Good morning.
Semis are the tape — MU, AMAT, European chips red in futures after Anthropic's $65B ARR missed the ~$80B bogey. (Unaudited, non-GAAP, private-company metric — that's the fragility.) 30Y AT 5.32%, HIGHEST SINCE 2007, "AI issuance" now the bond-selloff headline. KOSPI +3.1% THROUGH 7200, foreigners net buying — Seoul crowning the same AI story New York is discounting.
FN is the print of the window: DCI at $1B+ run-rate, Cisco 20% / Amazon 11% / Nokia 11% of revenue — AI demand broadening past NVDA. Xiaomi missed on memory costs — first big downstream casualty. HD beat; BABA +5% into Thursday. Macro: ZEW fourth straight month up, Moody's lifted Korea to 3.5%, Brent >$90, EM snapped a four-day win streak. Fed minutes Wednesday.
Theme one: the Anthropic "miss" is a 28-day × 13 artifact — base case still beats/raises, so the dip is positioning, not demand. Theme two: memory flipped from beneficiary to cost-shock — handset 2027 margins get repriced, MLCC bifurcates AI-spec vs consumer-spec. Theme three: the bond market now charges AI capex interest — circle financing's rate elasticity is the tail risk; watch the next infra bond spread. Theme four: China's AI capital-market window is open — Biren +18-21x, VeriSilicon backlog $1.85B, Unitree IPO the pricing test.
We'll hit up FN, MU, and AVGO first, then get to the optical complex — ANET, CIEN, LITE.
Five PT raises in one morning — UBS to $425, Citizens to $408, Evercore to $360, Stifel to $350, Jefferies to $310 — and they're all singing the same song: CoCo is real, consumption is inflecting, and Q2 is a beat-and-raise setup. The PT dispersion isn't disagreement. It's just where each shop lands on valuation patience for the only large-cap AI software name that's actually accelerating.
The market has already front-run the print. UBS flags buyer growth expectations at 36-37% for Q2 versus 30% guidance. Stifel says investors want >34%, implying a beat north of 3%. Evercore sees 300-400bps of upside to the $1.415-1.420B guide. So the street has priced the beat BEFORE the print. UP 87% OVER SIX MONTHS, 50% YTD, trading at 15x revenue and 62x FCF on CY27/FY28. There is zero discount in this tape.
The customer evidence here is genuinely strong — not hopium. Citizens surveyed 15 customers: 14 plan to increase spend over the next 12-18 months. Rep attainment median 131%, with 18 reps beating quota. Quota-carrying attrition collapsed to 4 from 15 last quarter. 88% of 49 data points positive versus 53% last quarter. That's a sequential improvement in the sales motion, not a seasonal artifact.
The best line comes from an industry source at one of Snowflake's largest partners:
"I absolutely see Snowflake continuing to do well as I do with Databricks. Agents are creating more and more queries and [with] more queries there is more data and that is how you charge in a consumption model."
That's the whole bull case in one breath, PMs. AI agents are a query multiplier. More queries → more data → more compute → more revenue. Snowflake gets paid on every loop of the flywheel. And it's not just a story — Stifel points out last quarter's CoCo usage spike wasn't in guidance. It IS in Q2 guidance now, but with one quarter of operating history, the assumptions around it are conservative. Conservative guidance + strong June consumption = another upside quarter likely.
Databricks adds color, not just comp. The $3B raise at $188B values a private, less profitable competitor at roughly 2x Snowflake's public market cap. Jefferies reads it as sector validation, and it's a useful mental anchor for how much room the AI data narrative has.
Bull: CoCo is a second growth curve with one quarter of history — meaning guided conservatively. Sales execution inflecting (attrition 15 → 4). AI agents structurally multiply consumption. UBS's Keirstead steelmans the macro: investors will pay up for the few accelerating AI software names, period.
"This leaves little room for error, but in our view investors are likely to be more generous with valuation multiples for the small number of accelerating AI-driven software stocks."
Bear: 15x revenue is a perfection price, and perfection is now the entry bogey. If Q2 lands at 31-32% growth — a beat, but not the 36% the tape wants — the multiple does the damage on the way down. And the CEO comp package ($448M tied to nearly doubling market cap) is a long-dated call option, not a near-term demand signal. One quarter does not make a trend line for CoCo. We don't know what usage normalization looks like.
Strong fundamentals. Improving execution. But the stock now trades on the AI multiple, not the fundamentals. The bull case works if you believe the agent flywheel compounds for several quarters. The bear case is simply that the market is already there, and Q2 is a "good news, priced in" risk. Own it into the print, but respect that the 87% six-month run means we're paying up for what we already know.
MDB has already traded like the guide-up happened. +41% in a month (IGV +10%), +93% over the year, $35.3B market cap. Buy-side expectations for 2H sit AHEAD of management's implied ~23% growth. The stock needs a 25%+ guide just to justify this multiple. Earnings in 14 days — that print is everything.
Six firms raised PTs into the $416–$560 range (Guggenheim $560, RBC and Tigress $515, Stifel $475, Needham $430, Cantor $416). All Buy. Stifel alone moved $435→$475 in one quarter — analysts are chasing the tape higher, not leading it.
The consensus: Atlas growth holds at 29% in Q2 and reaccelerates from there. Guggenheim sees Q3 guidance at 26% vs consensus 23%, with a full-year raise of AT LEAST 200bps to 26% at the midpoint. Stifel leans on hyperscaler consumption prints and says Q1 trends support ~3% upside to the quarter. Q4 Atlas did 29%. Buy-side needs a repeat against a harder comp. That's the bar.
The bull case is simple: agentic AI workloads need a document database, MongoDB is the default, and the consumption model maps 1:1 to hyperscaler spend. New ARR stacks through FY28, margins expand toward 22% non-GAAP, Atlas ultimately hits 30%.
"Management likely needs to move its third-quarter and second-half guidance to the 25%-plus range to support the stock at current levels." — Stifel
Bull: Product-market fit for agentic apps is real. Technical partnerships stack up. When hyperscalers rip, MDB rips harder. You're paying ~10x CY27 EV/rev for a company whose growth accelerates — not decelerates.
Bear: RSI overbought. Above fair value. 93% in a year means crowded positioning and an embedded expectation. You're not buying a beat — you're buying a raise. Guide Q3 at 24% instead of 26% and this gives back 15% in a week. Poor r/r into the print.
Watch the guide, not the beat. Q2 total revenue ~28% (Guggenheim, 3% above consensus) is already in the numbers. The real tells: Q3 Atlas guide (26% needed) and full-year raise (+200bps needed). Anything less and the 10x multiple compresses fast. Wait for the print. 14 days.
BofA added MU to the US 1 List alongside the reiteration. They're not hedging this one.
UBS's Tim Arcuri gets the cleanest articulation of the bull case:
Longer-term investors will see memory as capturing much more sustained value as the economics of AI token generation will be fundamentally more determined by memory than compute.
He's saying compute gets commoditized, memory gets monetized. That's the entire thesis in one sentence.
Bear: Memory has cratered every single time it looked "different." Investors are asking about capacity coming online, gross margin downside, and supply chain agreements struck at pricing one to two quarters ago — stale pricing that compresses when the spot market turns. Netlist's ITC patent complaint adds legal overhang. UBS admits MU stays a battleground near-term.
The r/r here is asymmetric — the narrative is shifting from cyclical to structural, and positioning hasn't caught up. At $1T+ market cap, that's an unusual combination.
UBS pulls the Buy tag but pushes the PT to €26 from €22.50. That's not a bearish call — it's a valuation call. The stock's up 84% off the March 2026 low and +8% YTD while peers are DOWN 0-41%. Risk/reward is balanced now, plain and simple.
The fundamentals actually came in hot. Management raised FY26 guidance to at least 13% cc revenue growth with 37.4% adj EBITDA margins. UBS bumps its cc revenue forecast to 13.7% vs 13.2% consensus, and sees SaaS/hosting leading H2 with evolutive support reaccelerating as implementation activity normalizes. Product dev +19.4% cc, config/engineering +25.2%. The soft spots: IT/digital transformation +1.6%, professional business apps -21.0%.
"Tailwinds including a favorable competitive environment and robust project portfolio management demand appear priced into shares."
One missing catalyst: capital management. €205M net cash on the books, but management wants fiscal visibility first — any special dividend unlikely before April 2027. So no near-term shareholder return kicker, and at these levels the multiple has to do the heavy lifting. Fine hold, not a fresh buy.
DB CUTS S TO HOLD — NOT ON THE TECH, ON THE PRICE. Stock's ripped ~90% off April lows, and the Street is paying for a Mythos-driven inflection that DB says isn't coming soon. PT goes to $24 from $17, which is the firm literally saying "great company, wrong entry."
Valuation is the whole ballgame here. PANW and CRWD now trade at ~4x growth on CY27 FCF. Everyone else gets single-digit revenue multiples. DB's latest reseller survey (51 participants) plus Black Hat field work shows a mixed picture — no clean demand signal that S is taking share in the AI era's "largest platforms win" dynamic.
"SentinelOne and Netskope are likely 'show me' stories at current levels and will need to show consecutive quarters of improving execution with sustained margin leverage to warrant further investor sponsorship."
That's the crux. This isn't a broken-company call — it's a broken r/r call. The market's pre-paid for a reacceleration, and new CFO Sonalee Parekh needs to print two or three clean quarters before PMs should chase. Watch the reseller checks and the margin line, not the narrative.
DA Davidson isn't moving numbers — reiterated Buy and $74 PT (33x 2026 EBITDA) — but the site visit to the San Jose eProbe manufacturing facility matters more than the rating itself. STOCK +163% OVER 12 MONTHS, +78% YTD. We're not early. But the 2027 book of business is actively building, with existing and prospective customers all engaged.
The Q2 beat was modest ($0.27 vs $0.26 est, $61.52M vs $60.96M). That's not the point. The point is the eProbe inflection hitting the P&L in 2027, and DA Davidson views that as "a primary source of upside to its above-consensus expectations." This is a narrative trade now, not a fundamentals trade.
"eProbe growth as a primary source of upside to its above-consensus expectations."
Bernstein also came in with Outperform and a $65 PT — same bull narrative, lower multiple conviction. So you've got two bulls, $65 and $74, both leaning on structural semis R&D spend and WFE requirements. The bear case: 33x 2026 EBITDA leaves zero room for a bookings slip. At $53.16, a 2027 miss doesn't just dent the stock, it compresses the multiple hard. But with the facility tour behind them and management reiterating the growth framework and margin trajectory, the near-term narrative has legs. Keep this one on the momentum tape.
Stifel reiterates Buy with $350 PT — the only real print on the tape this morning, and it's the right one. MRVL at $234 ($205B cap) after a 206% year. Multiple's rich (80x trailing vs 60x CY27 target), but the rate of change on guidance just stepped up.
July quarter beats the $2.70B estimate (11.7% q/q) — Data Center is 76% of mix and driving via optical interconnects plus lead XPU ramp. October guides above $3.00B (+11% q/q) on accelerating Interconnect, flagship XPU production, XPU-Attach scaling. THE BIG ONE: management ripped Interconnects to 70% YoY FY27 growth from 50% prior. Data Center total now +50% FY27 and +55% FY28.
Management now expects Interconnects to grow 70% year-over-year in fiscal 2027, up from a prior 50% forecast. Data Center revenue overall is projected to grow 50% in fiscal 2027 — and a further 55% in fiscal 2028.
UBS to $340 (CXL angle), Goldman to $195 Neutral (still behind), Erste to Hold on valuation. Bear case is just a P/E argument. With this trajectory, that's a dip-buy setup, not a short.
COMPASS POINT GOES BULLISH — BUT THIS ISN'T YOUR GARDEN-VARIETY AI INFRA STORY. Initiated Buy with $44 PT (stock at $14.95, street targets $24–40). The hook: GRRR claims a $5.5B commercial book across 40,856 signed GPUs in India and Indonesia since March 2026 — against just $101.4M of revenue in 2025. That's not a pipeline, that's a trajectory.
The thesis is a step-change, not an increment. Compass Point models revenue jumping from $222.2M in 2026 to $1.2B in 2027, with adjusted EBITDA going $46.4M → $772.7M. That's a 16x EBITDA swing in one year. The bull case rests on signed deployments converting, not on hopes and prayers — GRRR already guided Q2 2026 revenue to at least $44M vs. $33.9M consensus, and just announced a five-year AI compute contract worth ~$2.5B, with ~$1.3B tied to the NeutraDC Batam deployment in Indonesia.
What's differentiated here isn't the GPU buying — it's the full-stack positioning:
"Gorilla brings customer-specific infrastructure design, hardware integration, cybersecurity and Asian implementation to dedicated GPU deployments."
That's the wedge. Security intelligence, edge AI, and government work (the Egypt secure-network project) give them on-the-ground execution cred in Asia that pure-play GPU lessors don't have. Plus the Shackleton acquisition rebranded as Gorilla Tech Capital gives them a regulated financing platform — so they can own or finance the iron, not just resell it.
The risk is obvious: $101M of revenue doesn't normally become $1.2B in two years. Execution risk is enormous. Convertible notes ($125M at 7.50% due 2031) dilute over time, and the stock is already up huge off any base that existed before the AI pivot. But if even half those signed GPUs come online per schedule, the 2027 numbers make $14.95 look silly. This is a show-me story with real contracts underneath — worth owning for the optionality, not the comfort.
Verdict: valuation call, not a fundamentals call. Stifel raises PT to $90 from $65 (Buy), but the stock's already at $83.44 — implied upside of barely 8%. The bull case is simple: SaaS comparisons ease through FY27 and at ~13x CY27 EV/FCF you're not paying for perfection. Gross margins 76%, revenue +17% LTM. Nobody's arguing with any of that. They're arguing with the trajectory.
THE Q4 CLOUD PRINT IS THE WHOLE STORY. Cloud revenue $217.4M — DOWN $1.1M QoQ. First-ever Q4 sequential decline, and Q4 is supposed to be the seasonal peak. Growth decelerated to 20% YoY vs the ~$220M/~21% bogey. That one print drives everything: MS downgraded to Equalweight on growth inflection fears, Rosenblatt cut to $83 citing the restructuring (7% workforce reduction, CPO Ken Exner departing), UBS trimmed to $85. Piper's still constructive at $85 after CEO/CFO meetings. PT cluster now $83-90 — the stock is basically AT consensus.
"SaaS comparisons are easing over the course of fiscal year 2027."
That's the entire bull thesis in one sentence. If the deceleration was a one-off air pocket and comps ease as Stifel argues, then 13x FCF looks cheap. If Q4 marked the start of a structural slowdown — not a seasonal hiccup — then the valuation anchor is a falling knife. Two quarters of cloud growth data settles it. Until then, this is a low-conviction Buy with a tight PT range.
BofA stays Buy/$350 after meeting with NVDA management post the $105B OPENAI COMMITMENT. The story isn't the chip anymore — it's Nvidia morphing into the infrastructure banker for the entire AI buildout. Securing chips, land, power, and compute for frontier labs and neo-clouds is a deliberate hedge against the public hyperscalers who are all designing custom silicon in-house.
The bull case: GPU rental rates are solid, compute stays scarce, and the FCF machine funds the balance sheet expansion. The bear case: this is vendor financing at scale — if AI demand rolls over, Nvidia eats both the growth miss AND the credit risk. That's why the Aug 26 disclosure on off-balance-sheet commitments matters more than the revenue print.
BofA's math: 18x/15x CY27/28 EV/FCF vs a 36x/22.5x blended fair value — a 50%/34% DISCOUNT. P/E of 34.55 and PEG of 0.31 says the market still isn't paying for this growth. UBS separately sees a $3-4B revenue beat next week ($94-95B total), so the tape is set up for a good print. The real signal is what NVDA says about its liability side.
"Nvidia is committed to securing chip supply, land, power, and infrastructure for frontier labs and neo-clouds."
That's the strategy in one line — be the pick-and-shovel provider AND the lender to everyone who isn't a hyperscaler. Smart positioning, but it converts a pristine FCF story into one with real balance sheet leverage. Watch the filings on Aug 26, not just the guidance.
Stifel keeps Buy and $205 PT as TTM drops $1.1B cash on EPIQ Solutions — an AI-enabled software-defined radio / RF / space compute house serving signals intelligence and electronic warfare. Stock's up 237% in a year at $140.98, so the cheap stuff is long gone. This is a margin-mix compound story now.
EPIQ guides $160M revenue in 2027 with MID-30S EBITDA MARGINS and double-digit growth. Deal implies ~20x unadjusted EV/EBITDA, 17.4x synergy-adjusted — about where TTMI's own equity trades on FY26, for an asset carrying nearly twice the corporate margin. Net leverage ticks to 2.3x at close, down to 1.5-1.7x within 12-18 months. Straight-line execution of the "up the chain" directive from May's Investor Day. Q2 just printed $1.0B revenue — first time over a billion in a quarter — +37% YoY, EPS $0.99 vs $0.92 bogey.
"The EBITDA multiple is roughly where TTM Technologies' own equity trades on FY26 EBITDA, for an asset carrying double the corporate margin."
Defense-adjacent content, structurally higher margins, less cyclical grind. The risk is integration and whether Veritas squeezed every dollar of synergy out of the price — but at this multiple, with this margin step-up, the r/r still works.
Citizens keeps Market Outperform and $415 PT on PANW — no fireworks, but the setup is clear. The program expansion (partner-only motion for SMBs in JP/UK/AU/NZ) is the tell: direct sales stays anchored to large consolidation deals, channel absorbs the long tail. That's the right 3p allocation.
The broader tape is louder. AI security is the wallet-share winner. Goldman sees security budgets growing off AI spend; William Blair flags Mythos-driven urgency and expected firewall price hikes as demand accelerants. The PT cluster is now $400-430 (Tigress $430, Citizens $415, TD Cowen $400) — all anchored to the AI platform story.
Stock sits at $375.76, ~10% to the Citizens target, after a 113% year. Not a gift at these levels — but the rate of change on AI security spend is still positive, and in this tape, that's the currency that matters.
Benchmark reiterates Buy and pushes PT to $45 from $36 — 18% implied upside from the $38.02 close. The trigger: an absolute Q2 beat. REVENUE +20% Y/Y, 13% AHEAD OF MODEL. EPS $0.78, +15%, 11% ABOVE ESTIMATE.
The beat matters less than the forward raise. Benchmark bumps FY26 revenue growth to 11% from 6%. That's the real signal — they're saying the quarter wasn't a one-off, it's re-acceleration. Energy Services leads, Life Sciences/Data & Solutions improves, Specialty Health Care keeps compounding.
At $38 with an 89% six-month run already in the books, the easy money is made. But the PT keeps stepping up and Sutherland sees momentum holding into 2H. Modest upside, decent story — a solid hold for existing PMs, not a chase candidate at these levels.
Benchmark analyst Bill Sutherland said the firm expects solid momentum to continue into the second half of fiscal year 2026.
Market finally cares about the pivot. HIVE +11% on a $350M, five-year AI cloud deal with an investment-grade enterprise customer — validation that Buzz (the AI/HPC arm) is a real compute business, not a Bitcoin-mining side project.
$70M incremental AI cloud ARR, taking total contracted ARR to ~$180M. That's 90% OF THE $200M YEAR-END TARGET, and they got there in under nine months. Live cloud run-rate revenue sits at just $35M, so the gap between contracted and operational is the whole ballgame — ramp execution determines whether this re-rates further, not the press release.
Management also teased a potential colocation deal at the 32MW Boden site with an investment-grade Swedish tenant, possibly signing next month. HCW models another $45M ARR there — that puts contracted ARR at ~$225M, over target with months to spare.
Wainwright reiterates Buy and $7 PT. Stock trades $3.07, up 77% off the $1.73 low but still 60% below the $7.84 high — so the re-rating has room, but so does the drawdown if they stumble. Fundamentals back the story: FQ1'27 revenue $79.1M (+73% YoY), adjusted EBITDA flipped to +$13.4M from -$9M. The $142.9M GAAP loss looks horrid on a screen, but it's mostly a non-cash $84.7M Swedish tax provision — a headline bogey, not a cash burn signal.
Bull case: contracted ARR at 90% of target in nine months, IG counterparties, and cheap optionality on Boden. Bear case: $180M contracted is still only $35M live — Bitcoin remains the funding engine, and the deal flow has to stay this hot for the $7 PT to work. For PMs hunting small-cap AI compute exposure, this is one of the better asymmetric r/r setups in the tape.
CITIZENS IS THE LONELY BULL AT $86, AND THEY'RE PUTTING REAL FOOTPRINTS BEHIND THE CALL. Stock sits at $59.82 — that's a 44% implied upside if Walravens is right, and he's the only one with a target north of $55. The rest of the street is stuck at $50-55 (Jefferies, Wells Fargo) or just holding (Needham). That spread alone tells you the debate is live.
The thesis: federal business is inflecting. TWO AWARDS >$1M IN FQ2 FY27 ($3.3M DHS, $2.3M GSA) — that's FOUR large federal awards YTD vs ONE in all of last year. Combine that with sales attainment checks showing 16 of 18 reps beating plan at a MEDIAN 133% (up from 126% last quarter), and you have a rate-of-change story in the one segment that's actually predictable. Gross margins ~80%, revenue $3.29B LTM, Q1 beat the high end by 50bps and guided FY27 up $6M.
"Sales attainment checks based on publicly available sources identified metrics for 18 sales representatives in the second fiscal quarter, with sixteen exceeding plan at a median of 133%."
The bear case is right there in the same checks, though. Two private competitors told Citizens: one sees strength everywhere EXCEPT new enterprise deals — buyers are taking longer. The other had a genuinely disappointing quarter, driven by Europe weakness. That's not a noisy data point; that's a pattern. Enterprise is the margin driver for DOCU, and if deal cycles are stretching, the federal wins don't fully offset that drag. Q2 shows "modest improvement" vs Q1's enterprise disruption — modest isn't acceleration.
Net: this is a positioning play, not a momentum play. Federal wins give the floor, but the ceiling depends on enterprise digestion. At $59.82 with ~80% gross margins and an AI narrative (Iris, Perplexity integration, Slackbot), the equity's cheaper than the growth rate implies — but only if the enterprise booking environment cooperates. Citizens' $86 target is a bet that it does. The rest of the street is betting it doesn't.
UBS stays Neutral at $360 while the stock bleeds out — and honestly, the payroll price hike proves pricing power but also exposes the demand problem. INTU at $335.60, DOWN 49% YTD. The market doesn't pay up for price increases when hiring is the real bogey.
FY27 per-employee-per-month pricing: Core +8% (in line with FY26), PREMIUM +30%, ELITE +42% — a big step up from last year's +11%/+9%. No base subscription increase this year (vs 3-3.5% last year). UBS raises FY27 Online Payroll growth to 16.5% from ~13%, but that still marks a moderation from an estimated 18% in FY26. The new $5/employee/month Benefits Admin add-on is smart monetization without touching base pricing.
The offset? Gusto undercuts at $49-$80 + $6-$12 per employee. In a soft SMB hiring tape, the last thing customers want is a bigger per-head bill. Pricing power is real, but it's fighting a demand headwind.
The Street has already moved to "guide-down" positioning. Mizuho cut to $430 from $500 (Outperform), Truist downgraded to Hold at $350, Evercore holds Outperform at $400, TD Cowen nudged to $328 but stays Hold. The common thread: expect management to reset FY27 on the Q4 call. This isn't bull vs bear anymore — it's a debate over how deep the reset goes. UBS's 16.5% looks generous if the apps spending environment stays this horrid.
DB just told you the easy money's been made. Cut Netskope to Hold from Buy, slapped a $17 PT on a stock trading at $23.11 — near its $23.95 high. That's a 26% downside bogey. The only reason they raised the PT a buck off $16 is because the tape ran away from them.
The whole thesis is valuation and timing. NTSK is up ~90% off April lows, 72% in six months, on Mythos demand expectations. DB says the market's pricing in a demand inflection that isn't coming soon. Their reseller survey and Black Hat field work were mixed. Meanwhile NTSK trades only about one turn cheaper than Zscaler on CY27 EV/rev — for a company that still needs to prove sustained execution. That's not a premium worth paying when you're also hunting for a new CFO.
"Netskope will likely need to show consecutive quarters of improving execution with sustained margin leverage to warrant further investor support."
Translation: the next leg needs proof, not promises. The stock got ahead of the fundamentals — classic 90% rally, then the first sell-side guy blinks. The bull case is that Mythos is real and NTSK is the cheapest way to play SASE. The bear case is that this is a crowded, commoditizing market and you're paying near-ZS prices for an execution risk story. For now, the r/r says wait.
Stifel says SMTC beats again — and again. Buy reiterated, $188 PT, but the real call is the setup: July quarter lands above their $328M est, and the October guide clears too. The engine is AI infrastructure — 800G accelerating, 1.6T ramping — with Industrial (LoRa gen-4) and HighEnd Consumer adding sequential support on share gains and transient protection billings. The stock's up 182% in a year to $140, so you're not early. But the Street keeps reaching: UBS $225, Benchmark $230, TD Cowen $210, Needham $200 — all raised on data center growth into FY27. Stifel's $188 (10.8x CY27 EV/sales) is now the low end of that bull cluster. That's the whole r/r debate: AI tailwind is de-risked, the path to >60% gross margin via the Compal-backed cellular module divestiture is visible (52% today), but how much of the backlog is already in the price?
"October quarter revenue guidance may also exceed expectations, with upside potentially coming from Infrastructure based on an expanding AI infrastructure cycle."
Cantor says Overweight, PT to $550 from $510 — and the thesis is simple: estimates are still too low. Stock's at $389, up 70%+ in a year, and they still see a beat-and-raise coming for the October quarter.
October revenue of $4.25B+ and EPS of $3.75+ vs consensus $4.07B / $3.54. That'd be the 10th consecutive quarter of above-seasonal growth. AI-related revenue now ~19% of the mix, and the Empower Semiconductor acquisition closes the power-delivery gap for AI processors.
The setup is friendly. Analog peers guided 6-8% QoQ growth for the October quarter — ADI's normal seasonality is low-single-digit. Low bar, beat-and-raise, repeat.
"The analog earnings season has been challenging, with peers setting a bar for third-quarter growth of 6% to 8% compared to Analog Devices' typical seasonality of low single-digit percentage growth."
Gross margins at 73%+ (best-in-class, so limited upside — LTM GAAP gross margin is 64%, but nobody trades ADI on GAAP). Cantor sees CY27 EPS upside to $18 vs consensus $15.59, and CY28 closer to $20. At 28x that's the $550 target.
DA Davidson's the lower bogey at $498, also looking for strong Q3 FY26 prints — ATE and AI data centers doing the heavy lifting.
Earnings August 19. Options imply a 5.8% move. ADI has beaten the implied move in 6 of the last 8 prints. History says don't fade this one.
HC WAINWRIGHT REITERATED BUY/$12 PT WHILE THE STOCK SITS AT $2.94 (-38% YTD). That's a 4x gap — either the Street is blind or the backlog story is real. Revenue is horrid (-40% YoY to $5.0M) but this is an inflection play, not an earnings play.
The bull case lives in the balance sheet and the pipeline. Backlog grew 32% SEQUENTIALLY TO $18.4M, anchored by that $6M PRYMUS prime-power contract for a national package delivery company (deliveries 2H26). Gross margins jumped to 19.6% from 13.6% in Q1 — the e-Boost cost cuts are showing up. $10.7M cash, no debt, current ratio 4.11. They can survive while they wait for the converts.
The bear case is equally simple: revenues keep shrinking, net loss widened to $2.1M, and the e-Boost rental business — the original growth story — has gone quiet. At $2.94 you're paying for backlog conversion that hasn't shown up in the P&L yet. The Q2 top line beat expectations, but that's a low bar when you're down 40%.
Not worth a big allocation either way. Light coverage, single broker, wide PT dispersion. The PRYMUS deliveries in 2H26 are the swing factor — if those hit, the $12 PT has a pulse. If they slip, this is a $2 stock pretending to be a $12 one.
DB pulls the trigger on WDAY — downgrade to Hold from Buy, PT to $220 from $180. But this is a valuation call, not a fundamentals break. Shares ripped ~26% since June 1 vs. a ~4% AVERAGE DECLINE across SNOW, CRM, ADSK, ADBE, and INTU. Add the Silver Lake take-private chatter and the market is now pricing a strategic outcome — the r/r at these levels is simply more balanced.
The bull case intact: mission-critical HCM/FINS franchise, 76% gross margins, high switching costs, Aneel Bhusri back in the CEO seat with customer/employee morale reportedly buoyed. The bear case is the growth durability question and the AI overhang — adoption of agents still early innings, and the 6-month run is near +39% with RSI flashing overbought.
"Investors will continue to grapple with questions around the durability of growth and long-term impact of AI on enterprise applications."
Street's all over the place: BMO still OW, UBS Neutral at $220, MS down to UW at $145, CLSA initiating UW at $92. That spread tells you everything — this stock trades M&A optionality now, not just fundamentals. Smart move is to wait for the deal noise to fade and buy the pullback, not chase the headlines.