Good morning. Nasdaq futures flat-to-red, European chips mixed, Brent down a third day to $85 — all of it waits for NVDA tonight. The setup: implied move sits at its LOWEST SINCE 2022, the market sold the past four beats the next day — the bar hits the floor. The asymmetry tilts up, but only if margin guide and Rubin timing don't trip. Pre-print noise came in three rounds: OpenAI claims Jalapeño hits 2X PERF/WATT vs Rubin (vendor posture, but the attack narrative is real), Morgan Stanley's 2027 CoWoS model moves allocation toward AMD, SoftBank preps a $20B bond for OpenAI — the AI credit machine keeps adding lenders. Memory inflation is spreading, not peaking: MLCC inventories at RECORD LOWS with prices up 7%, and the HBM 12-high to 8-high shift reads thermal/yield, not demand destruction. Software takes hits from both directions: SAP is Europe's worst large cap on a UBS downgrade for slow AI adoption; INTU slides on TurboTax churn — consumer pricing power cracks. Asia: BABA swallows a $10.2B placement despite Jack Ma buying; Z.AI's Ox Alpha tops OpenRouter for free, open weights coming. China's price war has discipline and it's working. SMTC lifted the optical tape: street's 800G forecast jumps from ~50M to 80–90M units as 1.6T ramps concurrently — demand isn't waiting for the print. PCE hits 12:30 GMT, one hour before the main event, Warsh's Jackson Hole debut Friday the follow-through. We'll hit up NVDA, SMTC, and AMD first, then get to software (SAP, INTU) — two faces of the same anxiety.
INTU just did the most honest thing a beaten-down software stock can do: it reset. Management cut long-term targets, surrendered near-term tax ARPC, and guided to the slowest revenue growth in a decade — 9-10% FY27. FIRST SUB-10% PRINT SINCE FY2015. Stock's down ~45% YTD at $357.46, and the guide justifies the damage. Cheap at ~14x CY27 EPS with a 0.63 PEG — cheap for a reason. R/r only works if you trust the reset. I don't. Not yet. Nothing gets proven until the May Tax quarter.
Q4 was a clean beat: revenue +13.7% vs 11.5% consensus, non-GAAP EPS $4.03 vs $3.58, operating margin 33.3% vs 30.7% expected. That's all known. The market didn't care.
What's new is the guide, and it's ugly:
This is a strategy pivot, not a guidance stumble. It just costs them the narrative for 12-18 months.
The bull case: This is a reset-and-reaccelerate setup, not a structural break. Jefferies (Buy, $500 PT — the Street high by a mile) reads the guide as conservative and explicitly "sets a low bar" to clear. The mix shift underneath is real: mid-market, assisted tax, and money portfolio grew 34% in FY26 and are now ~30% of revenue. Enterprise Suite ARR hit $145M in Q4, UP 4X YOY. Total payment volume +32%. Management still guides high-teens non-GAAP EPS growth — cost cuts and the buyback do the heavy lifting. $5.5B repurchased in FY26 (2x FY25), $7.9B still authorized. At 14x CY27 EPS, the market is pricing permanent impairment, not a deliberate reset. BMO (O/P, $412) says successful execution of the customer-acquisition plan "could boost confidence in the durability of growth."
The bear case: The low end is structurally eroding, and AI is the accelerant. Even BMO — the bull — admits AI competition is hitting the lower end of INTU's market. TurboTax guiding below H&R Block in the same print is direct evidence of share loss, not a guide-calibration quirk. Online paying customer growth slowed 200bps YoY to 3% — the bleed extends beyond tax. Stifel (Hold, $300) has the cleanest framing: INTU needs to rebuild top-of-funnel go-to-market, and it will lag large-cap application peers while doing it. MS (EW, $315) says execution risk is unproven until May. The buyback is doing the EPS work; the revenue algorithm just broke. And there's zero evidence yet that giving up ARPC actually wins back the price-sensitive filer who can now do AI-assisted tax prep for free.
The PT range is a barbell: $300 (Wells Fargo, Stifel) to $500 (Jefferies). Middle: MS at $315, BMO at $412, Mizuho at $430. Simple average ~$376 against a $357 print — even the bull consensus sees less than 6% upside. That tells you everything about how this guide landed.
The movement matters more than the levels. The cautious cluster converged from three directions — WF cut $360→$300, MS cut $335→$315, Stifel raised $275→$300 — all landing in the same "show me" zone. The bulls held their lines. Rating-wise it's a 3-3 split: three constructive (Jefferies, Mizuho, BMO), three neutral (MS, WF, Stifel).
"Execution risk is high and unproven until the Tax quarter in May." — Morgan Stanley
"The changes will take longer to work through... Intuit will lag other large-cap application peers." — Stifel
"Conservative fiscal 2027 guidance that sets a low bar for the company to meet." — Jefferies
The tie-breaker is the multiple. At 21.89x trailing with a 0.63 PEG, the market has already de-rated INTU into value-software territory. Bulls call that the opportunity. Bears call it the correct repricing of a broken growth algorithm. The 6/26 date is the arbiter.
INTU is the canary for AI compressing the low end of software. The price-sensitive DIY tax filer is leaving for cheaper/free alternatives, and the tape is brutal — INTU guiding 2-3% against HRB's 4.8% is a direct share-loss tell. Watch H&R Block. Watch any consumer/SMB software with a self-serve low-end tier.
The counter-theme is just as important: INTU's mid-market, assisted tax, and money businesses grew 34% and now represent 30% of revenue. Value is migrating upmarket — the assisted, vertical, high-touch layer is where pricing power lives. That's a read-through for the entire application software complex.
For PMs: show-me into May. The $300 PT cluster is the technical magnet if the narrative keeps deteriorating. The buyback is a floor, but floors in software break when the multiple compresses. I'd rather own the upmarket compounders until INTU proves it can stop the DIY bleed.
Pricing transition is working. That's the entire ballgame. 48% OF ACV NOW ON PLATFORM PRICING (up from 40% in Q1), 12 of the top 20 US enterprise customers converted, and four top-20 accounts renewed EARLY at >10% average ACV uplift. Early renewal at a premium is the strongest signal a software company can print — customers locking in before prices rise further.
The analyst lineup reflects it: Needham Buy, $25 PT. Citizens raised to $25 from $23 (Market Outperform). BTIG stuck at Neutral — but even the hold is constructive, bumping FY27 growth +30bps and core sub growth +80bps. Three shops, one narrative: pricing transition ahead of schedule, and the Street hasn't fully modeled it yet.
Revenue $161M, +8% YoY — beat the $159.14M bogey. Core +12%. US mortgage -1% (the anchor drag). Adjusted EPS $0.05 vs $0.2658 expected — horrid, no way around it. The tape didn't care. Management RAISED FY27 revenue, op income, and FCF guidance above consensus. Q3 guide in line, with US mortgage trimmed on higher-for-longer rates. That's the one blemish — and the offset landed fast: the company signed its LARGEST DEAL OF THE YEAR in the first weeks of Q3, an international one.
Six quarters into the platform pricing rollout, and the data points compound. 48% of ACV converted (was 40% in Q1). 12 of the top 20 US enterprise accounts moved. 230+ customers paying for intelligence units — that's the AI monetization hook. Churn contained at ~$25M FY27, roughly a third mortgage-driven.
Citizens has the cleanest line on the moat:
"Financial institutions have told the company they do not want to rebuild the enterprise application themselves."
That's the whole thesis in one sentence. Core banking systems aren't getting ripped out. The pricing transition converts installed base into incremental revenue without incremental logos — which is why four top-20 customers renewed EARLY at 10%+ increases rather than waiting for the natural renewal cycle.
Bull: Platform pricing = self-help, and it's compounding — 12/20 top accounts converted, largest deal of the year booked, and buybacks are an active bid under the stock. At roughly 3x EV/FY28 revenue with margins inflecting, the r/r skews positive.
Bear: The EPS print was a whiff — $0.05 vs $0.2658 expected — and mortgage keeps shrinking with a trimmed outlook on rates. The stock's already up 25% in six months; even the bullish $25 PT is just ~20% upside, and there's no valuation cushion if the transition stalls.
Buyback floor at $17-19 — management has been the most aggressive buyer of its own stock: 15.8M shares at $18.97 avg since April, 4.7M of those at $17.40 in June, plus a fresh $100M authorization. Estimates are still catching up to the transition math (BTIG's +80bps core sub raise is a tell). AI monetization optionality via 230+ intelligence unit customers. This is a company doing everything right on the controllables — pricing, buybacks, capital allocation — while the macro (mortgage) is the only real negative.
Not a home run. But the market is still underwriting "transition in progress" while the company is printing "transition working." That gap is where the money gets made. As long as the early renewal pace holds, the multiple grinds higher.
SMTC remains a buy — and the beat-and-raise is real. The setup into FQ2 was the gift: stock +1% over the month, +38% over six months, while AI-connectivity peers ripped. Low hurdle. Then they printed revenue $341.9M vs $328.6M est., EPS $0.71 vs $0.61, TENTH STRAIGHT QUARTER OF GROWTH, and guided Q3 revenue 13.6% above Stifel's estimate with EPS $0.32 above. After-hours bid: +3% to ~$131.40. The stock is UP 117% OVER THE PAST YEAR, 73% YTD — and the market still gave you a discount on entry.
All three covering brokers — Benchmark ($230 PT), Needham ($200), Stifel ($188) — reaffirmed Buy. PT dispersion runs $155-$230. The multiple debate is alive, but the direction isn't.
This is the part PMs should care about. Benchmark's Cody Acree nailed the positioning edge: shares had gone nowhere for a month while AI-connectivity peers got extended, leaving a "lower expectation hurdle." Nobody was positioned for a blowout. When it arrived, the repricing was instant. That's an asymmetric print — modest expectations, real results, quick re-rating.
Two platforms, both firing. Data Center: 800G ramping, 1.6T CopperEdge and FiberEdge in "meaningful" ramp territory. LoRa: GUIDED +65% YOY in FQ3. Photonics remains the call option — gain chips and CW lasers, expanding customer base.
Margins are the sleeper. The Sierra Wireless cellular modules divestiture (closes FQ4'27) helps the math, but Stifel makes the sharper point: margin expansion is increasingly product mix, not just divestiture. That's higher-quality gross margin.
Bull: Two growth engines, one improving margin trajectory. Net leverage already down to 1.1x from 1.4x. The balance sheet is fixing itself while growth accelerates — that double-click sustains AI runs.
Bear: 35.5x CY27 P/E (Stifel's basis) leaves no room for slips. The low hurdle that made this print so satisfying now resets higher — Q3 bogeys are elevated. And the $155-$230 PT band tells you the sell-side hasn't converged on the frame yet.
Benchmark's Acree summarized the whole trade in one sentence:
"The shares had appreciated only about 1% over the prior month and 38% over six months... this left a lower expectation hurdle than many AI-connectivity peers."
That was the setup. It's played, partially. The fundamental story is still compounding — 800G, 1.6T, and LoRa are all acceleration stories, not one-quarter miracles. Longs stay long. New money should buy the next dip, not the after-hours pop. The r/r is decent, just less asymmetric than 24 hours ago.
Verdict: Street wants this print. PT cluster is the tell — most of the Street has marks in the $380-425 zone, and JPM just jumped to $384 from $326. The thesis isn't "beat the quarter." It's "platformization plus post-Mythos security urgency sets up FY27."
BTIG did the most useful work here: 11 channel contacts, including six partners with roughly $1.7B in combined PANW sales. All bullish. Partner core growth re-accelerating into the 15%-plus range as platform adoption and large-deal strength kick in. That's the rate-of-change you want to see into a print.
"This creates a path for management to guide modestly above current Street expectations when it reports earnings next week while leaving room for additional upside as the year progresses." — BTIG
Under the hood: LTM REVENUE $10.6B (+19.5%), GROSS MARGIN 72%, LEVERED FCF $3.79B. The stock sits at $339.90 after an 84.5% RUN, and this is the expensive part of the trade — 295x TRAILING P/E. OPTIONS MARKET PRICING AN 8.6% MOVE for the Sept 1 print. That's binary.
Bull: Channel checks support the narrative. Partners see 15%+ core growth, and BTIG's math has PANW exiting FQ4 at ~17% pro forma revenue growth, sustaining 16%+ in FY27. That gives mgmt room to guide above Street without pulling the "wait until next year" trick. Add Cantor's Frontier AI Critical Defense Program as narrative fuel, but the real bogeys remain NGS ARR and RPO.
Bear: Valuation doesn't leave a lot of room for "fine." 295x trailing earnings after an 84.5% run means you're paying for near-perfect execution. If mgmt guides in-line and defers upside to FY27, de-rate risk is real. The 8.6% implied move cuts both ways.
This is a "trust the channel checks into the print" situation. JPM sums up the narrative: PANW is in the right place at the right time. But PMs shouldn't conflate good field data with a cheap entry. Watch NGS ARR growth, RPO, and FY27 guidance. If the guide is modestly above Street, the stock likely works. If mgmt "saves some for later," the multiple does the talking.
BEAT AND DROPPED. That's the whole story in two words. Zoom cleared FQ2 — $1.55 adj EPS vs $1.48 bogey, $1.28B revenue vs $1.27B cons — and the stock still sold off ~4% after-hours. Why? Because the headline decel didn't stop: total revenue +4.9% YoY, down from +5.5% last quarter. The market just looked straight through the beat to the trajectory.
Enterprise revenue +7.8% — STRONGEST GROWTH IN THREE YEARS. That segment now prints 62% of total sales. This is the narrative shift PMs should care about: Zoom is becoming a software company with a comms tail, not a comms company with a software hope.
RPO confirms it: $4.5B vs $4.3B cons, +14% YoY, accelerating from +11% last quarter. Forward visibility is improving while Online just sits there.
Benchmark kept Buy and $125 PT on the enterprise evidence:
"Enterprise segment showed financial and KPI momentum, with second-quarter fiscal 2027 revenue rising 7.8%, the strongest growth in three years."
The drivers matter: wider product range, AI monetization, larger and longer contracts, push into higher-end markets, channel build-out, lower enterprise churn. That's a compound story, not a one-off print.
Citizens stayed Market Perform — same data, colder blood. Fair enough.
Bull: 62% of revenue growing 7.8% with RPO accelerating. AI monetization is a call option, not the thesis. At 14.89x earnings with a PEG of 0.14, you're paying nothing for the mix shift. If Enterprise gets to double digits, this rerates in a hurry.
Bear: The market isn't dumb. Total revenue still decelerating (4.9% vs 5.5%), Online flat-guiding for the year, and non-GAAP op margin came in light (40.0% vs 40.4% cons). Stock's +17% YTD vs +12% for the S&P — the good news was already in the tape. Beat-and-drop on quality numbers is the tell that PMs are discounting FY28 already, and FY28 doesn't scream growth yet.
Net: cheap, improving mix, but the market wants proof the acceleration holds before paying up. The $125 PT says a lot of upside if it does. The 4% drop says the market isn't convinced yet.
BofA says the quarter is a beat — the real catalyst is the commitment stack. Reiterates Buy, $350 PT into FQ2. Sees 3-4% sales beat and a raise on the Vera Rubin ramp. But the more important potential disclosure: multi-year, off-balance-sheet commitments for supply, power, models and demand.
That's the narrative shift. NVDA isn't just selling chips anymore — it's pre-funding the AI supply chain. BofA puts the worst-case funding burden at ~$500B, or 10% OF EV, with $150-200B tied to purchase commitments and cloud service agreements. PMs should watch for the word "commitments" in the prepared remarks. That's the multiple-expansion trigger.
Valuation is the hook. FORWARD EV/EBITDA DOWN 44% TO ~15X since the $100B OpenAI deal. AMD trades ~32x. NVDA's PEG is 0.29, P/E 32.5. A mega-cap growing like that at half the peer multiple — the bull case writes itself.
"The more important factor is Nvidia's potential disclosure of multi-year, off-balance-sheet commitments to secure supply, power, models and demand."
The FCF machine backs it up. $119B levered FCF over the last twelve months. Payout ratio sits at ~37% of FY27/28 consensus — a move to 50-75% puts a floor under the stock. BofA flags NVDA generating ~$1B DAILY FCF by next year. (Conveniently, that covers the worst-case commitment burden.)
Rest of the Street sings the same tune. BMO sees Q2/Q3 revenue beating consensus by $2-3B on data center strength. Raymond James raises PT to $352 (Strong Buy), CPU revenue growing from 3% to 5% of sales by 2028. KeyBanc Overweight at $330 — Rubin R200 mass production started in July. Cantor calls it a top pick, $350.
At 15x forward EBITDA with a beat-and-raise setup and a payout-ratio catalyst in play, this r/r skews positive into the print. The question is whether Jensen gives the market the commitment numbers. That's what moves the multiple.
Rosenblatt holding Buy + $83 into tomorrow's FQ1 print — and doing absolutely nothing else. No estimate changes, no PT move, just "in-line to marginal beat" expectations. That's a tell: they see a low-volatility print, not an inflection. Elastic Cloud decel to 19% (from 20%) keeps the narrative muted, even with a 76% gross margin and net cash balance sheet.
The street's split tells the real story. Stifel at $90 and Cantor at $91 on sales-led subscription strength and multiple expansion; TD Cowen only at $70; Morgan Stanley cutting to Equalweight on near-term growth concerns. That's a 20-point PT spread into a print. Nobody's convicted. The 7% RIF in June was the last meaningful event — now it's just execution.
Elastic remains well positioned to benefit from tool consolidation in log management, legacy SIEM modernization, and development of new AI and Agentic applications using the Elastic Search Relevance Engine with vector database capabilities.
That's the bull case — AI/Agentic as the re-rating catalyst, not the quarter. The Q4 beat was only 1%, so tomorrow is about guidance and whether Elastic Cloud's decel stabilizes. Light coverage, tepid positioning, wide PT dispersion. Let the print clear the fog.
Citizens reiterates Market Outperform, $315 PT — reaffirm, not upgrade. Thesis holds: logistics keeps retail margins sticky, AWS earns the premium multiple. Stock closed $261.06, implying ~21% UPSIDE. Valuation: 23.8x 2027 GAAP EPS of $10.99, 10.9x EBITDA of $262B. Fair, not cheap, not scary.
The real juice is in the other news: Rosenblatt initiated Buy at $335, and MS sees AI capex clearing 25% ROIC by summer 2026. That's the whole ballgame — bears treat the spend like a furnace, MS says it's an engine.
"The firm's analysis suggests these investments could yield a return on invested capital of at least 25%."
If MS is right, the bear case on margins collapses. (Shreveport's $18B data center campus is the tangible proof the spend is real.) AMZN stays a buy-the-dip, sell-the-rip until the capex narrative resolves.
Rosenblatt bumps PT to $105 from $95 into tomorrow's print (Q2 FY27, Aug 27 AMC), Buy maintained. Checks show healthy demand across backup/recovery/cyber-resilience. Models +28% YoY revenue to $397M, sub revenue +29%, sub ARR +31% — slight decel from Q1's 32% but still a gear most software names can't find. Stock at $92.65, so the street's pricing in ~13% upside on a beat-and-raise.
The setup is boring in a good way. 80.6% gross margin, 45.7% LTM revenue growth, 5.8% Q1 beat vs midpoint (8.9% TTM beat rate). Consensus already Strong Buy (1.11) with 18 upward EPS revisions. Not a crowded short, not a fight — a confirmation game. Guggenheim, Cantor, Oppenheimer, Scotiabank, Goldman all sit $106-120, all citing partner/VAR checks or AI-prep positioning. The multiple works if ARR holds 30%+. Watch the guide more than the print.
"Recent checks indicate demand for Rubrik's backup, recovery and cyber-resilience solutions was healthy in the second quarter."
Citizens is back on the bullhorn with a Market Outperform and $515 PT — not an upgrade, just a standing ovation. The math: 32x their 2027 GAAP EPS of $16.27. Stock closed at $346.96, which is a mere 21.3x that same number. That gap is the thesis in one line.
The bull case is boring in the best way. Search, YouTube, Android consumer distribution is still a cash cannon. Google Cloud is the growth engine, margins are best-in-class, and the market is effectively paying teens multiples for a duopoly with an AI call option. Even Rosenblatt came in this week with a Buy and $410, so calls are stacking.
The 68% one-year return tells you the easy money was made, but the PEG sits at 0.15 (yes, that's real). That's the kind of multiple that makes PMs who missed the move start chasing.
Recent product flow is supportive: Gemma past 1B downloads, Antigravity bundled into Gemini Enterprise, Search personalization rollouts. All small pieces, but they compound the "AI distribution" story — Google touches everything.
Not sure we need a new thesis here. This is a hold-your-position upgrade cycle, not a fresh catalyst. If you're long, you're riding the 2027 multiple re-rating. If you're flat, the $515 target is the north star but the risk/reward from $346 is no longer the layup it was a year ago.
UBS just pulled the ripcord on the AI story — cuts SAP to Neutral (PT EUR201 from EUR164). Not a valuation call: the stock ran, and now 27.86x P/E looks rich against near-term earnings growth. The problem is delivery. Agentic AI is landing slower than the bull case assumed.
SAP has shipped 17 out-of-the-box AI agents, with 15 in ramp-up. The 200-agent goal by year-end? UBS calls it "challenging" — that's polite for "not happening." The installed base is messy: large ERP customers run multiple instances across different versions, many on private clouds with custom code that limits SAP's visibility. Ready-made agents don't just plug in.
The size and complexity of SAP's customer base in core enterprise resource planning software creates challenges for delivering ready-made agentic AI solutions.
The growth story isn't dead. UBS still sees 19% earnings CAGR through 2028 on RISE migration. But they're flagging a highly likely cloud backlog deceleration in 2H, and FCF upside vs guidance looks less probable than the last two years. Migration credit utilization is eating cash flow. That's the sleeper issue for PMs.
The rest of the street is split. Bernstein and TD Cowen trimmed PTs on margin pressure but kept Buy/Outperform. KeyBanc cut on sector derating. BMO nudged up. Q2 cloud growth of 24% matched, backlog grew 26% beat — but operating income missed on M&A costs. The narrative's now a slugfest: RISE backlog = bull, AI agent delivery = bear. For now, the bear has the stronger tape.
APP trades at $310.54 — DOWN 54% YTD AND SITTING ON ITS 52-WEEK LOW ($297.50). Nobody's arguing the business is broken. They're arguing about the slope. Needham keeps Buy but cuts PT to $475 from $500, trimming 2026/2027 adjusted EBITDA estimates by 4.5%/3.3% after rebalancing Games vs Consumer segment splits. The pixel data underneath is the real tell: additions decelerated to +1% over the last 15 weeks, down from +3% in the prior period. No major brands joined in three months. The new logos — LiquidIV, WeightWatchers, No Bull, Quip — are small fish.
The one green shoot: Google Store pixel users jumped from 7,000 to 10,000 in a single month. That reads like existing advertisers pouring more resources in, not new demand. Nice, but not enough to offset the growth scare.
The rest of the Street has turned colder. BofA downgraded to Neutral ($400 PT) on sustainability of long-term revenue targets. BTIG cut to $408, Benchmark to $440, Piper to $325 — all landing on the same theme: maturing gaming, Commerce pivot taking longer, growth resetting post-Q2. Phillip Securities is the lone holdout at $610 Buy, acknowledging cost pressure.
"We remain disappointed by delays in the company's generative AI creative and lead generation model but continue to see opportunity for AppLovin to scale its Consumer business significantly."
That's the whole trade in one sentence. The bull case rests on Consumer scaling and AI upside that keeps slipping. The bear case says the pixel data is a ceiling, not a springboard. At 88% gross margins, quality isn't the question — timing is. I need to see pixel growth re-accelerate before stepping in front of this avalanche.
DA Davidson the only one in the booth today — reiterates BUY, $52 PT vs $36.49 spot. That's ~42% upside and they're not wavering. Kept FY27 estimates, added FY28 coverage. Thesis is simple: equipment volume ramps gradually through the year, recurring service revenue keeps compounding. Nothing sexy. That's the point.
The FQ4 print already backstopped it: EPS $0.50 on $55.8M rev, beating $0.39/$53.18M consensus, sales +10% YoY. Recurring revenue stronger, margins wider, dividend hiked. The market's lukewarm response ($36 shares) tells you the bar is now about durability of the RSR trajectory, not the quarter itself.
"Sell-through statistics remain healthy while the recurring service revenue run rate continues to rise, largely supported by demand for StarLink commercial fire radios."
The real kicker is the balance sheet — $138M cash, zero debt. Capital deployment optionality on top of an organic growth story. Networked locking/access is the call option beyond fire radios. At 42% upside with the P&L proving out, this is a decent risk/reward for a small-cap name PMs mostly ignore.
JPM just called it: downgrade to Neutral from Overweight, PT $120 (stock $111.85). The North America commercial reorg — EPAM's push to sell non-engineering skill-sets — is going to chew up near-term growth and margins. With the Digital ITS backdrop still challenging, there's NO near-term catalyst to counter AI narrative risk.
The tape already knows. Stock's up ~20% off the post-earnings low while peers did 5-11%. That's the re-rating done. JPM's right that the r/r is dead until growth inflects — this is a show-me story now, not a buy-the-dip.
Bear case (JPM, William Blair): North America stays soft, limited growth through 2026, and management needs to rebuild credibility with investors. William Blair already bailed to Market Perform. Q2 beat? Irrelevant — the full-year guide cut erased it.
Bull case (Guggenheim, the lonely one): PT cut to $140 but Buy maintained. The fundamentals still work — EPS $3.38 vs $3.14 est, rev $1.42B (+4.5%). Just not enough to outweigh the reorg overhang in the near term.
"JPMorgan does not expect the stock to outperform peers in a meaningful way until growth trends inflect."
STOCK DOWN 45% YTD. Watch for the reorg to show up in the numbers before stepping in. Plenty of cleaner growth stories in software right now.
Benchmark raises to $110 from $90, keeps Buy — on a name already up ~50% since the Aug 4 print. (Cantor's at $80, Evercore just cut to $95 from $120. The street is NOT aligned.) The SOTP refresh still says undervalued.
Core math: ex-Base44, the core Wix platform trades at 0.4x EV/2026E revenue. Base44 gets 12.5x EV/ARR — the same multiple Benchmark assigns to vibe-coding comp Emergent. Total EV to 2026E adjusted FCF: 2.0x. Q2 gave the model cover — EPS $1.59 vs $1.16E, adj FCF $61M vs $36M est, bookings +12% in 3p.
The updated analysis still implies an undervalued stock, though the company has not yet achieved clear visibility on near-term growth.
That last clause is doing the heavy lifting. Last week's Symphony launch is the bull's catalyst — Wix positioning into vibe-coding and LLM site-builder territory rather than defending against it. That's the re-rating story. But the visibility gap is real: at $86 you're paying for optionality on Symphony adoption, and the 0.4x core multiple only looks cheap if the platform business holds. Decent r/r versus AI names at 20x revenue, but the burden of proof sits with management.
Itau BBA cuts VTEX to Market Perform from Outperform, PT TO $4.50 FROM $5.00 (rolled to YE27 from YE26). The take: the 2H26 growth recovery is dead. Marketplace competition plus softer Brazil/Argentina macro killed the reacceleration thesis, so VTEX rolls into 2027 with weaker momentum and consensus estimates carry downside.
The valuation math is the tell. At 16.4x 2027 GAAP P/E (12x non-GAAP), the stock prices in a growth recovery that isn't showing up on schedule. The headline PEG of 0.15 looks like a value trap rather than a bargain — that's just what depressed near-term earnings do to the math.
"The current valuation of 16.4 times 2027 GAAP price-to-earnings or 12 times 2027 non-GAAP price-to-earnings appears to price in a growth recovery that may take longer to materialize."
Long-term story intact — LatAm digitalization, B2B and global expansion, AI initiatives, 79% gross margins. Q2 delivered an EPS beat ($0.05 vs $0.04) but a revenue miss ($64.38M vs $64.53M). FCF +79% YOY and non-GAAP op income +62% show the margin story works; growth is the missing piece. Itau's happy to wait for a better entry. So should you.
UBS stays with it — Buy, $225 PT — after NVT's $1.75B Maverick Power deal. That's ~11.5x 2026E EBITDA for a data center power distribution play. UBS sees ~6% EPS accretion on 2026 numbers. Not a steal, but the multiple makes sense if you're paying for the end-market exposure rather than the near-term arithmetic.
NVT's Q2 was already the proof point: EPS $1.45 vs $1.16 est, revenue $1.5B vs $1.26B, and 47% organic growth. Evercore also chimed in ($210 PT), calling out the $2.5B backlog execution and 23% EBITDA beat. 13 analysts bumped numbers since. Stock's up 72% in the last year — so the buy thesis is "still early in the AI power buildout," not "cheap."
"The deal is valued at 11.5x 2026 anticipated EBITDA." — UBS's Neal Burk, framing that as attractive given the demand backdrop.
Worth a spot on the watchlist, not just the newsflow. The risk isn't the deal — it's whether data center CapEx keeps its foot on the gas.
UBS just torched its BOX target from $29 to $37 (+28%) but left the rating at Neutral. That gap between price action and rating tells you everything: the acceleration is real, the durability is unproven.
The numbers back the optimism. Revenue growth RE-ACCELERATED to 11% constant currency from 7% a year ago. Billings +16% cc (vs 13% in Q1 FY27), cRPO +14% cc (vs 12%), and NRR ticked to 106% from 105%. Enterprise Advanced is the engine — pricing uplift plus seat expansion, against a near-80% gross margin. This is a quality problem to have.
"UBS views Box as one of the few software companies demonstrating tangible evidence that AI is contributing meaningfully to growth."
Bull case: AI is actually showing up in the financials, not just the marketing deck. FQ2 revenue beat ($321.1M vs $319.33M est), EPS in line at $0.40, and full-year guide raised. UBS sees H2 guidance as conservative.
Bear case: the H2 guide still implies moderation to 10-11% cc revenue and 8-9% cc billings growth. UBS wants to see momentum hold through FY27-FY28 before upgrading. Stock's already up 38% in six months and sits at ~$33 against a $37 target — the easy money has been made at this multiple. This is a "prove it for one more quarter" rating.
Needham staying long into Wednesday's FQ3 print — Buy, $600 PT. Thesis is simple: demand is broad and accelerating. Cloud hyperscaler AI capex and Telco MOFN projects both firing. Bookings already $1.38B last quarter, and Needham sees the sequential rebound at roughly DOUBLE that. That's a setup for a beat-and-raise.
The street is consolidating around the same story. Rosenblatt models $1.70B revenue vs $1.63B consensus, $1.90 EPS vs $1.72. Northland upgraded to Outperform on strong supplier read-throughs (component makers shipping = Ciena shipping). BofA trimmed PT to $550 from $660 but kept Buy — that's a valuation haircut on the optical group multiple, not a thesis break. 13 firms clustered PTs in the $500-600 range, for whatever that's worth this late in a momentum tape.
"The firm's industry checks indicate broad demand across both Cloud and Telco segments, driven by rising hyperscale AI capital expenditures and Telco MOFN projects."
The bull case is multi-year: coherent modem/line system leadership, optical backbone as the physical layer of AI infrastructure, backlog visibility, and operating margin expansion ahead. Add a 0.39 PEG and the growth arguably isn't in the price.
The bear case isn't demand — it's supply. Needham flags supply constraints capping upside, and Stifel sees the same across AI infrastructure names: orders aren't the bottleneck, shipping is. That's the kind of problem bulls actually like, but it does cap near-term print upside.
(Verizon dark fiber could add $100M+ in FY27/28 — not in this quarter's numbers, but a good 2027 narrative hook.)
Benchmark bumps SNOW to $360 from $290 ahead of the September 2 print. Dumb simple thesis: hyperscaler demand is ripping, AI workloads need compute, and SNOW is the cleanest liquid way to play the data layer. Stock's already up 66% in a year at $321.26 — basically parked at the 52-week high of $341.95. There's no need to catch a falling knife here. This is momentum WITH a fundamental backbone.
The macro setup is the story. OCI +93%, GCP +82%, Azure +43%, AWS +37%. TOP 3 HYPERSCALER AVERAGE: 54% — THE HEALTHIEST BACKDROP BENCHMARK HAS TRACKED SINCE 2022. Enterprise AI is infrastructure-constrained, not demand-constrained. That's a huge distinction. Customers aren't hesitating. They're waiting in line.
Benchmark expects SNOW to clear the consensus bogeys — 30% product revenue growth and 12.5% operating margin — and channel checks across data platforms and AI software back that up. The bull case writes itself.
The real debate isn't the quarter. It's the guide. Can management find the conviction to lift FY27 product revenue growth above the current 31% outlook? Post-June Investor Day, expectations are already running hot. That's the risk in the trade — not the print, but whether "beat and raise" is fully priced.
"The key debate has shifted from whether Snowflake can outperform quarterly expectations to whether management has sufficient visibility to raise fiscal 2027 product revenue guidance beyond the current 31% growth outlook."
Street's stacked alongside — TD Cowen at $370, Stifel at $350 (calling out CoCo usage up significantly QoQ), JMP upgraded on AI data demand. Even Guggenheim, the Neutral holdout, admits execution is solid on legacy migrations. MS projects ~34% product revenue growth — above consensus. Nobody's short this into the print. That's the tell.
AMD — The Morgan Stanley CoWoS report is the cleanest data point this tape has produced for AMD in months: TSMC's advanced packaging allocation tilts to AMD in 2027 on agentic AI demand, and that pokes the most important assumption inside NVDA's premium (exclusive access to leading packaging). Read the nuance — incremental allocation shift in an expanding pie, not a share collapse. But multiple math moves before the revenue does. Counterweight: NVDA's Vera CPU push targets AMD's server CPU turf. Both sides of that squeeze still net positive for AMD. The tape is buying it into NVDA's print.
ARM — Every hyperscaler building custom silicon — OpenAI's Jalapeño, Google's TPUs, Amazon's Trainium — pays ARM the royalty whether the chip beats NVDA or not. That's the structural bull case. The vendor performance claims are negotiation theater, but the shift from off-the-shelf GPUs to custom ASICs is permanent and royalty-accretive. Hot Chips' on-device/PIM direction only widens ARM's edge TAM.
BABA — Jack Ma's buy couldn't stop the bleed after the $10.2BN PLACEMENT — the market rejecting AI-driven shareholder dilution regardless of insider conviction. The two-week southbound digestion window is the tell: recover the placement discount and HK liquidity still works as an AI financing pool; fail to, and the China AI equity-raise channel closes. Stay cautious until the tape proves otherwise.
CRDO — Interconnect content has three straight weeks of supply-chain validation, and CRDO is the pure-play expression (AECs for accelerators). Problem: it's no longer a secret. Direction is right, but the r/r is crowded — price discipline matters more than the narrative here.
CRM — SAP just caught a downgrade for AI rollout being too slow — that's the template risk for every enterprise software AI story, CRM included. Agentforce is the right narrative, but the proof needs to land in bookings/P&L, not slides. Until we get quantified AI contribution in Cloud backlog, the multiple stays capped.
CRWD — No company-specific signal in the feeds today, but the AI infra buildout keeps expanding the attack surface — security stays a beneficiary of the same capex cycle. Quiet tape, no fresh data. Watch post-NVDA appetite for software before paying up.
CRWV — Nothing in the feeds. No coverage, no read-through, no position signal. Hard skip until something prints. (Don't force a view where there's no information.)
DT — Observability is the pick-and-shovel of the agentic AI deployment cycle — every new AI app needs monitoring. But the software AI-anxiety multiple applies to DT too. No hard data in today's feed; theme is right, timing signal absent.
EMR — BNEF's data point — on-site gas generation now cheaper than grid power for data centers — is a structural kicker for the whole power/thermal equipment complex, EMR included. Content-per-rack keeps climbing for power and cooling, and the supplier-level evidence strengthens weekly. No company print today, but the tailwind is real.
FROG — Dead tape. Nothing in the feed, no read-across from today's axes. Pass.
KDK — Opaque name, zero feed coverage. No signal. Skip.
MDB — The AI app layer thesis runs through MongoDB — every agentic workload needs a document/vector store. But SAP/INTU just punished the software complex for unproven AI monetization, and MDB trades on the same premise. Watch consumption disclosures for inflection; no fresh feed data today.
MRCY — Defense electronics sits off today's AI trading axes, but Pax Silica forcing 35 countries to pick sides is a slow-burn bid for trusted-computing suppliers. No specific feed signal. Too early to price bloc-alignment into MRCY, but sovereign AI budgets are the background story.
MRVL — The quietest big beneficiary of the custom silicon news flow. OpenAI, Google, Amazon — any verification of their ASIC programs, or just continued spend, runs straight into Marvell's custom silicon books. Stack the CoWoS allocation shift and interconnect content on top, and MRVL has multiple hard catalysts. NVDA's Q&A on custom silicon tonight is the next swing event.
MSTR — BTC STALLED AT $80K FOR A SECOND DAY and the debasement trade's marginal buyer is gone — that's MSTR's problem in one line. The stock is a leveraged BTC swap. If NVDA guides well, capital rotates back into AI hardware and out of BTC proxies. The Bessent bounce narrative is exhausted; the next leg needs a new bid.
NTNX — Local/on-device AI cuts both ways here. Private AI infrastructure is a Nutanix tailwind, but inference shifting to the edge challenges the data-center-centric model. No hard data in the feed. Monitor name for the token-mix disclosures, not a trade today.
OMCL — Clean miss. No feed coverage, no read-through. Pass.
PLTR — Pax Silica is the kind of geopolitical order shift PLTR's sovereign franchise feeds on — 35 countries forced to choose sides turns government AI budgets into a strategic defense line. The multiple debate is live, but the narrative tailwind from today's feed is tangible and durable. The tape wants to own sovereign AI into the NVDA print.
SBAC — Rates is the only signal that matters for SBAC today. Core PCE tonight and Warsh Friday set duration direction — hot PCE widens long-end yields and SBAC catches the de-rate. Dovish Warsh and REITs bounce. This is a macro trade, not a stock trade; no company-specific news in the feeds.
SKHY — Memory inflation keeps printing hard evidence (MLCC INVENTORY -8% TO RECORD LOWS, PRICES +7%) and SK Hynix owns HBM leadership — the tightest point in the AI memory chain. The Hot Chips fight with Samsung over PIM/on-device is a roadmap battle: HBM is today's cash cow, PIM is the attack on the next TAM. Samsung's GAIA (4nm, LPDDR5X-PIM, 2027) is the first mover to watch in AI PCs.
SNPS — Custom silicon proliferation is an EDA demand engine — every hyperscaler chip design needs Synopsys tools. Add advanced packaging complexity (FOPLP, chiplets) and the EDA content per design rises. The custom silicon narrative reads through directly, just with less beta than the chip names.
TTAN — No feed coverage. No read-through. Skip until there's a print.
VEEV — No signal today. Life sciences software isn't on this tape's axes. Pass.
WDAY — The SAP downgrade template applies directly: enterprise software without quantified AI contribution gets punished on the question. Workday's HR/finance workflows are AI-addressable and the agent narrative is real, but the evidence lives in the bookings disclosure. No hard feed data, but the multiple risk is tangible.
XPEV — Li Auto printed a SECOND STRAIGHT QUARTERLY LOSS and the China EV price war devours everyone without an AI narrative — XPEV sits in that bucket. Cleanest read of China tech's capital rotation: MiniMax +283% revenue vs. EV losses. Money leaves price-war EVs for AI. Until XPEV shows pricing power, the tape stays away.