Good morning.
AI optics and memory own the tape. AAOI pre-announces 800G/1.6T demand 20-40% ABOVE SUPPLY THROUGH MID-2027 — that's scarcity, not a beat. DDOG crashed the most on record on a margin miss despite a forecast raise — SaaS bifurcation is live. WOLF +5% premarket on the LITEON 800VDC SiC partnership. GOOGL's $25B IG bond 6x oversubscribed — $150B+ raised in 2026, the sentiment toggle is now AI-Math-On/AI-Math-Off.
NET printed +36% y/y, DBNR 120%, but gross margins down 320bps y/y — they're rejecting the capex arms race outright. HUBS is the bear: AI budgets disrupt spend while their GTM pivot eats the near term. Unity's July growth >100% per agency checks. Asia trades the same scarcity regime — memory and optics supply chain stay the bid.
Three themes frame today. First, memory: "peaked" is a pod de-grossing event, not a cycle top — HBM is insatiable and SNDK's $15B buyback says management agrees. Second, optical scarcity favors the semi layer (DSP/TIA) over assemblers through mid-2027. Third, software is a barbell — TEAM/FIG accelerate on usage, DDOG/HUBS/EPAM bleed on disruption.
We'll hit up SNDK, WDC, and HUBS first, then get to the optical/network semis — SITM, VIAV, CLFD, POWI.
HubSpot just killed its own reacceleration story. Q2 beat on the headline — 17% cc vs. 16% guide — but the internals were brutal: NET NEW CUSTOMER ADDS OF ~7,000 VS. A 9,000-10,000 BOGEY, LOWEST SINCE 1Q20. NRR DOWN 1PT TO 102%. Then management cut FY26 CC GROWTH TO 16.1% FROM 16.6% and walked back the promise that net new ARR growth would exceed cc revenue growth in 2H26. That was the load-bearing wall of the entire reacceleration thesis. Gone.
The street revolted. Stifel downgraded to Hold ($200 from $275), Bernstein to Market Perform ($220 from $381), Wolfe to Peerperform, plus Piper and Oppenheimer turning neutral. The tape did the rest — morning prints at $250.21, Wolfe had the stock at $198.74 by 3pm. A 20% intraday haircut on a bellwether is a signal, not noise.
Active PTs now span $200 to $300. The old $275-$381 cluster is history. The median lands around $220 — that's BofA, Stifel, Bernstein and Piper all clustered in the 200-220 zone. Wolfe is the most bearish with a fair value range of $145-$250. The bulls still standing: Raymond James at $250, Morgan Stanley at $287, Needham at $300.
The consensus is "good company, bad tape." Nobody screams sell at $200 given the franchise quality — but the burden of proof has shifted. Bernstein frames it best: "not fundamentally broken," but no near-term catalyst. Management guided Q3 to 15% cc from 17% in Q2 — the deceleration still has momentum.
Bull case: The company is not broken. Raymond James points out 16% cc growth is still stronger than most front-office software vendors, backed by 84% gross margins and a 0.25 PEG. Needham keeps the Street-high $300 target and argues the weakness is macro, not share loss: both direct and partner channels equally impacted, with the two weakest cohorts being small SMBs and PE-backed upmarket customers. Those are cyclical segments, not structural HubSpot-specific challenges. Morgan Stanley stays Overweight at $287, framing the quarter as "delayed AI-led reacceleration" rather than a write-off.
Bear case: The GTM changes were supposed to fix this. Management spent Q1 rolling out outcome-based agent pricing, free trials, and sales force retraining — the expectation was higher revenue per customer and NRR expansion. Instead, NRR dropped to 102% and existing customers asked for downgrades. The most damaging line: Stifel admits its channel checks did NOT flag this demand shift. That means the deterioration is faster and broader than any pre-earnings model captured. And Wolfe adds the kicker — WEAKNESS CONTINUED INTO JULY. Q2 didn't end with a fade; it ended with a trend that hasn't stopped.
"This marks the second consecutive quarter of deteriorating trends and its confidence in a near-term path to improved growth or organizational changes to unlock free cash flow has declined." — Wolfe Research
"The business is not fundamentally broken but is impacted by a difficult macro environment, uncertainty regarding the timing of AI adoption and maturation of the SaaS CRM market." — Bernstein
"The firm said its checks did not signal the demand environment shift that materialized in the quarter." — Stifel (translation: nobody in the channel saw this coming — that's macro overlaying the name)
New in this print:
This is the cleanest "AI budget squeeze" print in SMB software. Token-maxing — unpredictable AI costs consuming IT budgets — hit HubSpot's mid-market base directly. The read-through is immediate to MONDAY, ASANA, SMARTSHEET, and anything selling into PE-backed software companies. If HubSpot's 7,000 net adds is the new industry baseline, those names are facing the same wall.
The agentic AI narrative takes a gut punch too. The street upgraded HUBS last September on the thesis that agent pricing + GTM changes would drive reacceleration. That thesis is now formally retracted. "Delayed AI reacceleration" becomes the new euphemism across the software tape — expect it in CRM, ADBE, and every other AI-agent narrative that fails to ship. The $200 valuation floor is real, but stepping in front of a 15%-guidance train is how PMs give back the year. Let the tape find its footing first.
Blowout quarter, fading forward optics. Stock -6.7% to ~$1,260 after the print. The market isn't asking "was the quarter good?" — it's asking "what does the next ASP print show?" The guide says lower.
REVENUE $8.97B, +51% QoQ, ~$500M over consensus. GM 84.6% vs 79.6% expected. EPS $39.25 vs $34.96. Clean sweep. (After a 469% YTD run, the bar was high — and they still cleared it.)
Mix tells the real story: two-thirds of the sequential growth came from pricing, one-third from volumes. ASPs +33% QoQ, bits +13%. This is a pricing cycle, not a unit cycle. Data Center revenue +103% QoQ to $2.98B on TLC eSSD strength, with BiCS8 QLC Stargate just starting to ramp. Edge +48% to $5.43B. Consumer -32% to $556M. AI gets the supply, consumers get priced out. Classic.
Management sees the NAND market >$300B in CY26, approaching $500B in CY27. If that's right, this cycle runs for years.
PT moves span $1,750-$3,000. Jefferies takes the axe — cutting 42% to $1,750 — while maintaining Buy. Bernstein holds at $3,000 Outperform. Evercore trims to $2,800. Lynx actually raises to $2,550. Wide dispersion on the same quarter. The split comes down to one question: are 80%+ gross margins structural or cyclical?
The LTA book says structural. Contract coverage now exceeds 50% of FY27 bits and two-thirds of FY28 bits. $94B of minimum contracted value at floor pricing, with $16.5B of financial guarantees behind it. Bernstein flags that actual penetration now exceeds their 60% LTA scenario — the bull case is increasingly contracted, not aspirational.
"This coverage pushes actual penetration past the 60% long-term agreement scenario previously modeled by Bernstein."
The bear case lives in the guide. FQ1 revenue midpoint implies ~18% sequential growth — strong, but ASP growth decelerates to ~9-10% QoQ from 33%. GM guidance of 83-85% is still elite, but the mix shift toward Data Center SSDs carries higher DRAM costs. Jefferies flags moderating pricing and near-term Edge inventory build headwinds. The marginal dollar gets harder to find.
Bull: Contracted revenue plus 80% gross margins plus AI storage demand equals a structural repricing of NAND. The $94B floor de-risks the downside, HBF adds option value, and the Aug 13 Analyst Day in NYC could be a real catalyst. This is the first NAND cycle backed by multi-year contracts, not spot market hope.
Bear: You're paying peak-cycle economics right as pricing peaks. ASP growth decelerating 33%→9% QoQ says the easy money is made. Consumer demand destruction is already visible in the -32% print. Jefferies cutting its PT by 42% into a beat is the Street telling you the next twelve months look harder than the last three. (The floor pricing helps, but floor isn't spot — and spot is what's turning.)
Bottom line: exceptional quarter, decelerating forward curve. Analyst Day on Aug 13 — HBF details, tech roadmap, long-term margin architecture — is the next real catalyst. Until then, digestion is the trade.
Stock rips 17% on a beat-and-raise and the two PT hikes land at... $145. Right on the tape. Cantor and UBS both stayed Neutral and both independently arrived at fair value where the stock already trades. That's not a coincidence — that's the cautious camp's ceiling. Meanwhile JPMorgan says $185 and Evercore says $175. The debate isn't whether Q2 was good. It's whether you pay ~145x earnings for a growth reacceleration that's finally getting an AI assist.
GMV +30% y/y ex-FX — the FIFTH consecutive quarter above 30% — against a 4-5pt tougher comp. Revenue +32.5% LTM. Q3 guide calls for revenue up low-30s and gross profit up mid-to-high 20s, both above Street. OpEx ran 34% of revenue vs 35-36% guidance. FCF margin ~18% vs mid-teens guide. Clean sweep.
The nuance: Subscription Solutions margins bled ~60bps q/q. UBS attributes it to cloud/AI infrastructure costs. So the AI build is a margin tax in the near term, even if it's the growth flywheel later.
The signal that matters — AI traffic is small, but it's converting better than traditional discovery. That's the first hard data point bulls can hang their hats on:
AI traffic, although small, is showing growth and beginning to yield gains on conversion against traditional discovery paths.
Bulls: 30%+ growth on tougher comps, GMV beat by 3% and profit by 10%, Sidekick adoption driving pipeline for several quarters, and JPM/Evercore's $185/$175 targets imply 20-30% upside. The AI conversion tidbit is the narrative kicker — it turns "AI story" into "AI numbers."
Bears: 145x P/E with a Neutral consensus target sitting at the current price. The margin pressure from AI infra costs is real and visible in the subscription margin degradation. The guide was better than Street, but low-30s revenue growth isn't new — it's the same number, just against a tougher comp. You're paying a hypergrowth multiple for a stock that's already at fair value per the two most recent analysts to touch it.
The r/r here is genuinely split. Both Neutral shops raised estimates (Cantor FY27 rev/EBIT +5%/+6%) but refused to chase the stock higher. That's the tell: the fundamental beat is real, the multiple has simply run ahead of the near-term math. At $144 you're paying for perfection — and the stock just delivered a near-perfect quarter. The question PMs need to answer: is the AI conversion data the start of a re-rating, or the excuse the momentum crowd uses to mark up a 145x multiple one more turn?
Two raises, and they're chunky ones. Canaccord to $60 from $46, Scotiabank to $61 from $44 — both maintaining Buy-equivalent ratings. That's a 30%+ target bump in one morning. The quarter did the talking.
Clean beat: adjusted EPS $0.48 vs $0.44 bogey, revenue $554.5M vs $549.3M. But the number that matters — ORGANIC NET NEW ARR +41% YOY. That's not a re-rate. That's a step-change. Both shops point to the same drivers: one large deal plus the logs business ripping. Management also raised parts of the FY guide, and Scotiabank thinks that de-risks the Q2-Q4 ramp.
The AI story is doing real work here, not slideware work. Telemetry from AI workloads feeds the logs business, AI observability is an incremental TAM, and DT is monetizing its own AI/agents on top. Scotiabank flags a European energy company that upped spend specifically to observe AI/LLM workloads. That's a named proof point, not a hypothetical.
Bull: 41% net new ARR growth is the acceleration the street wanted. Customer checks show DT remains the go-to for large, complex hybrid environments — exactly where AI observability budgets land first. And at ~20x CY27 EBITDA with growth reaccelerating, the multiple hasn't run away. Not yet.
"Dynatrace remains the preferred vendor for large, complex, hybrid enterprise environments."
Bear: You're buying after a 51% six-month rip. The stock's at $50.86, essentially kissing the $53.28 high. Analyst targets at $60-61 imply ~18% more upside — fine, but not life-changing, and the easy money's been made. Also: 47x current EV/EBITDA. The CY27 math works. The path to CY27 is where it gets bumpy.
Worth a beat: DT repurchased $275M in the quarter. Aggressive for a company growing at this clip. And Jim Benson retires at year-end after a four-year tenure that included the ARR-disclosure pivot and this capital returns program. Canaccord calls the departure "limited risk." Fair — a CFO who cleans up disclosure, transitions the metrics, then hands off post-acceleration is usually a sign the hard part's done. But anytime the CFO walks, PMs should at least ask if there's a second shoe. Here, doesn't look like it.
Valuation check: Canaccord lands at $60 on ~6x CY27 sales / ~25x CY27 FCF. Scotiabank's at $61 on ~20x CY27 EBITDA. Both see double-digit upside. Neither sounds finished with the target. The stock's running hot, but the fundamental rate-of-change is finally matching the multiple. That's the kind of alignment that keeps momentum going.
MSI dunked on Q2 — EPS $4.41 vs $3.84 expected, revenue $3.13B ahead of bogey, guided up, and the record backlog kills the H2 ramp narrative that had bears camped out in the name. UBS goes to $520, BofA to $530, both Buy. Two houses, one collective thesis: core LMR is fine, and the growth stack is the real story.
REVENUE GREW 13.3% VS 8.6% EXPECTED. That's not noise. And it came with what UBS flagged as elevated short interest heading into the print — so some of this move is squeeze mechanics, but the fundamental hand is strong enough to hold.
"The slower growth does not indicate a material change in the business or a demand slowdown" — UBS, on LMR normalizing after supply-constrained years pulled demand forward.
UBS's model moves meaningfully: FY26 to $17.62 from $16.89, FY27 to $18.86 from $18.41. Silvus raised to $850M, Video +12%, Command +14%. The growth businesses keep compounding while LMR resets from an unsustainable catch-up pace — that's the bull case, and it's intact. At 35x it's not cheap, but after a quarter like this, the market is paying for compounding, not for hope.
Needham trims VIAV to $60 from $68, keeps Buy. This is a valuation trim, not a thesis break. The 287% run over twelve months stretched the multiple, and the PT reset just acknowledges that. Stock sits at $38.90 — still ~35% BELOW THE NEW PT. That gap tells you the sell-side isn't backing away from the story. The after-hours fade tells you PMs are taking chips off the table into strength.
The quarter was clean. FQ4 REVENUE $443.1M (2% ahead), EPS $0.34 vs $0.30 consensus. FQ1 guide better: 5% above consensus on revenue, $0.13 above on EPS. The engine is the AI-optic complex — DATA CENTER ECOSYSTEM DEMAND +>80% Y/Y in Network & Service Enablement, and guidance implies organic NSE plus Anti-Counterfeiting growth of ~44% Y/Y. Management says CPO testing demand has orders in hand. Optical circuit switch is broadening beyond a single customer.
So why did the stock fade on a beat-and-raise? 287% in a year. The set-up was crowded. A clean print was already in the price. Needham actually raised FY27 estimates and introduced FY28 — momentum intact. But at $39 the r/r for new money is mediocre. The bull case: real franchise in optical test and measurement, executing on both growth and operating leverage. The bear case: you're late to a crowded trade, paying for AI capex follow-through that's largely consensus.
"Viavi Solutions is executing well on both revenue growth and operating leverage, with strength in data center providing a positive indicator for optical coverage."
UBS CUTS GFS TO $55 FROM $61 — but the real headline is the margin rebuke. Street has CY28 EPS at $3.56; UBS sits at $3.13. That's a 12% gap, and UBS says the Street is over-modeling margins, not revenue.
The mobile guide is the culprit: FY handset outlook cut from high-single-digit decline to LOW-DOUBLE-DIGIT, which implies the Sep guide (midpoint ~2% below UBS est.) is just the warm-up for a harder step-down in Q4. Meanwhile the good stuff — comms infra/data center +62% YoY, FY growth 50-60% — is already in the numbers and priced in at 57.5x trailing P/E.
"The Street appears to be over-modeling margins."
Q2 itself was a beat ($0.46 vs $0.43; $1.79B vs $1.76B) with GM 29.9% and utilization high-80s. The problem is 2027/2028, not the quarter. Target lands near spot ($49.46) so the r/r isn't terrible — but there's no reason to step in front of a handset guide-down with a 57x multiple and zero margin flexibility.
Needham cuts CLFD to $40 from $52, keeps Buy — but the guide is the real story. FQ3 was fine (EPS $0.22 vs $0.19 est, revenue $43.9M roughly in line), yet FQ4 guidance lands 24% BELOW consensus with a $0.35 EPS shortfall. Management blames BEAD delays, fiber availability, and a $4.6M order cancellation from a tier 3 rural telco that also bruised gross margin in the quarter.
The offset: a $22M data center order for a hyperscaler project, shipping FQ1'27. That's the bull case — rural fiber is weak now, but data center diversification opens a second leg. The bear case: at ~148x P/E, a Buy-rated name still gets punished for guide-downs because the multiple leaves zero room for error. Needham trimmed FY26/FY27 numbers, so this isn't just a PT tweak — the growth re-rating out a year.
"The stock currently trades at a P/E ratio of 147.63, reflecting premium valuation expectations that make the guidance miss particularly significant for investors."
Segment split tells the rotation story: Community Broadband +11%, Regionals +57%, Cable -11%. The fiber patchwork is uneven, but the data center order is the first concrete proof the product portfolio travels beyond telco. Watching for follow-on orders — one $22M hyperscaler win isn't a pipeline.
William Blair walks away — cuts EPAM to Market Perform from Outperform. The stock's already down 41% in six months to $95.67 (13.85x P/E), and the downgrade just validates the tape. Q2 beat on every line: REVENUE $1.42B (+4.5% Y/Y), NON-GAAP EPS $3.38 vs. $3.14 EST. Nobody gets paid for beating the quarter when you cut the forward.
The problem is North America. WB calls out sustained pressure in EPAM's largest market with limited growth through 2026. This is a beat-and-lower — the classic tell that demand isn't inflecting. At sub-14x with a beat-down chart, the value bid will show up eventually, but the sell-side is saying wait for trajectory to stabilize before catching this knife.
"EPAM's second-quarter results exceeded our estimates for revenue, gross margin, operating margin, and EPS. However, lowered revenue guidance and commentary point to sustained pressure in North America and limited growth through 2026." — William Blair's Maggie Nolan
UBS cuts ORCL to $245 from $285 but KEEPS THE BUY — that's the tell. The stock's already DOWN 58% FROM THE HIGH at $144.39, including a 27% post-Q4 slide in May. Keirstead's message: the market is de-risking for him.
The concern list is every AI bear's greatest hits: capex spiraling, ROIC a question mark, credit markets tightening for new AI data center debt, a delayed New Mexico site, and OpenAI sitting there as massive customer concentration. (That last one keeps PMs up at night.)
UBS's counter: it's all in the price. Some issues — New Mexico especially — are overstated. PEG of 0.71 makes the growth-adjusted math work at these levels.
"Material risk is already reflected in the stock price."
Guggenheim's still out at $400 after meeting with execs, so the street's not aligned. The UBS cut reads as a de-risking move, not a thesis break. This is a "hate the narrative, love the math" setup at $144.
Raymond James throws in the towel — cuts OTEX to Market Perform from Outperform on the margin story, not the revenue story. FY27 adjusted EBITDA guide of 32-33% is a full 300-400bps below the 36% bogey, and the Q1 midpoint of $400M vs $462M consensus is a 13% miss. This isn't a speed bump; it's a structural reset from the ~40% margins and high-70s FCF conversion this name printed in 2021-22.
"Open Text guides conservatively and resets expectations lower each quarter."
That's the tell. When a management team lowballs and still misses, "conservative guidance" is just a euphemism for deterioration. Divestitures on hold due to macro — so the sum-of-the-parts catalyst is dead for now.
The bull case is alive but unexciting: 76% gross margin, 13% FCF yield, shares +3% on the print. Cheap gets cheaper, but value traps on Bay Street end careers. No r/r at Market Perform. Move along.
Needham trims PT to $80 from $90, keeps Buy. That's pure multiple compression, not thesis breakage — the stock is up 75% YTD, trades at 208x trailing earnings, and the new target is just 28x FY27 EPS of $2.87. Fair enough. The Buy is the signal here.
Q2 was a clean beat: $0.37 non-GAAP vs $0.32 est, revenue $118.9M (+10% Q/Q, +3% Y/Y). Q3 guide at $126M midpoint = +6% sequential. Needham models roughly $0.03 of EPS upside on the remaining guidance — not massive, but positive rate of change.
The real bet is the pivot. Management is pushing design wins into automotive, industrial, and AI data center — secular markets that should re-rate the multiple if the pipeline converts. That's the steelman for the Buy at 208x: you're paying up now for the earnings growth maturing within the design pipeline. If the pivot stalls, the multiple compression story gets worse before it gets better.
"Management is positioning the company to deliver meaningful earnings growth as the strategy matures and the design pipeline materializes."
Demand improving steadily, recovery intact, execution solid. At $61.95, this is a namesake stock — you're buying the optionality, not the current earnings. Needham keeps conviction, just recalibrates the math.
Raymond James just threw a $100 PT in the dumpster and printed a $175 target — a 75% hike. That's not a tweak, that's a thesis change. And it lands after the Q2 print showed exactly what we like: gross profit +18% YoY, operating earnings +31% YoY. This is operating leverage from a solutions business, not just a reseller top-line story.
The quarter itself was a blowout — adjusted EPS $3.86 vs $2.93 consensus, revenue $2.4B vs $2.17B. That's an 32% EPS beat and 10% revenue beat. Full-year guide raised. Yet the stock was a spectator in premarket. 72% YTD already, so maybe the tape's front-running the good news. Still, with the PT at $175 vs the current $145.41 print, RJ thinks there's another 20% left.
The bull case is simple: IT services and AI-adjacent hardware demand is accelerating, and Insight is one of the few players showing margin expansion while growing double-digits. RJ sees three key takeaways from the call aligning with its recent upgrade — likely around sustainability of the gross profit trajectory, operating discipline, and backlog visibility. The bear case? Valuation. The stock got flagged on some "Most Overvalued" screens — a classic bull-market objection that means little until the multiple actually gets in the way of the story.
I'd rather own the operator with 31% op income growth than argue about a forward multiple on a name with this kind of upside surprise momentum. PT hike of this magnitude says the analyst community is still catching up to the fundamentals.
Needham lifted PT to $19 from $18.50, keeps Buy — but the beat is the headline. MRAM just printed record Q2 revenue of $18.7M (+42% YoY) vs $17.3M Street, and non-GAAP EPS of $0.11 vs a consensus LOSS OF -$0.01. The stock already ripped 172% over the past year (74% YTD), now at $16.14. This print validates the move.
Non-product revenue spiked to $3.4M, +343% QoQ — that's initial recognition under the $40M Toggle MRAM subcontract. Product revenue grew a modest 8.6% QoQ. The beat is real, but the mix matters. Subcontract revenue is lumpy, not recurring. Strip it out and this is a steady, not explosive, product story.
Needham's new PT is based on 5x EV/Sales on CY2028 estimates — that's a long look for a $378M market cap name. The multiple only works if the CXL narrative materializes.
CXL keeps advancing: contract signed with a controller IP partner to build CXL controller IP for MRAM, plus an AMD collaboration on an UltraScale+ FPGA platform. That's the big optionality.
Astro Digital selected PERSYST 64Mb STT-MRAM as fail-safe boot memory for its Raven bus — the FIRST GEO satellite design win. Space-grade MRAM is a reference, not a revenue driver yet. But it builds the credibility stack.
The firm cited the company's quarterly beat and raised guidance above Street expectations.
Analysts see 27% revenue growth for fiscal 2026. Not cheap at these levels, but for a small-cap with CXL optionality, two live partnerships, and a space design win — the r/r still favors longs. Watch the non-product revenue cadence next quarter. That's the swing factor.
Raymond James says the GTA VI date is effectively locked. The Netflix premiere — six-hour exclusive window before Rockstar's own channel gets the footage — isn't just marketing noise. A cross-platform, fixed-date, negotiated distribution deal like that takes weeks of coordination. That's confirmation of the November 19 launch, and it pushes the delay bogey to "minimal."
The firm repeats Strong Buy, $300. Timing lines up: management telegraphed on the Q4-FY26 call that summer would bring the GTA VI marketing rollout. The Netflix hook has history too — they already carried GTA: The Trilogy on mobile in late 2023. And Netflix is a real distribution pawn here, not a dusty content library: live programming drove six of the top-10 sign-up days in the last five years, with engagement hours up 2% YoY in H1 2026. (16% revenue growth, 49% gross margin — the platform has scale and it shows.)
The only real bear case was a slip. RJ just kneecapped that.
"A crossplatform premiere with a fixed date, a negotiated exclusivity window, and an outside distribution partner is not assembled hastily."
The November 19 date is the macro call. Now it's just the countdown.
The margin scare is a distraction. Axon printed Q2 revenue 340bps ahead of Needham's model and the bookings engine keeps compounding — quarterly bookings +16.1% AGAINST a brutal year-ago comp. Needham reiterates Buy/$750 and says the Street is still underestimating the demand curve.
The number that matters: SOFTWARE REVENUE EX-EVIDENCE.COM +70%. That's the tell. Needham reads it as an acceleration signal for Axon Cloud into 2027 — the mix shift is real and it's compounding. ARR +38.5%, revenue +35% (10th straight quarter of 30%+ growth), future contracted bookings +41.1%. Cost pressures are a 2026 issue; the revenue trajectory is not.
Needham models mid-to-high 30% growth in five-year normalized bookings for 2026 and openly says the estimate is low.
"The firm believes its updated bookings estimates will prove conservative."
That's the whole bull case. $750 PT with bookings estimates that are "conservative" by the analyst's own admission — r/r skews positive here once the tape gets through the margin hand-wringing.
Needham bumps SITM to $900 from $850 and keeps Buy — 55x 2028 non-GAAP EPS on a story compounding faster than the model can capture. Fair, after the beat-and-raise: REVENUE +127% Y/Y TO $157.4M, EPS $2.34 vs $1.95 bogey, GM EXPANDED TO 67.1% FROM 58.2%. Every business unit grew >50% Y/Y.
The CED engine is the whole thesis. 1.6T REVENUE GROWING 100% Y/Y NEXT YEAR, and management carries 12-18 months of visibility on strong backlog. That's not hope — that's line of sight.
Renesas TPD (closed July 1) adds ~$85M/quarter near-term, supply-constrained with management bringing on additional supply partners. Incremental revenue stacked on top of organic acceleration.
"Growth is expected to sustain into fiscal year 2027 as management cited 12-18 month visibility and strong backlog across the business."
The bear case: $900 implies perfection — a $14.3B cap on a $157M quarter. The bull retort: operating margin at 34% vs 10% a year ago, non-GAAP net income +400%, and a TPD division yet to fully ramp. The multiple is rich. The rate of change is richer.
UBS trims PT to $525 from $560 — but the real tell is the multiple cut to 9x while CY28 EPS goes UP to $56.76. That's the entire bull/bear fight in one print: earnings power marching higher, multiple compressing anyway. Stock sits at $519 after a 606% YEAR-OVER-YEAR RUN, so the market already prices in something structural.
Firm calls the quarter solid, guidance merely "slightly better" on revenue, with enterprise/branded shipments lagging Seagate. UBS buys management's mix explanation and sees it as transient — HAMR ramp starts soon, and they see a STRONG LIKELIHOOD GROSS MARGIN PUSHES ABOVE 75% within eight quarters (current LTM: 45%). Management knows exactly what it ships for the next 4-5 quarters, so the beat-and-raise machine stays intact.
But the divergence is the story: Baird lifts to $630 on the drive tech (40TB ePMR shipping now, 44TB HAMR next); Summit goes to Hold on HAMR transition risk. Same catalyst, two directions. The bull case says structural change in storage. The bear case says you're paying 9x for a cyclical peak that's fully credited.
"This is the same cyclical story that results in the multiple compressing over time."
Watch the TDK external head procurement and whether Toshiba actually gets its act together — those are the two cracks in the structural-change narrative.