Good morning.
Futures are flat to slightly positive, QQQ fractionally in the green, SOX +1.8% — not bad for a morning after the Iran ceasefire trade and a Gopinath bond market warning. Dell's +31% AH print is dominating the tape, NVDA holding +1.5% near highs, memory names (MU, AMAT, LRCX) continuing their run. Taiwan weighted and Korea Kospi AI supply chain index both hitting fresh highs for the year — these are the reads that matter, not the headline indices.
Dell FY guide to $167B, AI servers $60B — that's the physical confirmation the AI capex cycle just got pulled forward another quarter. Short interest on NVDA still ~2.1%, tight float means any dip gets absorbed by algo buying. The bear case keeps dating the top; the cycle keeps extending it.
Asia: Taiwan revised GDP up on AI export surge running at 50-year fastest clip, China defending CL at 7.19 with FX Reserves draw (down $150B over four months), Japan just spent record $73.6B defending the yen — first intervention since 2024. Japan's in a bind: energy importer getting squeezed by FX strength while defense spending goes to record highs. That's not a TMT story directly but watch the yen carry impact on semiconductor equipment names (AMAT, LRCX, ASML make ~30-40% of revenues in JPY terms). US-Iran ceasefire chatter pushed Brent down ~5%, energy stocks up, AI stocks also up — siloed pricing confirmed, these are different money pools and they're not cross-exiting.
Three themes structuring today:
One — Supply chain institutionalization is the meta-trade. Anthropic took $65B from SK Hynix, Samsung, and Micron (equity stakes, not just LTAs), closing the circle on AI lab memory supply security. This is the move that re-rates MU structurally long-term even as it creates short-term allocation risk for non-Anthropic customers. Snowflake's Cortex Code hit fastest growth in company history, agentic workflows eating data platform consumption — that's the downstream signal that data infrastructure demand isn't a cycle, it's a platform shift.
Two — AI now prices like infrastructure, not software. Anthropic raised at $965B with 40% debt in the structure, SpaceX IPO target dropped from $2T+ to $1.8T (less inflation, more honesty). Both moves are the market accepting utility-style depreciation for AI compute assets. If ECB actually hikes in June, the tension between "AI as long-duration growth" and "rates going up" becomes the macro trade of the summer.
Three — Power is emerging as the real bottleneck nobody's pricing yet. DataBridge acquiring ArcLight (largest US data center owner + largest private power operator) and MHI expanding gas turbine production 50% by FY31 are the physical signals. Silicon supply gets solved, power delivery doesn't — 1 GW of data center needs ~1.4 GW of power capacity, $3-4K per kW to build, and there's a 3-7 year grid queue standing in the way. This is the trade that develops over next 18 months, not tomorrow.
We'll hit up Dell first (squeeze behavior + backlog read-through), then get to the memory chain (MU, AMAT, LRCX) and the AI lab infrastructure names (ARM, SNPS)... then circle back to the compute layer where NVDA and AMD are clearing the tape between earnings.
Bears: RPO growth decelerated to 38% from 42% QoQ — not alarming, but worth watching. Free cash flow of $233M missed $282M consensus — that divergence between earnings beat and cash generation needs monitoring. Snowflake remains unprofitable on a GAAP basis (EPS -$3.95), and while analysts project $1.94 EPS in FY2027, that's a big swing to price in. Databricks is still a formidable competitor — Scotiabank's channel checks said no current threat from OpenAI or Anthropic, but the data platform wars are real. And at $175-$240, SNOW is pricing in a lot of execution perfection. A 30% stock decline over the past six months (per JMP) means positioning is cleaner, but it also means the bar was lower. Now the bar is high.
"We think this momentum is durable and will continue to grow, underpinning our belief that SNOW is one of the best-positioned names in our coverage to benefit from AI adoption" — Piper Sandler, $295.
"The ability of the newer AI and ML products to drive increased consumption across the platform should enable Snowflake to sustain at least 25% or more annual product-revenue growth in the coming years" — Stifel, $300.
"The increased use of coding tools is facilitating migration from legacy data platforms to modern platforms like Snowflake" — Goldman Sachs, $278.
For peer read-through: if SNOW is winning on governance + AI, that's a negative read on pure-play ML platforms that lack the enterprise data control plane. Watch whether DBT, Confluent, or Databricks see similar AI-driven consumption acceleration in their prints. SNOW at 13x EV/2027E sales (Cantor's new multiple) vs Databricks at 20x+ — there's a valuation gap that's narrowing if the growth rates converge.
The bull thesis is straightforward: Marvell owns the connective tissue of AI clusters. Optical interconnect, switching, Serdes, I/O — all of it. KeyBanc just raised its PT to $260 citing optical demand that pushed FY27 revenue growth guidance to 70% from 50%. Needham bumped to $270 and called out scale-up optics now expected to add $300M in FY28 (doubled from prior expectations). Custom silicon? Management guided to $10B in FY29 from $4B currently. That's the number hanging over this trade — a 2.5x revenue ramp on a single segment in two years.
"We're encouraged by these results and as such are raising ests and our PT to $260." — KeyBanc
Data center is the engine. Barclays and Benchmark both at $275 flagged the segment growing 50% in FY27 and 55% in FY28. Interconnect specifically tracking to 70% growth. XPU doubling in FY28. This isn't a one-product story anymore.
The bear case exists, and it's honest: the stock trades at 65x earnings and 29x CY28 consensus ($7.65). Cantor Fitzgerald called it out directly — NVDA and AVGO sit at 14x and 17x CY27 respectively, and MRVL is asking for a premium that requires flawless execution on custom silicon ramp and optical expansion. Goldman Sachs bumped its PT to $180 (still Neutral) and said it best: lower visibility into custom silicon customer expansion versus peers keeps them cautious. TD Cowen held at $200 (Hold) citing networking as the driver but flagged that XPU visibility remains murky for investors. The stock has more than doubled in two months — at some point the bar for "priced in" gets real.
"The central institutional debate is how much of Marvell's $16.5 billion fiscal 2028 revenue framework and fiscal 2029 custom-silicon target is already embedded in the share price." — Benchmark
What's new versus what was known: the FY27 guide moved to 40% growth from 30%. FY28 to 45% from 40%. The optical segment revision (doubled FY28 outlook) is fresh. XPU doubling in FY28 is a step up from "doubling." The $10B custom target for FY29 is new framing. Everything else — data center dominance, AI infrastructure leverage, 42% LTM growth — was in the price.
Read-through: Marvell's optics upgrade is a thumbprint of broader AI cluster buildout. If MRVL is doubling its optical outlook, that points to density increasing at hyperscalers — which has legs into INTC's data center, AVGO's networking, and AMD's server CPU story. Custom silicon at $10B in FY29 also validates the ASIC wave that's been building. This isn't just a Marvell trade — it's a bet on the pace of AI infrastructure spending, and the street is saying the pace is accelerating.
PTs span $175 (DA Davidson) to $315 (JMP Securities) — a $140 range that tells you everything about the conviction gap. Most names cluster above $240 (Stifel $250, Cantor $250, TD Cowen $240, Truist $280, KeyBanc $290), but the herd has been dragged down by multiple compression cuts rather than fundamental downgrades. BMO ($215), Piper ($215), and DA ($175) are the outliers on the bearish end — not because the quarter was bad, but because the second-half acceleration thesis hasn't been earned yet.
The setup is more compelling than the price action suggests. Agentforce ARR crossed $1B ($1.2B actually), growing 205% YoY — that's not a pilot program, that's a product with traction. Slack drove 50% of $1M+ deals. cRPO was a non-event miss ($33.6B vs $33.7B). Gross margins held at 77.68%. The stock trades ~11x EV/FCF with a 10% FCF yield and a PEG of 0.86 — that's cheap for a company growing revenue double-digits and deploying $27B in buybacks.
"Salesforce opened FY27 with a clean headline beat that defends but does not advance the bull case. With shares already deeply derated, we believe the risk/reward is asymmetric to the upside." — Freedom Broker
Truist calls Headless 360 a long-term game changer — enabling Salesforce agents and workflows to operate inside external apps and AI-native interfaces is a wallet-share lever that doesn't show up in current numbers but could re-rate the story in 18-24 months. Stifel notes Agentic Work Units and Token Consumption growing >200% — usage metrics that precede monetization.
"Wait and see" is polite. "Stalled" is more honest. Q1 revenue guidance for Q2 implies 7-8% constant currency organic growth — that's deceleration, not acceleration. Tableau underperformed. cRPO missed for a second consecutive quarter. The Informatica contribution (~4.5pp of the beat) is acquisition math, not organic momentum. And the AI monetization model? Still being built.
"The earnings beat was largely attributable to strategic-investment gains and the $25B accelerated share repurchase program." — Freedom Broker
DA Davidson cut to $175 specifically because the updated guidance did not provide support for a more positive outlook or validate expectations for top-line reacceleration in H2. That's the crux: management's credibility on the H2 call is now on trial. Cantor notes investors are "left having to trust management that second-half acceleration is around the corner" — which is a fragile foundation for a re-rating.
New: Headless 360 as a discrete product/offering (Truist called it a "potential long-term game changer"). Agentforce ARR crossed $1B threshold. Informatica contribution was quantified at ~4.5pp of growth. Agentic Work Units and Token Consumption metrics now disclosed with >200% QoQ growth.
Already known: FY27 Rule-of-50 framework, Slack as a deal driver, the buyback program, Agentforce as the strategic centerpiece. The script hasn't changed — the numbers just haven't caught up to the narrative yet.
CRM's stall is a microcosm of the broader SaaS/Enterprise AI trade: great consumption metrics, murky monetization path, and a market that's done waiting for "it's coming." The multiple compression (KeyBanc explicitly cited software multiple pressure as the reason for their cut) reflects a sector-level reset — not a CRM-specific story.
If Agentforce tokens, work units, and Headless 360 drive H2 acceleration, this re-rates. If they don't, you're left owning 11x FCF with a 10% yield — fine, but not a catalyst. For PMs running relative value in TMT: the risk/reward skew is asymmetric to the upside if you believe the H2 narrative. The market is pricing in skepticism. Conviction is everything here.
Street sentiment flipped constructive after fiscal Q3 results showed Nutanix outperforming across the board—EPS $0.47 vs $0.36 expected, revenue $703M up 2.4% to consensus. But the afterhours dip tells you something: investors are nervous about the conversion runway, not the demand signal. Bookings surged past 20% growth, yet revenue conversion keeps getting pushed out by server supply chain friction. That's the trade-off framing PMs need to internalize right now—
Bull case: Management raised full-year revenue and FCF guidance for the first time since cutting it earlier in fiscal 2026. 87% gross margin is elite, NRR holding at 106% (decelerated 100bps QoQ, but flattish in context), and they added 700 net new logos versus 620 a year ago. Contract duration extended to 3.4 years—mix shifting to larger, longer-dated deals across land, expand, and renewals. Bookings acceleration above 20% signals competitive wins holding. BofA called it out: "Nutanix delivered a solid quarter with robust demand. Total contract value bookings expectations for the fiscal year are higher relative to the last earnings call."
Bear case: Pipeline conversion risk remains the coiled spring. Server lead times are normalizing but pricing stays elevated, and management guided fiscal 2027 renewal base could slow. Fourth-quarter revenue guidance came in 1% below Street—Middle East headwinds and extended booking-to-revenue lags cited directly. Piper Sandler cut PT to $60 from $63 despite keeping Overweight, flagging fiscal 2027 renewal softness as a real concern. 14x FCF multiple looks cheap on paper, but only if you trust the conversion timing.
What's new today vs. what's already known: The beat and guidance raise were incremental positives, but the real signal is bookings accelerating above 20% for the first time in recent quarters. That demand trajectory was the unknown variable. Supply chain constraints were known—which is why the gross margin durability (87%) and FCF generation (37% growth) matter more now. NRR decelerating to 106% from 107% is worth watching; not a cliff, but the direction matters for a company guiding on subscription ARR.
"The company raised its fiscal 2026 revenue and free cash flow guidance after cutting both metrics in the first and second quarters." — Goldman Sachs
"We remain Buy-rated and raise our Price Target to $60 (from $55)." — Needham
"BofA Securities raised its price target for Nutanix to $65 from $60, maintaining a Buy rating, citing strong demand and improved contract value bookings." — BofA
Thematic read-through: Nutanix is the hyperconvergence play antibiotics—legacy VMware migrations are still the majority of new logos, OEM partnerships (Dell, Cisco, AMD) are expanding the total addressable market, and the NC2 product is gaining traction in hybrid environments. This feeds into the broader "enterprise infrastructure modernization" narrative where HCI beats on simplicity and TCO against traditional virtualization stacks.
Positioning: Book is heavy in client facing mixed positioning—some PMs are building positions on the dip, others waiting for Q4 guide to print clean. Consensus says the selloff is overdone; supply chain noise is temporary, bookings momentum is durable. But the fiscal 2027 renewal risk Piper flagged isn't sitting well with anyone who models cash flows out two years. AT $47, you're looking at 20-35% upside to PT range—decent risk/reward IF conversion normalizes. If not, this stays a range-bound operation until server lead times clearly inflect.
Bottom line: NTNX is a high-conviction call on execution timing. Demand is there. Margins are strong. The question is whether the bookings-to-revenue conversion window closes before investors lose patience again.
Street view in one sentence: Five firms touched SNPS, with PTs clustering $450-$600 — Stifel, KeyBanc, and Rosenblatt at the high end ($575-600, Buys), Morgan Stanley dragged up to $525 (Equalweight), and Piper Sandler sitting at $450 (Neutral). The consensus leans constructive, but the spread tells you something: the bull case is priced in at current levels, and the remaining upside requires you to buy the IP recovery narrative.
Elliott Investment Management getting a board seat (Jesse Cohn, effective June 1) is the subtext nobody's writing about explicitly. That's activist capital signaling comfort with the status quo — or at minimum, a thumbs-up on the Ansys integration path. Combined with the September 30 investor day, there's a narrative hook coming. Stifel cited "accelerating synergies from the Ansys acquisition" as a key driver for their $600 target. That's the bet.
"Nothing Can Stop Me, IP Will Be Sequentially Up; Synopsys reported better F2Q results, despite the Ansys accounting benefit. Importantly, the Company noted IP bottomed in F1Q and expects sequential q/q growth throughout the year driven by existing pipeline and engagements around a new monetization model, which we find encouraging." — KeyBanc
That's the bull thesis in one quote. IP bottomed. Sequential growth incoming. New monetization model in FY2027. KeyBanc models IP exiting FY2026 at mid-teens growth rates. Rosenblatt sees IP at 12% of revenue with recovery already underway — up from a trough in Q1. The structural transition to a royalty/cost-of-technology-aligned model next fiscal year is the long-game call, and the street is giving management the benefit of the doubt.
Morgan Stanley's Lee Simpson is the honest skeptic — not calling the stock a sell, just saying the medium-term case rests on three things: (i) co-design offering, (ii) silicon-to-systems strategy, (iii) turning joint Synopsys-Ansys products into repeatable, large-scale customer workflows. That's a 12-18 month execution story, not a Q2 beat story.
"While 2Q26 was in line, we think the medium-term case for SNPS rests on (i) the co-design offering, (ii) the silicon to systems strategy and (iii) turning joint Synopsys-Ansys products into repeatable, large-scale customer workflows." — Morgan Stanley
Piper Sandler frames it plainly: "recent momentum in shares may have come in advance of a truly validating recovery in fundamentals." They're Neutral at $450, raised from $430, and their point is the stock ran 33% since end of March on what is still a work in progress. The after-hours decline despite the beat suggests the market agrees — this was a "sell the news" moment for short-termers, not a capitulation.
Bull vs. Bear:
Peer/thematic read-through: SNPS is the proxy for AI-driven chip design complexity. As hyperscalers push custom silicon, EDA vendors with full-stack offerings (SNPS + Ansys) capture钱包. The co-design trend — where chip design and system-level simulation converge — benefits SNPS more than any other pure-play EDA name. CDNS and MRVL don't have the same Ansys adjacency. If the September investor day delivers on synergies and FY2027 model clarity, the $600 PTs look cheap. If not, the 82x multiple compresses.
Verdict: The beat is fine. The guidance is better. The IP inflection is the swing factor. At 82x P/E, you're paying for the story — which means execution has to be flawless. Hold or add on dips. Don't chase at current levels unless you're sizing for the investor day catalyst.
"The firm maintained high conviction over the next 12 months, stating that core underlying demand remains strong. Cantor Fitzgerald characterized the short-term noise as a buying opportunity for longer-term investors."
"The firm cited multiple contraction as the reason for the reduced price target despite strong underlying business performance."
What's new vs known: The acceleration narrative held — that's the most important data point and it held. Raised guidance is incremental positive. AI lab win is new and hadn't been priced. But the margin deterioration is worse than the street was positioned for, and the Decisioning Studio miss suggests the AI product rollout is supply-constrained, not demand-constrained — that's a nuance that wasn't in prior quarter commentary. Professional services mix shift was a known accounting change but the magnitude surprised.
Read-through: Customer engagement platforms (Braze, possibly Amplexd, Iterable plays) are in a "prove the AI moat" phase. The stock getting hit on margin despite beating revenue is a classic SaaS valuation tension — markets are currently rewarding margin expansion more than growth acceleration. If BRZE can't convert 30% revenue growth into operating leverage, the 2.0x revenue multiple is fair, not cheap. Watch whether FY27 guidance includes any margin inflection language when full-year numbers print. No position change recommended — but this is a name to revisit on the next pullback below $22.
The investor day was the catalyst. Management laid out AI-driven vertical integration across data center networking, mission-critical infrastructure, and A&D. Reaffirmed 13-15% adjusted operating margin (old target was 11-13%) — that's the key call. Stifel called the 15-20% CAGR target "conservative," which tells you they're thinking the numbers can run hotter.
The A&D pipeline is where it gets interesting. $7.2 billion. Golden Dome embedded, 30+ munitions programs, +$500M capacity coming online by 2028. They've also secured a fresh $1B revolver and hinted at European A&D M&A. That's not a small-cap PCB story anymore — that's a defense contractor ecosystem play.
THE TENSION: Stock pulled back modestly after the investor day despite the positive guidance. It's sitting at $188-190, just below the 52-week high of $200.68. Truist called the messaging "less than expected" but framed it as a buying opportunity. InvestingPro's fair value model says overvalued. Investors are asking the obvious question: how much juice is left?
Bull steel: 536% gain in a year, but the PEG is 0.7 — cheap relative to the growth trajectory. A&D pipeline hasn't even started converting meaningfully. M&A optionality and margin expansion are still in front of the story.
Bear steel: Momentum names get whacked when expectations reset. "Less than expected" at a $4B revenue company with this multiple is the kind of thing that snaps your r/r real quick. Fair value concerns aren't sell-side propaganda.
The quote from Stifel captures the bull case:
"TTM Technologies' PEG ratio of 0.7 suggests attractive valuation relative to its growth prospects."
If you're long, you hold. The AI vertical integration story is real, the A&D pipeline is massive, and the margin step-up changes the math. If you're flat, the pullback after investor day is the entry point you're looking for. Not sure we can read too much into the modest selloff — this thing has been runner, and runner don't stop on one muted day.
SMTC is on an absolute heater — up 324% over the past year, trading near its 52-week high ($168.29) at $157-$166 depending on the tape, and the analyst community just punched a bunch of PTs higher to the $200-$230 range. UBS lifted to $225, Benchmark went straight to $230 (from $120 — a bigger-than-consensus jump), TD Cowen to $210, Needham to $200, Stifel to $188, Mizuho to $225. That's a tight cluster of conviction around a name that was written off as a legacy analog supplier not that long ago.
The beat-and-raise story is real. Q1 prints showed EPS of $0.51 (versus $0.45 expected) and revenue of $291M that came in 2.67% above est. Q2 guidance came in ahead of Street as well. But the real spark is data center networking — now nearly 50% of revenue — with strong 800G optical demand and active copper cable volume ramping for Google Ironwood racks. The backlog gives them confidence to guide 35% QoQ and 85% YoY acceleration into Q2.
The thesis playing across the buy-side set: Semtech is moving up the optical food chain. UBS makes the clearest case — growth transitions to 1.6T ecosystem this year, then active copper cable wins, coherent TIAs/drivers, and HieFo fab product expansion push into CY2027 and FY2028 before 3.2T networking matures. Benchmark drills into the board-level product set supporting 800G TIAs, 800G LPO, 1.6T FiberEdge/FRO, CopperEdge ACCs, plus laser content. That's a wider socket win than most had modeled.
"Data center networking is the fastest growing part of the portfolio and now represents nearly 50% of revenue. Beyond this year, additional active copper cable wins, coherent TIAs/drivers, and expanding the product set at the HieFo fab are positioned to drive incremental growth."
UBS (Timothy Arcuri) is essentially saying the company has threaded the needle — LoRa+ exposure to iPhone and wearables and IoT holds the base while data center becomes the re-rating lever.
Bull case — the leverage to AI-driven infrastructure buildout is underowned, 1.6T cycles haven't even hit in earnest yet, and Semtech's product breadth across linear equalizers, TIAs, drivers, and active copper cables makes it a "one-stop shop" for hyperscaler optical interconnect needs. 324% gain and still a $14-15B market cap looks cheap if datacenter，迈向 $1B+ in just that segment alone.
Bear case — a lot of good news is in the stock at these levels, the optical supply chain is competitive (Credo, Macom, InnoLight have their own angles), and if hyperscaler capex rolls over or 1.6T deployments push out a quarter, estimates reset hard. The stock has basically tripled in a year with limited margin of safety for disappointment.
Bottom line: the fundamental trajectory is legit and the analyst community is aligning around higher estimates and PTs. The risk is momentum overshooting near-term. Not a sell at $160, but maybe not chasing here either — let it digest some of that post-earnings pop before adding aggressively.
UBS bumped their PT to $355 from $310, calling it a vote of confidence in the multi-year trajectory. They're pointing out that the Climate business is being valued at roughly 16x FY28 EBITDA while the stock trades at 32.3x LTM EV/EBITDA — meaning the market isn't giving credit yet for where this thing inflects. That gap is the bull case in a sentence.
KeyBanc went further, punching their PT to $370 from $250. Overweight. They're citing margin pressure in Climate Solutions as the fly in the ointment, but apparently not enough to derail the thesis. D.A. Davidson sat tight at $265 — notably lower than the Street's new consensus — and flagged that CS segment margins got whacked by weather downtime, mix headwinds, and supply chain noise. They're not wrong to flag it.
The stock's up roughly 220% over the past year, 110% YTD. It printed around $293 Tuesday. That's not a value trap — that's a momentum name that just got fundamental confirmation.
"The $4 billion long-term data center agreement announced today is supporting continued upside in the equity value." — UBS
Analysts are coalescing around a thesis: usage revenue is troughing and rebounding, MAX is gaining traction as an upsell vehicle, and Commercial continues to plug along as the growth engine. Stifel sees earnings that "exceed first-quarter fiscal 2027 estimates through execution across trades, continued momentum in Commercial, and early results from MAX." They're calling for a full-year revenue raise, not just a one-quarter pop. Piper Sandler similarly thinks low-20% growth is achievable, calling out Voice Agents, commercial, roofing, and the HARDI revenue inflection (10.4% in March) as proof that transaction value trends can stabilize.
The Bulls: Growth recovery + MAX program acceleration + commercial traction + AI adoption = potential re-rating if the company guides up. The stock was left for dead after the multiples compressed; now you're looking at 58% upside to BMO's $92 target and 90%+ to the higher-end estimates ($125 from Stifel, $120 from KeyBanc). For a business doing 24.5% revenue growth with a path to 20%+ and a differentiated model that isn't seat-based? That math gets people's attention in this market.
The Bears: Software multiples are getting hammered and TTAN was a rich stock. BTIG cut to $90 from $105 specifically citing software multiple compression — basically saying the fundamentals might be fine but the market won't pay for them. Truist did the same ($100 from $130). And there's genuine uncertainty around whether usage revenue headwinds were seasonal/calendar or something more structural. If MAX doesn't accelerate fast enough, the beat-and-lower pattern continues.
Bottom line: The June 4 print is the test. If they raise full-year revenue guidance — which KeyBanc and Stifel are both penciling in — the stock re-rates. If they print in-line and guide flat, the compression story wins. $65 is either a gift or a value trap. Position size accordingly heading into the print.
Mizuho and DA Davidson both out raises today, and the thesis is tightening: memory isn't a commodity anymore — at least not HBM, and the market is still pricing MU like it is.
DA Davidson's Gil Luria lays out the valuation disconnect cold. AMD and Intel trade at 40x+ earnings. MU sits at $928 and runs a P/E of 35x — but with a PEG of 0.1. That spread exists because the market still thinks memory is bought on a spot screen, fully fungible. Luria's call: that's outdated. HBM is now co-designed into hyperscaler architecture, customers are signing multi-year strategic agreements (Barclays flagged a new 5-year SCA as a pivotal development), and at the margin that changes the pricing dynamic permanently. DA Davidson took their target to $1,500, up from $1,000.
Mizuho's numbers reinforce the urgency. HBM projected to hit 23% of fiscal 2028 revenue with potential price increases of 70% to 100% year-over-year in calendar 2027. Non-AI customers remain 30–50% undersupplied. No meaningful capacity ramps through calendar 2027. Mizuho's fiscal 2028 EPS estimate sits 41% above consensus. They target $1,150 (5.3x price-to-book vs. SanDisk at 5.4x — basically in line). UBS went bolder: $1,625 PT, projecting EPS above $100 through 2029. Barclays at $1,175.
Let me steelman the bear here. The stock is up 868% over the past year and sits within $12 of its 52-week high. A lot of the thesis is in the price. AI capex cycles are uncertain — if hyperscalers pull back on training clusters, HBM demand assumptions get shredded. And historically, memory cycles brutally. PEG of 0.1 looks great until it doesn't.
But the bull case has structural legs. HBM isn't a commodity. Supply stays tight. And at roughly 10x fiscal 2027 earnings, you're not paying for the AI tailwind yet — you're paying for the trough with optionality on top.
"The codesign of HBM into data center architecture appears to be changing that dynamic, as are longer-term deals. Conventional wisdom holds CPUs are not commodities while memory remains one — this may have been true in the past." — Gil Luria, DA Davidson
Bottom line: the commoditization discount is shrinking. MU has crossed $1T market cap for the first time. If HBM repricing plays out through 2027, current levels aren't expensive — they're early. Watch supply discipline and any signs of capacity ramp announcements from the Korean duopoly — that's the key risk to the bull case.
Three firms (Cantor at $575, Mizuho at $540, TD Cowen at $525) all reinforced Buy/Outperform this week. Mizuho led the charge with the most substantive numbers: they nudged WFE estimates to $153B in 2026 and $190B in 2027 — that's 23% and 24% YoY growth, respectively. The driver is AI logic and memory, backed by TSMC running at the high end of its $52B-$56B capex range, Samsung/Micron ramping HBM, and total memory WFE hitting $112B in 2026. That's not speculative — that's what the foundries are confirming.
Cantor's framing resonates: AMAT sees Agentic AI and increased CPU/DRAM as "20% accretive to overall wafer fabrication equipment spending." They doubled capacity to support $10B per quarter in semiconductor revenue ($180B-$200B WFE run-rate). Management is talking eight-quarter order book visibility and conversations extending beyond 24 months — unusual language for a cap equipment name.
The EPIC platform is doing real work too. Broadcom just signed on to co-develop advanced packaging for AI systems, leveraging AMAT's global network of innovation centers. That's positioning for chiplet and high-bandwidth packaging demand — not a theoretical bet.
Earnings revisions tell the story: FY2027 now seen at $42.8B revenue and $16.90 EPS (up from $40.1B/$15.31), FY2028 at $49.9B/$19.86. Lam's $40B node transition spending flag (majority before end of 2027) supports the multi-year ramp thesis.
The real question isn't whether the numbers hold — they do. It's whether the next leg comes from multiple expansion or continued earnings beats. Right now, both are in play. Maintain conviction here; the risk/reward skews favorably as long as AI infrastructure spending stays on track.
Q4 guidance came in slightly below consensus. More importantly, preliminary FY2027 guidance implies revenue and ARR growth decelerating to 16%-17% — a meaningful step down from the 24% LTM topline growth. Analysts are flagging a cluster of headwinds: Red Canary acquisition comparisons creating a tough comp, elevated churn in the MDR business, weak new logo trends, sales force restructuring underway, and hardware inflation driving higher capex. That's a lot of noise in the pipe at once.
The sell-side is doing what sell-side does — ratcheting PTs down in clusters. Freedom Broker took it to $230 from $270 (still Buy). FBN cut to $175 from $200 but kept Outperform — Shelby Seyrafi citing "strong top-line metrics, record next-gen operating margin, and strong ZFlex growth" as reasons to stay constructive. Cantor Fitzgerald dropped to $225 from $300 on capex concerns, Rosenblatt to $200 from $223 on FY2027 concerns, Stephens to $200 from $225. Piper Sandler went further, $160 from $185, and has a Neutral — the most cautious of the bunch.
"Box is starting to prove it can sustain a double-digit growth profile."
That's the DA Davidson line and it's the crux of the bull thesis. Enterprise Advanced demand driving seat expansions and pricing premiums — i.e., they're not just selling more seats, they're taking price. When cloud companies get pricing power, models re-rate. UBS frames it differently: the growth inflection is real, the stock at ~11x CY26 EV/FCF deserves some relief, but the margin picture hasn't budged. FY27 EBIT margin guidance essentially unchanged — just 20bps of improvement. The H2 FY27 guide of ~9% CC growth looks conservative by UBS's own math.
Bottom line: if you believe AI is actually lifting total growth at Box (not just AI attach, but AI-driven displacement of legacy ECM), this is a 40%+ upside name with a 79% gross margin business. If you're waiting for margin expansion to confirm the thesis, neutral is the right call. The stock at $25.18 vs $27.38 fair value (per InvestingPro) tells you the street hasn't priced the AI narrative yet. Positioning opportunity — but you'll need the seat count data to confirm it's real.
The analysts split on what to do with that. DA Davidson stuck with their Buy and $14 PT, citing the post-earnings 19% pop and arguing the stock trades at 7x FY2027 EBITDA against a $11.70 fair value. They're not wrong that valuation has been crushed — down 61% over the past year. Truist went the other direction: Hold, PT lifted to $12 from $11, calling it a premium multiple at 27.71x EV/EBITDA that "reflects premium valuation expectations despite near-term headwinds." They're not upgrading because the business still isn't firing on all cylinders.
The bear case in one line: fiscal 2027 will print Rule of 30s, not Rule of 40. Management themselves acknowledged they're not hitting that longer-term target this year. Truist put the inflection timing right there:
"The key factor for a return to stronger growth is the normalization of pharmaceutical spending, particularly in Network Solutions. Management views fiscal 2028 as the earliest potential inflection point for the business model."
That's a meaningful caveat — you're buying a story that requires pharma budgets to loosen, and you're doing it against a stock that's already halved in six months. The AccessOne payment solution expansion with PNC Bank is noise at this point; it doesn't change the top-line trajectory.
Bottom line: Q1 gave bulls a catalyst, but it's a small beat on thin revenue. The $14 vs $12 PT split tells you analysts aren't aligned on whether the valuation reset is an opportunity or a value trap. If pharma normalizes, this works. If it doesn't, you're holding a Rule-of-30s software comp with no obvious re-rate catalyst until FY2028. Not sure we can read too much into the post-earnings pop — that's often just short covering.
HOOD pops today on the agentic AI launch — not enough to call a bottom, but enough to refocus the narrative on what's next for a name that's been left for dead. The company's rolled out two products: Agentic Trading (beta, equities-only at launch, customers hand a separate account to their own AI agent) and an Agentic Credit Card (virtual Gold Card, AI-controlled spending). Citizens reiterated Market Outperform at $155. The Street's thinking around $86-$94 on the other side.
The bull case is straightforward: HOOD is early-moving on a structural shift where retail finance gets mediated by AI agents rather than manual human workflows. Citizens frames it exactly that way — "customer activity on Robinhood is increasingly initiated by permissioned agents rather than manual human workflows." That's the thesis. Platform assets at $345B (+49% YoY), 27.6M funded customers, $6B in net April deposits running at a 23% annualized growth clip — the underlying business is healthy. And at 36x earnings on a $68.7B market cap, you're not paying up for the new stuff.
The bear case is equally clear: this is early, it's beta, and nobody knows the TAM. Goldman's own note concedes "it remains challenging to assess the total addressable market and success of the new products, given agentic banking and brokerage products do not exist today." The CTO just walked out — Jeffrey Pinner gone with no successor named. And the stock is still down 40.67% over six months. Some PMs will call this the dip. Others will call it a story with no near-term fundamental catalyst.
Goldman on the strategic angle:
"The announcement represents an early effort at embedding agents into retail brokerage and consumer finance."
That's the right frame. Early. Not proven. But HOOD is one of the few names actually shipping product in this space — which matters when your customer base skews young, mobile-first, and comfortable delegating decisions to software.
The regulatory angle is interesting too. SEC's expected innovation exemption for tokenized stocks could open decentralized trading on crypto platforms without company backing. HOOD's got crypto exposure — this could be a tailwind or a disintermediation risk depending on execution.
Verdict: HOOD isn't breaking out on fundamentals today, but the AI narrative gives the story a refresh. Not a buy-at-this-price for us, but worth watching closely if the agentic products show any traction in customer adoption metrics. The $345B AUM and deposit flows tell you the core business isn't broken. The question is whether HOOD can turn AI novelty into sticky platform revenue before the next 6-month -40% run.
OOMA +71% over six months. +64% YTD. That kind of move gets people's attention — and this morning two more analysts piled on.
Freedom Broker lifted its PT to $20 from $18.30, citing the Q1 beat and AirDial momentum. Benchmark went further, raising to $24 from $23, calling out the "remarkably broad product bouquet" still undervalued. Citizens stayed put at Market Perform — not bearish, just not convinced the story deserves more credit yet.
The quarter did the work: EPS $0.35 vs $0.32 estimate, revenue $81.1M vs $79.8M, adjusted EBITDA $11.8M vs $11.0M. Management raised FY2027 guidance on the back of AirDial installations accelerating — they're seeing real lift from copper line shutdowns, especially on the AT&T side. The FluentStream and Phone.com acquisitions (late 2025) are now compounding into the business segment numbers. Ooma AI is the next lever, positioned to drive ARPU on the SMB side.
Bull case is clean: PEG of 0.43 makes this look cheap relative to 18% revenue growth, AirDial is early innings on a massive POTS replacement TAM, and the residential base (Telo, MyPhone) is a steady floor. Bear case is just as honest: 71% in six months means you're chasing a lot of the good news, Citizens isn't moving, and small-cap VoIP doesn't get multiple expansion without scale or a catalyst the market can't ignore.
The stock trades at $20.18. At Benchmark's $24 target you're looking at roughly 19% upside — decent but not screaming. Freedom's $20 target leaves almost no room. The real question is whether AirDial keeps scaling through the year, because that's the story and it's not a sure thing yet.
"The near-term numbers are less important than the company's comments on TPU."
$525 at 25x forward earnings implies the street's already baking in meaningful AI infrastructure spending — question is whether the beat-and-raise narrative holds or we get a "fine, what's the next leg" response. AVGO's networking + custom silicon combo gives it a floor most semis lack, but the bar here is high heading into the print.
The numbers anchor everything: $253B LTM revenue (71% growth), 74% gross margin. Those margins in an AI infrastructure buildout at this scale are the tell. Most infrastructure plays compress margins as they scale; NVDA is expanding them. That suggests pricing power that hasn't peaked.
Jensen's announced $150B annual Taiwan investment is the underreported item here. That's not just supply chain hedging — that's locking in manufacturing relationships while competitors are still designing chips. Custom silicon from hyperscalers (Google, Amazon, Meta) gets the headlines, but Tape Out cycles and ecosystem lock-in don't reverse on a quarterly basis.
The bear case — stretched multiple, geopolitical concentration, AI capex cycle risk — isn't wrong, just early. At current multiples, you're paying for the next 2-3 years of outperformance. Not insane given the TAM, but worth acknowledging. The positioning question for PMs: do you trim on consensus getting this bullish, or hold because the underlying revenue revisions haven't peaked yet?
"Nvidia is positioned to capture an outsized share of global AI infrastructure spending in a multi-trillion-dollar market." — Tigress Financial
Take: Hold positions. The PT cluster at $270-$500 tells you the sell-side hasn't reached capitulation — there's still dispersion, which means room for further revision. Watch for Q2 guidance cadence; any softness there would crack the narrative. Otherwise, this stays a core AI infrastructure overweight.
ARM trading at $302 — up 194% YTD — and Mizuho just pushed their PT to $360 (from $290) on AI CPU momentum. Outperform.
The thesis is straightforward: Agentic AI workloads are creating a new compute tier, and ARM's custom CPU portfolio (Grace, Vera, Axion, Cobalt, Graviton) puts them at the center of it. Mizuho models internal CPU ramps into CY2027, with a PEG multiple stepping up to 2.7x FY2028 estimates from 2.2x — meaning they're paying up for the growth curve, not just the current print.
Other names piling on: Rosenblatt (PT $270) flagged the licensing beat (up 29% YoY) as durable, Jefferies (PT $290) highlighted AGI CPU demand doubling over the next couple fiscal years. Citigroup's long-cycle view puts agentic CPUs at $59.4B by 2030 — that's the TAM frame. Evercore noted ARM gained 140bps of server CPU unit share last quarter while Intel slipped from 59% to 55%.
"Arm Holdings stands to benefit from Agentic AI CPU ramps under three scenarios it outlined."
The Cerebras probe (ARM/SoftBank approached, got rebuffed) is noise — worth noting as a signal of strategic intent, but no near-term catalyst.
The move puts ARM at roughly 19x forward earnings on 23% revenue growth. Not cheap. But if the agentic compute cycle plays out, the multiple holds. If it doesn't, you're paying up for a story. Stock's running hot — keep size disciplined.
The thesis centers on wafer fab equipment spending — Mizuho now sees 2026 WFE at $153B and 2027 at $190B, representing 23% and 24% YoY growth respectively, up from prior calls of $142B and $163B. The driver is AI logic and memory capex. TSMC anchoring the logic side with 2026 capex at the high end of $52B-$56B, including 2nm ramps, and a follow-on 2027 spend of $74B (+28% YoY). Memory isn't slacking either — Samsung, Micron, and others pushing HBM and DRAM with total memory WFE near $112B in 2026. MKSI rides both legs of that buildout.
The estimates tell the story. Mizuho lifted FY2026 EPS to $11.77 (from $11.20, vs. $9.94 consensus) and FY2027 EPS to $15.68 (from $13.88, vs. $12.18 consensus). That's a meaningful beat-and-raise dynamic, not a one-quarter pop. Q1 already delivered — EPS $2.30 vs. $2.04 expected, revenue $1.08B vs. $1.05B expected. Consistent execution against a rising tide.
The bull case: if WFE estimates hold, MKSI at $330 trades at roughly 22x forward '27 EPS. For a company exposed to both logic and memory capex with leading position in thin film and plasma processes, that's not crazy if the cycle has legs into '28. The risk: you're buying a 284%-mover into a 52-week high with a PT that's 21% above the current print. Momentum is real, but at some point the bar gets raised. The stock needs the WFE numbers to actually print — there's no cushion if capex gets pushed a quarter.
Bottom line: the fundamental case is tight. The positioning is the debate. If you're already in, the Mizuho raise validates the thesis. If you're building a position, you want a pullback to $310-$315 to work the r/r. Not a must-buy here, but nothing to sell either.
Morgan Stanley upgraded to Overweight last week on NAND systems growing 59% above prior peaks. UBS ($310 PT) and Stifel ($325) are singing the same AI-capex song from different corners. 24 analysts have revised earnings upward for the upcoming period — that's not a split market. Stock at $319, one percent off its 52-week high, and Mizuho is modeling FY28 EPS at $10.63 against current consensus — implying the multiple barely has to do anything to justify $380.
One honest caveat: LRCX already returned 287% over the past year. The risk reward on today's run is tighter than it looks. But if you're running a fund without meaningful LRCX exposure and you believe the capex supercycle is real, being underweight into this is the uncomfortable trade. Not sure we can get cute about valuation here — the orders speak louder.
"The initiative represents a multi-billion dollar revenue opportunity," Rosenblatt says, citing Snap and OpenAI traction as the comps.
Worth flagging: the VT AG addiction lawsuit proceeds after SCOTUS declined to hear the appeal. Not the end of anything, but another data point on regulatory exposure adding friction to the engagement flywheel. JLens also pushing a content moderation transparency vote, Glass Lewis in favor. Not material to the quarter, but the narrative war doesn't stop. Stock continues to grind through 2025 resistance — this subscription pivot could be the catalyst that reframes the multiple if adoption numbers print. Watching early conversion metrics more than PT models right now.
"The cybersecurity environment remains robust, driven by AI adoption increasing demand across network security, cloud security, identity, Prisma SASE, Cortex XSIAM, and AgentiX." — Benchmark analyst
Q3 setup looks clean. Platformization deals and NGS ARR growth remain the bogeys. Watch for any guide-down risk on macro-driven large deal pushouts — that's the bear case if it surfaces post-print.
CRWD catching another upgrade before the June 3 print. Benchmark moved PT to $700 from $500, Cantor did the same to $700, TD Cowen to $625. Stock sits at $646, just shy of the $677.50 52-week high — up 43% YTD. That's a lot of juice already priced in, but the Street keeps saying there's more.
The thesis is straightforward: AI adoption is creating a broad-based demand cycle across network, cloud, identity, SASE, and agentic AI — and CRWD is landing in all of it. Benchmark expects Cloud Security to approach $1 BILLION at 30% YoY, LogScale SIEM to clear $600M at 50% YoY, and Next-Gen Identity to hit $600M at 30% YoY. That's product-level detail that signals the AI tailwind isn't just narrative — it's flowing through the numbers. The company is guiding Rule of 50 (23% revenue growth + margin expansion), 75% gross margins, 32% FCF margin. Moody's upgraded to Baa2. Claude AI integration is live.
Bull case is clean: beat-and-raise on June 3, AI cycle extends the growth runway, multiple re-rates higher.
Bear case lives with Guggenheim: stock at 52-week highs with the print seven days out leaves little room for error. Miss or guide soft and you're down 10% on massive volume.
"The cybersecurity environment is fairly robust driven by AI adoption uplifting demands across network security, cloud security, identity protection, SASE, security operations, and agentic AI." — Benchmark
Bottom line: this feels like a high-conviction long into a known catalyst. Size the position based on your risk tolerance for a beat or miss on a stock that's already up 43%. If you're underweight, June 3 is the re-entry point if it sells off. If you're long, trim into the print if you're nervous about execution. Not a place to be size-blind.
UBS: "Amazon appears undervalued based on its Fair Value assessment."
Other housekeeping: Amazon's on track for the £40B UK investment by 2027, with drone delivery now live. Delta picking Amazon over Starlink for in-flight Wi-Fi is a small but meaningful enterprise reference. Berkshire selling its stake and rotating into Delta is... interesting. Not a bearish signal on AMZN necessarily, but a data point. Wolfe and TD Cowen both reinforcing at Buy with the grocery expansion story (30-min Amazon Now) — that's the under-the-radar margin play.
Bottom line: The AWS AI mix shift is the story, and consensus is still catching up. $333 feels achievable if Bedrock backlog conversion holds. Watch for AWS margin expansion in the next print — that's where the beat vs. miss lives.
Mizuho's thesis leans into Agentic AI server ramps — specifically Dell-made Vera-only racks, expanding AI server backlog, and storage continuing to scale. They're out above consensus on both FY2027 ($144B vs $142B) and FY2028 ($159B vs $158B), with EPS estimates that also beat. New PT implies 23x FY2028 earnings, up from 20x. Current P/E sits at 35.5, but the PEG of 0.93 suggests the growth justifies the premium — for now.
Other names in the conversation: Evercore ISI reiterated Outperform, calling out AI infrastructure momentum. Bernstein went to $280 on agentic AI demand driving server market. Morgan Stanley raised to $170 but held Underweight — their concern is memory inflation and supply shortages eating into margin. That's the bear case hiding in the bull narrative.
Also notable: IREN just signed a $1.6B deal with Dell for air-cooled Blackwell systems. That's a real contract, not a pipeline mention. Points to delivery visibility.
Earnings print tomorrow. Options showing stock movement expected. Given the run, the bar is high — any beat needs to be clean, not just headline OK. But the AI infrastructure theme isn't fading, and the backlog depth gives them some cushion if growth decelerates a touch.
The stock sits at $1,589, near its 52-week high, after an absurd 4,063% run over the past year. Yes, four thousand percent. At that entry point, the 9.9x multiple Mizuho's anchoring to (up from 9.4x prior) is the only thing keeping this from looking stretched — but they're not backing away from it, which tells you how confident they are in the margin trajectory.
The thesis hasn't changed, it's just gotten more confirmed: NAND supply stays tight through 2027, HBF and CMX rack deployments add incremental demand pressure, and SanDisk's new contracting model ($42B in NAND bit obligations across five customers with financial guarantees) locks in both volume and pricing. Data center revenue up 191% YoY and ASPs up across all segments — that's the pricing power story everyone wanted to see.
Barclays going to $2,300 with an Overweight, S&P bumping to BB+ on the net-cash transition, Cantor at $1,800, Bernstein at $1,700, Jefferies at $1,400 — the buy-side is rotating higher, not trimming. At this stage of a run, you're not chasing fundamentals, you're managing positioning risk. But when the beat-and-raise cadence keeps coming and the multiple doesn't compress, the tape is telling you something.
IBM caught a double-fister today — RBC sticking with Outperform at $300 and a $1B Chips Act wafer for Anderson, IBM's quantum foundry play. The thesis hasn't changed: open-source vulnerability surface area exploding as AI accelerates exploit discovery, IBM sitting on the largest enterprise open-source install base with both the incentive and chippies to protect it.
Project Lightwell is the anchor — leveraging Anthropic's Mythos model to surface nearly 3,900 critical-severity OSS bugs. That's the wedge into SecDevOps workflows at the F500. Revenue +10% and a PEG of 0.24 is the valuation argument sitting underneath. Not exciting, but credible.
The quantum wafer play is the longer-dated option. $2B total Chips Act initiative, IBM getting $1B to spin up Anderson as a standalone quantum foundry in Albany. That's a decade-plus horizon bet — not changing the multiple today. War contract extension ($46M modification, now $155M total) is steady Eddie maintenance work.
Stifel at $290 versus RBC's $300 — not a meaningful spread, both assuming the Red Hat/consulting momentum holds through the year. The stock at $265 is doing what IBM always does: grinding, not galloping. The risk is that hybrid cloud tailwinds plateau while the quantum investment drag persists.
Bottom line: stable, well-positioned on the AI security narrative, valuation not insulting. Keep it on the book but don't chase. The Chips Act wafer is interesting narrative but too early to model.
NDSN catching a pickup here — DA Davidson bumps PT to $345 from $335, keep Buy. Stock's already up 47% over the past year to $288.78, sitting just below the 52-week high at $305.28. Not chasing, but the setup has legs.
Thesis is straightforward: semiconductor dispense demand accelerating and T&I orders picking up — that's feeding the ATS segment hard. Management flagged synchronized organic growth across all three segments (IPS, MFS, and the legacy stuff) which is rare and suggests this isn't a one-pocket story. EFD and medical components doing the heavy lifting in MFS. Automotive and PPS demand improving too — confirms the cross-sector demand thesis.
Q2 beat on both top and bottom ($741M vs $728.78M expected, adj EPS $2.86 vs $2.82) gives them the confidence to raise FY2026 and FY2027 estimates. Not revolutionary but solid execution.
The balance sheet is the underrated part: 1.9x net leverage gives them room for both buybacks and M&A. That's the optionality piece if the semiconductor cycle keeps extending. Net leverage at 1.9x is a long way from distressed — management has firepower.
Not a lot of noise in the name but the trajectory is clean. If you're underweight, this is a name to revisit on any pullback. Not a blockbuster catalyst story but the fundamentals are grinding higher and the multiple probably hasn't fully repriced the semicaps upside yet.
UBS lays out the bull case for Retail POS: currently ~12% of GMV, could hit 20% by 2035. That's 300-500 basis points of incremental GMV CAGR support over the next decade on top of the 15-21% base. The TAM punch lands — in-store volumes are multiples larger than eCommerce, and nobody owns this market the way Toast owns restaurant POS (30% volume share in US). Fragmented, no modern leader, Shopify with the ecosystem advantage.
"This business represents an important part of the longer-term growth algorithm and a relatively underappreciated portion of the company's terminal value."
That's the bull. Steelmanning the bear: stock's down 32.8% in six months, and Cantor just trimmed to $115 on margin concerns. Revenue growth looks hot (31.85% LTM) but GM margin at 47.97% leaves questions. When a $100M Thrive Capital check (same crew backing OpenAI) is your main recent catalyst, that's a vote of confidence in AI potential, not in current fundamentals.
Bottom line: The POS long thesis holds. But with the stock in the penalty box and multiple PTs ranging from $115 to $170, you're not getting a clean signal here — you're getting a coin flip on timing. Hold neutral until either margins inflect or the valuation compresses further.
The GAAP EPS miss ($0.12 vs $0.27 expected) is the obvious complication — "revenue exceeding expectations not enough to offset the earnings challenges" per the writeup. But that's non-GAAP vs GAAP noise, and the non-GAAP $0.35 beat (+30% vs consensus) tells you the operational picture is cleaner than the headline implies. Non-cash charges are muddying the water. Not unusual for SaaS names in investment mode.
24-month RPO at $871M (+5% YoY) is the one metric I'd want to watch — that deceleration from +8% prior quarter could be a signal of demand softness or just digestion in the pipeline. Could read too much into it given the headline beat, but it's the one number that doesn't scream reacceleration.
Bottom line: If you're looking for a cloud banking name that's showing velocity improvement with real margin leverage, NCNO fits. The GAAP miss is a red herring for anyone who's been watching this name — the non-GAAP picture is strong. Citizens likes it at $23. That's a bet worth tracking.
Jefferies and Piper Sandler both reinforcing the same near-term bull case: political advertising in H2 2026 plus World Cup coverage (104 matches on The Roku Channel via the new FOX One $19.99/month tier) should lift Platform revenue. Improved fill rates also in the mix. Three firms, same theme.
The catch: Fair Value analysis suggests the stock may be slightly overvalued at current levels. $19.5B market cap with 100M+ households is impressive scale, but ROKU has always traded on the gap between platform reach and monetization efficiency. This Home Screen is the most concrete attempt in years to close that gap. Worth a position if you believe the personalization flywheel can accelerate ad yield — just don't chase if you're late to the name.
The gross margin story (33%) and the acceleration narrative are what bulls are hanging their hats on. Bears will point to the PT trims as evidence the street is losing conviction, and the Idelic integration risk is real — $28M up front, up to $12M contingent on revenue targets is meaningful earnout exposure.
"The risk-reward profile as attractive, noting that the company's valuation is near 10-year lows." — RBC's Paul Treiber
Not sure we can read too much into the mixed PT action — BMO and MS both raising earnings estimates while cutting PTs suggests they're marking down multiple expansion assumptions, not the earnings story. That's a nuanced signal. Q1 print will be the next real test.
ON getting love from Mizuho this morning — PT lifted to $150 from $130, Outperform intact. Valuation math: 32x FY2027 EPS now vs 28x prior. That's a meaningful multiple expansion justified by what? AI server ramps. Agentic AI, training, inference — ON's 800-volt architecture is landing in the right places. Lead times stretching past 40 weeks on some products, capacity genuinely tight on AI server MOSFETs. Nexperia continuing to serve as a tailwind in the near term — ON picking up share. 27 analysts have revised earnings upward. Stock up 193% over the past year.
Here's the tension: stock trades at 88.56 P/E. That's not cheap. That's pricing in a lot. BofA naming ON a top pick in AI power semis, citing multi-year content gains not yet in estimates — that's the bull case, and it's credible. AI power infrastructure buildout is real, and ON has the right exposure.
Cantor Fitzgerald came in more cautious — PT $100, Neutral. "Turning cycle" plus data center revenue uptick, but not enough to justify the premium. That's the bear steelman: the cycle turn is real, but the bar is high.
Near-term headwinds from slower low-voltage power products exist, but Mizuho says offset by the AI stuff. Pullback in chip names post CPI print is noise — it's the setup that matters.
Verdict: ON is in the right thematic box at the right time. The valuation is stretched but not absurd given the forward trajectory if AI infrastructure spending holds. Not chasing the 88x P/E here at current levels, but the long thesis is intact. Watch for any softness in AI server buildout signals — that's the binary for this name.
"We believe in the company but do not view the stock's current risk-reward ratio as attractive enough to warrant a buy recommendation at present levels." — Freedom Broker
Verdict: the spin-off narrative is real and the BofA upgrade at $180 reflects institutional conviction on it. But a 49% run in three weeks on a name that hasn't even filed spin-off documents yet? That's a lot of pre-positioning. Watch for the $120-$130 zone Freedom flagged as a true entry. If you weren't around the May 5 announcement, you're chasing. The margin expansion and shareholder value framing from BofA is the right bull case — just want to see the spin-off details before treating it as a done deal at 62x.
Stock's already bloodied, down 71% over the past year, trading near $22. The Street's been quietly trimming: BUY-side holders like D.A. Davidson ($29) and Benchmark ($30) just lowered bogeys despite Q1's "slightly exceeded" print — 12% revenue growth, EBITDA 5% above estimates. Q2 guidance of 8% YoY came in light though, and that's what got folks worried. NEUTRAL shops (Cantor $20, Wedbush $21) have been trudging down PTs too, with some flagging spillover from Publicis' LiveRamp deal drama. Publicis is paying $38.50/share for LiveRamp — that's reshaping the ad-tech map, and TTD's sitting in the blast radius.
Bull case still breathing: Platform scale, data clean room momentum,CTV tailwind. But Rothschild's not buying it — says competitive moat has been "exposed." That's the line I'll throw you:
"Trade Desk's competitive challenges have been exposed."
If you're long here, you need a clear catalyst for the competitive dynamics to reverse — and right now, the narrative's controlled by the bears. $11 feels extreme, but the direction of travel is ugly. Watch for Q2 print to eitherconfirm the guidance weakness orprove the concerns overblown. Until then, not a place I'm adding risk.
Morgan Stanley models 20bps annual conversion rate expansion → revenue/EBITDA ~50% above consensus by 2030. Each additional 10bps = 17 points of net revenue growth. The 88% gross margin gives them the firepower to keep investing there.
The Q1 print was clean: EPS $3.56 vs $3.42E, revenue $1.84B vs $1.78B. Q2 guide 4-6% QoQ despite seasonal softness. Piper Sandler moved to $665, Wolfe to $580, Goldman to $585, Jefferies at $700. The PT cluster is now $580-720 — that's a wide range but all bullish.
The real debate: is conversion rate expansion a durable moat or are we watching a law of large numbers problem start to bite? Scale helps — more data, more optimization cycles — but at $170B+ market cap the bar is obviously higher. Not sure we can read too much into the 30bps expansion over 18 months as a rate-of-change signal given the massive base. Still, APP has the margins, the cash generation, and the adjacent adjacencies (non-endemic ads, e-commerce) to keep the story intact. Conviction buy for now.
The long and short: a deeply wounded loan book in a consumer credit environment that still isn't clearing. The dividend and buyback commitment provide a floor, but until revenue trajectory inflects — and there's no evidence of that yet — you're collecting yield while hoping for a turn that keeps getting deferred. QFIN at current levels isn't broken, but "not broken" isn't a thesis.
Here's the trade: GLBE is paying $350M upfront (50/50 cash/stock) plus up to $75M in 2026 earnouts for Passport, an e-commerce fulfillment player expected to add roughly $100M of revenue in calendar year 2026. Deal closes early July subject to approvals. Management says it'll be neutral to slightly positive on adj EBITDA in H2 2026 — basically break-even while they integrate. The seasonal pattern mirrors GLBE (heavy Q4), and critically, most Passport revenue flows through fulfillment revenue, not GMV, so it should be high-quality, high-margin. GLBE has more cash than debt, so no balance sheet stress.
The bull case: this pushes annual revenue guidance higher post-close, Q1 already showed strong momentum (GMV +40% YoY, revenue +33%, adj EBITDA +59%), and Q2 guidance beat consensus. UBS flagged that adj EBITDA margin exceeded expectations. Needham separately reiterated Buy/$47. The coverage is coalescing around a $47-$60 PT range — that's a wide spread but reflects uncertainty on how much Passport contributes and when margins inflect.
The bear: at $29.60 you're paying ~0.7x sales for a company acquiring another business at a point where cross-border e-commerce sentiment is trashed. The earnout math only works if Passport hits 2026 targets — not guaranteed. And if integration takes longer than expected, "neutral to slightly positive" could slip to negative.
No estimate changes until close. If you wanted to play the deal, you're betting on GMV acceleration sustaining and Passport synergies arriving faster than the "neutral EBITDA" guidance suggests. Not a bad bet given the Q1 numbers, but not a screaming signal either.
"Passport revenue is expected to follow a similar seasonal pattern to Global-E Online, with greater weighting toward the fourth quarter and holiday season." — Benchmark, reiterating Buy/$60
Passport looks like a solid tuck-in, not a transformation. Street likes the deal terms; stock is just waiting for confirmation the integration goes smooth.
“Monetization upside from expanding owned supply with richer audience signals and improving targeting and ad pricing driven by the integration of more first‑party data.” — BofA analyst
Bull case: execution cadence now consistent, the 56% PT implies a re‑rating if FY27 visibility holds. Bear case: ad‑tech macro still iffy, competition from larger DSPs could cap pricing, and the stock’s recent +5.7% pop could be a short‑term head fake if macro headwinds resurface. Not sure we can read too much into that one upgrade without more coverage, but the beat‑and‑raise pattern and the PT give a decent floor for a re‑test of $6‑7 if execution sticks.
"The integration includes Walmart's closed-loop measurement tied to actual sales outcomes and is designed to reduce workflow friction, improve signal fidelity, and provide advertisers a more scalable path to commerce-driven campaigns across omnichannel environments."
That's the core thesis. Scale through partnerships, not owned inventory. RBC acknowledges the ramp will be gradual — not a binary catalyst — but sees it as incremental upside and sentiment support given MGNI trades below InvestingPro Fair Value. The Q1 beat (EPS $0.13 vs $0.11 est, revenue $164.4M vs $159.2M) didn't move the stock, which tells you the market wants visibility on this Walmart relationship converting to revenue. That's the thing to watch — not the headline partnership, but whether it shows up in the numbers over the next couple quarters. Not a story stock right now, but the PT implies a lot of value embedded at these levels if execution holds.
The bull case is crystallizing: AEC (active electrical cable) ramps at multiple hyperscalers are showing up in the numbers. Stifel thinks the April quarter prints above their $430M / $1.02 EPS estimates, and July guidance could beat $455.8M. That's the near-term setup.
The bigger story is the $500M fiscal 2027 optical-linked revenue target anchoring 75% corporate revenue guidance. Stifel frames it right: Credo isn't a copper play or an optics play — it's a systems-level connectivity provider. The DustPhotonics acquisition (Jefferies and Mizuho both flagged this) is the catalyst driving the +75% YoY fiscal 2027 revenue thesis. Rebellions partnership for ZeroFlap active electrical cables into AI inference clusters adds credibility to the systems angle.
Broader coverage picture: Jefferies at $225 (Buy), Mizuho at $220 (Outperform), Rothschild Redburn initiated at $206 (Buy). That's a $206-$250 cluster — previously $200 was the ceiling. Consensus is firming up, not fraying.
Bull steelmanned: hyperscaler CapEx staying elevated, new product ramps, and that $500M target looks achievable if AEC penetration holds. At 20.6x CY27 EV/Sales (Stifel's multiple), you're paying up, but Credo's positioning in AI cluster connectivity has duration.
Bear steelmanned: 248% in a year means a lot of good news is priced. Valuation leaves no margin for error — any softness in hyperscaler spend or ramp delays hits hard. The $500M optical target is an anchor, but execution risk on the DustPhotonics integration is underappreciated outside the detail.
Bottom line: the narrative is tightening from "AI connectivity play" to "systems-level infrastructure provider with visible FY27 numbers." Not sure we can extrapolate the rerating much further without a print to confirm the beat, but the positioning argument holds. If you're long, hold. If you're watching, the next print is the test.
Key driver is Cash App. Canaccord flagged 38% gross profit growth and better monetization — that's the swing factor. Square's reacceleration in GP is the other leg. Truist models Square GP +1%, Cash App GP +4-5%, and nudges adj EPS mid-single-digit. Not a dramatic revision, but consistent with a model that's quietly inflecting.
Block's trading around 0.6x PEG on 24% earnings growth into 2028 (Truist's math). That's not demanding given the momentum. The Hedy situation remains a known overhang for PMs tracking narrative risk, but the street's looking through it — for now.
Square for Drive-Thru is fine, product增量 not catalyst.
Bottom line: consensus is tightening above $80, Cash App is doing the heavy lifting, and there's no data point in the last four quarters that argues against the setup. If you're underweight XYZ, you're fighting the tape on this one.
Hearing Anthropic's $65B Series H isn't just a financing — it's a supply chain equity play. Samsung Electronics and SK Hynix as investors means Anthropic's HBM procurement shifts from external transaction to internal coordination. That's STRUCTURAL BULL for SK Hynix (stable demand anchor) and STRUCTURAL BEAR for MU (loses a major external customer).
Word is Musk's timing on the Anthropic S-1 data center协议质疑 (180 days vs 3 years) is either deliberate IPO positioning or he actually found something material in the disclosure. Either way, the SpaceX IPO timing and the Anthropic S-1 challenge in the same news cycle is too coordinated to be accidental. Watch for SEC commentary.
Channel checks suggest TSMC's CoWoS capacity is effectively sold out through H1 2027 — the Taiwan export data (50-year fastest growth) and Computex pipeline signals confirm the supply constraint is physical, not speculative. Any new entrant (AMD MI355X, custom silicon programs) trying to get CoWoS allocation is looking at a 12-18 month queue.
Hearing the Pentagon $9.7B software contract in Dell's guidance is being questioned by some PMs — is this recurring revenue or a one-time large deal? If it's the latter, DCF terminal value needs a haircut. Dell hasn't clarified the nature of the contract in the earnings transcript yet.
Word is Japan's $73.6B FX defense intervention was the largest since 2024, and USDJPY pulled back from 160+. For Taiwan and Korea AI exporters, a stronger yen against dollar means their dollar-denominated revenue gets a translation headwind — marginal but worth watching if BOJ continues exit.
Channel checks suggest HBM spot market premiums are tightening — LTA customers are getting allocation priority, spot buyers are paying up but finding supply scarce. MU's pricing power in the memory cycle is stronger than consensus thinks, but SK Hynix's vertical integration with Anthropic creates a two-tier HBM market.