Good morning.
Tape set up green. SPX closed 7,580.06 Friday — NINTH STRAIGHT WEEKLY GAIN +1.4%, four fresh ATHs, 22 YTD vs 18.5 historical average, +19% off the March 30 low. Q2 tracking best midterm Q2 ever. Bonds bid, gold ripping, energy absolutely slammed (oil $109 → $87 in two weeks, -20%). Hartnett's frame doing the rounds: long bonds, long humiliation, short hubris. Yardeni bumped YE SPX to 8,250.
The signal of the day is DELL +33% Friday — LARGEST ONE-DAY GAIN IN COMPANY HISTORY — on the first confirmed shipment of an NVDA Vera Rubin NVL72 rack (72 Rubin GPUs, 3.6 exaFLOPS FP4) to CRWV. Hyperscaler '27 capex >$1T per LSCC CEO. Compute supply chain is real and shipping.
Memory got fresh ammunition — Goldman hiked Hynix 2027 op profit +21% to $266B, Samsung +21% to $352B, peak now pushed to FY ending March 2029. NAND tightness into 2028. MU has a $1,600 bogey floating around with the 6/24 print, biggest volume trade of '26. Calls at $150 each.
Three threads framing the tape:
We'll hit up CRM and PANW first, then get to semis and the cyber setup.
Verdict: Show-me story now, not the AI victim trade. Stock at $200, down 47% YOY, down 13% on the week, 20% MTD, kissing the $218 52-week low. The print itself was fine (organic +12% vs ~10% Street). What killed it: a $500M cut to 2H ARR guidance tied to a freemium pivot, plus a CFO departure on top of an open CEO search. Two exec seats empty at once is the kind of thing that makes PMs wait for the other shoe.
PT range is WIDE: $195 to $379. That's a 95% spread between the high and low — basically the Street is fully split on whether this is a broken franchise or a generational entry. The cluster below $250 is where the freemium skeptics and growth bears live; the $285-379 tier is the "wait it out" camp.
Here's the consolidated PT map (all cuts, all post-print):
Bull case (Bernstein/RBC/DA Davidson, $285-379 PTs, ~25-90% upside from $200): The print WAS good. Organic +12% beat Street by 200bps. AI-first ARR tripled to $500M+. Firefly ARR +50% QoQ. Acrobat/Express MAUs at 850M vs 700M last year — that's a real distribution asset. Margins still 89.4% gross. Trades at 7x CY27 EV/FCF and ~13x P/E, which for a franchise with this kind of FCF and brand is historically cheap. Freemium is a land-grab, not a surrender — they're buying user base now to monetize later. CEO/CFO transitions are messy but the underlying engine is intact. RBC's framing: the bad news is already in the print, the multiple has already done the work, 47% down in a year is a lot of pessimism priced in.
"The consensus view, which we share, is that creative AI technology has impaired Adobe's pricing power in the down-market/individual segment, a major dent to the narrative. We are Neutral-rated." — Karl Keirstead, UBS
Bear case (KeyBanc/BMO/TD Cowen/UBS, $195-245 PTs, downside from here or marginal upside): Freemium is not strategy — it's an admission that AI-native competitors (and changing user behavior in AI search) have broken pricing power in the consumer/SMB segment. Adobe is voluntarily giving up near-term ARR ($500M in 2H) to defend a user base that may not convert at historical rates. KeyBanc/BMO/TD Cowen are all modeling negative net new ARR growth (-18% to -25% depending on timeframe) — that's not a valuation reset, that's a re-rating of the entire growth algorithm. Two C-suite seats open simultaneously during a strategic pivot is genuinely unusual. CFO leaving for Marvell mid-print raises questions. Jackson Ader at KeyBanc: "Between the good but not stellar organic performance and guidance and the news that CFO Dan Durn is departing, it is hard to step in... we expect shares to continue to lag the sector." The downside case is Adobe becomes the next CRM — a once-great franchise that AI commoditizes over 3-5 years.
This is the cleanest "AI disinflation in SaaS" data point since CRM got hit. ADBE is admitting — in the guide, not in a slide — that creative AI tools are changing the demand curve for individual/SMB subscriptions. The bear case on the entire SaaS complex (CRWD, MSFT, NOW, even CRM's consumer-ish motion) is that AI-native challengers compress pricing power faster than incumbents can retool. Adobe's guide is the first major vendor to put numbers on it. If you're long the SaaS basket, this is uncomfortable. If you're short the SMB/consumer-facing names with freemium exposure, this is the receipt.
Bottom line for PMs: $200 is technically cheap on FCF (7x CY27) but "cheap" is not the same as "bottom." Need to see at least one of: (1) new CEO/CFO named, (2) freemium conversion data, (3) stabilization in net new ARR. Until then this is a fade-the-bounce, not a buy-the-dip, until proven otherwise. The $195-225 zone is the bear fair value cluster; a flush below $195 with no strategic update changes the math.
Verdict: The Street's buy-and-hope-2H-hits camp is leaning in. Five firms in one week, all Buy, PTs $220-280 vs $164 print — that's 34-70% implied upside from a stock that's already down 34% in six months. The trade is the M&A spree + Agentforce inflection. Risk is nobody believes the 2H reacceleration until it's in the print.
Five reiterations this week, all Buy/Outperform. PTs cluster $220-250 with Truist the bull outlier at $280 and Wolfe the floor at $220. Consensus feels like ~$240 (roughly 45% upside). Wolfe/Stifel/Jefferies/Canaccord/TD Cowen all weighed in, with Truist's $280 lurking from a prior note. That's a tight, bullish cluster — not a wide distribution. When everyone lands Buy with PTs inside a 30% band, the real question isn't direction, it's sizing the 2H beat.
CRM is buying Fin (formerly Intercom) for $3.6B CASH — that's ~9x trailing ARR on $400M, or 6x CY27 EV/Sales per Jefferies. Fin's AI Agent alone is $100M ARR growing 3.5x with a 76% autonomous resolution rate. Deal closes Q4 FY27, no change to FY27 guide, no impact to buyback. Cash balance post-close: $6.7B at end of CY26.
This is the 15th M&A deal since May 2025 (Informatica $8.3B, Qualified $1.5B, Contentful $1.5B, m3ter, now Fin — roughly two dozen bolt-ons). Stifel framed it cleanly:
"Salesforce has completed roughly two dozen acquisitions since the beginning of 2025 to support its Agentforce platform... [building] data capabilities and [expanding] agentic features across service, sales, and marketing functions."
The bull read: Fin plugs a time-to-value gap Wolfe and Jefferies both flagged in channel checks — SMB/mid-market customers need packaged, fast-deploy agents, not heavy implementations. Canaccord's read is the sharper one:
"Customer service appears likely to be one of the first enterprise functions where AI meaningfully replaces human labor, and Fin gives Salesforce one of the stronger and more quickly scaling assets at the low-end of that market."
Translation: CRM is buying its way into the autonomous-labor transition before ISVs get disintermediated. Smart if the thesis holds, expensive if it doesn't.
BULL (TD Cowen, Stifel, Jefferies, Canaccord, Truist): Agentforce at $1.2B ARR up 200%+ YoY — that number is doing real work here. Q1 showed inflecting AI momentum, FDE (Forward Deployed Engineer) deployment model differentiates vs pure-play AI natives, headless architecture + Slack as the conversational layer is a real moat. 2H reacceleration, Rule of 50 by FY30, trades at ~11x EV/FCF on CY27 with a 0.49 PEG. Cheap on cash flow, growth re-accelerating, buyback intact. TD Cowen's investor dinner takeaway was a tone call: bullish into 2H and beyond. 24 analysts have revised FY27 earnings upward per data referenced this week.
BEAR (the price action is the bear case): Stock at $164, 52-week low $161, down 34% in six months, down 8.9% in the past week alone. The market is pricing structural AI displacement risk — software-assisted workflows → autonomous digital labor means SaaS seat counts get crushed. CRM laid off Agentforce/Mulesoft/Marketing Cloud staff this month (ironic given the AI buildout). Integration risk on a 15-deal M&A spree in 13 months is real — Informatica alone is $8.3B. Time-to-value concerns haven't gone away; Fin helps but doesn't solve the enterprise deployment tax. The bear doesn't need to say anything — the tape is saying it for them.
This is the cleanest expression of the "old-software-buys-AI-native" trade in the tape right now. CRM is paying 9x ARR while private AI customer-service comps trade at multiples multiples of that. Whether that's a bargain or a value trap depends entirely on whether autonomous agents expand the TAM or just cannibalize the existing seat base. The Fin deal skews SMB/mid-market — that's the segment where agentic AI is shipping fastest and where incumbents are most vulnerable. CRM is trying to be the incumbent that buys its way out of disruption. Watch the next 2-3 quarters for: (1) Agentforce ARR (need to see $2B+ run-rate), (2) deal integration noise on margins, (3) any 2H guide-down — that's the kill shot. Until then, cheap on FCF, heavy M&A, and the Street is willing to wait.
The DASH trade worked. Now what? Stock's up 70%+ YTD sitting at $231, the Street just clustered PTs around $260, and the only question is whether you're fading the multiple or riding the AI complexity tailwind. We're in the camp that this is a positioning story more than an estimates story right now — BMO literally said the AI narrative is driving the stock, not revisions. That cuts both ways.
DASH was a tour de force on paper. CEO Pomel, CTO Lê-Quôc, and CPO Li walked through 1,000+ new features since June 2025, including 100+ launches at the show. The headliner is Bits AI — pitched as an agentic workflow engine that's morphing into an "autonomous operations layer" across observability, security, and dev. AI breakout featured Anthropic, OpenAI, Hugging Face, and Vercel. Customer logo wall looks strong.
The hard data: LLM Observability spans up 30x+ YoY. That's not a rounding error — that's a product category going from zero to something. Management is leaning into the complexity narrative hard, and frankly, the message is landing because the world IS getting more complex (AI agents, multi-cloud, model sprawl).
PT cluster is $260 (BMO, Benchmark, TD Cowen), with Wolfe the high at $295 and Cantor at $226. Bernstein the lone holdout at $180 — that's a -22% downside tag on a stock that just ripped. Spread is wide, which tells you conviction is anything but uniform.
BMO's mechanics are worth a look — they took their multiple to 18-19x EV/FY27E rev (from 15-16x) and 68-69x EV/FY27E FCF (from 57-58x). The (EV/FCF)/growth ratio went from 2.1-2.2x to 2.5-2.6x. They're paying up for durability. When your bull-case broker is raising the multiple by 3 turns of revenue, you're not buying estimates — you're buying narrative.
"AI Ushers in Increased Complexity, a Positive for DDOG" — the framing from the post-DASH note, and it's the right way to think about DDOG's wedge against the bears.
Bull: AI complexity is structural, not cyclical. 30x in LLM Observability, platform expansion into security and dev, 80% gross margins, and the only true observability OS story out there. Wolfe's $295 implies ~28% upside; if Bits AI lands as the "autonomous ops layer" narrative suggests, this multiple is sustainable through FY27.
Bear: 54x NTM EV/FCF is priced for a quarter of perfection every quarter. Multiple expansion did the work in 2025-26, not estimates. InvestingPro fair value flags it as overvalued. Bernstein at $180 is the canary — if growth decelerates a single beat from the 29% LTM print, air pocket is violent. And the AI-leader basket (DDOG, MDB, SNOW) trading at 54x FCF vs. 13x for ADBE/CRM/WDAY means the regime trade unwinds ugly if the AI capex narrative cracks.
We're not fighting the tape here — the trend is your friend and AI complexity is a real tailwind. But at 54x FCF post a 70% YTD run with Bernstein as the designated skeptic, the r/r skews neutral. The 260 cluster is base case already priced; you need Wolfe's 295 view to make new money work, and that requires believing the multiple HOLDS through 4Q earnings. We'd rather buy a 5-8% pullback into 2Q prints than chase here.
Verdict: Street just re-rated the framework, not just the numbers. PTs aren't moving 20% — they're moving 130-200%. That's a regime change, not a recalibration. Still, at $1.2T mkt cap with the stock already +751% YTD, the easy money is behind us. Fresh capital here is paying for CY27/CY28 perfection.
MU trades $981.61, $1.2T mkt cap, +751% over the past year. LTM rev growth 85.6%. PEG of 0.11. This isn't a stock story anymore — it's a memory cycle / AI capex story. The Street just made a fundamental shift in how it values the name, and the price action into that shift is what's creating the melt-up.
Six firms raised PTs in a tight window, and the magnitude is what jumps out. Prior cluster was $525-$660. New cluster: $1,200 (RBC) to $1,750 (Susquehanna), with a $1,500 median (TD Cowen, DA Davidson) and Aletheia/Wolfe in between. That's a 130-200% lift in targets in a single print cycle — almost unheard of.
What's actually driving it: sell-side migrated from peak P/B to a P/E framework anchored on 10x CY27 EPS. That's the structural change. Aletheia and RBC both explicitly call out the 10x CY27 multiple. TD Cowen's $1,500 implies roughly 10x their $150 CY27 EPS estimate. The Street has decided memory is no longer a cyclical commodity to be valued off trough book value — it's an AI infrastructure comp.
Underlying assumption set across the board: DRAM prices +30-40% Q/Q in 3Q cal '26, HBM ASPs doubling YoY in cal '27, memory content in AI hardware rising from mid-40s today to >70% by cal '27 (Aletheia — certain racks north of 90%). RBC calls out the DRAM upcycle is now in its 12th quarter vs 8-9 in 2014 and 2018, and they see a credible path for 5-6 more. Supply muted until late '27 on clean room constraints + HBM conversion absorbing wafer capacity.
TD Cowen with the framing line: "the role of memory in artificial intelligence is structural rather than cyclical." That's the entire bull case in one sentence — and it's coming from the firm that also flagged we're closer to peak than not. Take it seriously.
Bull — Memory is a tax on AI capex and Micron is the best-positioned non-Korean HBM supplier. DRAM content per gigawatt is rising structurally (even after SOCAMM de-specing per TD Cowen). Agentic AI + inferencing shifts the workload mix toward memory-heavy. Aletheia models $350-400B cumulative FCF FY26-28 — a cash generation profile that justifies 10x forward earnings for a capex-light mature cycle. 9 analysts revising EPS higher in a single week.
Bear — TD Cowen themselves said it: in a typical cycle, memory stocks peak 3-8 months BEFORE server pricing YoY peaks, and they expect server pricing to peak ~3Q cal '26. That window is closing now. You're paying 10x CY27 EPS, which means two more years of execution priced in, on a stock that's already done 7.5x in a year. Susquehanna at $1,750 is the high-end outlier and even that implies the market is buying the structural re-rating. PEG of 0.11 isn't "cheap" — it's a tell that the multiple has fully captured the growth. New long money here is buying peak EPS into a cycle that's already the longest on record.
Management commentary on long-term agreements. TD Cowen flagged this as the key catalyst — the Street needs visibility on locked-in HBM pricing to sustain the re-rating thesis. Bechtel partnership on the NY fab (Clay, White Pine Commerce Park) advancing is a tailwind for the supply story but a 2027+ capacity unlock — doesn't help near-term tightness, which is bullish for prices. Dr. Alexis Black Björlin (ex-NVIDIA, Meta, Broadcom, Intel) to the board is a quiet signal on AI infra positioning.
Constructive into the NDX inclusion with fresh Overweight calls, but this is a Neutron-binary name now. The June 12 selloff (-11%) into a catalyst cluster is the kind of setup that rewards PMs who can size for beta 2.5.
KeyBanc made the bigger call today - upgrade to Overweight from Sector Weight, $135 PT (~40% upside from $102.39). That's a RATING CHANGE, not a PT tweak, and it carries more weight in our book. Stifel already moved to $132 from $110 (Buy maintained) on revenue momentum + backlog. Cantor reiterated Overweight, leaning on the NDX inclusion as material - the index is tracked by 200+ products with $800B+ AUM, so the passive bid starts June 22 whether you like the stock or not.
"Gained additional visibility on the Neutron program over the past six months and a January testing anomaly has been fully resolved." — KeyBanc
That line is the actual story. Read it twice.
Two analysts converged on the same point: Neutron stays on track for first flight later this year, and the Jan test anomaly is behind them. That's structural, not narrative. Neutron is what unlocks the medium-lift TAM and justifies the premium after a +303% YTD run. Until it flies, this is the bogey - everything else is wallpaper. The combined satellite design + manufacturing + launch stack is the bull case moat; KeyBanc even floated the optionality of RKLB building its own constellation down the road. Cantor echoed the same vertical-integration thesis.
Bull: Neutron flies on schedule → instant TAM expansion into the medium-lift market. NDX inclusion = mechanical passive bid starting June 22. 90+ Electron missions = proven cadence. Vertical integration is unique among small/mid-cap space. SpaceX IPO at $2T+ valuation is rerating the comp set, and RKLB is the cleanest beneficiary with public liquidity.
Bear: Beta 2.5 tells you the buyer base is fast money. 11% one-day drop on no obvious catalyst means the floor is thin if Neutron slips. After a 300% run, multiple compression cuts both ways. SpaceX shadow is real - any update from them reframes the comp discussion overnight. (CFO Adam Spice joining Syntiant's board reads as networking, not heads-down on Neutron - small tell, probably noise.)
Bullish lean into the June 22 inclusion with a Neutron catalyst window opening later this year. Two Overweight calls in two days isn't nothing, and KeyBanc's rating change is the one that matters. Just don't forget the beta.
Needham kicked off coverage with a Buy and a $55 PT — 70x CY27 EPS, ~20% upside from $46.03. The angle is Physical AI, positioning CEVA as the picks-and-shovels play beyond compute and memory: cars, drones, robots. Stock's already up 131% over the past year, so the "easy money" initiation is arriving after the trade. Rich setup.
"If Physical AI becomes a reality, upside potential remains for the stock despite recent appreciation."
That hedge tells you everything. Needham likes the name, but they're not pretending this is undiscovered.
Q1 was solid and licensing did the heavy lifting:
Bull: 68% global share in wireless connectivity IP. Sensing and DSP/NPU exposure gives you a basket of Physical AI adjacencies before the category is even real. Royalty stream is the long-tail compounder. 87% GM means operating leverage once volumes hit.
Bear: 70x CY27 is a lot of Physical AI priced in for something >2 years away from adoption. Stock's up 131% YTD-ish — positioning looks crowded into a name that just got initiated by a shop whose PT is the high on the Street. If licensing normalizes next quarter (Q1 had pull-forward potential from those 14 deals), the multiple gets ugly fast.
Read: The licensing quarter was real, the Needham call is a tailwind, but chasing here with 131% already printed and 70x forward feels like paying for a movie that doesn't come out until 2028. Tactical add on weakness > strategic position. Not sure we can read too much into one quarter of licensing strength, but the Q-Q trajectory is what we need to watch next.
Mizuho sticks with Outperform, $320 PT — and the capex number is the relief trade, not the worry. ORCL guided FY27 gross capex of $90-95B, netting to ~$70B cash outlays. Incremental financing need: $20B on top of the $50B disclosed in Q3. So ~$70B total raise vs the $100B Street worst case management dismissed back in Q2. Equity holders exhale.
Abilene supercluster came online on schedule, IaaS accelerated to +90% YoY in Q4. Q1 alone brings 1GW of new capacity — matches all of FY26. Mizuho calls FY27/FY28 the peak capex years. That's a ceiling on the financing pressure. Bring-your-own-cloud + prepayment model is the self-funding angle; if it works, equity dilution risk shrinks materially. (And keeps ORCL off the "death spiral" narrative that haunted the stock into the print.)
Mizuho: FY27 revenue guide is a "conservative starting point." Revenue +34% in FY27, building on +17% last year. Classic hyperscaler J-curve.
"Mizuho views the fiscal 2027 revenue guidance as a conservative starting point and sees the bring-your-own-cloud and prepayment model helping Oracle reach a point where it self-funds its growth."
RBC at $190 (Sector Perform) reads it the other way — flagged "mixed results in cloud revenue growth." And SaaS printed +9% vs the 12% expected. The legacy software drag is real. Bears will argue IaaS is a rocket ship bolted onto a decelerating $190B-ish revenue base, and the multiple compresses the second FY28 numbers wobble.
PT cluster post-print: TD Cowen and KeyBanc at $300 (cloud momentum), BMO at $220, Piper at $225 (RPO expansion), RBC at $190 (lone holdout), Mizuho at $320 anchoring the high. Wall Street consensus sees ~33% upside.
+90% IaaS YoY is a number you have to respect. The question is whether you fade the SaaS drag or chase the cloud build. Longs at the $300+ cluster look air-tight on a 2-year view. Bears own the multiple if FY28 prints disappoint — and the $20B incremental raise keeps leverage in the conversation either way.
The take: MS steps in with an Overweight at $371 PT and the bull case essentially writes itself — process control names have been working, ONTO sits at the intersection of advanced node spend AND advanced packaging AND a margin re-rating story. Three different buckets all firing at once is rare. Street's now clustered $350-371 (DB $350, Stifel $350, MS $371) — that's the bogey band for this name.
The outgrowth story is the core pitch. MS models ONTO advanced node revenue +30% in 2026 vs WFE +27%, then accelerating to +39% in 2027 vs WFE +28%. They're outgrowing the spending pool they're swimming in. Specialty Device and Advanced Packaging is the rocket ship here — +46% in 2026, +38% in 2027 — driven by Dragonfly G5 traction and Semilab inorganic contribution.
Gross margin path: 54.6% (2025) → 58.0% (2027) → 58.3% (2028). That's a ~400bps expansion in 24 months. Current LTM is 54.19%, so this isn't fully baked. MS pins it on Asia manufacturing shift plus stronger front-end and software mix. And here's the multiple math that matters:
"Semiconductor equipment companies with 60% gross margins trade at a 24% valuation premium to those with 50% gross margins." — Morgan Stanley
ONTO going from low-50s to high-50s is literally the multiple-expansion flywheel. Software mix is the part to watch — that's where the stickiness and the real margin dollars come from.
ONTO priced a $1.3B convert (UPSIZED from $1.1B planned), 0.00% coupon, due 2031, initial conversion $381.80 — a 50% premium to the $254.53 reference price. Zero-coupon, long-dated, upsized demand. That's management telling you they see a path through $381. Translation: they're funding M&A or capex into a setup where they think the stock re-rates. Worth noting the convert ref is well above current $323 — implies convert arb holders are underwriting a move higher.
DB and Stifel both Buy $350 — Stifel's specifically called out ACCELERATED Dragonfly G5 adoption. That's the product cycle nobody's arguing with.
Bull: Three growth vectors compounding (advanced nodes, advanced packaging, Semilab) + 400bps margin expansion + multiple re-rating. The convert tells you insiders see it too.
Bear: Stock at $323 already prices in a lot of the good news. WFE itself could disappoint if memory capex rolls over. Advanced packaging adoption can stall if AI accelerator architectures shift toward alternative packaging (CoWoS-L variants, etc.). Not sure we can read too much into the Semilab ramp until it prints for a couple quarters.
Positioning read: ONTO's been a steady compounder, not a meme stock — but the convert and the MS initiation clustered with DB/Stifel at $350+ is the street getting more constructive together. The $381 convert strike is the real tell on where smart money thinks 12-18 months out.
Verdict: UBS reiterates Buy, $485 PT, but the interesting stuff is what's NOT in the print — Google TPU v9 slipping on back-end supply and the SPV is just formalizing the existing book, not net new.
UBS came out this morning backing the bull case — $485 PT vs $381 print, 27% upside. The headline driver is the 20GW SPV with Apollo/Blackstone for OpenAI and Anthropic, which sounds explosive but on closer read is largely a financial repackaging of engagements already in the model. UBS explicitly says these are "already existing engagements" they helped size for CY27. So don't confuse this with new TAM.
The real meat is the 2028 math. UBS pegs 13GW of OAI/Anthropic shipments in CY28 at $10-15B per GW for ASIC + networking, implying $130-190B of revenue from those two customers alone. Add $73B for GCP TPU and $20B for META and you're at $223-283B in 2028 AI revenue. For context, current quarterly revenue is running $22B stepping to $29.4B next quarter. The trajectory is real but that's a LOT priced into $1.81T market cap (52% YTD, btw — positioning is not light here).
"UBS believes the SPV helps formalize capacity and financing structures but is capturing already existing engagements with OpenAI and Anthropic that Broadcom helped size for calendar 2027."
Quick take: stock's had a monster run, UBS math on 2028 is aggressive on $/GW, and the Google v9 delay is the first real crack in the narrative. Erste upgraded today too (Q3 AI semis at $16B), and BNP is the outlier at $640 PT. Consensus clearly still catching up — 22 upward EPS revisions per InvestingPro. Trade it, but respect the size of the move and the fact that the SPV is financial engineering, not new revenue.
Verdict: FROG keeps punching higher into a governance-for-agents narrative — and the Street is leaning in. Cantor reiterates OW with an $80 PT, sitting just above the $77.74 print. Stock is up 91% YOY already, so the easy money's been made, but the thesis is evolving beyond pure devsec tools into agentic AI governance. That's the right side of the trade for now.
The catalyst: JFrog dropped a plugin for Anthropic's Claude Code this week, giving Claude agents direct access to its artifact repo, security scanning, compliance controls, MCP registry, and skills registry. In plain English — every dep, package, and AI asset gets governed in real time before it ships. Smart land grab while the agentic dev stack is still forming.
Cantor views the announcement as reinforcing JFrog's effort to position itself as "the system of record and governance layer for agentic software development."
That's the right framing. If agents become the primary code author, you NEED a governance layer, and FROG is planting the flag early. Not a zero-to-one product launch, but a strategic moat extension into a TAM that's expanding fast.
PT cluster is tight and bullish: Cantor $80, Stifel $85, KeyBanc $86 — all Buy/OW. Stifel raised to $85 off a customer survey flagging cloud migration + security adoption; KeyBanc at $86 cites cloud growth durability and AI-native cohort potential. Street's in the $80s, stock's in the high $70s, so there's some room but not a lot. Russell 3000 inclusion June 26 is incremental mechanical bid — index funds need to own it, small impact but real.
Bear case (and it's real): stock is up 91% in a year, valuation is stretched, and the platform's own Fair Value model flags overvaluation at current levels. The Anthropic integration is strategic but not yet revenue-accelerating — this is positioning, not a print. If agentic dev adoption slows or competitors (GitHub, GitLab, Snyk) bridge the gap faster, multiple compression hits a 90%+ YOY winner hard.
Positioning take: FROG is a quality compounder with a fresh narrative. Trim risk into strength rather than chase here. The governance layer thesis is legit but the easy re-rating is done. Wait for a pullback or a print that proves the AI agent story is translating to ARR.
Street chasing a runaway. B.Riley just took PT to $345 from $240 (Buy) — Stifel $321, Benchmark $275, Cantor $220 (Neutral). Whole cluster trying to catch a name up 230% YTD vs SOX +86%. After-hours -1.5% reads as digestion, not thesis break.
"Jensen Huang called Marvell the next $1 trillion company."
That's the tape. When your biggest customer publicly frames you as the next $1T company, you're either printing or you're the bag. Given the NVDA partnership spans Custom silicon enablement, NVLink Fusion, and Celestial in-house optics — MRVL is basically the picks-and-shovels for the entire NVDA franchise — we lean print.
CFO swap (Durn in June 15, Meintjes out after 10 years) is noise. S&P 500 inclusion on June 22 is a real mechanical bid — passive trackers have to own it. 26 analyst EPS revisions higher, PEG 0.15, options vol 605K contracts with heavy call skew. Extended? Sure. But momentum doesn't care about your feel. Not selling into this bid.
Verdict: Stock at $511 against a $450 Wolfe PT — somebody's got this wrong, and it ain't the tape. Wolfe reiterated Outperform but the math here is awkward. Either they lift fast or they're marking to a print that already left their model behind. After a 340% run, AMD is the cleanest expression of the "AI content per rack expands every gen" thesis — and July's MI500 event is the next test.
July event = MI500 details. Then MI-600 deployment 12-15 months after that. The pitch is straightforward: PRICE GOES UP EVERY GEN because rack performance improves by a multiple with each successive generation. Power draw rises too, but AMD's claim is that dollar per gigawatt still expands — so hyperscalers pay more absolute dollars per rack, and AMD's slice gets fatter. That's the bull case in one sentence.
CPU is the underrated leg. Wolfe calling for 30% AI-DRIVEN CPU MARKET GROWTH THROUGH 2028, with TSMC capacity as the gating factor. Whoever secures the wafers wins. AMD positioning as the #2 behind INTC (such as it is) in AI-exposed CPU is real optionality that the market under-discounts vs. the GPU story.
"AMD's central processing unit opportunity adds to this content expansion."
— Wolfe Research, the most important throwaway line in the note. The street still frames AMD as a GPU duopolist. The CPU angle is the sleeper.
Stock at $511 with street PTs clustered in the $400s (at least Wolfe's). The July MI500 event needs to deliver, not just tease. If hyperscaler order commentary underwhelms or pricing leverage narrative gets challenged, the multiple compresses fast. NOTHING CHEAP HERE. Positioning likely crowded given the run. R/R into July event is OK but not screaming buy. BTDs beware.
THE READ: ANALYST DAY WAS A PLATFORM REPOSITIONING, NOT A QUARTERLY UPDATE. RBRK used the inaugural analyst day to pivot the narrative from data backup/recovery to cyber resilience + identity, with agentic AI as the next S-curve. Wall Street bought it — 4 known PTs cluster at $90-110, all Buy/Overweight. Baird most aggressive at $110. Fundamentals still backstop: 80.58% GM, 45.73% LTM rev growth.
The "we're a data security company" pitch is dead. RBRK is now telling the street it's a platform play spanning data + identity, with Rubrik Flex (new licensing structure) designed to lower friction for multi-product adoption. Multi-product attach is the key — that's what turns a backup vendor into a platform. Management called out "consistent execution and strong multi-product expansion" alongside the financial update, which is the real tell. Revenue growth hasn't broken, margins haven't broken, and they're using the analyst day to plant a flag on a bigger TAM.
The agentic AI angle is interesting but still aspirational. RBRK Agent Cloud for Anthropic's Claude (agent rewind, immutable codebase recovery) is a real product, but the "S-curve" framing is forward-looking. Worth flagging — this is a 2027+ story, not a near-term catalyst.
"The firm came away incrementally positive on the company's broadening opportunity and overall strategy." — Truist's Junaid Siddiqui
JPM just went to EUR18 from EUR12 — 50% raise, OW maintained. Easy money's been made (NOK +170% in 12 months to $14.88), but JPM's not done. This is a "consensus is still wrong" call, not a chase.
The guts: JPM is now modeling 2028 EBIT 58% ABOVE THE MIDPOINT of Nokia's own Nov-2025 guide (EUR2.7-3.2B). That's the headline number. When your 3-year EBIT is ~60% above company guidance, either you're right and they're sandbagging, or you're going to get clipped when reality hits. JPM's argument is that the optical market strength wasn't on the radar in November, and NOK hadn't yet locked in new switch design wins. Fair — but that's also why the stock ran from €6 to €15 already. The market's already pricing a lot of this.
"JPMorgan said the strength of the optical market was not anticipated in November and the company had not obtained new switch design wins at that time."
Sales estimates raised 1.7% / 5.2% / 10.2% for '26/'27/'28 — back-loaded, as it should be on an optical ramp. EBIT revisions are 1.8% / 6.0% / 40.1% ahead of consensus, so the real delta is operating leverage in the out-year. PEG of 1 at 29x P/E is the framework — implies they're treating this as a growth name now, not a 4-5% yield/value telecom gear story. Big regime change in how the Street is supposed to model this.
Argus also chimed in separately, upgrading Hold→Buy, $15 PT, citing Q1 EPS beat and AI networking capex. Different analyst, same narrative — optical + AI datacenter buildout is the trade.
Bull case: NOK is a legit optical beneficiary of the AI datacenter build (coherent, 800G/1.6T, DWDM). New switch wins are the underappreciated catalyst. Restructuring/mix shift drives margin expansion into '27-'28. JPM isn't alone — Infront consensus is still way behind.
Bear case: Stock up 170% in a year and still people want to buy it? Q1 revenue MISSED (EUR4.49B vs EUR5.32B expected) — yeah, constant currency was +4%, but that print was horrid on the top line and the market's looking past it. 29x P/E for a gear vendor with lumpy cycles is rich. And if optical/AI is the thesis, why not own COHR, LITE, or even MRVL where you get cleaner exposure? NOK is the conglomerate play on this theme, which means you get the optical upside diluted by mobile infra and fixed networks drag.
If you're long, you hold — the JPM print validates the AI/optical narrative into a quarter where the revenue number was genuinely weak. If you're flat, hard to chase a name up 170% even with a 50% PT raise, especially when your upside from spot to PT is ~20% on a stock that just did 6-bagger move. Not sure we can read too much into Argus's $15 PT (basically spot) — that's a hold-in-disguise relative to JPM.
The real question: does JPM have the right to be 58% above company guide in 2028? If yes, €18 is conservative. If no, this is a sell into strength situation by year-end. Watching the optical order book and any new hyperscaler design win announcements as the next catalyst — that's what moves the stock from here, not incremental PT raises.
UBS nudges PT to $5.50 from $5.00 and stays Neutral — but the stock's at $5.93, so the Street's bogey is now BELOW the print. Translation: even the most constructive analyst update can't catch up to the 477% YTD rip. That's your first tell.
Two things, neither of which is the fundamentals:
1. Narrative pivot to AI services. Rackspace announced a Riyadh regional HQ (UBS flagged it as a sign of regional enterprise cloud demand), an AMD MOU for governed Enterprise AI Cloud infrastructure aimed at regulated/sovereign workloads, and the existing Palantir Foundry/AIP partnership. UBS called the shift "encouraging" — moving from commodity cloud infra to AI implementation and managed services.
2. Short covering + low float dynamics. The article explicitly cites short covering as a contributor to the move. With a beta of 3.05 and the stock up ~5x YTD, you don't need to invent a fundamental story — this is mechanical.
"UBS said it is encouraged that Rackspace is shifting to AI implementation and a managed services partner from commodity cloud infrastructure provider. The firm remains Neutral-rated amid limited near-term revenue contribution and an uncertain environment."
That's the bear case in one sentence from the bull themselves.
Bull: AMD sovereign AI cloud + Palantir distribution + Riyadh hub = real optionality on Rackspace becoming the "boring plumbing" layer for regulated AI workloads. BMO sees it (PT to $5.00, Outperform). Cheap stock, real partnerships, narrative tailwind.
Bear: Q1 was ugly under the hood — revenue $678M beat the $675M whisper but EPS came in at -$0.06 vs -$0.03 expected. Unprofitable, no path to profitability this year per consensus. The AI deals are MOUs and press releases, not revenue lines yet. And the stock trades above UBS's PT, which means even the bulls who updated are underwater on their targets.
Momentum trade running on narrative and short covering. Partnerships are real but small in the revenue mix. UBS staying Neutral with a sub-print PT after a 477% move tells you the analyst doesn't believe the multiple here. Not a long — it's a fade candidate or a watchlist name for the next leg down. If you missed it, you missed it; if you're in it, size matters more than thesis.
Northland moves to Outperform, $25 PT (from a prior level not fully disclosed) — call it ~29% upside to the $19.37 print. The upgrade is a look-through name, not a quarter. They're saying Q3 FY26 was a write-off and the real setup is FY27.
THE ORDER MATTERS. $25-30M order from a current customer, Northland reads as very likely Verizon FWA MDU build. If that read is right, this is a multi-quarter revenue tailwind, not a one-shot. Combined with BEAD funding starting to flow and private network expansion, the FY27 acceleration story has real legs. Aviat's microwave backhaul is exactly what fixed wireless operators need as densification picks up.
But let's not get cute about the print — Q3 was awful. EPS of $0.06 vs $0.48 expected (87.5% NEGATIVE SURPRISE), revenue $100M vs $107.53M. That's not a soft quarter, that's a broken one. Anyone underwriting the FY27 ramp needs to figure out what went wrong in March and whether it was timing, project slippage, or a structural issue. The product expansion into all-indoor microwave (Pasolink + IRU600) and the new long-haul international play targeting a $250M+ TAM is incremental optionality but doesn't fix a quarter that badly missed.
"Northland believes this order is very likely in support of major customer Verizon's fixed wireless multi-dwelling unit rollout."
R/R: $19 in, $25 PT, downside to maybe $15 if FY27 doesn't inflect. That's not a core long — this is a small, illiquid name where a 30% move requires the narrative to actually deliver. The setup is there, but the stock is a SHOW ME trade into FY27 prints, not a buyer here on size. Needham conference on June 23 is the next catalyst window.
Not sure we can read too much into one order being a marquee customer — small base effect distorts the optics.
Verdict: Roadmap sells the story, but the chart says wait. Mizuho lifted PT to $35 (from $29) post-Analyst Day, holding Outperform — now you've got a street cluster of $35-43 (Stifel $35, Rosenblatt $43). Stock at $23.37 is STILL 50% OFF ITS 52-WEEK HIGH OF $46.75. The name trades on roadmap, not earnings, and the roadmap just got a refresh.
D-Wave is repositioning as a full-stack, dual-platform player — annealing (where they lead) PLUS gate-based (where they're playing catch-up). Updated gate roadmap: 10 LOGICAL QUBITS BY 2030, 100 BY 2032. They demonstrated error correction cycles under 5 microseconds, which is the kind of detail that gets the physics PhDs on Twitter excited and does nothing for next quarter's revenue. TAM framework they cited: $450B-$850B BY 2040, with gate-based addressing ~75% of that. Big number, long timeline — that's a 2030+ trade, not a 3-month trade.
The classical simulation challenge to their supremacy claims is the elephant in the room. D-Wave defended, said the sims didn't replicate the full scope. Not sure we can read too much into that until peer review settles it — but it's a real overhang for a name priced for technological inevitability.
"Mizuho cited D-Wave Quantum's leadership position in annealing quantum computing and its roadmap to fault-tolerant gate-based quantum computing as reasons for maintaining its Outperform rating."
That sentence is the whole bull case in one line. Lead in annealing, credible path to gate-based, TAM is massive. The question is whether you're getting paid to wait — at $23.37 vs a $35 Mizuho PT, you're getting ~50% upside to the low end of the PT cluster, but you're fighting a stock that already proved it can give back 50% in a heartbeat. Size accordingly.
WIX at $45.91, need to talk about the FCF math because that's the whole ballgame right now. Needham just reiterated Buy at $80 (analyst transfer to Stefanos Crist), and their core argument is simple: fears about Core Wix ceding share to AI/vibe coding tools are overdone, and even in the downside scenario where that happens, Base44 re-rates higher vs private comps and the FCF machine still chugs. The 2027 numbers do the talking — Needham modeling $608M FCF, or $12.74/share, which is a 28% FCF YIELD OFF CURRENT PRICE. That includes a -$266M drag from Base44 investment. Strip that out and the core business is printing closer to a 33-34% yield. Either you're underwriting a melting ice cube or you're getting paid absurdly to wait.
The PT spread right now is ugly though. Benchmark $115 (still Buy, raised FCF outlook despite the guide-down), Scotiabank $90, Needham $80, UBS $58 (Neutral, cut on workforce reduction and slowdown), and RBC at $45 (cutting on AI headwinds and partner weakness). That's a $70 range from high to low — street can't agree if this is a value trap or a deep value setup. RBC's $45 is essentially the bear case being AI disintermediates the legacy editor, growth stalls, FCF doesn't materialize at scale. UBS is closer to "show me." The bulls are saying the cash conversion is the moat, not the ARR growth.
"Wix maintains a competitive advantage based on its drag and drop model and backend solutions." — Needham
Stripe Projects integration is a small positive — devs can now tap Wix's backend (commerce, bookings, CRM, events) through Stripe's CLI. Doesn't move the needle near-term but expands the developer surface area, which matters if Base44 is the long-term growth vector.
Setup into Tuesday's print looks crowded. Stock at $384.82, kissing the 52W high at $386.64, and the Street has sprinted to chase it — Stifel just ripped their PT to $430 from $290 (Buy), Raymond James at $425 (Strong Buy), BofA at $354. That's a $354-$430 cluster with the bullish end clearly winning the air war. JBL has DOUBLED since December, and the multiple has done the same — trading 25x FY2 P/E vs high-teens a year ago and high-single/low-double digits before that. The re-rate is real and now you're paying for it.
KeyBanc reiterates OW, $120 PT, top pick. Stock at $68.24, down 35% YTD — that's the setup you want to underwrite. Light coverage today but the tape is pretty one-sided constructive. Sell-side PT cluster sits $103-125, implying 50-80% upside from spot. Clean risk/reward if Max AI delivers.
TD Cowen $125, Truist $110, Freedom Broker $105, BMO $103 — all Buy/Outperform. Mid-$100s cluster, fresh off Q1 where revenue grew ~25% vs. 19% guide. Sub growth stepped up to a record (TD Cowen called it out). LTM revenue now $1.01B at 24% growth. Durability is not the question here.
KeyBanc hosted CFO Dave Sherry and IR head Jason Rechel. Came away incrementally positive on Max — the AI upsell product. Two things matter: (1) it drives meaningful ROI for customers, so the sales motion works, and (2) customers ramp 2x on subscription spend once they adopt it. That's the bull thesis in two lines. If Max converts even a fraction of the install base, the multiple expands fast.
"ServiceTitan is a durable software grower with upside potential to mid-20s% growth in FY27 and FY28." — KeyBanc
Bull: 25% grower with a fresh AI product cycle, sub acceleration, and a 35% drawdown that's created an entry. Mid-20s% growth visibility in FY27/FY28 per KeyBanc.
Bear: 35% YTD says the market is worried about something — duration, SMB exposure, multiple compression. The insider tape isn't helping either. CAO Michele O'Connor sold 10,000 shares at $69.14 last week. Small print, not a red flag, but not a C-suite buyer at the lows either.
Software grower, AI product story, sub growth accelerating, PTs imply 50-80% upside. Bear case is mostly multiple/sentiment, not fundamentals. Not a lot of coverage today but the setup is obvious — $68 with a mid-$100s PT cluster and a top-pick designation from KeyBanc. Easy to see why someone starts a starter position here.
Setup looks disconnected to us. Stock got hit 8% in a week despite a clean Q1 beat ($1.008B vs $965M consensus, +18% YoY) and a constructive setup into Q2. Guggenheim reaffirmed Buy/$24 this morning — and that's actually on the LOW end of the Street. Benchmark sitting at $33, Cantor/UBS just moved to $30. So you're getting ~50-65% upside to the high-end PTs while the stock acts like a broken name. Not a great look for a name with 10 consecutive quarters of double-digit global user growth.
Guggenheim's note leans hard on the monetization story, and it's the right call. Performance+ now represents 30% OF LOWER FUNNEL REVENUE, and here's the key data point — ADOPTERS GROW SPEND AT NEARLY 2X THE RATE OF NON-ADOPTERS. That's the kind of cohort behavior that compounds. SMB traction is real too — UBS flagged Performance+ gaining share with small/medium advertisers, which is exactly the bucket that's been undermonetized historically. tvScientific closes later this month and pencils in ~2 POINTS TO QUARTERLY TOP-LINE for the balance of the year (U.S.-only, some seasonality — call it a modest tailwind, not a home run).
Bull case: UCAN growth holds double-digits even as large retailers absorb tariff noise (huge tell on resilience). Performance+ becomes the default SMB buying tool, expanding TAM. Intl is the option value — currently a drag given GTM restructuring and leadership changes, but set up for a 2H reacceleration. Meta's premium subscription push is actually a relative tailwind for PINS — takes ad pricing pressure off the long tail of social.
Bear case: International is decelerating NOW, not later. That's a real top-line headwind in Q2 print. Meta launching ad-free tiers globally is a structural question mark on social CPMs broadly — PINS may dodge some of that given intent-driven search-like behavior, but the read-through is uncomfortable. Stock down 8% on no real news tells you positioning is soft.
"Performance+ adoption remains a key growth driver, currently representing 30% of lower funnel revenue, with adopters growing spend at nearly twice the rate of non-adopters." — Guggenheim
Tape is wrong, fundamentals are fine. We'd be buyers of weakness here — $20 with $24 floor and $30+ upside on Performance+ compounding. Watch intl commentary on next print as the tell on whether 2H reacceleration is real. Not a portfolio-killer if you're long; an add if you're not.
Bottom-fish setup, but the market's still right to be skeptical. TTD at $19.39, down 72% Y/Y, hugging the 52-week low ($18.31). Truist (Buy, $35 PT) flagged the Publicis settlement as a net positive with ad spend resuming in 2H26 — that's the bull case in a line. New CFO Nate Olmstead (Logitech/HP pedigree, starts July 9) is a clean operational add. But 12% Q1 revenue growth is a 2/3 haircut from the old algo, and Redburn's Sell initiation plus DA Davidson's PT cut to $29 say the supply-chain competition thesis has real teeth.
Q1 in brief: rev +12% Y/Y, +1.5% vs Street. Adj EBITDA ~5% above. Beat, but on an easy bar. Not sure we can read too much into a 12% growth print after the multiple compression this name has already taken.
Verdict: Guggenheim bumping PT 50% to $15 FROM $10 is the headline, but the 25% probability of success on ABS-201 tells you this is still a lottery ticket — albeit one with a $5B+ US TAM if it hits. Stock's already up 145% OVER THE PAST YEAR to $6.80, so a lot of optimism is partially priced.
The bull case: hair loss market is genuinely underserved, ABS-201 has multi-indication optionality through the prolactin franchise (endometriosis, immunology via ABS-202), and the asymmetry is real — $5B+ US / $8B+ EX-US unadjusted sales potential on a sub-$1B market cap. 13-week interim POC data drops 2H26, 26-week in early 2027. That's the catalyst path.
The bear case: 25% POS is honest, not generous. Q1 revenue was a trainwreck — $215K VS $1.43M EST, AN 85% MISS. This is a platform story, not an earnings story. And at $6.80 with 120% implied upside to Guggenheim's PT, the market is already sniffing something out.
"The hair loss market is underserved by treatments that deliver limited efficacy or have concerning side effects." — Guggenheim
Two things to watch: (1) any pre-readout data leaks or KOL chatter that moves the implied POS higher, and (2) cash burn — every quarter without data is a quarter of dilution risk. BTIG just initiated Buy at $9, so there's dispersion: STREET PTs RANGE $9-$15 and that's a wide gap for a name with one asset in the clinic. Position sized for binary, not conviction.
Specs or it didn't happen. That's the trade into tomorrow's Spiegel keynote at AWE — and the setup's decent. SNAP at $5.51, ~$9.2B mkt cap, profitable this year (~$0.61 EPS forecast), and B.Riley's $10 PT implies ~80% upside IF the AR Specs consumer launch lands. That's a big "if" but the asymmetry is real: B.Riley is out there saying a successful launch is "transformative" and adds a growth vector "not contemplated in the firm's or Street expectations." Translation — this is a call option on AR glasses that the Street is giving you for free right now. The stock's been dead money, sentiment is awful, and the long-term product is potentially mispriced. That's a setup PMs should at least be looking at.
The bull case is clean: SNAP has been building AR for years (lenses, developer ecosystem, now hardware via Specs), the stock trades like the optionality is worthless, and the Meta premium sub news shows the social ad market is still expanding into new monetization vectors. If Spiegel comes out tomorrow with actual consumer pricing, ship dates, and a credible dev ecosystem story, this is a re-rating event.
The bear case is the bear case: SNAP has disappointed on every "transformative" product hype cycle for 5+ years. Specs have been "coming" forever. The market cap is $9.2B for a reason — the company has struggled to monetize beyond ads, competition from Meta is unrelenting, and AR glasses as a consumer category remains unproven (see: Apple Vision Pro flop). B.Riley hedges with "cautiously optimistic" — they want credit for the call if it works, cover if it doesn't. Smart.
Bottom line: Light position, define your risk into the keynote. If you believe AR is a real consumer category in 24-36 months, SNAP at $5.51 is the cheapest way to play it. If you don't, this is a fade into any pop. Not the kind of name you want to be flat into a binary catalyst.
Verdict: Bull/bear dispersion is THE story here. Post-Investor Day PTs cluster $250-$300 from a $136 bear outlier — that 2.2x spread tells you conviction is paper-thin despite the platform tailwinds.
Citizens reit Outperform $270 (12% upside) following the June 9 Investor Day at the NYSE. Stock's had a rough go of it lately — DOWN 7.8% PAST WEEK — but still +16% YTD, handily beating S&P +9%. The week-long drawdown reads more like rotation/TMT chop than thesis break given what came out of the analyst day.
The post-event PT cluster is bullish on the surface: Truist and UBS both $250 (Buy/Neutral split), TD Cowen $265 Buy, KeyBanc $300 Overweight, Citizens $270. So 4-of-5 name $250+ with constructive framing — platform expansion, AI agent security narrative, rule-of-50 by 2027. That's the bull steelman done in one breath.
The bear is Bernstein SocGen at $136 Market Perform, and they're not just grumpy — they're flagging Vercel as a real AI inference/edge competitor. That's a 56% discount to the Street median and a 2.2x bear-to-bull multiple spread. When your bear is THIS bearish relative to consensus, either they're wrong or the next 3 quarters matter more than the last 6. Would want to see RPO growth, Workers/AI product attach rates, and any Vercel displacement data before fading the bear.
"Cloudflare's role in securing AI agents through its cloud-native architecture" — TD Cowen, $265 PT
Positioning question: stock got crushed 7.8% into/around the Investor Day, which is unusual for a "platform narrative intact" event. Either the June 9 print disappointed the buy-side in person, or the 3p tape just caught up. Watch the $136 bear case as the real floor marker — that's your "thesis is broken" level, not the recent low. Net-net, asymmetric setup favoring the longs at current levels IF you believe the AI agent security wedge is real. Not sure we can read too much into a 1-week move post-Investor Day, but the action is doing the opposite of the PT cluster, which bears watching.
VERDICT: Stock down 11% in a week into a clean beat and PT raises. That's positioning pain, not thesis pain. Trading $27.79 vs. a $35 Cantor PT, and the Street cluster is basically $32-36 (UBS $32, DA Davidson $35, BTIG $36, Cantor $35). PTs went UP after the print. The stock went DOWN. Reads like weak hands dumping into strength — or someone knows something we don't. Not sure we can read too much into a one-week move, but the 11% drawdown with no obvious catalyst is the tell.
Cantor reiterated Neutral at $35 with the stock at $27.79, basically conceding they're not paying for a re-rating here. The interesting stuff was the Transcend event announcements — four new capabilities, but the two that matter:
Nobody's saying it out loud, but the 14% RIF the Street keeps citing as a "positive efficiency story" is also a growth deceleration signal. Q1 revenue +23.1% beat by 3.6% — solid, not spectacular for a stock that traded at 15x sales at the highs. 87% gross margins (genuinely elite) but margins on revenue growth that may be re-accelerating OR may be peaking. Hard to tell from one quarter.
With 9-figure books I would've been trimming into the print, not adding. The setup is a stock that needs to prove agent monetization (Duo, Flex) actually drives a re-acceleration into FY27 numbers, not just narrative. Until then, it's a show-me story at a still-lofty multiple — and the 11% week suggests the market's patience is thinner than the sell-side community wants to admit.
BofA inits Buy, $34 PT on 9.5x EV/2028E revenue. PT cluster now $32-46 (Citizens $32, Bernstein $46) — street has re-rated WULF as AI infrastructure, not a miner. Stock at $26.06, up 548% over the past year, hugging the $27.78 52-week high.
The BofA note is the headline but the real tell is the $3.2B HY bond print — largest single-bank-led junk deal in 30+ YEARS, with Google stepping up to guarantee the debt once Lake Mariner comes online. You don't get a Google credit wrapper for vapor. Bernstein's framing on the capital-light lease model is the right way to think about this name now.
1.8-3.0 GW critical IT load target by 2030, Lake Mariner build by YE26, Muskie (Kentucky) initial deliveries H2 2028. Near-term catalyst is the Justified Data customer announcement — that's the next leg. Lake Mariner milestones are the receipts. Watch for the leveraged loan market exploration with Morgan Stanley too (more debt, more capacity, more capex risk).
"TeraWulf's strong order book and capital-light lease model... projected to drive rapid AI-revenue growth" — Bernstein
Bull: 1.8-3.0 GW pipeline monetizes faster than models with Google as anchor tenant; capital-light lease structure means ROIC scales without the typical data center capex drag. Bear: you're paying 9.5x forward 2028 revenue for a story that's two years out, the stock is up 548% in a year, and any AI infra rotation (and there will be one) hits crowded names like this hardest. Not chasing it here, but the data points are trending the right direction. Setup favors patient adds on weakness into the Justified Data catalyst, not a breakout chase at the high.
Setup looks interesting here. Stock's been destroyed — DOWN 48% YOY to $102 — and you have Benchmark bumping their PT to $130 (from $125) post-Investor Day chatter, calling it a "top large-cap value pick." Oppenheimer separately reiterating Outperform on AI reaccel into 2H26. The ask is whether the software de-rating is over.
NOW's K26 Investor Day laid out a 19.4% revenue CAGR from 2026-2030, landing at $30-32B topline, with 100bps of op margin leverage annually starting 2027. For a $100B+ mcap SaaS co, that's a credible growth-plus-margin algo — and it comes against a 76.6% gross margin and 21.7% LTM growth print, so it's not aspirational. The drivers: agentic AI, cyber, data, workflows.
"describes ServiceNow as a top large-cap value pick in its coverage universe." — Benchmark
PT bump is small ($5) but the framing matters — Benchmark is leaning into the "this got cheap enough" narrative rather than chasing a higher multiple. Reads as a sentiment tell more than a model update. Stock needs to absorb the Agentic AI monetization story (still TBD on actual revenue contribution) and prove the reaccel into 2H26 that Oppenheimer is underwriting.
R/R setup: ~27% upside to the Benchmark bogey, downside to the Lows if AI monetization slips. We've seen this movie with large-cap SaaS — but NOW's operating model gives it a margin of safety the zero-margin AI plays don't have. Worth a starter position on weakness. Not a chase here.
The article bundles NowVertical Group (different co entirely, small Canadian) — skip it. Also: OpenAI poaching CMO Colin Fleming from NOW is a one-line data point, not a thesis changer.
Wedbush came out Neutral with a $99 PT — basically telling you to sit on your hands here. Stock's at $95, so we're talking <5% upside on a name that's already been HAMMERED 55% IN SIX MONTHS. The thesis is clean and uncomfortable: AI is a cannibalization risk for the whole IT services space, EPAM's pedigree doesn't insulate it, and until organic CC growth re-accelerates to the ~7% peer average, the multiple stays compressed. Goldman's already downgraded this thing from Buy to Neutral, so Wedbush is late and confirming.
Call it a show-me story. Q1 print was a modest beat ($2.86 EPS vs $2.75, $1.4B rev vs $1.39B), but the marginal investor doesn't care about the print — they care about the growth trajectory and whether EPAM's 33-year engineering DNA and early ML adoption actually translate into differentiated AI monetization. Right now the answer is no. Peers have similar technical depth with BIGGER NA sales footprints, and at ~7x FY27 P/E (vs current 13.6), Wedbush is pricing in a structural reset, not a cyclical dip. PT cut from prior $215 to $110 by Goldman tells you the street's been doing the same math.
"We do not view the argument that pedigree alone insulates EPAM from AI commoditization as genuine... We view EPAM's AI positioning as competitive rather than differentiated until organic constant currency growth moves in line with the peer average of approximately 7%." — Steven Wahrhaftig, Wedbush
KeyBanc moves to OVERWEIGHT, PT $135. Stock $31.87. So the call is roughly 4x upside on a name that just got DRAMATICALLY cheaper. Read on.
The setup here is weird. KeyBanc is the upgrade — citing the $75M MoonFall NASA contract on the Elytra spacecraft as a marker for where the launch business can go. Consensus is already at 1.78, so this isn't a contrarian call, it's a lagging one. Fine. Revenue growth of 71% LTM accelerating to 175% projected FY26 is the real story. Execution on Alpha Flight 07 was solid, Eclipse in 2027 is the medium-lift option.
But here's the tension the upgrade doesn't solve: the stock is DOWN 12% IN A WEEK for a reason, and that reason is the secondary. Firefly sold 12M shares at $48 — 4M primary, 8M secondary — and the stock now sits at $31.87. That's a ~35% GIVEBACK from the offering print. Insiders/early holders monetized at $48, public buyers are sitting on a mark-down, and there's still a 1.8M greenshoe hanging out there. Classic post-secondary drift, not a thesis problem per se, but it explains the tape.
"Firefly ranks among the top commercial space companies with exposure to NASA's lunar initiatives... execution on the launch business could yield long-term upside." — KeyBanc
The bear case is the dilutive loop: capital-intensive launch business + early cadence ramp + multiple secondary offerings ahead = share count creep. Every dollar of revenue needs capex behind it, and that capex needs equity. The bull case is that Elytra and Eclipse eventually hit commercial scale, defense-adjacent wins (SciTec's $5.5M USAF option under the $24M ABMS program is a small but real proof point) layer in margin, and the stock re-rates to comps.
Not investment advice. Not a recommendation to buy or sell securities.
Roth/MKM kicks off coverage with a BUY, $35 PT — clean initiation, no consensus to fight yet. The angle: BETR is morphing from a cyclical DTC mortgage shop into an AI-native origination platform (Tinman, their proprietary stack) with distribution legs through Credit Karma and ChatGPT partnerships. Lower-cost origin story vs. legacy lenders if the platform thesis holds.
LTM EBITDA still -$94M (UGLY, BUT KNOWN) while revenue +54% LTM — Roth sees a credible path to adj EBITDA breakeven by Q4 2026. That's the pivot. Stock actually DROPPED in premarket despite a solid Q1 print, which tells you positioning is messy and the market isn't giving credit yet. R/R is genuinely asymmetric here IF you believe Tinman actually moves the cost-to-originate needle.
The Coinbase/Bitcoin collateral mortgage story is cool marketing, noise for the thesis. The thing that matters is unit economics on the AI platform — that's the Q4 2026 breakeven test.
"Better is positioned to be a structurally lower-cost winner as it transitions from a cyclical direct-to-consumer mortgage lender into an AI-native origination platform with distribution partners." — Roth/MKM
Not sure we can read too much into one initiation, but the $35 PT implies real upside from here and the catalyst path (Q4 breakeven) is dated — PMs who want the mortgage/AI platform trade should at least put it on the watchlist.
Verdict: TD Cowen sticking with the Buy on a name that's already cheap and starting to inflect. Stock at $21.87 vs $30 PT — that's ~37% upside on a multiple that looks optically wrong relative to the comp set. Reiterating a rating on a name you've been bullish on isn't exactly a catalyst, but the framework here is the thing worth stealing: BRZE at 1.9x EV/CY27 sales vs small/mid-cap SaaS peers at 3.7x. Almost a 50% discount for a name growing 20%+. The math closes by itself if the growth holds.
The setup is getting more interesting, not less. NRR rising, new logo velocity accelerating, and — the one that matters for the next leg — AI monetization through new product offerings expected to begin NEXT YEAR. That's the optionality nobody's paying for. The print itself was a clean beat-and-guide on the top line (30% YoY rev growth vs 26% consensus, top-line guide raised to 22% YoY), with the usual margin hiccups (5.0% op margin vs 5.2% expected, gross margin 67.4% vs 68.8%). Margin noise, not a thesis-breaker.
Street PT cluster is $30-38, with Mizuho the recent skeptic — cut to $32 from $40 citing valuation. So even the bear frame is "expensive, but not that expensive." Cantor sits at the high end with $38, Stifel and TD Cowen anchor at $30. Mizuho's cut is the one to watch because it means even the bull camp is recalibrating. If BRZE can string together two more clean prints and start showing the AI monetization slip into the model, that $30 street anchor looks light.
"TD Cowen noted improving fundamentals, including rising net revenue retention and increased velocity of new customer wins... the absence of seat-based pricing headwinds to help close the valuation gap."
Bull case: 1.9x vs 3.7x is the kind of dislocation that closes fast on a re-rating. Add in NRR inflecting, AI monetization catalyst into '27, and a path to double-digit operating margins — this is a multiple expansion story waiting on confirmation. Legacy competitor displacement (Salesforce/Adobe MAP displacement) is a real tailwind, not a slideware slide.
Bear case: Mizuho's right that valuation got ahead of fundamentals on the prior run. Gross margin MISS of 140bps isn't nothing — it suggests some pricing pressure or mix shift. The AI monetization story is a 2027 catalyst; through year-end you're trading on fundamentals that, while improving, aren't yet at the level the prior $40 PT assumed. And seat-based headwinds may be "absent" but the competitive backdrop in martech is brutal — everyone's bundling AI now.
The optical/InP trade is the cleanest read in the feed today. NVDA-TSMC customer concentration is now confirmed. Intel 18A is the foundry re-rate. HDD is back as a bottleneck. Every semi name worth owning has a 2027 datapoint in this print.
NVDA — still the best show in town. Confirmed >20% of TSMC revenue (surpasses Apple as #1 customer), first Vera Rubin NVL72 rack shipped via DELL to CRWV, Jensen GTC Taipei with N1X + Feynman today. Bears weaponizing customer concentration as the crash-50% trade; bulls note 401(k) target-date flows dump $2.87B/month into NVDA alone as the structural floor. Forward PE 13 with "insane FCF" is the bear-killer bogey.
TSM — every capex decision is now an NVDA decision. Apple 95% volume still on TSMC near-term but Apple-Intel 18A-P evaluation is the real diversification tell. Taiwan AI economy thesis is intact.
COHR — preferred name in InP. $125M (2025) → $4.3B (2030) = one of the fastest growers. 6-inch wafer + laser capacity is the moat. NVDA wants 20x scale-up, vendors committing 12x — supply-constrained through 2030.
LITE — InP volume leader. $600M → $9B is the biggest absolute ramp in the optical chain. 50% supply gap persists in 2030 = multi-year pricing power. Own it.
AAOI — smaller-cap pure-play with highest revenue % growth. $60M → $2.1B at ~10% UHP CW CPO share. Highest beta to the InP trade.
VIAV — Goldman framework preferred name in optical/photonics. Test/measurement + optical incumbent. Multi-cycle demand visibility. Lowest beta of the optical cohort.
INTC — 18A-P production milestone hit. EMIB revenue tracking ahead of UBS estimates. Lip-Bu Tan: yields rising fast, internal demand filling, Apple/Google/Nvidia/AWS are the only external targets. Foundry re-rate is the trade; cash flow proof is 6-12 months out. Two parallel signals (production + packaging) compounding the narrative.
QCOM — Snapdragon C for $300+ laptops with Acer/HP/Lenovo. Cristiano Amon opens Computex 2026. QCOM is the incumbent ARM-on-Windows; N1X is the challenger. Show us the silicon at Computex.
ARM — architecture backbone across N1X, Snapdragon C, Apple M-series. ARM-vs-x86 lock-in is the structural question; Apple Silicon proved the lock is breakable.
LSCC — Lattice CEO pegs 2027 hyperscaler capex at $1T+ ("just the hyperscalers"). FPGA exposure to the AI capex super-cycle. Take the print and run.
FOTO — US-listed pure photonic ETF finally listed. Basket play for the optical trade. Less stock-picking risk if you don't want to pick the winner.
WDC — led S&P on Friday. Storage/NAND is a critical AI capex leg. Combined with DELL COO HDD shortage call, positioned as storage beneficiary with HDD transmission alongside NAND.
META — channel analysis says most incremental training GW in 2026, may exceed the ~6GW OAI/Anthropic target by EoY. If MSL team ramps and META's lower inference demand frees capacity for training, becomes a credible frontier challenger in 12 months. Monetizes via training-driven ad inventory — wins the marginal rev/marginal capex ratio if inference pricing compresses. The trade is "META is the dark horse frontier."
MSFT — rejected $3B ORCL cloud contract on security/compliance. Maia + NVDA path. N1X collaboration hedges silicon. The Copilot critique ("lowest presence, lowest usability, zero buzz" per Chinese-language source) is real — DeepSeek "feels usable and reliable." Enterprise is fine; consumer is a problem.
AMZN — cup and handle breakout. Project Rainier ~$500M tax breaks in Indiana. AWS re-acceleration + AI capex buildout in parallel.
DELL — +33% FRIDAY, LARGEST ONE-DAY GAIN IN COMPANY HISTORY. +43% on week. 4 consecutive ATH. First Vera Rubin NVL72 integrator. COO EXPLICITLY called HDD the next component shortage — Resonac, TDK, Nidec (~80% global share), Minebea Mitsumi (>80% share) are the Japanese beneficiaries. HDD is back as a bottleneck trade. AI server demand is real; DELL is one of two box-builders that ship at scale.
HPE — +86% since referenced post. After-hours Monday is the next catalyst. Juniper networking integration is the long-term optionality. Second AI server beneficiary with networking.
VRT — 2026 guide $13.75B revenue, 34% growth. Labor build = the real signal. AI power/cooling infrastructure play with capacity-constrained growth.
CRWV — first recipient of Vera Rubin NVL72 via DELL. Bought at $105-108. Flagship neocloud for next-gen NVDA. Csquare Brookfield IPO will benchmark the cohort. Crusoe/Helix/Prometheus/IREN/Csquare — fifth public exit = private capital disposing of mature AI infra assets to public.
NBIS — Nebius Token Factory "built around optimization, orchestration, and agentic deployment, not basic GPU rental." Inference-optimized alternative to CRWV/IREN. Right place for the next capacity fight.
IREN — neocloud + bitcoin mining hybrid at $63 reference. Part of the Crusoe cohort. Pure beta to AI infra deployment.
SNOW — +48% ON WEEK. Q2 is the binary. Databricks margin compression from "swarm of AI agents" is the structural risk: every agent = additional compute cost, not free. The "AI agent = infinite margin" narrative is the lie. If margin guidance moves down, the entire AI infra cohort re-rates lower. Q2 commentary is the cohort-defining print.
CRWD — next week's print is the test. Iran using Western AI (ChatGPT, Gemini) for phishing/malware per FT = structural tailwind for the defender side. AI-cyber beneficiary.
PANW — FIRST QUARTER WITH CYBERARK. Integration validation is the catalyst. Iran AI cyber signal compounds the bull case. Identity security expansion via CyberArk is the new growth vector.
AAPL — smart glasses DELAYED to late 2027. Bull frame: Apple doesn't ship until perfect. Bear frame: giving up 18+ months to META Ray-Ban + Snap AR Specs. Evaluating Intel 18A-P — $3-5B annual revenue for Intel foundry at 2-3 SKU volume; Q3-Q4 quality test + LTA negotiation start. iOS 28 "Bell," macOS 28 "Poppy" — pair internally "Boppy." No mainstream AI product yet; OS rebuild is the tell but the product isn't here.
IONQ — optics supply chain strategy. $2B CHIPS Act commitment to quantum foundries. Pure-play with optics + foundry tailwinds. Low conviction but in the theme.
RGTI — listed in quantum optics supply chain map. Alternative pure-play.
GFS — foundry exposure to quantum + advanced packaging.
TSEM — analog/specialty foundry with quantum exposure. Tower Semi.
IBM — futures reference, quantum exposure via CHIPS Act. Incumbent IT services + quantum optionality.
TMHC — Berkshire acquisition at $72.50 all-cash, $8.5B. First big BRK.B deal "in forever." Housing exposure post-rate-cut. "Hopefully the premium proves better timing than other public homebuilder purchases."
BRK.B — first major deal under Greg Abel. Cash pile thesis being monetized. Clayton prefab homes on TMHC land speculation. Deploying capital again.
MSTR — Berkshire bracket reference. Bitcoin proxy with leverage.
TTWO — bottom-fish idea, down only 4% YTD. Value name within AI-driven market.
DIS — deep-value with streaming + parks + ESPN spin-off optionality.
TKO — UFC White House event "may lose money directly, but the real play is attention." Media/attention play with sports monopoly economics.
TMDX — Natan's LARGEST position (19.5%). High-conviction medical device bet (TransMedics). Organs-on-demand commercial model.
HOOD — 15.5% of Natan portfolio. Retail trading cycle trade. Gaming consumer flows.
PDD — 14.8% of Natan portfolio. Temu/Shein-style global discount retail. China consumer + global expansion.
PRCT — 9.5%. Surgical robotics pure-play. Procedural TAM expansion thesis.
BABA — Qwen open-source AI beneficiary of Anthropic Fable ban. "1.5B people in China use these models and they are getting better by the second." China AI + commerce compounder. 7.7% of Natan portfolio.
MNDY — 4.2%. Work management software. Monday.com with project management cycle.
WIZZ — 6.3%. European ultra-low-cost airline. Travel recovery.
PYPL — 5.7%. Payments/fintech turnaround.
MRNA — 2.4%. mRNA biotech turnaround with pipeline optionality.
REGN — 3.1%. Defensive biotech with Eylea franchise + pipeline.
NFLX — targeting $90-$91.50 by Friday. "1% in 2003 NFLX = rich. 1% in 2003 Blockbuster = irrelevant." Long-term streaming winner.
HIMS — $26 reference. Telehealth/DTC pharma. GLP-1 distribution angle.
NVO — $45 reference. GLP-1 duopoly leader. Obesity TAM expansion. The real compounder if the supply scales.
LMND — $58 reference. Insurance tech disruption.
ROOT — $51 reference. Auto insurance tech turnaround.
GRAB — $3 reference. SE Asia super-app.
SOFI — $18 reference. Consumer fintech turnaround.
NKE — chart mention only, no directional view. Post-COVID athletic retail reset.
ADSK — AEC software 17% revenue CAGR since 2020. RPO 15% annual growth. Sticky compounder with multi-year visibility.
CSGP — Andy Florance sales culture + LoopNet strategic logic. Commercial real estate data platform with quality sales culture.
LOAR — 2-3x latent pricing power post-acquisition. Niche industrial roll-up.
CAR — Avis locks in physical workshops; rivals only built brand loyalty. Asset lock-in thesis.
BLBD — school-bus customer lock-in. Multi-year state/local contracts.
ARXS — former Qnnect VP of Sales credibility. Small-cap industrial with credible sales culture.
CVLT, GEV, ATEN, PLTR, ACN, BE, DT, COHU, BDC, DOMO, RRX, ROKU, USAR, GILT — no fresh signal in today's feed. Sitting on hands until next catalyst.
Hedge fund IT positioning is the macro risk nobody's pricing. Cross-asset tape says hedge fund gross and net exposure to IT is "running out of buyers." Short interest in the median S&P 500 stock is at the highest level since 2012. Net-net: the rally is narrow, the squeeze is real, and the unwind risk is asymmetric. (Biguniverse Twitter feed)
Yardeni raised year-end SPX target to 8,250. Index is up 19% from March 30 low. Tracking best Q2 of any midterm year ever. 9 straight weekly gains. Hartnett (BoA) framing: "post-bubble investor roadmap since 1929 is long bonds — long humiliation, short hubris." (Biguniverse)
Polymarket now sees 30% chance of US-Iran permanent peace by June 30 — half of where it was a week ago. Oil fell from $109 (May 18) to $87 (Friday) on this. Energy "absolutely slammed" while everything else rallies. Watch for reversal.
Anthropic raised $65B at $965B valuation, $47B ARR vs. $9B at end of 2025. That's a 5x ARR ramp in 6 months. Capital availability for AI capex is structurally intact — the funding channel isn't the bottleneck, it's power and silicon. (Biguniverse)
Anthropic's "Fable" ban is pushing Chinese users to Qwen. "1.5 billion people in China use these models and they are getting better by the second." Open-source is winning the China AI distribution game. (Twitter default)
BYD Ti7 ships with DeepSeek driver-assist — China AI-auto commercialization moving into mass production on a different cycle from US frontier labs. Watch the China auto AI supply chain as a parallel capex theme.
META may lead incremental training GW in 2026. Channel analysis: META could exceed the ~6GW OAI/Anthropic target by EoY 2026. If MSL ramps and inference demand stays low for META, training capacity frees up — and META becomes a credible frontier challenger in 12 months. The dark horse is real.
xAI reportedly negotiated 90-day reclaim optionality on short-term capacity — structural advantage if inference demand accelerates faster than lease buildout. Channel check suggests this was the lever that let xAI move fast on Memphis.
Hearing DELL COO's HDD shortage call is not noise — the call came with Nidec and Minebea Mitsumi share data (~80% global). If DELL is forecasting HDD constraints into 2027, the storage bottleneck is real and the Japanese beneficiaries are the play. (NVDA optical crowd is now watching HDD.)
Word is Intel 18A-P production milestone triggered Apple Q3-Q4 quality test and LTA negotiation start. External customers remain META, GOOG, NVDA, AWS only. 18A packaging projected "$80B+ a year business" — that's the bull case in one line.
Vista Equity insurance claims workflow dropped from ~$8M (manual) to $200K via proprietary inference system. The unlock is context (data structure + orchestration), not model capability. Data-silo bottleneck is the real enterprise adoption gate — confirms the data infrastructure layer is the next spend cycle. (Twitter default)
"AI harness wars 2027" framing from one source — platform battle shifts to user-owned memory and context portability across AI tools. GBrain v0.42.1 ships a SkillOpt-derived auto-improvement loop. GStack entering GitHub Top 100 signals agentic tooling infrastructure is maturing at the developer level. The "harness" layer (Claude Code, Cursor, Codex) is becoming a moat.
Pretraining threshold debate — per one channel source, labs have NOT used 300T+ tokens for pretraining or paid $4/hr GB300 long-term. The $1B pretraining threshold has not been crossed. This contradicts the bear case on compute demand scaling — and supports the inference-vs-training bifurcation thesis.
Sora team transitioned to robotics. OpenAI Robotics hiring actively. "Week of next-gen models" framing from one source — frontier releases incoming. If robotics is the next platform shift, the implication for NVDA, QCOM, sensor chain is material.
"SPVs at the heart of the AI bubble" has entered the narrative. Worth tracking — if the SPV funding structure breaks, the capex visibility narrative breaks with it.
Elongation is now visible in the feed: NB ODM 2H26 production cuts -15-18%, MLCC hoarding. AI supply allocation is crowding consumer electronics out. Optical positioning is exhausted (most preferred names already public). "Fabulous earnings momentum" (FEMO) framing — earnings-led, not P/E-led. (Biguniverse)
Iran using Western AI (ChatGPT, Gemini) to turbocharge cyber operations per FT — phishing attacks, malware development. Cyber is structurally bid as AI becomes a weapon for adversaries. CRWD/PANW are the defender beneficiaries.
SoftBank €75B / 5GW French commitment — Europe's largest single AI infrastructure announcement. But the article's own framing: European energy prices surging on US-Iran war will force DC migration to lower-cost-power regions. Winners and losers across the continent. Watch regional power markets.
MSFT rejected $3B ORCL cloud contract on security/compliance. MSFT is signaling it can build vs. rent. Negative for third-party DC operators, neutral for NVDA, bullish for Maia silicon thesis.