Good morning. Tape feels heavy — quants just suffered their WORST DAY SINCE AUGUST, hedge funds notching their third "never since COVID" stress day of 2026, and the AI unwind chatter is getting louder with Goldman calling this "Phase Two" as a $106B DEBT BILL COMES DUE. Not a panic, but the marginal bid is getting stingier — the days of +20% pops on recycled press releases are gone.
Overnight earnings reactions: nothing blockbuster on the tape from our names, but the AH chatter belongs to the AI complex. TSLA DOWN 3% ON THE DAY IT REVEALS A CAR THAT FLIES (Roadster read-through is Optimus/Cybercab validation, not P&L — halo, not earnings). Semis CDS keeps widening — AMD CDS IV at 99% — but JPM's CDS desk says it's hedging, not default betting; CRWV's spread move is convert accounts buying protection against their positions, which means selling CDS vol is arguably the better trade than panicking on equity.
Asia: Samsung breaking ground on Taylor fab 2 in November and sniffing around silicon photonics M&A — foundry arms race on two fronts. MOFCOM now gating Chinese model hosting deals (Nebius, Baseten), which gates distribution for GLM 5.5 and Kimi K3.1 just as they drop. Keep an eye on the China open-source channel.
Four themes framing today:
1. Supply, not demand, is the bull case now. Jensen pegs NVDA content at $40B+ PER GW (up from ~$18B) and names substrates and silicon photonics as constraints that "could persist for years." That's the TSM/ALAB/CRDO/CLS long thesis in one quote — scarcity pricing up and down the chain. Memory too: 80% of capacity under LTAs effectively fully booked, and the binding constraint is capacity, not bandwidth.
2. The credit microscope. Goldman's Phase Two framing means financing gets harder and math gets scrutinized. CRWV is the poster child — advance rates and customer prepayments are the variables to track, not blended debt cost. Deep-pocketed stakeholders keep the death-spiral narrative at bay for now. We're constructive but watching the funding variables like hawks.
3. NVDA's moat anxiety is showing. Locking neoclouds out of GTC Berlin for adding AMD/TPU capacity is a pressure-tactic tell at exactly the customer layer where switching is easiest. The "GPU converging toward TPU" bear thesis is crystallizing — watch whether exclusion tactics accelerate the quiet multi-vendor shift rather than stop it.
4. SaaS positioning is a knife fight. The "vibes trade SaaS AI blowup" short basket is on (CRWD flagged as cleanest short), but our checks say TWLO, SNOW, NOW screen better on AI than the short thesis implies. Meanwhile WDAY stays on the wrong side of the AI monetization question — the MNDY read-across (only ~$20-30M AI ARR vs. Cognition at $1B run-rate) is the bogey for every legacy SaaS name claiming AI traction.
We'll hit up VRNS, SMTC, and BE first, then get to the broader SaaS and AI infra complex...
The equity story is done. Proofpoint (Thoma Bravo) is taking VRNS out at $49.75 ALL-CASH, $6.3B deal, and the Street responded exactly the way it always does: a mass downgrade wave straight to the deal price. DA Davidson, Needham, RBC, Jefferies, Piper — all collapsed to Neutral/Hold at $49.75. Even TD Cowen, who had raised to $56 Buy just days earlier on 3Q in-line expectations and AI demand, capitulated and joined the herd at $49.75 Hold.
Stock at $47.34 vs. $49.75 offer. Classic arb spread.
That's the strategic read — Thoma Bravo bolted Proofpoint's email/people security onto VRNS's data security platform. Proofpoint's Data Security segment is ~25% of its $2.44B ARR and growing mid-teens; VRNS accelerates that.
Bull (for deal holders): All cash, definitive agreement, PE buyer with deep pockets and zero antitrust overlap. TD Cowen's Eyal — the last bull standing before folding — "expects the transaction to close." Spread to $49.75 is your return.
Bear: 1H 2027 close is a hike away, and a 98% six-month run means the deal price was negotiated off a stock that had already re-rated hard. The 6% headline premium to announcement close tells you the market had priced this in. Not sure we can read TD Cowen's brief $56 bump as a genuine competing-bid signal — they folded within days.
This is an arb position now, not a TMT position. If you own it, the r/r is the spread vs. deal-break risk on a long close timeline. If you don't, there's nothing to do here. Next real catalyst is a competing bid — don't hold your breath at 53x FCF.
Bloom keeps ripping on the AI power trade and the Street still can't agree on what it's worth. Street PTs now span $240-351 — Truist went to $240 from $218 (Hold), Baird's the bull at $351, RBC at $335, Bernstein $282, Roth/MKM $260 — and even the raisers sound reluctant. Stock's up 172% YTD but has given back 16% in the past week into the print.
"Upside bias on near-term numbers, but still debating the opportunity size through 2030."
That's Truist, and it captures the whole debate. Nobody doubts the near-term setup — 91% LTM revenue growth, strong 3Q bogey, improving profitability. The fight is over terminal market size: what the behind-the-meter data center TAM actually is and what happens beyond 2GW of capacity. At these levels, valuation already prices a lot of the long-term answer.
Earnings Oct 22. Options imply a ~9% move, and BE has beaten the implied move in half its last eight prints — this is a name where you respect the setup or you don't touch it. (We're in the former camp on positioning, but the bar is high after 172%.)
Buy the AI durability story — that's the Street's take after Snowflake's NYC World Tour. Truist reiterated Buy at $425 PT, coming away "incrementally more positive" on the AI consumption tailwind, with CoCo demand the standout — customers combining models with Snowflake-native context, and actually managing costs via quotas, eval and routing rather than throttling deployment (that's the part bulls needed to hear on consumption sustainability).
Fresh initiations pile in behind it: PhillipCapital at $423, Yorkville at $400, StoneX reiterating at $400 — a clean $400-425 cluster, all citing AI-driven product momentum (Q2 product revenue +37% Y/Y, best in three years). Street high sits at $525. With the stock UP 133% in six months and 39 analysts revising earnings upward, the easy multiple expansion is done — this is now a durability bet, and the sell-side keeps saying yes.
"CoWork is beginning to extend Snowflake beyond technical users and traditional dashboards."
That natural-language analytics push broadening the user base is the second derivative worth watching — if Snowflake cracks non-technical seats, the consumption TAM argument gets real teeth. Bear case is simple: you're paying up for a stock that's already priced a lot of this in, and "AI consumption tailwind durability" is doing heavy lifting in every note.
TSM delivered the cleanest kind of print: beat, raise, and a 2nm ramp that's running ahead of schedule. Q3 revenue came in at $46.7B, UP 16.2% SEQUENTIALLY, with LTM revenue +36% y/y to $155B. The real tell is the mix shift — 2NM HIT 8% OF WAFER REVENUE VS 3% LAST QUARTER. That's a node ramp pulling forward, which only happens when customers are banging on the door for leading-edge capacity (read: AI accelerators).
The one soft spot: gross margin ticked down to 66.4% and management flagged continued pressure from the 2nm ramp plus overseas fabs. That's the known tax on this story — new node plus Arizona/Japan dilution. Capex guidance held steady, which tells you they're not panicking on either side of the AI debate.
Street response was a chorus of the same note: D.A. Davidson's Luria reiterated Buy/$500 calling it "beats and raises," Needham pushed to $560 from $530, and Lynx held NT$3,000 while arguing the five-year AI accelerator CAGR is likely HIGHER than prior models. The collective thesis is simple — CEO says demand is "extremely robust" and there's a real supply/demand gap, so estimates still chase the stock, not the other way around.
"The five-year compound annual growth rate for AI accelerator revenue is likely higher than previously anticipated." — Lynx Equity
R/r stays favorable as long as the 2nm mix keeps climbing faster than margin dilution bites. The bear case is the same one it's been for six quarters: AI digestion + overseas margin drag. Nothing in this print feeds it. Watch GM trajectory next quarter — that's the bogey.
UBS is fading the rally into a 52-week high — and the tape isn't agreeing with them yet. They reiterated NEUTRAL with a $296 PT (stock at $344.67, just off the $348.15 high), citing Evidence Lab data showing iPhone 18 wait times running SHORTER y/y — Pro down ~4 days, Pro Max roughly flat — and rolling over week-over-week (Pro -1 day, Pro Max -5). Their read: declining wait times against stable supply = demand elasticity bite from this year's price hikes, and limited upside to iPhone estimates.
"Declining wait times against a stable supply backdrop are an increasing concern."
Context matters. This isn't consensus — it's one corner of a genuinely split street. Evercore sits at OUTPERFORM/$380 reading the same delivery-time data bullishly, KeyBanc is UNDERWEIGHT/$250, and Bernstein is Outperform but nervous on Dec-quarter gross margin and units. So you've got a stock at 39x earnings, near highs, with no clean signal on iPhone demand and a margin fight brewing for the guide.
The wait-time bogey is worth tracking week to week — last year at this point times held flat, this year they're compressing. But UBS's $296 vs. a $345 stock tells you how much the market cares right now. Not sure the availability data alone gets you to shorting this into the print. Watch the Dec-quarter GM guide — that's where Bernstein's worry and the IDC memory-cost headwind (global smartphone shipments -22% in Sept Q) actually collide with the multiple.
Street keeps chasing this one higher. BTIG lifts PT to $315 from $285 (Buy, top large-cap software pick), joining a $290-325 cluster from Rosenblatt, TD Cowen, and Stifel. Stock at $273.40, UP 157% over the past year and +16% in the last 1.5 months on the back of S&P 500 inclusion (announced Oct 1) and the agentic AI narrative getting real teeth.
The thesis: Twilio becomes the communications infrastructure layer for personal AI agents. Near-term revenue contribution stays modest — mostly Verify use-cases — but the medium-term unlock is two-way messaging, outbound AI voice, and cross-channel orchestration. MS adds that consumer AI agents multiply interaction volume, which feeds directly into usage-based monetization. That's the right way to think about it — this is a picks-and-shovels play on agent volume, not an AI app bet.
"Recent meetings with management bolster BTIG's confidence in Twilio's positioning as the communications infrastructure layer for the agentic layer. The firm continues to see shares as offensive and defensive in the current backdrop."
Bear case is valuation and expectations. Up 157%, bogeys now assume organic growth keeps inflecting — Stifel's call is explicitly a top-line beat. If agent revenue stays "modest" longer than the market's patience, the multiple compresses fast. But third-party data says demand stays healthy, and positioning post-S&P inclusion keeps a bid under it. R/r still favors the bulls here, just with a higher bar. Not the place to add aggressively after this run, but no reason to fight it either.
Analyst Day went badly — stock closed DOWN 10% after trading off as much as 16.6% intraday, capping a 23% weekly bleed (still up 103% over the past year, so this is a positioning unwind, not a broken story). The culprit: management's FY2030 targets ($3.2B rev, $12.40-13.00 non-GAAP EPS) implied a growth deceleration in FY29/30, CopperEdge detail was thin, and in this tape AI names get zero benefit of the doubt. StoneX held the line at Buy/$230, calling it a growth-expectations reset — and the street broadly agrees, with BMO at $230, MS at $205 on data center strength, and Stifel trimming to $200 (Northland, meanwhile, upgraded then downgraded within weeks, so take them with a grain of salt).
StoneX said the selloff reflects a growth-expectations reset... the implied slowdown in fiscal years 2029 and 2030, limited detail on CopperEdge products, and a skeptical AI trading environment drove the disappointment.
Worth noting the company actually RAISED FQ3'27 bogeys to $417.5M rev / $1.14 EPS at midpoint — near-term numbers fine, long-term slope is what got punished. After-hours bounce of +2.1% suggests the washout found some buyers. At a $13.3B cap after a 23% weekly drawdown, r/r is more interesting than it was Monday, but we'd want to see the FY29/30 bridge before calling the reset done.
Macquarie joins the bull camp with an Outperform and $780 PT — right in line with where the street already sits (UBS $804, Truist $800). Call it $800 fair value vs. $611 last. The thesis is the remix: CACI isn't your father's federal IT services shop anymore. Electronic warfare, space, counter-UAS — higher-margin, higher-growth revenue, and the ARKA deal deepens the space exposure.
The valuation case is the kicker. FY27 FCF yield around 7%, compressing to 5.2% at target — still nearly double the S&P's ~3%. Add a $32B+ backlog with funded backlog up 29% y/y and 7%+ organic growth, and this screens like a cheap way to own defense tech without paying defense tech multiples. Not a new idea (three firms circling the same $780-804 bogey says consensus), but the r/r still works if execution holds.
"CACI’s evolution beyond its legacy Federal IT Services business toward higher-value technology and products... this provides additional valuation support alongside continued execution in higher-growth end markets." — Macquarie
Street's piling in and the stock keeps confirming. ZETA at $33.64 sits just off its 52-week high of $34.92 after ripping 93% in six months, and the PT cluster has migrated to $38-40 — Needham at the top ($40 from $30), Truist, DA Davidson, and KeyBanc all at $38. Truist's reiterate today comes off an expert call with a digital marketing vet who's worked the Zeta platform since 2018, covering Athena/AI opportunity, the Palantir tie-up, and the martech replacement cycle.
The through-line across all four desks: Zeta Live 2026 product drops (AthenaOS, Athena MCP) plus customer feedback that AI marketing budgets are real and broadening — this isn't pilot-program spend anymore. Consensus is clearly long here, so the question is less direction and more whether $38-40 bogeys are stale given the stock's run. R/r gets tighter near highs, but nobody on the Street is stepping in front of this one. Only knock: the coverage is all momentum-and-product narrative — not much on the actual numbers.
StoneX defending the Buy ($450 PT) into the October 26 print, and the setup here is more interesting than the rating itself. CDNS has UNDERPERFORMED the SOX and NDX YTD despite posting the strongest revenue growth since FY22 — Q2 revs came in at $1.584B, +24% Y/Y, with EPS $2.11 vs $2.05 consensus and a big FY guide raise. Muted reactions to back-to-back beat-and-raises. Classic case of a stock where fundamentals keep accelerating and nobody's paying for it yet.
The bull case StoneX is making: EDA is the stealth AI beneficiary. Chip design activity inflects with every new compute cycle, and 20%+ EPS growth plus multiple expansion should close the gap vs. peers. They've already outperformed IGV YTD and SNPS by ~450bps — so relative performance is working, it's the absolute re-rating that's lagging.
"StoneX expects Cadence Design and Synopsys shares to outperform the software sector as investors recognize these software tools vendors can benefit from AI and new compute hardware."
The honest caveat: StoneX itself admits the CDNS/SNPS choice got harder after SNPS's Sept 30 investor day signaled real momentum. And at a PEG of 1.8 with the Street spread from $300 to $470, this isn't a consensus screaming buy — Piper's sitting at Neutral with a $349 PT. R/r into the print leans favorable given the YTD underperformance vs. SOX, but if you're picking one EDA name for the AI design cycle, the SNPS debate is live. Watch October 26 AMC.
FIVN keeps grinding higher — shares UP 124% over six months and UBS finally catching up, lifting its PT to $48 from $36 and staying Buy after doing the conference circuit (10 customer/partner checks on AI adoption and the Muse personal agent chatter). The real thesis isn't the AI hype, it's the two LARGE FINANCIAL SERVICES DEALS ramping — UBS thinks one alone adds ~2pts of incremental growth in both FY26 and FY27, taking revenue growth from 10% in Q2 to ~14% and putting upside on the Street's 10% FY27 bogey. Valuation still screens cheap for a re-acceleration story: ~10x CY27 EV/FCF, 8% FCF yield (new PT is just 11x CY28).
The honest hedge from the same note: most checks say personal agents are unlikely to drive a meaningful uptick in volumes or spend near term — so don't pay for the inbound-AI narrative, pay for the signed deals. The rest of the Street sits below UBS: PTs cluster at $40 (Cantor, Needham, Truist, FBN) with DA Davidson the outlier at $33/Neutral, all anchored on the AI business now past a $150M run-rate (~15% of subscription revenue). Setup into Q3 looks fine, but after a 124% run the r/r hinges on those fin-services ramps actually showing up in the print.
"Most checks suggest personal agents are unlikely to drive a meaningful increase in Five9 volumes or spending in the near term, despite renewed investor interest around the potential for higher AI-generated inbounds." — UBS
NFLX INTO THE PRINT TUESDAY AND THE TAPE IS UGLY. Stock down 32.6% over six months to $72.61 — a long way from the $124.86 52-week high — and BMO's Brian Pitz just trimmed PT to $125 from $135 while keeping Outperform. The worry is the same one that's been building: UCAN growth slowing, engagement weak, and Sensor Tower app data pointing to a muted 3Q. Options are pricing a 7.3% move on the October 20 report, which tells you positioning is already nervous and the bogey isn't high.
The bear case writes itself. Content chief turnover (Peter Rice replacing Bela Bajaria), a 6% workforce cut, and open questions on the long-term growth algorithm. When a company swaps its Chief Content Officer two weeks before earnings, that's not nothing.
The bull case rests on ads. BMO's expert call flagged advertising as the clearest path to reacceleration in UCAN, and the firm's proprietary survey still shows strong consumer positioning despite the selloff.
"Questions about the company's long-term growth algorithm" — BMO keeping Outperform but acknowledging the core debate.
Not sure we can read too much into one PT trim on a $125 target when the stock sits at $72 — that's still 70%+ of upside embedded in the number. The real tell is whether ads revenue and the engagement data show any inflection Tuesday. Jefferies initiating Buy with a 14%+ decade-long EPS growth call and KeyBanc at $85 (FX headwinds, tempered margin expansion) bracket the debate nicely: this is a "prove it" print, not a "miss it" one. The 2030 World Cup rights in Spain are interesting strategically but don't move the 3Q needle. Risk/reward skews interesting only if you believe the ad tier inflects — otherwise, wait for the print.
Street is piling in ahead of the OCT 28 FQ1 print, and the setup looks forgiving. StoneX goes to $600 from $450, and that's actually the LOW end of the recent cluster — UBS at $675, TD Cowen at $625, Yorkville Ives initiating at $615. Consensus range now spans $440-870 with ~11% average upside. The bogey that matters: Azure, and StoneX calls the bar "highly reasonable." That's a good r/r framing into a print.
"Fiscal 2026 proved Microsoft can build AI infrastructure at extraordinary scale, while fiscal 2027 increasingly becomes about converting that scale into greater AI yield — useful intelligence, revenue and profit per dollar and watt deployed."
That StoneX framing is the right lens. Build phase is done and proven; now it's yield phase, and every one of these notes (OpenAI growth at TD Cowen, enterprise AI monetization at Yorkville, Copilot agent redesign) points the same direction. The real tell on the 28th isn't the print itself — it's whether Azure in-line gets rewarded or met with a shrug. Given where positioning sits after this PT parade, shrug risk is nonzero. But nothing here changes the core long.
DTE is a data center story wearing a utility costume, and the Street can't agree on what it's worth. Truist cut to $150 from $160 (kept the Buy) while Mizuho went $152 to $136 (kept Outperform) and BMO nudged up to $134 (Market Perform) — so you have cuts across the board but nobody abandoning ship. That's a "valuation reset, thesis intact" tape, not a broken story.
The thesis: Truist expects the upcoming plan update to bake in the Google contract approval (MPSC conditionally approved the special contracts, 15 conditions to accept within 30 days) and model 7% growth in 2027 off the 2026 guide midpoint, then 8% annually thereafter. There's 2GW in advanced discussions behind that, with the firm expecting more small and large load announcements — possibly with capital and upside detail on the Q4 call.
"Truist anticipates the plan will show 7% growth in 2027 off the 2026 guidance midpoint plus 8% growth thereafter off the 2027 7% midpoint."
Near-term watch items: Q3/Q4 is the window Truist flags as key, and the February rate case order determines stayout potential (Michigan may shift to multi-year rate plans — that's a real regulatory tailwind if it lands). Truist's FY26 estimate of $7.72 sits at the top of the $7.59-7.73 guide, so they're leaning high-end. Caveats: Q2 was a miss ($1.32 vs $1.47 consensus) and Vantage monetization remains an open question. At $129 with a ~3.6% yield and 56 years of dividend payments, you're getting paid to wait — but the r/r hinges on that 2GW pipeline converting into signed contracts, not just "advanced discussions."
Street's stacking up ahead of the print. SIX firms raised or initiated this week — Piper to $175 (from $140), RBC $176, Citi $170, TD Cowen $165, Oppenheimer $165, Yorkville Ives initiating at $160 — with the stock at ~$147. The collective thesis: sentiment bottomed in 1H, and 2H26 has real catalysts. Security crossed $1B ACV in Q2, the Armis and Veza deals are outperforming, and Q3 is the second quarter of a SKU transition carrying a 20-30% average price lift. Consumption revenue should build as NOW pushes toward its $1.5B AI ACV bogey, and current FY26 guide implies a modest organic Q4 acceleration. Setup looks clean — 22% growth, 75% gross margin, and the AI-as-threat narrative flipped to AI-as-catalyst.
Yorkville Ives: ServiceNow could benefit from AI as a catalyst rather than a threat.
The bear case writes itself at this multiple if the SKU transition price increases hit seat counts or the consumption ramp slips, but nobody on the street is making that call right now. Watch the print for whether Q4 organic acceleration actually shows up in guide — that's the bogey that matters.
UBS stayed Neutral at $10 PT after hitting Asana's conference — nine customer checks plus a management Q&A — and the takeaway reads like a shrug with a thesis. The multi-product pivot (new AWM packaging plus three agentic apps: Service Management, Client Management, Command) gives ASAN growth levers beyond seat expansion, and UBS calls that a positive step. But customer AI adoption is still early, and nobody at the event dropped new metrics or guidance. Meanwhile the stock's already RUN 78% in six months to $10.12, which at ~15x CY27 EV/FCF UBS calls "relatively full" for a high-single-digit grower. Fair.
"UBS views the evolution from a single-product offering to a multi-product platform as a positive step that provides additional growth levers beyond seat expansion."
Bull case: cRPO grew 11% (up from 8% in Q1), NRR improving, Citizens sits at Outperform with a $15 PT. Bear case: Q3 guide had PLG headwinds and pricing-model noise baked in, and the AI agent monetization story is checks-based, not numbers-based. At $2.3B market cap with 87%+ gross margins, the r/r here hinges entirely on whether agentic apps actually move the growth rate off ~9-10%. UBS is comfortable waiting. So are we — this is a "show me the AI revenue" story, not a position you chase after a 78% run.
WDAY's Rising conference gave the Street what it needed — hard agentic AI numbers — and the stock's +57% RUN OVER SIX MONTHS now has receipts to point at, even if BTIG stays Neutral. The headline: AGENTIC AI ARR ~$600M, 25% of the customer base on an AI product vs ~10% a year ago, and AI now MORE THAN 50% OF INCREMENTAL CB ARR in Q2 with management steering toward ~70% by FY28. Buyers of at least one agent saw a 12% average ARR uplift — that's the attach/expand math PMs care about, not the keynote demos.
"Agentic AI annual recurring revenue of approximately $600 million... AI represented more than 50% of incremental customer base annual recurring revenue in the second quarter, with that mix expected to move toward approximately 70% in fiscal year 2028."
The plumbing data matters too: 300 Flex Credit deals signed, 5,500+ customers on organic agents, Sana approaching $100M ARR, and Extend AI/HiredScore/Paradox each clearing $100M. Over 50% of customers on unified multi-sub agreements, with visibility to 75% over the next few quarters — good for revenue durability. Management also floated potential upside to the ~11% FY28 subscription growth framework.
The Street is split, not sold. BMO stays constructive (Outperform, $230 PT) on faster product innovation, while Stifel nudged up to $175 on AI momentum. The bears' case is simpler: KeyBanc downgraded to Sector Weight on AI budget allocation concerns, and UBS trimmed to $210. BTIG's Neutral says the quiet part — after a +57% move, the FY28 bogeys (11% sub growth, 33%+ non-GAAP op margin) are already the expectation, so r/r from here is about beat-and-raise, not multiple re-rating. Watch whether that "upside to FY28 framework" comment converts into an actual guide raise.
Morgan Stanley initiates at EW with a $56 PT — and the subtext is the story: the stock is UP 221% YTD and MS thinks valuation has already caught up to the improved growth. "Valuation as already having calibrated to improved growth expectations." That's a polite pass at $49.60.
The fundamentals aren't the problem. Q2 beat clean — $0.37 EPS on $219.9M revenue (+22% y/y), EBITDA +27% to $28M, record 18.3% margins, raised FY guide across the board. Global Voice Plans (the biggest cloud comms piece) has accelerated from 4% growth in Q1 FY25 to sustained 10% over the last three quarters. Incremental gross margins at 82% in FY25. Carries traffic for every UCaaS/CCaaS Magic Quadrant leader — MSFT, GOOG, ZM, AMZN, Genesys. Real infrastructure, real customers.
The bear case is mechanical, not existential: political campaign messaging adds $13M to FY26 that won't repeat in FY27, so there's an optical decel baked in regardless of execution. Management says double-digit cloud growth persists beyond FY26 ex-campaigns, but you have to believe them. (Also worth clocking: stock sold off hard premarket on a beat-and-raise quarter — classic "what do you do for an encore" positioning after a 3x run.)
Street's split. B.Riley at $85 and Needham at $60 (Voice AI angle) vs. Piper's $52 raise, BofA Neutral, and now MS EW — the bulls are underwriting the Voice AI / Agentforce CX optionality, the neutrals are saying the multiple already did. With PTs scattered from $52 to $85, nobody actually knows what this is worth yet. R/r feels balanced-to-squishy at $49.60; you'd want the FY27 guide to confirm the ex-campaign growth story before paying up here.
AMZN — The Muse-blocking story is the tell. Amazon blocked Meta's agent within days of it hitting #1, and a third of US shopping/travel sites did the same. This isn't privacy theater — it's defense of the ~$68B ads business, which agents bypass by design. Agent neutrality is now a real regulatory/regime question for the ad-supported commerce model. AMZN's moat here is structural (they own the marketplace), but watch Shopify/Walmart/Best Buy cutting connector deals with Meta — that's the flank.
CRWV — CDS widening is convert hedging, not default betting — don't fade the death-spiral narrative just because spreads moved. The real variable is advance rates: hold 75–85% secured advances plus customer prepayments and the liquidity math works; compress toward 55–65% and you're staring at $10–15B residual funding holes. Too many deep-pocketed stakeholders (NVDA, MSFT adjacency) to let it spiral. Also note the GTC Berlin exclusion story — NVDA pressuring neoclouds to stay single-vendor is a competitive moat tell, and CRWV is ground zero for that fight.
GFS — Nothing fresh on tape. The adjacent read: Samsung targeting November groundbreaking on Taylor fab 2 and sniffing around silicon photonics M&A keeps foundry capacity/expansion narratives warm, but that's not a GFS signal. No bogey to work with here.
CRDO — No direct signal, but Jensen naming silicon photonics as a multi-year constraint is the cleanest thematic tailwind in the feed for connectivity/retimer names. Supply-constrained content-per-GW going from $18B to $40B+ means every dollar of interconnect content gets pulled. Worth flagging for positioning even without a print catalyst.
BSY / PCOR — Quiet on both. Infra software (construction/engineering adjacency) has no fresh checks in the tape. Jensen's repeated emphasis on land, power, permitting, and skilled construction as the binding constraints on 43 GW is directionally friendly for the vertical — somebody has to design and manage these builds — but that's thesis, not signal. Not enough to act on.
CLS — Nothing direct. EMS names stay levered to the AI buildout cadence, and the NVDA content-per-GW math ($40B+/GW) plus "allocations below requested quantities" is supportive of order durability into 2027. No new alpha on the tape today.
SPCX / CCC / ECHO / ALAB — No fresh signals. ALAB gets the same silicon photonics/connectivity read-through as CRDO (CXL/retimer content per rack rising with NVDA's content math). SPCX, CCC, ECHO — tape is silent. Don't manufacture a view.