. A quarter in which price erosion offsets more of the volume gain would remind investors that the obesity market is expanding partly by becoming more accessible.

Sandisk’s contracts show how flash storage can evolve into infrastructure. Its long-term customer agreements total at least $93.9 billion, with $91.1 billion in remaining performance obligations. That converts a historically spot-priced commodity into contracted supply.

AI inference and retrieval workloads need fast, high-capacity storage close to compute. Enterprise Solid-State Drives (SSDs) are replacing hard disks in parts of the data centre and NAND supply has been disciplined after years of losses.

Investors will focus on the durability of pricing within those agreements, whether they carry take-or-pay or prepayment terms, the pace at which the industry adds wafer capacity and the mix shift towards enterprise and data-centre products.

Sandisk sits in Memory and Storage, the second layer of our seven-layer AI architecture and offers the clearest test of memory’s shift from a cyclical commodity to contracted supply. A print that meets the guidance with firm contract commentary and restrained capital spending would support the structural view.

Monday

Palantir reports Monday after the close. The Artificial Intelligence Platform (AIP) has turned Palantir's long sales cycles into faster, workshop-driven deployments in which customers move from pilot to production in weeks.

Investors should watch US commercial revenue growth, customer count, remaining deal value and net dollar retention, which indicate whether growth is coming from new logos or deeper adoption.

US government is the second leg. Defence modernisation, the Maven Smart System and the Army's enterprise agreement give the business multi-year visibility. International commercial remains the weakest region and the one that would show whether AIP's appeal extends beyond the US.

At its Palantir’s current valuation, Chief Executive Alex Karp well-touted Rule of 40 seems the bare minimum. Palantir needs sustained acceleration across segments, not only a beat.

What changed last week

Markets began pricing the cost of growth

Last week's results drew a clear distinction. Companies that showed growth with visible returns on capital were rewarded. Companies whose growth required more capital without a clear payback were not.

Meta made the trade-off explicit. Revenue of $64.601 billion grew 26.1% y/o/y and comfortably beat consensus, with ad impressions up 10.0% and average price per ad up 14.0%. Meta also raised the floor of its full-year CapEx range to $135 to $145 billion, reported negative FCF of $2.120 billion and told investors it would ‘continue investing aggressively’ in 2027 without giving a figure. A $9.89 billion legal charge related to its agreement with 52 state attorneys general cut EPS to $3.83, against $7.40 excluding the charge. The shares fell almost seven percent the following day.

Tesla gave the opposite answer and still did not win outright. Automotive gross margin excluding credits improved to 18.1% from 16.3% in Q2, FCF was positive at $1.530 billion against an expected outflow and management cut full-year capex to about $20 billion from more than $25 billion. Operating margin fell to 2.6% as R&D rose 69.4%. CapEx discipline helped, but investors still want evidence on autonomy and robotics.

Netflix showed what happens when guidance trims the next growth driver. Revenue of $12.874 billion and EPS of $0.87 beat expectations, but the advertising revenue target was cut to $2.7 billion from $3 billion. Q4 EPS guidance came in below consensus and FCF guidance was lowered, which contributed to weigh on sentiment towards Netflix ability to write its next growth chapter.

Semiconductor equipment confirmed demand and supply constraints in the same breath. Lam Research reported revenue of $8.22 billion, guided Q4 to $8.6 billion, below expectations, and said 2027 wafer fabrication equipment (WFE) spending would exceed $200 billion, with DRAM growing fastest. It also said supply-chain constraints had limited its guidance. Lam’s shares fell after hours.

Aerospace and defence had the strongest week. GE Aerospace reported orders of $19.0 billion, up 49.0%, and expanded operating margin by 220 bps to 22.5%. RTX reported a record $340 billion backlog, with Raytheon's book-to-bill at 6.54x on Tomahawk and AMRAAM awards. About 160 bps of its margin gain came from IEEPA tariff cost recovery. Lockheed Martin's backlog reached a record $292.0 billion, including a $52.5 billion PAC-3 MSE award, and it raised FCF guidance to $7.8 to $8.0 billion. Northrop Grumman raised EPS guidance by $0.70 on a record $110 billion backlog. Conversion, not demand, is now the story.

GM showed that pricing power and capital return can coexist. Record revenue of $50.015 billion, an EBIT-adjusted margin of 8.2% and a new $20 billion buyback accompanied guidance raised to $14.5 to $15.5 billion of EBIT-adjusted.

Commodity producers turned windfalls into cash. Valero reported adjusted EPS of $23.90, with elevated crude freight widening heavy-sour discounts, and paid out 52.0% of cash flow. Newmont reported all-in sustaining costs (AISC) of $1,545 per ounce, well below expectations, and FCF of $1.924 billion.

Bloom Energy grew revenue 138% and still fell 8.8% the following day, with operating cash flow negative and sell-side analysts’ concerns on margin pressure.

Taken together, the week showed that beats now matter less than the quality of what supports them and the company’s ability to raise guidance beyond the Street’s estimates to sustain momentum. This week's docket moves the same test from the advertising platforms to the clouds and from the factories to the grid.

While every effort has been made to verify the accuracy of this information, EXT Ltd. (hereafter known as “EXANTE”) cannot accept any responsibility or liability for reliance by any person on this publication or any of the information, opinions, or conclusions contained in this publication. The findings and views expressed in this publication do not necessarily reflect the views of EXANTE. Any action taken upon the information contained in this publication is strictly at your own risk. EXANTE will not be liable for any loss or damage in connection with this publication.