| Ticker | Name | Research | View | What's said | Source |
|---|---|---|---|---|---|
| SNPS | Synopsys | QT · SA · STK · FA | Positive | EDA reaccelerates — the organic proof that was missing in Q2. Q3 revenue +42% Y/Y to $2.5B ($40M beat) with non-GAAP EPS of $3.91 ($0.24 beat). "The Ansys acquisition still drives much of the headline growth, but core EDA revenue grew 8.5%, with management now expecting double-digit organic growth in Q4." Design IP returned to growth, +11% Y/Y to $474M after falling 6% in Q2 — AI infrastructure is helping, with Synopsys winning more than 95% of PCIe 7 opportunities and its die-to-die IP business "on pace to double this year." The Ansys integration is moving into commercialization: the first Multiphysics Fusion products launched and the agentic AI platform now has 30+ active customer engagements, with meaningful revenue contribution expected in FY27. FY26 revenue guidance raised $50M to ~$9.715B. Bottom Line: "Q2 left investors waiting for stronger organic growth. Q3 delivered it, with EDA accelerating and Design IP back in growth mode. The September Investor Day now becomes the key test for how much Multiphysics Fusion and AI can lift FY27 growth." | article ↗ |
| ADSK | Autodesk | QT · SA · STK · FA | Positive | MaintainX lands cleanly. Q2 revenue +16% Y/Y to $2.05B ($40M beat) with non-GAAP EPS of $3.30 ($0.18 beat); constant-currency growth was softer at 14%, and operating margin expanded 4pp to 29%. The core held up through the ongoing go-to-market reorganization — cRPO +12%, total RPO $7.4B, underlying growth expectations raised, with construction and emerging markets "particularly strong." The $3.6B MaintainX acquisition closed just after quarter-end, expected to add roughly $60M of FY27 revenue, and Autodesk says "operating leverage and sales efficiencies should fully offset the acquisition's non-GAAP margin dilution." The first concrete validation of the "design-to-operate" strategy also emerged: a major global retailer adopted Forma for construction data and Tandem for digital twins. FY27 revenue guidance raised ~$135M to $8.295–$8.345B with the non-GAAP margin outlook intact; Q3 revenue guidance came in ahead of consensus, EPS slightly below. Bottom Line: "The core business remains resilient, margins are improving, and MaintainX has closed without changing the long-term profitability trajectory. The next test is whether Autodesk can turn that operational-data acquisition into meaningful cross-sell rather than simply a larger TAM story." | article ↗ |
| VEEV | Veeva Systems | QT · SA · STK · FA | Positive | Falcon lands early adopters — and agentic AI reaches production. Q2 revenue +18% Y/Y to $928M ($23M beat) with non-GAAP EPS of $2.35 ($0.13 beat); subscription revenue +16% to $767M, and shares jumped after Veeva raised its full-year outlook. Veeva Falcon now has five early adopters, with first go-lives expected this year; Veeva also launched Falcon MLR for automated marketing-content review and acquired Copli — but "importantly, this is still early adoption rather than meaningful revenue contribution." Vault CRM had its best quarter yet, crossing 180 live customers from 150+ in Q1, with five top-20 biopharmas live and 12 top-20 commitments as of August; one top-20 customer "deployed Veeva's Agentic Call Report across its entire US field team, providing a more tangible example of AI moving from product roadmap to production." Momentum extends beyond CRM: Veeva Safety crossed 100 customers, Quality added 30+, and newer R&D products are "increasingly taking over growth from mature applications such as eTMF and QualityDocs." FY27 revenue guidance raised again to $3.682–$3.687B (from $3.635–$3.645B). Bottom Line: "Vault CRM remains the near-term growth engine, but Veeva's agentic-labor opportunity is starting to move beyond the concept stage." A disclosed author holding. | article ↗ |
| AFRM | Affirm Holdings | QT · SA · STK · FA | Positive | Growth meets leverage. Q4 revenue +~34% Y/Y to $1.17B ($90M beat) with GMV +36% to $14.1B, "comfortably ahead of the $13.4B consensus," and adjusted operating margin expanding to 30% (from 27%). GAAP EPS of $4.77 "was distorted by a $1.45B tax benefit, so it isn't particularly meaningful." Growth was broad rather than single-product: direct merchant integrations generated roughly half of GMV growth, with direct-to-consumer (including the Affirm Card) driving much of the rest; card penetration is ~19% of active consumers, and Pay-in-4 keeps benefiting from merchants adopting evergreen 0% financing. Active consumers +21% to 27.8M, funding capacity $30B, credit performance described as stable — and the stated discipline is that Affirm "would slow originations before allowing underwriting standards to deteriorate." FY27 guidance: >$64B of GMV (from $50.2B in FY26), revenue above ~$5.4B, adjusted operating margin above 30.5%. Bottom Line: "Affirm is showing it can sustain momentum as Pay-in-4 comparisons became harder. The story is increasingly less about one product carrying growth and more about building a broader payments network across merchant checkout, Card, and new geographies." | article ↗ |
| OKTA | Okta | QT · SA · STK · FA | Positive | Agent deals arrive — but the payoff comes through the existing platform. Q2 revenue +11% Y/Y to $805M ($12M beat) with non-GAAP EPS of $1.05 ($0.09 beat) and free cash flow of $227M; shares surged on a raised FY27 outlook. Okta for AI Agents closed dozens of deals in Q2, including several worth more than $1M, with average AI deal sizes above the company average — "but management was equally clear that AI-specific revenue is still immaterial and is unlikely to become meaningful in FY27. Instead, the bigger near-term benefit is that AI-agent security concerns are prompting customers to modernize their broader identity infrastructure." That platform strategy is working: new products were 30% of bookings, up from 25% in Q1, deals including them carry ~40% higher ACV, customers spending >$1M annually grew over 20% to more than 600, making Q2 "Okta's strongest non-Q4 bookings quarter ever"; Identity Governance is the largest contributor among newer products and current RPO accelerated to 14% from 12%. FY27 revenue guidance raised to $3.216–$3.226B (10%–11% growth) with FCF margin to 28%–29%, still absorbing "roughly a one-point revenue headwind from moving professional services to partners." Bottom Line: "The more important AI payoff may arrive first through Okta's existing platform rather than a standalone agent-security revenue line." | article ↗ |
| RBRK | Rubrik | QT · SA · STK · FA | Positive | ARR picks up — and it accelerated before the acquisition contributes anything. Q2 revenue +38% Y/Y to $427M ($31M beat) with non-GAAP EPS of $0.20 ($0.16 beat); subscription ARR +33% Y/Y to $1.66B, accelerating from 32% in Q1, net retention above 119%, and customers spending $100K+ in ARR up 23% to 3,084. Rubrik Agent Cloud added support for Claude Code, and five major consulting firms joined Project Hourglass to deploy agent-security capabilities for enterprises. Rubrik also acquired Strata Identity — technology "designed to keep identity systems operational during cyberattacks" — but "importantly, Strata contributed no ARR in Q2 and is excluded from FY27 guidance." Profitability improved alongside: subscription ARR contribution margin +~460 bps to 14%, free cash flow $66M at a 15% margin. FY27 subscription ARR guidance raised ~$25M to $1.88B (~29% growth), revenue guidance raised ~$46M to $1.69B, FCF now expected at $323–$333M. Bottom Line: "Q2 was encouraging… The next test is whether these newer products can keep ARR growth near 30% as the base approaches $2 billion." A disclosed author holding. | article ↗ |
| NTNX | Nutanix | QT · SA · STK · FA | Positive | Hardware workarounds scale — the constraint became a timing problem. Q4 revenue +16% Y/Y to a record $757M ($19M beat) with non-GAAP EPS of $0.60 ($0.11 beat) — though "the reported performance was polluted by a massive non-recurring $1.2B tax benefit." ARR growth accelerated to 16% at $2.5B, free cash flow margin reached 37%, and Nutanix added more than 3,000 customers in FY26. The server shortage is still real, "but customers are increasingly routing around it": external storage and NC2 bookings both jumped sharply Q/Q, letting customers adopt Nutanix "without replacing existing hardware or by moving workloads into the cloud," with NetApp support already helping win large deals and management expecting these alternatives to accelerate in FY27. The headwind persists — "server lead times are stabilizing, but Nutanix expects shortages and price increases to persist throughout FY27, with more customers consequently delaying contract start dates." FY27 guidance: revenue $3.18–$3.23B (~12%–13% growth), non-GAAP operating margin 24%–25%, free cash flow $850–$950M. Bottom Line: "Supply remains a headwind in FY27, but increasingly looks like a timing problem rather than a demand problem." | article ↗ |
| ESTC | Elastic | QT · SA · STK · FA | Positive | AI penetration jumps. Q1 revenue +15% Y/Y to $478M ($8M beat) with non-GAAP EPS of $0.70 ($0.12 beat); sales-led subscription revenue grew 18% to $399M, while cRPO accelerated to 21% and total RPO grew 27%. The AI story became more measurable: Elastic ended Q1 with 670+ $100K-ACV customers using AI, up roughly 70 Q/Q, and AI penetration within that high-value cohort reached 37%, from ~21% a year ago — "management says these AI users continue to grow faster than customers who are not using AI." Large-customer momentum strengthened too, with 80+ net additions to the $100K ACV cohort, taking the total above 1,800. Product-wise, Observability gained a new Columnar Mode that cuts time-series storage costs by roughly 20%, and the acquisition of Deductive AI adds automated incident investigation for an emerging "agentic SRE" offering. FY27 revenue guidance raised $12M to $1.998–$2.010B (~15% growth) with sales-led subscription revenue expected ~+17% and non-GAAP operating margin up to 19%. Bottom Line: "Q1 showed growth reacceleration might be back-half weighted… The next test is whether that higher AI penetration can push overall revenue growth meaningfully above the mid-teens as FY27 progresses." A disclosed author holding. | article ↗ |
| MRVL | Marvell Technology | QT · SA · STK · FA | Neutral | Custom silicon breakout — and a beat-and-raise that still fell. Q2 FY27 (ending August 1) revenue +37% Y/Y to a record $2.74B ($30M beat) with non-GAAP EPS of $0.94 ($0.01 beat). Data Center revenue jumped 46% to $2.17B, 79% of total sales, "as demand for AI infrastructure remained exceptionally strong." The next leg is custom: management expects "a significant acceleration beginning in H2 FY27, followed by custom silicon revenue more than doubling in FY28 as hyperscalers increasingly design their own AI chips," and Marvell expanded its Google partnership across AI accelerators, storage, networking and near-memory compute, "backed by a six-year warrant agreement." Connectivity is the second engine — optical interconnect demand accelerating as AI clusters need more bandwidth, with the CXL memory-expansion business called a "home run" across multiple hyperscalers. Guidance moved up on both years: FY27 to ~$12B (from $11.5B) and FY28 to $18B (from $16.5B), with Q3 revenue of $3.15B well ahead of consensus and implying "another ~15% sequential jump." Bottom Line: "But expectations are already enormous: the stock trades over 50x forward earnings, and much of the Google opportunity through FY28 was already embedded in guidance. That helps explain why shares fell despite Marvell raising both FY27 and FY28 outlooks." | article ↗ |
| INTU | Intuit | QT · SA · STK · FA | Neutral | DIY price reset — a good quarter and a materially lower bar. Q4 revenue +14% Y/Y to $4.35B ($130M beat) with adjusted EPS of $4.03 ($0.44 beat); Global Business Solutions +14% and Consumer +14% (primarily Credit Karma). "Yet shares fell sharply as FY27 guidance called for revenue growth of just 9%–10%, down from 14% in FY26." The reset is TurboTax: "management acknowledged that price has become the #1 reason DIY customers leave, particularly among filers earning around $50,000. Intuit now plans to accept lower revenue per customer to win those users back," potentially including more free offerings — so TurboTax revenue is expected to grow just 2%–3% in FY27. The same acquisition push is reaching QuickBooks: QuickBooks Free launched with 20,000+ customers already using it or converting, "the idea is to acquire businesses cheaply upfront and monetize them later through payments, payroll, lending, and other services." Mailchimp remains the weak spot, flat to −1% in FY27. Guidance: FY27 revenue $23.3–$23.5B, below consensus, GBS 13%–14%, Credit Karma 11%–13% — with the caveat that the huge FY27 EPS shortfall "is less concerning than it looks because Intuit will begin including stock-based compensation in its non-GAAP results, making the new EPS guidance incomparable with prior estimates." Bottom Line: "FY27 now becomes a test of whether this is a deliberate reset before reacceleration or evidence that competition is structurally weakening Intuit's pricing power." A disclosed author holding. | article ↗ |
| WDAY | Workday | QT · SA · STK · FA | Neutral | AI mix jumps — adoption is real, the meter isn't running yet. Q2 revenue +13% Y/Y to $2.65B ($10M beat) with non-GAAP EPS of $2.75 ($0.14 beat); subscription revenue +14% to $2.47B and operating margin slightly expanded, ahead of expectations. The AI traction "became much more tangible": AI products generated more than $100M in new ACV, over 25% of all new ACV in the quarter; more than 5,500 customers now use at least one Workday agent, up 35% Q/Q; and AI SKUs are approaching $600M in ARR, up more than 200% Y/Y. "The catch is the timing of monetization. Only a little over 200 customers have signed up for Flex Credits, despite thousands already using agents, as Workday is deliberately prioritizing adoption before metering some AI usage" — management still sees Flex Credits as the answer to seat compression but calls it a "delayed gratification" model "whose revenue impact may take months to appear." The core is decelerating underneath: FY27 subscription growth held at ~13% with adjusted operating margin guidance +0.5pp to 31%, but management expects FY28 subscription growth of roughly 11% (with ≥2 points of margin expansion) and total subscription backlog grew only 8% to $27.4B. Bottom Line: "The next test is whether today's AI adoption turns into enough consumption revenue to offset that deceleration." | article ↗ |
| ZM | Zoom Communications | QT · SA · STK · FA | Neutral | Enterprise momentum returns — on a business still growing 5%. Q2 revenue +5% Y/Y to $1.28B ($10M beat) with non-GAAP EPS of $1.55 ($0.07 beat). Enterprise revenue accelerated to 8% growth, its fastest pace in three years, and is now 62% of total revenue; Online grew just 1% with monthly churn roughly steady at 2.9%. AI is driving the enterprise expansion: licensed monthly active users of AI features +125% Y/Y, Zoom Virtual Agent customers +256%, the Customer Experience suite growing at a high-double-digit rate with AI included in 9 of Zoom's 10 largest CX deals, and Zoom Phone taking share — "all 10 of its largest deals involved competitive replacements." Beyond meetings: Workvivo crossed $100M in ARR and the Common Room acquisition adds buyer intelligence to Zoom Revenue Accelerator, with the goal "to connect collaboration, customer service, sales, and employee experience through a common AI layer." FY27 revenue guidance raised $23M to $5.090B at the midpoint and adjusted EPS from ~$5.98 to ~$6.10; "Q3 guidance was less exciting, with revenue roughly in line with consensus and EPS slightly below." Bottom Line: "But Zoom is still a roughly 5% grower overall. The next step is proving that these newer AI and CX products can become large enough to meaningfully accelerate the top line." A disclosed author holding. | article ↗ |
| BBY | Best Buy | QT · SA · STK · FA | Neutral | PC prices do the work. Q2 revenue +4% Y/Y to $9.8B ($190M beat) with adjusted EPS of $1.47 ($0.09 beat); comparable sales accelerated to 4.1%, well above the ~1% outlook, with positive comps "across nearly every major category despite lapping last year's Switch 2 launch." But the mix matters: computing was the biggest contributor while average PC selling prices rose by the mid-teens as memory costs increased and unit volumes fell by the high single digits — growth bought by inflation, not demand. Home theater also strengthened, and emerging categories such as AI glasses and collectibles more than doubled, contributing roughly one point of comp growth. The higher-margin businesses keep scaling: Best Buy Marketplace reached $300M in quarterly GMV with FY27 expectations raised to $1.3B, and Best Buy Ads contributed to gross-margin expansion; Best Buy also completed its OpenAI commerce integration and launched Ask Blue, its own conversational shopping assistant. FY27 comparable-sales guidance raised to 1.9%–3% (from −1% to +1%), with Q3 comps expected at 1%–3% "even as the company laps stronger Windows 10-driven computing demand." Bottom Line: "PC growth is increasingly price-led rather than unit-led, making Marketplace, Ads, and newer hardware categories more important as computing comparisons get tougher in H2." | article ↗ |
| HQY | HealthEquity | QT · SA · STK · FA | Neutral | Margins keep climbing — while the asset engine cools. Q2 revenue +8% Y/Y to $351M ($1M beat) with non-GAAP EPS of $1.24 ($0.05 beat); adjusted EBITDA +11% to $167M, "pushing margin to a record 48% from 46% a year ago. Shares nevertheless fell sharply after the print." The HSA flywheel is healthy but slowing: HSA assets +14% to $37.9B (moderating from Q1), accounts at a record 10.7M with record Q2 new-account sales, and custodial revenue +10% to $176M. Technology is doing the margin work — mobile monthly active users +62% Y/Y to 1.4M, and "AI now resolves 85% of routine chat inquiries and handles 55% of card-related phone contacts without an employee." Marketplace active members rose to 14,000 (from "10,000+" in Q1), "although management says revenue remains immaterial." FY27 guidance nudged up to $1.411–$1.421B of revenue with adjusted EBITDA to $628–$636M. Bottom Line: "HSA asset growth has cooled from 19% to 14%, so the next test is whether these efficiency gains can keep earnings compounding faster than the high-single-digit top line." A disclosed author holding. | article ↗ |
| PDD | PDD Holdings | QT · SA · STK · FA | Negative | Growth slows again. Q2 revenue +8% Y/Y to $16.6B ($0.3B miss), "decelerating from 11% in Q1," while non-GAAP EPADS of $2.85 beat by $0.12. "Both revenue engines remain subdued": Online Marketing Services grew just 3% to $8.5B and Transaction Services grew 13% to $8.1B, slowing from 20% in Q1, with "domestic competition… intense as Pinduoduo battles Alibaba, JD, and social-commerce platforms for increasingly cautious Chinese consumers." Temu's structural edge is eroding: "the end of duty-free treatment for low-value parcels in the US and new European import fees are increasing fulfillment costs and reducing the appeal of shipping inexpensive products directly from China" — management warned of slower fulfillment and higher costs in affected markets and is responding by building local warehousing. The Q1 transformation continues (merchant support, supply-chain investment, platform governance, R&D) while "its new first-party brand initiative is progressing more slowly than initially expected." The balance sheet is not the issue: cash and short-term investments reached $67B and operating cash flow +19% to $3.8B. Bottom Line: "PDD has enormous financial capacity to fund the transition, but investors are still waiting for evidence that all this spending can restart the growth engine." | article ↗ |
"View" is App Economy's analytical framing in this PRO recap, taken from each section's Bottom Line — positive where the print delivered the thing that was in doubt and the guide went up cleanly (SNPS's organic reacceleration, ADSK's margin path surviving MaintainX, VEEV's Falcon adopters + raised guide, AFRM's GMV beat and FY27 step-up, OKTA's raised outlook and record non-Q4 bookings, RBRK's ARR accelerating pre-Strata, NTNX's ARR accelerating through a supply constraint, ESTC's cRPO + AI-penetration jump); neutral where the Bottom Line is explicitly two-sided (MRVL's beat-and-raise against 50x forward and a pre-embedded Google guide, INTU's chosen price reset framed as "deliberate reset or weakening pricing power," WDAY's AI adoption against 8% backlog and ~11% FY28, ZM's three-year-high enterprise growth on a 5% grower, BBY's price-led PC comp, HQY's record margin against 19%→14% asset growth); negative only for PDD, the one section whose Bottom Line has no offsetting positive. App Economy Insights is financial-analysis journalism, not a buy/sell stance — BUY/SELL/HOLD ratings are shared only with App Economy Portfolio members; the author discloses owning ESTC, HQY, INTU, RBRK, VEEV and ZM. Research: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis. The "Source" link opens the newsletter (no per-name timestamps — it's a written post). Named only in passing and not given rows: Google (Marvell's expanded AI-accelerator partnership and six-year warrant), Alibaba and JD (PDD's domestic competition), NetApp (Nutanix external-storage partner), OpenAI (Best Buy's completed commerce integration), Anthropic/Claude Code (supported by Rubrik Agent Cloud), Nintendo (the lapped Switch 2 launch), Microsoft (the lapped Windows 10 refresh), and the products/units described inside their parents' rows — Temu and Pinduoduo (PDD); TurboTax, QuickBooks, QuickBooks Free, Credit Karma, Mailchimp (Intuit); Ansys, Multiphysics Fusion (Synopsys); MaintainX, Forma, Tandem (Autodesk); Flex Credits (Workday); Veeva Falcon, Falcon MLR, Copli, Vault CRM, Veeva Safety (Veeva); Affirm Card, Pay-in-4 (Affirm); Okta for AI Agents, Identity Governance (Okta); Agent Cloud, Project Hourglass, Strata Identity (Rubrik); Marketplace, Best Buy Ads, Ask Blue (Best Buy); NC2 (Nutanix); Columnar Mode, Deductive AI (Elastic); Workvivo, Common Room, Zoom Virtual Agent, Zoom Phone (Zoom).
A jargon-free summary of the read behind each name. (Plain-language companion to the table above; renders on each ticker's consolidated page.)
Synopsys sells the software chip designers use to design chips. Nobody builds a modern processor by hand; they build it inside tools from Synopsys or its rival Cadence, which makes the business a toll booth on essentially every chip project in the world. It also licenses ready-made building blocks — "IP" — that customers drop into their own designs rather than engineering from scratch.
Revenue grew 42%, but that number is misleading and the newsletter says so immediately: most of it came from buying Ansys, a simulation-software company. When a company grows by acquisition, the honest question is how fast the part it already owned is growing. That figure is 8.5%, and it is the number that matters here, because last quarter the complaint was precisely that the underlying business had gone quiet. Management now expects the pre-acquisition business to grow at a double-digit rate next quarter.
The second recovery is in the building-blocks business, which had been shrinking 6% and grew 11% this quarter. AI hardware is the reason: Synopsys says it won more than 95% of the opportunities for PCIe 7 — the next-generation standard for moving data between chips inside a machine — and its "die-to-die" blocks, which let two chips in one package talk to each other, are on track to double. The Ansys purchase is also starting to produce joint products rather than just combined revenue. Guidance went up. The author flags the September investor day as the moment management has to quantify how much of this reaches next year's growth. Analysis, not a recommendation.
Autodesk makes the design software architects, engineers and manufacturers use — AutoCAD and its descendants. Revenue grew 16% and, more importantly, operating margin improved by four percentage points to 29%, meaning a bigger share of each dollar of sales is now profit. That happened while the company was in the middle of reorganizing how it sells, which usually costs a quarter or two of disruption.
The event of the quarter is the $3.6 billion purchase of MaintainX, which closed just after the quarter ended. MaintainX software runs maintenance for physical facilities — the work-order system a factory or hospital uses to keep equipment running. The strategic idea is that Autodesk already owns the design of a building or machine, and now owns the record of how it is operated and repaired; connecting the two is what Autodesk calls "design to operate."
The reason this counts as a clean deal rather than a worry is the margin arithmetic. Acquisitions normally dilute profitability for a while — you absorb the target's costs before you earn its revenue. Autodesk says efficiencies elsewhere will fully offset that, and it raised revenue guidance while leaving the margin outlook untouched. It also produced first evidence the strategy sells: a large global retailer bought both Forma (construction data) and Tandem (digital twins — a live software model of a real building). The open question the author leaves is whether Autodesk can actually sell MaintainX into its existing customers, or whether the deal just makes the addressable market look bigger on a slide. Analysis, not a recommendation.
Veeva sells software built specifically for drug companies — the systems that manage clinical trials, regulatory documents, product-safety reporting and the sales reps who visit doctors. Because pharma is heavily regulated, this software has to be validated and audited, which makes it very hard to replace once installed and very hard for a generic competitor to enter.
Two different things happened this quarter, and it is worth keeping them apart. The first is Falcon, Veeva's attempt at "agentic labor" — AI that performs a whole job function rather than assisting a person doing it. Falcon has five early customers with the first live deployments due this year. Five is a pilot programme, and the newsletter says so plainly: this is early adoption, not revenue.
The second is the part that is already paying. Vault CRM — the system reps use in the field, which Veeva built to replace a Salesforce-based product it used to resell — crossed 180 live customers from 150-odd three months ago, with five of the twenty largest drug companies now running it and twelve committed. And one of those top-20 customers put Veeva's "Agentic Call Report" — AI that writes up a sales visit automatically — across its entire US field force. That is the difference between a pilot and production, and it is the strongest evidence in this issue that agentic AI is doing real work inside a real company. Growth is also broadening beyond CRM: the safety product passed 100 customers, and newer research products are taking over from the mature ones. Guidance was raised again and the stock rose. The author owns it. Analysis, not a recommendation.
Affirm is a "buy now, pay later" lender: it pays a merchant when you check out and collects from you in instalments, earning a merchant fee and sometimes interest. The headline volume measure is GMV — the total value of purchases flowing through it — and this quarter GMV grew 36% to $14.1 billion, well ahead of what analysts expected, with revenue up about 34% and margins improving from 27% to 30%.
Ignore the reported earnings per share of $4.77. It contains a $1.45 billion tax benefit — an accounting event, not a business one — and the newsletter says outright that it "isn't particularly meaningful." This is the routine adjustment worth internalising: when a number is startlingly large, find the one-off inside it before reacting.
What makes the quarter notable is where the growth came from, because the worry with these companies has always been dependence on one product. Roughly half the growth came from deals struck directly with merchants and most of the rest from selling to consumers directly, chiefly the Affirm Card — a physical card that turns any purchase into an instalment plan, now used by about 19% of active customers. That matters because the original product, the four-instalment split at checkout, was facing harder year-over-year comparisons.
For a lender, growth is only good if the loans are good. Affirm said credit performance is stable, has $30 billion of funding lined up, and stated the discipline explicitly: it would slow lending before loosening standards. Next year it expects volume above $64 billion, up from $50 billion, with margin also higher — volume and profitability guided up together. Analysis, not a recommendation.
Okta runs identity: the system that decides who — or what — is allowed to log in to a company's software. Every employee, contractor and application authenticates through it, which makes it plumbing that is painful to rip out.
The AI story here is unusually honest, and that is what makes it useful. Okta sold "dozens" of deals for its AI-agent product, several worth over a million dollars each — and management said in the same breath that AI-specific revenue is immaterial and will probably stay immaterial next year. Most companies would let the deal count do the talking.
The actual mechanism is second-order. When a company starts deploying AI agents that can read files and call systems on their own, each agent effectively becomes a new user who needs permissions, and the security team discovers its existing identity setup was not built for that. So the agents do not generate agent revenue — they generate an identity modernization project, which is Okta's core business. The evidence that this is happening: newer products are now 30% of bookings (up from 25%), deals that include them are about 40% larger, customers spending over $1 million a year grew more than 20% to over 600, and the pipeline metric that leads revenue — current remaining performance obligations, essentially contracted revenue due in the next year — accelerated from 12% to 14%. It was the best bookings quarter Okta has ever had outside a January quarter.
Guidance went up, cash flow guidance went up, and the shares surged. The author's framing is worth keeping: the AI payoff may arrive through the existing product line rather than as a new revenue line. Analysis, not a recommendation.
Rubrik does backup and recovery — but the modern version, "cyber resilience": keeping a clean, immutable copy of a company's data so that when ransomware encrypts everything, the business can be restored rather than ransomed. It sells by subscription, so the number that matters is ARR (annual recurring revenue) — the annualized value of contracts currently running.
ARR grew 33%, and the important word is that it accelerated — last quarter it was 32%. Recurring-revenue businesses almost always decelerate as they get bigger, simply because the base is larger, so acceleration at $1.66 billion is unusual. Existing customers also spent more than 119% of what they spent a year earlier, meaning the average customer expands rather than churns.
There is a clean-accounting detail that strengthens the claim. Rubrik bought a company called Strata Identity during the quarter, and normally you would suspect an acquisition of flattering the growth number. It didn't: Strata contributed zero ARR and is excluded from next year's guidance entirely. So the acceleration is organic.
Profitability improved alongside growth — the margin on that recurring revenue rose about 4.6 percentage points and free cash flow reached $66 million — and guidance went up on ARR, revenue and cash flow. The question the author leaves is the arithmetic one: it gets progressively harder to grow 30% as the base approaches $2 billion. The author owns it. Analysis, not a recommendation.
Nutanix sells software that turns ordinary servers into a private cloud — companies run it in their own data centres instead of renting from Amazon or Microsoft. It has been the main beneficiary of customers fleeing VMware after Broadcom raised its prices.
Its problem for the past year has been physical: to adopt Nutanix the traditional way, a customer buys new servers, and servers are in short supply and getting more expensive because AI data centres are consuming the world's memory. That turned a demand story into a delivery story.
This quarter showed customers going around the obstacle instead of waiting. Two alternatives grew sharply: connecting Nutanix to storage hardware customers already own (helped by a partnership with NetApp), and NC2, which runs Nutanix inside a public cloud so no new hardware is needed at all. Recurring revenue growth accelerated to 16% — during a shortage — which is the evidence that the demand was always there and only the path was blocked.
Two cautions. The reported profit was inflated by a one-time $1.2 billion tax benefit, so ignore the earnings headline. And management expects shortages and price rises to continue through all of next year, with customers pushing back contract start dates. The author's conclusion is the right shape: this reads as a timing problem, not a demand problem. Analysis, not a recommendation.
Elastic makes search technology. Its original product indexes a company's logs and documents so they can be searched instantly; the same machinery turned out to be exactly what AI applications need, because to answer a question about a company's own data, an AI model must first retrieve the relevant documents. That put Elastic in the AI stack almost by accident.
Revenue grew 15%, but the interesting numbers are the ones that lead revenue rather than report it. cRPO — contracted revenue expected to be recognized over the coming year — accelerated to 21%, and total contracted revenue grew 27%. Both growing faster than reported revenue means the business signed more this quarter than it billed, which is what an inflection looks like before it shows up in the income statement.
The AI metric here is better constructed than most. Rather than counting anyone who touched an AI feature, Elastic reports penetration within its large customers — those spending over $100,000 a year. That figure went from about 21% a year ago to 37%, on a cohort that itself grew past 1,800 customers. And management says the AI users grow faster than the non-AI users, which is the claim that turns a feature into an expansion engine.
Two product moves support it: a new storage mode that cuts the cost of keeping time-series data by about 20% (in observability, cost is the reason customers throw data away), and the purchase of Deductive AI to build an "agentic SRE" — software that investigates an outage the way a site-reliability engineer would. Guidance rose, though the author notes the reacceleration is weighted to the back half of the year, which is always the softer kind of promise. The author owns it. Analysis, not a recommendation.
Marvell designs two things for AI data centres. First, the chips that move data around — optical interconnects that carry information between racks as light, and memory-expansion parts that let a server address more memory than it can physically hold. Second, and increasingly, "custom silicon": when Google or Amazon wants its own AI chip rather than buying NVIDIA's, it hires a company like Marvell to help design and produce it.
The quarter was excellent. Revenue grew 37% to a record, the data-centre business grew 46% and now makes up 79% of the company, and management raised its forecast for both this year and next — to $12 billion and $18 billion respectively. It expanded its partnership with Google across accelerators, storage and networking, backed by a six-year warrant agreement, which means Google took the right to buy Marvell stock. A customer accepting equity exposure to its supplier is a genuine commitment signal.
The stock fell anyway, and understanding why is the transferable lesson. Marvell trades at more than 50 times expected earnings, which means the price already assumes years of the growth just described. And the newsletter identifies the specific problem: much of the Google opportunity through 2028 was already in the guidance before this announcement. So the news added enthusiasm without adding numbers — investors learned nothing that changed the forecast they were already paying for.
That is why the view here is neutral rather than positive despite an unambiguously strong print. The business is doing well; the question is whether anything is left over after the expectations are subtracted. Analysis, not a recommendation.
Intuit owns TurboTax (do-it-yourself tax filing), QuickBooks (small-business accounting), Credit Karma and Mailchimp. The quarter itself was strong — revenue up 14%, earnings well ahead — and the stock fell sharply anyway, because of what management said about next year.
Intuit made a striking admission: price is now the number one reason people stop using TurboTax, especially filers earning around $50,000. Rather than defend the price, it is choosing to lower it — accepting less revenue per customer, possibly giving more away free — to win those users back. The consequence is stated plainly: TurboTax revenue will grow just 2–3% next year, and group growth drops from 14% to 9–10%. It is doing the same thing at QuickBooks, launching a free tier that already has 20,000 users, on the theory that you acquire a small business cheaply and then earn from it through payments, payroll and lending later.
Whether this is good or bad depends entirely on a question the numbers cannot yet answer, and the newsletter frames it exactly right: is this a deliberate reset — sacrifice a year of price to rebuild the customer base, then grow again — or is it an admission that free competitors have permanently taken away Intuit's ability to charge what it used to? A voluntary price cut and a forced one look identical in a guidance table.
One technical warning attached: the apparent collapse in next year's earnings guidance is partly an artefact. Intuit is changing its definition of adjusted earnings to include stock-based compensation — a real cost most software companies exclude — so the new number simply is not comparable to the old analyst estimates. Notably, this redefinition makes Intuit look worse, which is the more credible direction for such a change. Mailchimp, meanwhile, remains a problem: flat to slightly down. The author owns it. Analysis, not a recommendation.
Workday runs the HR and finance systems of large companies, charged per employee. That pricing model is the key to this quarter, because it is exactly what AI threatens: if AI agents do work that employees used to do, a vendor charging per employee has a shrinking bill. The industry term is "seat compression."
The AI adoption numbers are genuinely impressive. AI products drove over $100 million of new contract value — more than a quarter of everything Workday sold this quarter — and more than 5,500 customers now run at least one Workday agent, up 35% in three months. AI product lines are approaching $600 million of recurring revenue, triple a year ago.
Now the catch, and it is a big one. Workday's answer to seat compression is "Flex Credits," a system for charging by usage instead of headcount. Just over 200 customers have signed up for it — against those 5,500 using agents. Fewer than 4% of the users are on the meter. Workday is doing this deliberately, prioritising getting agents adopted before it starts charging for them, and management calls it a "delayed gratification" model. That is a reasonable strategy and a real risk: the giveaway is certain, the later payment is not.
Meanwhile the core business is slowing in the places that show the future first. Total contracted backlog grew only 8% — well below the 13–14% revenue growth being reported now — and management guided next year's subscription growth to about 11%. It offset that by promising higher margins. A company guiding growth down and margin up is telling you which number it intends to defend. That combination — impressive adoption, minimal monetization, decelerating backlog — is why the view is neutral rather than positive. Analysis, not a recommendation.
Zoom is trying to stop being a video-meeting company. The meetings business is mature and its consumer half barely grows, so the strategy is to sell adjacent products to the same customers: a phone system, a customer-service suite, an employee-communications app, and AI features layered across all of them.
This quarter gave the best evidence yet that it is working, at least at the enterprise end. Business revenue grew 8% — its fastest in three years — and is now 62% of the company. The AI usage numbers are large: customers using AI features more than doubled, and customers of the Virtual Agent (an AI that handles support conversations without a human) grew more than three and a half times. AI was involved in nine of the ten biggest customer-experience deals, and every one of the ten largest phone deals was won by displacing a competitor, which is the cleanest form of proof that the product is good — someone chose to leave a working system.
The problem is arithmetic. All of that sits on top of a company growing 5% overall, because the consumer half grows 1% and the base is large. Even Workvivo crossing $100 million of recurring revenue — a genuine milestone — is about 2% of Zoom's revenue. The full-year forecast rose by $23 million on a $5.09 billion base, and the next quarter's guidance was unremarkable.
So the view is neutral: the direction is right and the evidence is real, but the new products have to become several times larger before the headline growth rate changes. The author owns it. Analysis, not a recommendation.
Best Buy beat expectations and raised its outlook. Sales at existing stores grew 4.1% against a forecast of roughly 1%, and nearly every category grew — while lapping last year's Nintendo Switch 2 launch, a hard comparison. Full-year guidance went from "possibly negative" to 1.9–3%.
The reason the view is neutral anyway is what is inside that growth. Computers were the biggest contributor, but the average PC sold for a mid-teens percentage more than a year ago while the number of PCs sold fell by a high-single-digit percentage. Best Buy is not selling more computers; it is selling fewer computers at higher prices. The price rise is not Best Buy's doing either — memory chips have become expensive because AI data centres are buying them, and that cost is passed through to laptops. It is worth noting where that shows up: the same memory shortage discussed in chip earnings arrives, two steps later, as a retailer's comparable-sales figure.
Price-led growth is fragile in a way unit-led growth is not: it stops when prices stop rising, and it can reverse if shoppers eventually balk. Which is why the other developments matter more than the headline. Best Buy's Marketplace — letting third-party sellers list on its site, earning a fee without owning inventory — reached $300 million of quarterly volume with the full-year target raised to $1.3 billion, and its advertising business helped lift gross margin. Both are far more profitable per dollar than selling a television. Newer categories (AI glasses, collectibles) more than doubled and added a full point of growth. Best Buy also finished integrating with OpenAI's shopping flow and launched its own conversational assistant, Ask Blue.
The author's summary is the right one: as the computing comparisons get harder in the second half, Marketplace, advertising and the new categories have to carry more of the load. Analysis, not a recommendation.
HealthEquity administers health savings accounts — the tax-advantaged accounts Americans use to pay medical costs. Its business has two halves: fees for administering the accounts, and a cut of the interest and investment income on the money sitting in them. That second half means the pool of assets under management is the real engine, and it compounds as long as balances keep growing.
The margin story this quarter is excellent and unusually well-documented. Adjusted EBITDA margin reached a record 48%, up from 46%, and the source is automation rather than price: AI now resolves 85% of routine chat questions and handles 55% of card-related phone calls with no employee involved, while customers increasingly self-serve on mobile (active mobile users up 62%). This is one of the clearest examples in the issue of AI showing up as cost removed rather than revenue added — which is a perfectly real benefit, and often an easier one to verify.
The shares fell sharply anyway, and the reason is the engine. HSA asset growth slowed from 19% to 14%, reaching $37.9 billion. Account numbers hit a record 10.7 million with record new-account sales for a second quarter, so the business is not losing customers — the balances are simply compounding more slowly. That matters because cost savings are finite: you can only automate the same phone call once, whereas asset growth compounds indefinitely.
Hence neutral. The author's framing is the test to carry forward: whether efficiency gains can keep profits growing faster than a high-single-digit revenue line, once the easy automation is done. The author owns it. Analysis, not a recommendation.
PDD owns Pinduoduo, a Chinese bargain-shopping app, and Temu, the international version that flooded Western markets with very cheap goods shipped directly from Chinese factories. Both halves slowed this quarter: total revenue grew 8%, down from 11%, and missed. The advertising business that monetizes Pinduoduo grew just 3%.
The domestic problem is competitive and familiar — Alibaba, JD and social-commerce apps fighting over Chinese consumers who have become cautious about spending.
The international problem is more serious because it is structural rather than cyclical. Temu's original advantage rested on a rule: parcels below a certain value entered the United States duty-free, so a $12 item could be shipped individually from a Chinese warehouse without paying import duty or being processed like a normal import. That exemption is gone, and Europe has added its own import fees. The economics that made the model work have been legislated away. PDD's answer is to build local warehouses and fulfilment in the destination countries — which is the right response, but it converts an asset-light business into an asset-heavy one, and management has already warned about slower delivery and higher costs.
Money is not the constraint: PDD holds $67 billion of cash and short-term investments and generated $3.8 billion of operating cash flow in the quarter. It is spending heavily on merchant support, supply chain and a new first-party brand effort — which is itself running behind schedule. The author's verdict is the reason this is the one negative view in the issue: investors are still waiting for evidence that any of this spending restarts growth. Analysis, not a recommendation.
Key points & figures extracted from the PRO App Economy Insights newsletter (article text in transcript.txt) for personal study. Not investment advice. © App Economy Insights for source material.