Thursday, 5 November 2026 | Information cutoff: 06:45 GMT
Morning View
The overnight earnings set sharpens the central AI investment question: demand is broadening, but the distribution of economics is becoming less uniform. Qualcomm reiterated a $5bn fiscal 2027 data-center revenue target and Arm lifted evidence of server share gains, while Coherent and GlobalFoundries showed the optical and packaging layers accelerating. Yet Qualcomm’s warning that wafer, memory, packaging and test costs will depress chip margins until pricing catches up is the clearest evidence that supply-chain inflation can coexist with booming AI revenue. The next phase of estimate revisions will depend more on gross-margin ownership than on exposure labels.
Software results reinforce the same distinction. AppLovin combined 55% growth with an 84% adjusted EBITDA margin and a materially better outlook, while Datadog’s 35% growth and strong beat were overwhelmed by conservative assumptions for its largest customer. Figma produced unusually concrete evidence of agent and credit adoption, but the market focused on fourth-quarter deceleration; HubSpot’s strong quarter similarly failed to offset a below-consensus outlook. AI monetization is becoming measurable, but investors are no longer accepting usage growth without visibility into durable bookings, inference costs and forward revenue.
The morning is incrementally positive for technology fundamentals but mixed for positioning. The strongest read-throughs favor intellectual-property owners, optical suppliers and platforms with demonstrated pricing or operating leverage. The weaker read-throughs fall on handset-exposed semiconductors, software companies carrying demanding forward multiples, and vendors whose top-line growth requires continued spending before cash conversion improves. The market is rewarding evidence, not narratives.
1. Qualcomm’s $5bn data-center target is real, but margin compression changes the quality of the diversification story
Qualcomm reported fiscal fourth-quarter non-GAAP revenue of $10.44bn, down 7% year over year, and non-GAAP EPS of $2.17, broadly matching expectations, while renewing long-term licensing agreements with Apple, Huawei and Transsion and reiterating a $5bn fiscal 2027 data-center revenue target. Automotive revenue reached a record $2.0bn, up 89%, but handset revenue fell 26% as industry shipments declined and Apple continued moving toward internal modems. The genuinely new information is the margin warning: broad-based inflation across wafers, assembly, test, advanced packaging and memory is expected to push the chip business’s operating margin to 20%-22% in the December quarter from 24%, with pricing actions not fully benefiting gross margin until the second half of fiscal 2027. Shares fell about 7.5% after hours after rising 5.3% in the regular session, a rational reversal because diversification is arriving with lower near-term economics. Bulls can point to Amazon collaboration, data-center scale and non-handset revenue expected to grow more than 60%; bears will argue that custom silicon, automotive and IoT cannot yet offset handset erosion or the loss of Apple content. Advanced Micro Devices, Arm, Marvell Technology, Broadcom and hyperscaler silicon teams are the competitive read-throughs, while Taiwan Semiconductor Manufacturing Company, memory vendors and packaging suppliers gain pricing leverage. The next test is whether first-quarter revenue of $10.4bn-$11.2bn holds while the chip margin troughs, and whether data-center revenue carries acceptable incremental margin. Sources: Qualcomm results; margin analysis and market reaction.
2. Arm is moving from architecture supplier to silicon vendor, increasing both upside and channel conflict
Arm reported fiscal second-quarter revenue of $1.44bn, up 27%, with licensing revenue up 50% to $773m and royalty revenue up 8% to $670m; non-GAAP EPS of $0.52 exceeded the roughly $0.47 consensus. Data-center royalties doubled, and management expects Arm-based server share at the largest hyperscalers to exceed 60% by the end of calendar 2026 after surpassing 50% in 2025. The strategic change is Arm AGI CPU: the company now has purchase orders from multiple customers through calendar 2027, first commercial deliveries remain on track for the March 2027 quarter, and confidence has increased in generating as much as $2bn across fiscal 2027 and 2028. Third-quarter revenue guidance of about $1.54bn was only modestly above consensus, and shares slipped roughly 1% after hours as royalty growth missed expectations. Bulls see a higher-value capture model layered onto an unmatched ecosystem; bears see direct silicon sales competing with licensees, adding inventory, supply and customer-concentration risk to an asset-light royalty story. Intel and Advanced Micro Devices face further server pressure, while Nvidia, Amazon, Microsoft, Alphabet and Oracle gain from a common power-efficient architecture. Investors should watch royalty acceleration from the current licensing wave, AGI CPU gross margin and whether customers view Arm as a partner or an increasingly direct competitor. Sources: Arm shareholder letter; consensus and market reaction.
3. AppLovin’s rebound resets the AI-disintermediation debate in advertising software
AppLovin delivered third-quarter revenue of $2.18bn, up 55%, net income of $1.44bn and adjusted EBITDA of $1.84bn, equivalent to an 84% margin; free cash flow was $1.4bn. Fourth-quarter revenue guidance of $2.37bn-$2.40bn was roughly 6% above the $2.25bn consensus midpoint, and the board added $3.2bn to the buyback authorization after the company repurchased or withheld $1.4bn of shares during the quarter. Shares rose more than 20% after hours following an 11% regular-session gain, a large move but understandable after the stock had fallen nearly 60% during 2026 and the result reversed the prior quarter’s guidance shock. The bull case is that Axon’s targeting economics, self-service expansion and proprietary data remain defensible despite frontier models lowering the cost of creative generation; the bear case is that one quarter does not eliminate regulatory, platform-dependence and customer-acquisition risks. The result is a negative competitive read-through for Unity and smaller mobile-ad networks and a more nuanced one for Meta Platforms and Alphabet: cheaper creative can expand ad supply, but performance platforms with superior conversion data retain the economics. The decisive catalysts are self-service adoption outside gaming, advertiser concentration, regulatory scrutiny and whether the 84%-85% guided adjusted EBITDA margin persists as the addressable market broadens. Sources: AppLovin results; consensus and market reaction.
4. Coherent turns the copper-to-optics transition into both growth and margin leverage
Coherent reported fiscal first-quarter revenue of $2.37bn, up 50% year over year and above the roughly $2.31bn consensus, with non-GAAP EPS of $2.09 versus an estimated $1.99 and non-GAAP gross margin of 40.6%, up 184 basis points. The company guided the December quarter to $2.55bn-$2.75bn of revenue and $2.28-$2.48 of non-GAAP EPS, expects fiscal 2027 pro forma revenue growth above 60%, and now targets its first quarter above 45% non-GAAP gross margin before the end of calendar 2028. Shares rose about 8% after hours, reversing part of the regular-session decline. The change is that the optical bottleneck is no longer only a unit-growth thesis: pricing, manufacturing scale and product mix are translating into operating leverage. Bulls see 800-gigabit, 1.6-terabit, co-packaged optics and increasingly optical scale-up networks supporting multiple years of growth; bears worry that customer concentration, capacity additions and eventual normalization will pressure pricing. Lumentum, Marvell Technology, Broadcom, Arista Networks and Fabrinet receive positive demand read-throughs, while copper interconnect suppliers face content displacement. Watch 1.6-terabit yields, customer concentration, capex discipline and whether the path from 40.6% to more than 45% gross margin survives rapid capacity expansion. Sources: Coherent results; consensus and market reaction.
5. Datadog’s 35% growth is not enough when the largest customer shapes the guide
Datadog reported third-quarter revenue of $1.20bn, up 35%, and non-GAAP EPS of $0.81, above consensus of roughly $1.15bn and $0.64; non-GAAP operating margin expanded to 28%, and free cash flow reached $251m. Customers generating at least $100,000 of ARR increased 20% to about 4,860. The shares nevertheless fell sharply, by roughly 12%-19% in premarket indications, because fourth-quarter revenue guidance of $1.19bn-$1.20bn implies little sequential growth and assumes that the company’s largest customer consumes only its contractual minimum after renewing above minimum commitments in the third quarter. That conservatism creates upside potential, but it also exposes how much a single AI-native account influences quarterly expectations. The bull case is that model training, inference, logs and agent observability structurally expand consumption and that Scout reduces Datadog’s own inference costs; the bear case is that hyperscalers, open-source tools and customer optimization constrain net expansion as AI workloads mature. Dynatrace, Elastic, Splunk, New Relic and cloud-native monitoring services are the primary read-throughs. The market reaction appears excessive if the guide is deliberately de-risked, but justified if large-customer concentration masks weaker broad-based consumption. Watch organic net retention, customer concentration and whether fourth-quarter usage tracks above contractual floors. Sources: Datadog results; consensus and market reaction.
6. Figma provides unusually concrete agent monetization evidence, but growth still decelerates
Figma reported third-quarter revenue of $391.4m, up 43% and $16.4m above the high end of its prior guidance, with 131% net dollar retention, an 82% non-GAAP gross margin and an 11% free-cash-flow margin. More than 80% of customers above $10,000 of ARR were consuming AI credits weekly by the end of October, 65% were using the Figma agent, and Model Context Protocol usage tripled sequentially; the share of work powered by Figma’s own post-trained design model more than doubled, improving inference economics. The company raised full-year revenue guidance by $33m to about $1.50bn, yet fourth-quarter guidance of $402m-$404m implies growth slowing to 33%, and shares fell more than 7% after hours. The central debate is whether code generation commoditizes design or makes the collaborative design layer more valuable. The adoption data favor the latter, but investors need proof that credit consumption becomes durable revenue without eroding gross margin. Adobe, Canva, Atlassian, Microsoft, GitHub and AI coding vendors are exposed as design, prototyping and production code converge. Watch paid credit expansion, Code Layers adoption, seat growth versus usage pricing and whether the company can narrow the 30% GAAP operating loss while sustaining growth above 30%. Sources: Figma results; market reaction.
7. GlobalFoundries is repositioning mature-node capacity around AI optics and packaging
GlobalFoundries reported preliminary third-quarter revenue of $1.90bn, up 13%, non-IFRS gross margin of 31.4% and adjusted free cash flow of $136m. The headline IFRS loss of $270m was driven mainly by a $441m non-cash charge associated with shares issued to the U.S. government, while underlying gross margin benefited from a richer mix of AI-related applications. The strategic development is more important than the accounting noise: an amended CHIPS Act award reallocates $955m toward silicon photonics, silicon germanium and advanced packaging, and a $2bn agreement with Taiwan Semiconductor Manufacturing Company will establish U.S. production of silicon interposers for CoWoS beginning in the first half of 2028. This gives GlobalFoundries participation in leading-edge AI systems without competing directly at the leading logic node. Bulls see differentiated domestic capacity, optical demand and government support lifting utilization and mix; bears see long lead times, policy dependence and heavy capital intensity before revenue ramps. Taiwan Semiconductor Manufacturing Company, Marvell Technology, Monolithic Power Systems and U.S. packaging suppliers are the clearest beneficiaries, while existing Asian interposer capacity faces diversification pressure. Monitor capital spending, milestone receipts, photonics yields and binding customer commitments before the 2028 ramp. Source: GlobalFoundries results.
8. HubSpot’s strong quarter cannot offset a sharper fourth-quarter slowdown
HubSpot reported third-quarter revenue of $947.6m, up 17%, and non-GAAP EPS of $3.81, above consensus of roughly $925m and $3.27; non-GAAP operating margin expanded 270 basis points to 22.6%. Fourth-quarter guidance of $945m-$947m and $3.85-$3.87 of non-GAAP EPS fell below market expectations of about $965m and $4.09, sending shares down roughly 8% after hours. The issue is not current execution but the durability of growth among small and mid-sized customers as demand remains uneven and AI monetization is still early. Bulls see a unified customer platform with expanding agents, data and commerce products supporting consolidation and margin leverage; bears see generative AI reducing differentiation in marketing content while Salesforce, Microsoft and low-cost point tools compete for the same budget. The lower share count from $237m of third-quarter repurchases supports EPS, but it does not answer the organic-growth question. Salesforce, Adobe, Intuit, Klaviyo and private customer-engagement vendors are the principal read-throughs. The next catalysts are net customer additions, seat and credit monetization, partner-channel productivity and evidence that fourth-quarter guidance reflects prudence rather than a structural deceleration toward low teens. Sources: HubSpot SEC filing; consensus and market reaction.
9. Rapid7’s rally prices stabilization before growth has actually returned
Rapid7 reported third-quarter revenue of $212m, down roughly 3%, and ARR of $820m, below $824m in the prior quarter, while non-GAAP operating income reached $46.2m and free cash flow was $26.5m. Adjusted EPS of $0.60 and revenue exceeded consensus of about $0.45 and $209m, and fourth-quarter guidance of $208m-$210m was broadly in line. Shares rose about 16% after hours as investors rewarded early evidence that the restructuring is stabilizing the core business and preserving the approximately $130m full-year free-cash-flow outlook. The bull case is that a smaller, more focused exposure-management and managed-detection portfolio can return to growth with materially better cash economics; the bear case is that flat core ARR and continued non-core churn show share loss to platform vendors is not over. CrowdStrike, Palo Alto Networks, Tenable, Qualys, Microsoft and managed-security providers compete for the same consolidation budget. The rally is primarily a sentiment reset rather than an estimate-revision event because total ARR still contracted sequentially. Watch core ARR, customer retention, managed-service bookings and whether reinvestment in AI and service delivery can produce durable growth without reversing margin gains. Sources: Rapid7 results; consensus; market reaction.
10. Fastly’s edge turnaround is becoming measurable, but concentration is rising
Fastly reported record third-quarter revenue of $190.2m, up 20%, with security revenue up 32%, non-GAAP gross margin of 65.8% and free cash flow of $25.2m. Net retention improved to 119%, its highest level since 2022, RPO increased 24% to $334m and adoption of four or more product suites among large customers rose 15 points to 34%. Fourth-quarter revenue guidance of $198m-$202m was slightly above the roughly $197m consensus, while EPS guidance of $0.16-$0.18 was well ahead of approximately $0.09. The change is that the turnaround now combines accelerating usage, better mix and operating leverage rather than relying solely on cost control. The caveat is concentration: the ten largest customers generated 39% of revenue, up from 32%, which increases volatility and negotiating risk. Bulls see AI inference, API security and real-time edge control driving a durable reacceleration; bears see Cloudflare, Akamai Technologies and hyperscaler networks constraining pricing and enterprise share. The next evidence should come from large-customer diversification, security attach, RPO conversion and whether gross margin can remain at record levels as traffic grows. Sources: Fastly results; consensus.
What to Watch
Before the next briefing, Cloudflare, Atlassian, Akamai Technologies, Lumentum, Microchip Technology, JFrog and Bill.com are scheduled to report, providing direct tests of edge demand, developer spending, optical capacity, analog inventory and software consumption. Investors should also monitor whether Qualcomm’s after-hours weakness spreads to handset and memory suppliers, whether Coherent’s guide produces follow-through across optics, and whether Datadog clarifies the contribution and contractual floor of its largest customer. Over the following week, the most important confirmation would be sustained bookings and margin guidance from software and infrastructure vendors after the strong October equity rerating.
Bottom line
AI demand remains strong across compute, optics, software usage and advertising, but the earnings set makes the economics increasingly company-specific. Arm, Coherent and AppLovin show the value of scarce intellectual property, manufacturing know-how and proprietary data; Qualcomm shows how input-cost inflation and legacy exposure can dilute even a credible data-center ramp. In software, reported beats are no longer enough when the forward guide decelerates or concentration rises. The preferred exposure remains businesses that can convert AI demand into pricing, margin and free cash flow without relying on a single customer or distant capacity promises.