Good morning.
S&P futures down another 0.5% after the Mag7 got crushed — GOOGL -7%, TSLA -14% (that's a 5-year worst for the basket). But INTC flips the script: +11% AH on a Q2 beat with 15-year high sales growth and a $20B capex guide for 2026. OIL ABOVE $100 on Trump/Iran escalation — defense and energy are the only green today. Asia mixed; KOSPI setting up for a technical low (good for MU, memory names).
Three themes framing the tape: 1) INTC is the narrative inversion. Wall Street was ready to bury the foundry story, but Lip-Bu just said 14A is a "go," volume in 2028, and they're heading for a capital raise because the business is working. This is a real re-rate if the market buys the capex plan. 2) GOOGL is the battle between FCF math and Cloud velocity. Consensus Cloud rev went from ~$65B to ~$100B in a year, but the $8110B in future commitments (up $500B in a quarter) and negative FCF are crushing the multiple. This is circular — either Gemini 4 saves it, or the compression continues. 3) AMD is speaking the PMs' language. MI350P targets enterprise on-prem inference, air-cooled, 4x tokens per dollar vs competition. That's not a CUDA killer today, but it breaks the "no one wants AMD GPU" narrative. 4) Power supply chain scares are noise. 800V HVDC slipped to 1Q27 on design validation, not demand loss — Delta ramps 2Q27. AMKR's $1.5B prepaid deal from NVDA confirms packaging is the real bottleneck.
We'll hit up INTC, AMKR, and MU first, then get to semis and power infrastructure — GEV needs a look after the selloff.
THE TWO-TONE NARRATIVE. Cloud accelerating (82% y/y) is the headline. But the stock dropped ~4% after-hours because the bill for that growth just got a lot bigger. Capex guidance raised $15B for 2026 to $195-205B. 2027? "Significantly higher" from a higher base. The bull case is a multi-year AI infrastructure buildout with monetization visible. The bear case is that cost curve steepens faster than revenue, and FCF turns negative.
13 analysts in the blender. Price targets range from $350 (DA Davidson, Neutral) to $515 (Citizens, Market Outperform). The middle cluster is $395-425. Consensus is still Overweight/Buy-heavy (8 Outperform/Buy, 3 Neutral, 1 Hold upgrade to Buy), but the direction of revisions matters more than the level. Six firms cut PTs post-print. The bulls are more about reaffirming the thesis. The bears aren't selling — they're trimming targets on higher capex.
BULL: This is the AI infrastructure buildout playing out in real-time. Cloud grew 82% y/y to $24.8B, backlog surged $52B q/q to $512B, and margins hit a record 36%. TPU monetization has begun — Wolfe estimates $100B+ in TPU sales already in backlog, mostly hitting 2027-2028. AI Mode crossed 1B MAUs, Gemini hit 950M MAUs with DAU tripling. Management is framing AI as an "expansionary moment" for Search. The 80% return over the past year says the market was early, not wrong.
"More than $100 billion in TPU sales within fiscal year 2026 backlog... Google expects to realize the vast majority beginning in 2027, with likely 2-3 gigawatts of external TPU capacity and 4-5 gigawatts for internal usage." > — Wolfe Research
BEAR: The cost side is getting ugly. Services margins dropped 345bps q/q as AI queries go undermonetized (Oppenheimer). Capex for 2027 is now $25B higher than prior expectations (Wolfe), pushing 2027 FCF to a $50B loss per Evercore. UBS flags that organic EPS — ex equity investment mark-to-market — is heading for downward revisions. And the stock is at "near-peak multiple" per UBS, who prefers the r/r on AMZN. The EU slapped a $1B fine. Currency headwinds of 200bps are coming in Q3.
Bigger capex, deeper commitment. The $195-205B 2026 guide was the second hike this year. 2027 is now "significantly higher" from a $200B base — not $180B. That means the FCF cliff gets steeper before it gets better. Negative FCF in Q2 was not a surprise but sustained negative FCF through 2027 is now the base case for several models.
TPU monetization is starting to size. Wolfe's $100B+ backlog number is new. If 2-3 gigawatts of external TPU capacity comes online in 2027, that's a new revenue stream that didn't exist 12 months ago. Changes the Cloud unit economics discussion.
Search deceleration is telegraphed. 17% y/y in Q2 vs 19% in Q1. Management flagged tougher comps and 200bps of FX headwind for Q3. The "AI expansion" narrative is being stress-tested by a slowing base.
Negatives are concentrated in Services margins and hiring. UBS flagged "ongoing higher hiring needs" as a structural opex headwind. SBC came in higher than expected.
AMZN / MSFT: The Cloud narrative is now fully extended. All three are in a "spend-to-win" capex cycle. The question becomes: who has the cleaner FCF profile on the other side? AMZN reaccelerating AWS + retail margins vs GOOGL's Services margin erosion is a live debate. MSFT's Azure guide next week will be the tiebreaker.
META: Advertising stability is confirmed. YouTube acceleration (13% y/y) and Search holding 17% y/y both point to a healthy digital ad market. No red flags here for META's print next week.
NVDA / DC chain: This print is a green light for DC equipment suppliers. $200B+ capex in 2026, $330B+ in 2027. The Trump pledge by utilities (NextEra, Duke) to cap AI energy price pass-throughs is a thematic negative for the utility space but validates the narrative that hyperscaler demand is real and persistent.
The "AI under-monetization" theme is a double-edged sword. GOOGL's Gemini queries today are margin-dilutive (Oppenheimer), but that's also the call option on future monetization. Similar architecture to META's "AI investment now, monetization later" story, except META has cleaner Services margins to absorb it.
Verdict: Q2 was a margin miss wrapped in a narrative win. Revenue beat ($28.2B vs $25.6B est) but auto margins cratered to 16.9% (Street 19.4%) on lower regulatory credits, tariff drag, and pricing. EPS $0.33 vs $0.49 consensus. The bulls are leaning hard into FSD acceleration, Cybercab production, and the Optionality™. The bears see cash burn, capex spiraling, and no near-term profit catalyst. Stock at $374 (-17% YTD) – the narrative is now the only thing holding the tape together.
Consensus PT range: $400-$485. Morgan Stanley cut to $400 (Equalweight), Cantor to $485 (Overweight), Mizuho to $450 (Outperform). BofA held at $460 (Buy). Needham and Oppenheimer sit on the sidelines at Hold/Perform. The cluster says: "we like the long-term story, but near-term earnings are a wreck."
Bull case: The autonomous thesis is real and accelerating. FSD attach rates are inflecting, Cybercab is live, and the capex spend buys a leadership position in physical AI. "Tesla is building an asset-light, high-margin revenue stream on top of a vehicle business that's still growing deliveries." The revenue beat shows demand is fine – the margin compression is temporary (tariffs, warranty noise). BofA: "Buy the miss, own the transition." Cantor maintains overweight on autonomy optionality.
"FSD attach rate of 55% with 1.5M subs is a massive pivot point. Recurring revenue at scale changes the earnings algorithm even if auto margins stay compressed." — Mizuho
Bear case: Margins are structurally eroding, not just cyclical. $0.33 EPS on $28B revenue is horrid – shows how little flows to the bottom line. The capex plan is a money pit with no guaranteed payoff. "Tesla is spending $25B+ to build robots and robotaxis that don't generate profit yet, while the core auto business faces its first real competition (Rivian R2, legacy EV ramp)." Morgan Stanley's cash burn math is ugly. Oppenheimer: "Management commentary on Optimus and Robotaxi was muted given the technology challenges." The $30B borrowing authority is a red flag.
"The capex step-up extends cash burn through the end of the decade. We're not sure the market is pricing in a multi-year FCF negative stretch for a stock at 30x 2030 EBITDA." — Morgan Stanley
New (incremental):
Tesla's narrative is now completely bifurcated from its near-term financials. That's fine for momentum traders but dangerous for PMs with a 12-month horizon. Read-through to mega-cap tech: The market is rewarding spending on AI/autonomy infrastructure (NVDA, AMZN, GOOGL all got passes on capex). TSLA is trying to play the same game – but it doesn't have the cloud revenue to lean on. The capex-to-cash-flow math is worse here.
Peers: Rivian (R2 entering Model Y's core segment) is a direct threat. GM's 33% delivery decline shows the legacy EV transition is still messy. But TSLA's valuation (344x P/E, 124x EV/EBITDA) means it trades on narrative, not relative value. The bull/bear debate here is really about time horizon – can you hold through the cash burn to the robotaxi payoff? That's a conviction call, not a valuation call.
TXN printed the cleanest analog quarter of the cycle — stock still got taken out back.
Revenue $5.46B (beat by ~$220M), guided Sep to $5.90B (consensus $5.61B). Gross margins exploded +340bps QoQ to 61.4%, with another +150bps baked into Sep guide (62.9% est.). Data center again doubled, now 12% of rev (~$676M). Industrial +10% QoQ, Auto +8%. Broad recovery, no asterisks.
Yet shares fell ~3% after hours. That’s the setup.
Three shops updated — consensus? Good print, bad multiple.
| Analyst | PT (Prior) | Rating | Key Line |
|---|---|---|---|
| Benchmark (Acree) | $360 ($315) | Buy | “One of the cleanest analog & embedded prints of the current recovery” |
| Cantor | $340 | Neutral | Gross margins surprising, but capex tracking ahead – wait for more margin conviction. |
| Mizuho | $305 ($300) | Neutral | Data center growth impressive, but 31x CY27 P/E is full. |
Benchmark’s Acree had the most conviction. Said the selloff was “positioning and valuation, not fundamental” — the stock ran 70% YTD into this, so anything short of a heaven-sent beat was vulnerable to profit-taking. He’s right.
“One of the cleanest analog and embedded prints of the current recovery.” — Cody Acree, Benchmark
Cantor and Mizuho both stayed Neutral. Cantor’s $340 target is 30x CY28 EPS discounted back — fair value, but no urgency. Mizuho barely moved the PT ($300 → $305) despite the beat. Their message: we see the recovery, but the multiple already prices it.
🟢 Bull case: This is a textbook cyclical trough-to-recovery inflecting into a secular data center tailwind. Gross margins have room to run back toward mid-60s as utilization picks up and price increases stick (TXN has already started hiking). $5.90B Sep guide is for real — backlog growing, lead times extending slightly. At 30x forward earnings with a 70% YTD return, the stock is expensive, but winners stay winners when fundamentals are this clean.
🔴 Bear case: P/E of 50 on trailing earnings. 31x CY27 consensus. That’s priced for perfection in a commodity analog business. Capital spending is tracking ahead of expectations — the 300mm wafer buildout is a cash incinerator in the near term. Data center is only 12% of rev, and the “doubling” comps get harder from here. One clean quarter doesn’t justify this valuation multiple expansion without margin sustainment.
TXN’s recovery is real, broad, and accelerating. Gross margins are inflecting exactly when you want them to. But the stock already ran 72% YTD before this print — that’s 3x the SOX return. The after-hours dip tells you the multiple cap is tight. We’re constructive on the fundamentals but cautious on upside from here. If you’re long, you’re betting the multiple holds (or expands) through a capex-heavy 2027. Shorts: not yet, but watch the next guide for margin momentum.
VERDICT: BUY-SIDE VERDICT IS CAUTIOUSLY BIASED LONG. Stock at ~$96 after a 50% drawdown from highs, and Q2 had enough juice to keep the AI platform trade alive. But the beat quality matters — and the street is talking about deceleration, not just beats. Positioning is tricky: fundamentals solid, narrative intact, but the multiple (57x P/E on reported numbers) leaves zero room for error.
Three firms all reiterate Overweight/Outperform/Market Outperform — no downgrades, no cuts. Price targets cluster around $140-157 (Cantor $141, Oppenheimer $140, Citizens $157). That implies 45-65% upside from current levels. But nobody is chasing the print higher — the stock is down 50% over the past year, so the question is whether this quarter changes the narrative.
The headline numbers were fine:
Bull case is all about AI platform consolidation. ServiceNow is the workflow layer where enterprises are shoving their AI agent experiments into production. Oppenheimer calls out: "The cybersecurity business exceeds $1B and organic growth is accelerating." That’s a new disclosure — NOW is now a legit cyber platform play, not just ITSM.
Cantor sums it up: “View ServiceNow as a beneficiary of secular AI growth trends as enterprises consolidate AI-enabling spend on the company’s platform.”
Bear case: beats are shrinking, growth is decelerating, and AI is a margin headwind, not a tailwind. Oppenheimer flag: “The second-quarter beat magnitudes are smaller than over the last twelve months, and the company results and guidance show signs of continuing decelerating growth.” Combine that with the pull-forward noise and a 57x P/E, and the risk/reward gets tight.
This is the open question for the next 6 months. NOW is running a classic playbook: flat headcount, aggressive M&A (Armis ahead of plan), and AI-driven revenue acceleration. But AI usage is eating into gross margins — 50bps drag in sub GM alone. If AI ACV keeps growing faster than total revenue, margins stay under pressure. PMs trying to model FY27 need to decide whether this is a temporary cost hump or structural margin compression.
Citizens (the most bullish at $157) is comfortable. Others are watching.
Blockquote of the day — Oppenheimer, capturing the mixed read:
“We reiterate our Outperform rating on ServiceNow after most of its 2Q KPIs beat consensus estimates. The AI business momentum remains strong and is on track to exceed the 2026 $1.5 billion annual contract value target. However, the beat magnitudes are smaller than over the last twelve months, and the company results and guidance show signs of continuing decelerating growth.”
That’s the whole trade in one sentence: quality business, secular tailwind, but the rate of change in beats is slowing. Buy the franchise, hedge the multiple.
Three downgrades/cuts in 24 hours. Stock -21% in a week, now $25.99, kissing the 52-week low of $25.10. This is not a garden-variety miss – it’s the second consecutive quarter of ACV and FCF underperformance, and management just yanked its 2026 guidance. The narrative is broken. Customers are in “wait and see” mode on AI, deals are slipping, and the one bull case left is “P/E of 12.77, maybe you get paid to wait.”
Q2 ACV growth (constant currency) printed 8% – well below the 15% full-year target that now looks unreachable. FCF of $84M missed Loop’s estimate. Cloud revenue grew 28% YoY to $213.9M (51% of total) but still came in ~$4M short of Rosenblatt’s forecast. Term license revenue beat modestly. The headline EPS of $0.35 missed consensus of $0.43.
Management kept the 2028 FCF target of $700M alive but withdrew the 2026 FCF target of $575M. Translation: “We don’t know when this stops.”
Three firms moved – all in the same direction, just different degrees of pain.
“remains unconvinced that Pegasystems’ Blueprint AI offering is a decisive business driver or competitive differentiator, rather than a useful tool for improving internal pipeline generation and sales-cycle efficiency.”
That’s the crux of the bear case – PEGA is using AI to sell faster, but customers aren’t convinced AI is worth buying.
Bull case (Rosenblatt, D.A. Davidson) – Cloud revenue still growing 28%, FCF generation remains “robust” (not quantified but the 2028 $700M target implies ~$500M+ annual by then). At $26, PEGA trades at 12.77x P/E and a PEG ratio of 0.25 – cheap if you believe the AI confusion is a few quarters of noise. D.A. Davidson kept Buy, $55 PT. The bull says “deals are delayed, not lost.”
Bear case (Loop, William Blair) – Two consecutive quarters of underperformance is a pattern. ACV growth decelerated from 14% in FY25 to 8% now. The willingness to withdraw 2026 guidance signals management sees more downside risk than upside optionality. The Blueprint AI narrative is an internal efficiency tool, not a wedge to win new business. The buyer confusion around AI “may reflect something more durable than a temporary pause.” When the stock is already at the 52-week low and still getting downgraded, that’s the price discovering a lower floor.
GE Vernova is a beast right now, and both analysts are just pulling targets higher into the slipstream. This is not a debate about whether the story is real — it’s about how much of the 57% (Guggenheim) to 97% (Oppenheimer) 1-year return is already priced in. Answer: not enough, per these two.
Revenue beat ($11.1B vs $10.73B), orders EXPLODED 88% YoY to $24.2B, backlog hit a record $176B. But EPS missed — $2.47 vs $3.04 consensus. The market didn't care. Neither do we. Free cash flow guidance RENEWED to $11.5-12.5B, revenue guide raised to $45.5-46.5B. This is a growth and cash generation machine, and the earnings miss was noise from wind segment mix.
Guggenheim moved PT to $1,450 (from $1,300), Buy. Oppenheimer to $1,338 (from $1,303), Outperform. That’s a 2026-2028 thesis: management guidance is the floor, not the ceiling.
Orders are the headline, but the real story is backlog quality and duration. Power & Electrification are on fire — medium voltage, SST solutions, cross-sell synergy. Management plans to scale turbine capacity to 30GW by 2030. Cash flow is funding organic and M&A investment. Guggenheim explicitly sees upside to capital returns beyond guidance.
"Other elements of the investment story are sufficient to drive additional improvement in the stock price." — Guggenheim
Oppenheimer adds the SMR angle: a reported agreement with the U.S. around fuel enrichment. That’s a wildcard call option on small modular reactor deployment.
The 2Q power bookings number was a spike — Guggenheim says it's "unlikely to be repeated." And shares are currently above InvestingPro's fair value estimate at $985. So the easy multiple compression trade is done. Concerns about peak orders are real — Oppenheimer flags that as the overhang weighing on the stock post-earnings. Also, wind continues to drag on margins, though it’s a shrinking percentage of the mix.
Still, the bear case feels like a fade of momentum, not a fundamental thesis break.
Hold on — $176B backlog with a 10% rev growth run rate and free cash flow guidance that implies >25% conversion. This is an industrial that acts like a software company on cash generation. Net cash balance sheet too. Oppenheimer calls financial health "GREAT." We’d say: undervalued relative to the free cash flow trajectory, even at $985.
Buy the dips. The narrative is intact.
Verdict: HUT’s power-first model is printing. The lease stack is getting absurd — $9.8B Phase 2 at Beacon Point with the same investment-grade tenant doubles contracted capacity to 704MW. That’s 70% of the 1GW campus spoken for, and the market is only starting to re-rate the pipeline.
Benchmark just yanked its PT to $195 from $165 (still Buy) after that second 15-year lease. Needham followed to $145, Rosenblatt stuck at $124 (Buy). Morgan Stanley initiated Overweight with a $263 PT — that’s the highest on the street and implies 2x+ from here.
The collective thesis: HUT has flipped from a Bitcoin miner with a side hustle to a high-investment-grade data center developer with a 2.7GW exclusivity pipeline. The TeraWulf comp at Justified Data Campus ($18.50/watt) set the floor. Beacon Point Phase 1 + 2 alone are worth ~$17B in contracted lease value.
Morgan Stanley’s base case values the 2,680MW under exclusivity at $11.6B — discounted to recent comps for execution risk. Their REIT endgame framework on the three signed leases drives the $263.
"Until TeraWulf announced a $19B deal at Justified Data Campus, Hut 8’s $18.50 per watt in estimated equity value creation from the River Bend Campus deal represented the most attractive terms to a high-performance computing company." — Morgan Stanley initiation
Bull case: The power-first model is self-reinforcing. Each lease unlocks another tranche of cheap, pre-permitted power. The 1GW expansion opportunity at River Bend and Batavia (50MW clearing construction) are next catalysts. If HUT executes, it’s a $20B+ company.
Bear case: Execution risk is real — 409% in a year means the narrative is already pricing in a lot of the pipeline. The $70 floor in analyst targets (yes, someone has that) says not everyone buys the REIT framing. And Bitcoin mining revenue is still a non-trivial chunk of earnings. A Btc drawdown or hyperscaler capex pullback could take the air out.
HUT now trades at $112, $12.6B market cap. That’s 0.5x Morgan Stanley’s base case pipeline value. If they hit 2027 milestones (Beacon Point energized, Batavia live), the upside to $263 is 135%. The rate of change in lease signings is accelerating — two massive deals in five months. PMs should watch for the next exclusivity announcement (likely River Bend expansion or a new site). That’s the trigger for the next leg.
TDY is printing. Stifel and Needham both bumping PTs this morning (to $775 and $750, respectively) after a Q2 that beat on both top and bottom line by a mile. Revenue of $1.66B vs $1.58B consensus. Organic growth of 9.0%. EPS of $6.28 crushing $5.79. Book-to-bill of 1.23x — that's ELEVEN straight quarters above 1.0x. Funded backlog at ~$5B.
Raised FY26 sales guide by $120M to ~$6.535B (now implies ~7% growth vs prior 5% consensus). EPS guide going up 12% at the midpoint. Paid down $450M in debt in the quarter — leverage now at 1.1x. Balance sheet is a weapon here.
Needham highlighting the dual-engine narrative: long-cycle defense orders accelerating AND short-cycle Commercial business inflecting. That's the setup PMs should care about — rate of change in both directions.
"We are more optimistic about prospects for upside in the second half of 2026 and 2027, given strength in the long-tail defense business and the inflection in the short-cycle Commercial business." — Needham
Digital Imaging organic growth in the quarter. A&D high single-digits. Everything working.
Bull case: 1.23x book-to-bill means visibility for days. Defense is a multi-year super-cycle, Commercial recovery is real, and the balance sheet lets them do M&A when they want. 27% YTD but still room — growth re-rating isn't done.
Bear case: P/E of 31.3x on a mid-single-digit grower (pre-this guide). If short-cycle Commercial falters or defense budgets get squeezed, that multiple compresses fast. Not sure we can read too much into one quarter's inflection.
Verdict: VRT is the purest data center infrastructure play in the space, and the Street is still sleeping on 2027 capacity additions. KeyBanc initiates Overweight with a $360 target (~19.5% upside from $301) – but that's not even the aggressive call. Bernstein and Baird both went Outperform this week with PTs of $416 and $370 respectively, all citing the same thesis: ~85% of sales tied to data centers, end-to-end power and cooling portfolio, and a service moat underpinning recurring revenue.
The real edge here is the rate of change in data center capacity. KeyBanc explicitly says 2027 consensus estimates look too low as deployment accelerates. VRT trades at a headline P/E of 76x, but the PEG ratio of 0.56 tells the story – growth is underpricing the multiple. The stock is up 86% YTD and 132% over the past year, but the narrative is still building, not cresting.
"Vertiv's product mix and data center exposure provide a competitive advantage compared to peers" – KeyBanc analyst David Tarantino Jr. (on why they expect continued earnings beats and guidance raises)
The premium valuation is the obvious bogey, but given the growth trajectory (new Malaysia facility, ThermoKey acquisition in thermal management, 800VDC innovation flagged by Baird), the r/r still leans bull for long-duration positioning. Just don't expect a smooth ride at this altitude.
ORCL is pricing in existential risk at $125 – near 52-week low, 0.63 PEG, Guggenheim still sees $400. The message from CFO Maxson after investor meetings: delays are isolated, not systemic, and cost-plus contracts protect margins. The credit downgrade is a manageable concern, not a structural problem.
"The build-out is at a materially de-risked stage. Reported delays are neither systemic across Oracle’s many data center sites nor financially meaningful."
The bear case exists – CLSA just slapped a Hold on this, flagging debt-funded capex that could require $500B by 2030 with internal cash covering a fraction. Wisconsin's utility regulator also hit ORCL with a $7B collateral requirement for one data center. That's a headline risk, not a thesis-breaker. If you believe the AI build-out is real and ORCL is the landlord, this is a decade stock trading like a distressed asset.
META IS STILL THE AI OPTIONALITY PLAY. Bernstein reiterates Outperform / $850 PT (current $630, $1.6T market cap). Thesis: core ad biz is firing on all cylinders (82% gross margins, 26% revenue growth) but the real story is AI optionality — one product-market fit hit changes the whole narrative. The easy comps are behind us (tougher 3Q ahead), but that's already priced into the multiple compression.
"Meta only needs to get one idea to hit product-market fit to shift investor sentiment."
Bernstein bumped revenue and capex estimates. This year's spend moving higher by another $5-10B purely on bill-of-materials inflation (not structural). Capacity comes online in 2027 — that's the real payoff timeline. The July 29 earnings call is the next catalyst: look for specifics on Meta AI, Muse Spark monetization, and short-term infra deals. That's where the narrative flips from "ad company with side bets" to "AI platform with a cash cow funding it."
Valuation is a blended 50/50 of 7x 2027 EV/Sales and DCF (10% WACC, 3.5% terminal growth). At $630 you're paying for the ad business and getting the AI call option for free. Mizuho and Raymond James also in the bullish camp (PTs $835-850) — no dissenting voices this morning.
Bernstein holds the line — Outperform, $646 PT — as MSFT enters earnings next week (July 29). The core debate: is Azure AI actually good business, and will AI cannibalize SaaS? Answers are trickling in: Azure growth and Copilot adoption accelerating, AI gross margins stabilizing, capex growth (ex component inflation) slowing relative to Azure revenue. Street expectations are muted — "good but not breakout" — but the firm sees little downside at 23x P/E, 0.79x PEG. They're not calling this quarter, but believe a breakout is coming within a few quarters.
"We believe a breakout will come within the next few quarters and that Microsoft will be an AI winner for many years."
Other shops echo the vibe — Truist at $575, Oppenheimer at $515 — but none louder than Bernstein on the structural thesis. Options pricing implies a 6.7% move post-earnings. If the AI gross margin narrative holds, that's the circuit breaker for the stock's next leg.
The CoreWeave deal changes the narrative. Citizens raised its PT to $16 (from $14), Market Outperform, pushing the stock that's already up ~188% YTD. The $335M five-year multi-exabyte storage agreement with CoreWeave is the key catalyst — validation that Backblaze can play in the AI storage layer for neocloud providers. F1Q26 B2 Cloud revenue beat by 4%, EBITDA $2.9M ahead. The question now is repeatability: one large deal or a pipeline? Citizens is buying the thesis that more neocloud opportunities follow.
"The CoreWeave agreement increases confidence in the company’s AI storage relevance and improves the probability of additional large neocloud opportunities."
The broader analyst crew (Needham, William Blair, Craig-Hallum) all moved higher post-deal — PTs clustered around $14-16, with upgrades from Market Perform to Buy. No one is bearish here, but the stock's 188% YTD run means the easy money is in the price. Next catalyst: F2Q26 earnings on August 3. PMs should watch for any color on deal pipeline beyond CoreWeave — that's the rate-of-change vector.
This is a micro-cap lottery ticket with a real narrative. Maxim Group starts CYAB at Buy, $1 PT (from $0.28 — >250% implied upside). The thesis: AI-powered disinformation detection for gov/enterprise clients (State Dept, NATO, PepsiCo, e.l.f., WarnerMedia). ARR sits at $7M as of March 2026, with Maxim modeling 53% YoY growth this year and 46% next.
But holy cash burn. Current ratio of 0.23, working capital deficit of $12.6M, only $3.1M cash post-de-SPAC — they just raised $6M gross via a July private placement to stay alive. Management and board (including Mike Pompeo, Sonny Vu) participating in the placement gives some alignment, but this is a capital-intensive story until scale hits.
Maxim’s DCF target is a decade-out bet. Gartner sees the broader misinformation-fighting market hitting $30B by 2028. CYAB is banking on being a pure-play AI layer in that. The early customer wins are real (renewed two-year contract, six-figure Fortune 500 annual deal), but the runway is measured in quarters. Worth watching for the narrative — not for sizing.
"According to Gartner, spending on battling misinformation and disinformation will surpass $30 billion by 2028."
Bottom line: If you believe the AI disinfo wave is real and CYAB survives long enough to ride it, the r/r is insane. But this is a position-size-as-option kind of name, not a core hold. Don't confuse $0.28 with "cheap."
Morgan Stanley throws cold water on the narrative — initiates Equalweight at $36.50, calling the 175% run “enthusiasm” that’s pricing in a lot of growth without the returns to show for it. They see unlevered yields on capex below 15%, equity value creation meaningfully less than peers. That’s a direct challenge to the bull case.
The firm values the book on six leases at Polaris and Delta Forge (~$36B contracted revenue over 15 years). They explicitly exclude the “up to a gigawatt” expansion potential. That's the wedge: if you believe that capacity converts, the stock is cheap. If not, current price already reflects it.
“Unlevered yields on capex below 15% reduces equity value creation relative to deals by other Powered Shell Providers.” — Morgan Stanley
The counter is loud and fresh. Compass Point reiterated Buy, $70 PT. Northland’s Mike Grondahl raised to $82 after the 810MW lease spree with a new hyperscaler. Those are 2x+ upside from here. The divergence in opinion is extreme — and it’s all about confidence in the expansion pipeline and whether the lease economics eventually deliver.
Bottom line: $30 stock with a $36 (MS), $70 (CP), $82 (Northland) PT cluster. That’s either a fat risk premium or a flag that MS sees something the bulls are ignoring. The cash burn is real (InvestingPro flagged it) – they’re unprofitable at $0.46 LTM loss per share. If the leases start generating cash, the short squeeze potential is massive. If they don’t, the equity gets torched. This is a pure r/r bet on execution vs. yield compression.
Nokia’s AI order pipeline is the story — not the soft Q3 guide. JPMorgan reiterates Overweight / EUR18 PT after Q2 EBIT beat by 15.6% vs Infront consensus. The headline: AI & cloud orders hit EUR2.8B in Q2 (vs EUR1B in Q1), with half expected to convert to sales in the next 12 months. Q3 guidance came in below consensus (sales +3-7% QoQ, op profit flat), but JPM sees that as noise relative to the order trajectory.
"The guidance should be viewed in the context of the second-quarter beat and the AI and cloud order intake."
Stock up 131% over the last year — now trading above InvestingPro fair value. That’s the hangup for PMs who want to chase it here. The order momentum is real (EUR2.8B is a step-function), but the guide suggests near-term P&L won’t linearize. Wait for a pullback or a guidance raise to add.
GUGGENHEIM GOES LONG — $25 PT ON WHAT THEY CALL THE CHEAPEST NAME IN SOFTWARE. That's 134% upside from $10.07. The thesis: Freshworks is not dead money despite the AI-induced selloff. Yes, seat-based models face headcount compression risk. But the stock already trades at just 2.3x EV/NTM recurring revenue. Assume hyperefficient (no growth, no decline) and that multiple gets to 5.4x on future cash flow — call it $15+/share baked in.
The real lever is the strategic pivot from SMB to mid-market/enterprise. CEO Dennis Woodside (ex-Servicenow board) brings upmarket GTM discipline. The Freddy AI Agent Studio launch and acquisitions (Fire Hydrant, Device42) give the story some narrative juice.
Blockquote:
“Freshworks trades at approximately 2.3x EV/NTM recurring revenue, which Guggenheim said is well below the intrinsic value of its future cash flow of 5.4x assuming it was run hyperefficiently and not growing or declining.”
But don’t kid yourself — AI risk is real. Early data inconclusive? Sure. But if seat compression accelerates on the enterprise side, the pivot gets harder. JMP and Needham still bullish (buy/outperform), Cantor reiterated overweight but called the last quarter “adequate, not exciting.” The r/r here is binary: multiple compresses further if AI narrative worsens, or you get a re-rating as upmarket execution bites. At 2.3x, the put is cheap.
Verdict: Fortinet's pre-earnings narrative is a tug-of-war between AI-driven demand and a sky-high multiple. Cantor raises PT to $165 (Neutral) off a partner survey, but the beat mix softened — 53% above plan vs 62% in Q1. Stock up 95% YTD, now at 14.2x EV/S (peer median 5.6x). "The stock needs a meaningful revenue and billings beat, along with a broad guidance raise, to move higher."
Bull side (TD Cowen PT $215, Buy) sees durable product demand; bear side (HSBC downgrade to Reduce, PT $102) calls valuation extreme. Morgan Stanley upgraded to Equalweight from Underweight. Moody's upgraded credit to A3, projecting 15% revenue growth in 2026. Earnings July 29, options market pricing a 12% swing. The premium is fully paid — execution alone gets the next leg.
OpenAI's sandbox escape is actually a positive signal for JFrog. Raymond James keeps Outperform / $80 (stock at $82.18, up 105% in a year) after OpenAI disclosed a zero-day in third-party software proxying/caching package registries. The vendor is unnamed, but RJ thinks it's JFrog – the description matches Artifactory's core function. The incident shows leading AI labs still depend on JFrog's infra. The exploit was novel, not a fundamental flaw in Curation or supply-chain controls.
"We would not view the incident as overly negative for JFrog... the exploit targeted a specific proxy inside an isolated environment rather than package-admission controls." – Raymond James
Other analysts mostly constructive despite valuation fatigue. Morgan Stanley downgraded to Equalweight but raised PT to $95 (from $80). Guggenheim to $105, UBS to $110, Cantor reiterates $80. The collective view: cloud revenue re-accelerating (50% YoY), RPOs up 36%, margins 77% – but the 105% run makes the r/r less obvious. The OpenAI narrative adds a new "why you still need a third-party binary repo" catalyst. Not pricing in yet.
Evercore stays Outperform, $250 PT — the lone bull in a sea of cuts. IBM printed Q2 revs of $17.2B and EPS $2.93, in line with the negative pre-announcement and below consensus ($17.86B). Revenue grew 1% YoY, dragged by Infrastructure (-7% cc), but Software (+5% cc) and Consulting (+1%) held. The stock is pinned near its 52-week low of $204, down 28% YTD. Evercore calls this “better-than-feared” — guidance was trimmed to 4-5% cc revenue growth from >5%, but FCF guide ($~1B YoY increase) and 100bp margin expansion target were left untouched. Red Hat accelerated to 11% cc growth (vs. 10% last q), Data grew 18% cc, and Distributed Infrastructure jumped 37% cc on POWER/storage.
The rest of the Street is far less charitable. Five firms cut PTs to a $235-280 cluster — RBC to $270, Stifel to $235, JPM to $250, Bernstein to $280 (Market Perform), Argus to $280. The collective thesis: mainframe weakness is structural, and the CapEx pivot to AI is compressing IBM’s legacy hardware cycle. Evercore’s defense? A PEG of 0.2x — the stock is pricing in zero growth — plus $10B in quantum capex over five years and a fault-tolerant quantum computer target by 2029. That’s a narrative that requires a 3-5 year hold, not a PM’s typical horizon. The risk/reward is defined: $204 floor (52-week low) vs. Evercore’s $250 target, but the rate of change is negative. Not a conviction long here unless you’re buying the quantum call option for free.
Guggenheim initiates with Buy, $245 target — 16% upside from here, but that’s the floor not the ceiling. Multiple other analysts (Stifel, UBS, RBC) already sitting at $285–305, so the bull case is well-supported. Thesis hinges on: (1) 92% gross margins + 18% revenue growth accelerating, (2) AI positioning in design (Autodesk meets 3 of 5 of Guggenheim’s AI framework criteria — context-rich systems of record, multi-stakeholder exposure, regulated industries), (3) the new direct-billing commercial model that cuts out reseller friction, and (4) MaintainX acquisition ($3.6B, largest ever) extending into operations workflows. The macro headwind is the contraction in the Architectural Billings Index, but Guggenheim sees upside to FY27 revenue and billings anyway.
“Autodesk began work on AI in 2018 with generative design in its Fusion manufacturing product… meets three of five criteria in our framework.”
Multiple PTs float between $245 (Guggenheim) and $305 (RBC), implying 6.4x EV/NTM recurring revenue vs. the current 5.4x — not expensive for a 92% margin, 18% grower with a clean new go-to-market. Stifel, UBS, and RBC all reiterated Buy after the Q1 beat (revenue +18.4%, billings +18.4% vs. street at 9.5%). The MaintainX deal is the real swing factor — if it accelerates the ops workflow narrative, the re-rating gets serious.
Mixed signals. Mizuho cuts PT to $34 (Neutral) citing revenue at low end of $11-12.5B guidance on component/memory shortages and shell construction delays. But gross margin comes in WAY better than feared — 15-17% vs guided 8.2-8.4% — a massive beat on mix shift toward traditional/agentic AI compute and liquid cooling. Needham and Rosenblatt both raised PTs to $46 and $45 (Buy), seeing the capital raise as an attractive entry point. Record backlog with $60B+ in new June quarter orders backs the fiscal 2027 outlook, but competition and financing pipeline remain real overhangs.
"The improved margin reflects an increasing mix of traditional and agentic AI compute servers and direct-to-chip liquid cooling systems." — Mizuho
Bottom line: Revenue disappointment is real, but margin inflection is huge and orders are off the charts. The bulls have a stronger narrative on trajectory; the bears have a better short-term tape read. Watch the Aug 11 print for guidance — that’s the real catalyst.
Cantor stays Neutral at $150, but the real signal is the setup. Stock’s down 29% in six months, 44% off the $224 high. That’s a lot of pain priced in. Cantor sees Q2 (print July 30) as the trough of a go-to-market disruption – they need to see proof the disruption is behind them, not just a decent headline. Long-term story still intact: Infinity adoption, CTEM ramp, M&A optionality. But 33 analysts cutting estimates into this print says the near-term consensus has already moved lower.
"The bar for the stock to work depends less on the headline earnings print than on evidence the disruption is behind the company."
(Also worth noting: Raymond James downgraded to Market Perform on weak channel checks, while Guggenheim went the other way to Buy at $188 – so the Street is split on the turn timing. Options activity spiked to highest call volume since 2021, implying some hedge fund conviction that the setup is asymmetric.)
SNAP GETTING HAMMERED. Bernstein slashes PT to $5 from $7 (Market Perform), joining a growing consensus cluster at $5-6. DAU softness persists across both US and international — alternative data confirms the bleed. Revenue looks reasonable on easier comps starting 2Q26 and Snap+ still chugging, but the growth rate on that sub line is slowing.
The bull case rests on AR glasses (SPECS), new developer tools, AI ad suite — but Bernstein isn't buying any of it as a catalyst. They cut their EV/Sales multiple to 1.5x from 2x, bumped WACC to 11%, and dropped terminal growth to 1.5%. The result: PT cut and lowered long-term revenue growth.
"Longer-term competitive pressure from Meta, TikTok, and ChatGPT remains a structural overhang on sentiment."
Cantor and Mizuho both went to $5/Neutral yesterday. DA Davidson started Neutral. The only bright spot: Cantor sees 2Q revenue at $1.55B (15% YoY, high end of guidance). But with DAU trending soft and a legal settlement over teen addictiveness (terms undisclosed, Meta left holding the bag), the r/r here is still awful. Not chasing.
DA Davidson likes the Domo deal. Raised PT to $45 from $40 (low end of Street’s $45-$83 range), keeps Buy. Thesis: this is a margin expansion story, not a revenue synergy play.
Buying substantially all of Domo’s assets for $355M effective cash price — ~1.1x DA’s CY26 revenue estimate for Domo. Close expected before PRGS fiscal year-end (Nov 30). The key: Progress can wring huge cost savings out of Domo’s sales & marketing spend.
"The firm views the acquisition favorably primarily due to the significant margin improvement that Progress Software can deliver from Domo, particularly around sales and marketing."
Product fit is clean — Domo slots into Progress’s data platform. And PRGS already prints 85.6% gross margins (InvestingPro). The deal doesn’t change Q3 guidance (report Sept 30 backdrop). Q2 beat and raised FY outlook add cover. At current levels, stock still looks cheap relative to sum-of-parts — but the $45 target limits upside until we see Domo margin progression.
Guggenheim kicked it off with a Buy/$45 target (24% upside from $33.33), adding to a consensus that’s already running hot — Wolfe at $50, KeyBanc/$41, Piper/$40, RBC/$42. Thesis is straightforward: Samsara rode the ELD mandate into trucking, but the real story is the platform’s expansion into video safety, physical asset tracking, and now AI that reinforces the asset-based pricing model rather than disrupting it. The greenfield is still massive — international penetration is low, and they’re moving beyond vehicles into “Connected Operations” for entire physical estates.
“AI brings Samsara's multi-year ML investments to the forefront at a time when enterprises seek practical AI use cases with demonstrable ROI.” — Guggenheim
The numbers back the narrative: $1.73B revenue, +30% growth, 76% gross margins. Not cheap, but the execution is real. The single-use Bluetooth tracking label launch adds a consumable revenue kicker. Risk is valuation at 6x+ sales for a hardware-adjacent model, but the rate of change in recurring attach and big customer wins (RBC noted 30% ARR growth) keeps PMs comfortable.
Benchmark just went to $685 from $485 on AMD — that’s a 41% PT hike — and the stock is already up 248% in the past year, sitting at $899B market cap. The call is clean: Microsoft’s Azure expansion around Helios (announced July 20) gives AMD a near-term production validation point for MI455X, EPYC Venice, Pensando DPUs, and ROCm. Then Anthropic throws in up to 2GW of MI450 GPU capacity in Helios rack-scale, with the first GW starting H1 2027. That’s a demand-side anchor from a frontier-model lab.
The firm projects 36% EPS growth in FY28 and a 3yr CAGR of 66.7% (FY26–28). The $685 target is 36.7x that FY28 $18.67 estimate. Not cheap, but the narrative is shifting from “can AMD win any AI wallet share?” to “they now have a cloud deploy and a lab commit in the same ecosystem.”
"The combined developments are stronger than either announcement alone because the pairing includes cloud-platform deployment with frontier-model customer engineering."
Bottom line: This is a rate-of-change signal. Not just a token hyperscaler nod — Microsoft is integrating Helios into Azure production, and Anthropic is writing big dollar checks for MI450. The bear case is still Nvidia’s insurmountable software moat and CUDA lock-in. But Benchmark is arguing the ROCm/AWS/Anthropic flywheel is real. Watch for the next catalyst: any MI400 tape-out leaks or EPYC Turin share gains.
Canaccord cuts PT to $12 (from $17) but keeps Buy — call it a "show me" story. Competition from Chinese players is real, and traction beyond VW remains minimal. Western OEMs are "sitting on the sidelines trying to figure out their own identity." That’s not a knock on the tech — it’s an ecosystem timing problem. But the thesis still hinges on execution: VW/Porsche/Audi ramps in 2027, plus Mobileye’s own robotaxi fleet as a proof-of-concept. The Q2 beat ($0.19 vs $0.06 consensus, revenue $508M vs $482M) is a reminder the core business isn’t dead — but the market yawned because the near-term guide is cautious.
"While the company wasn’t profitable over the last twelve months, analysts predict profitability this year with EPS forecast at $0.27."
That’s the base case. The bull case needs FSD take-rates (Tesla as proxy) and VW stepping off the sidelines. The bear case: China eats their lunch and Western OEMs never get serious. Stock at $7.44 — 45% off last year — pricing in most of the bad news. But the r/r only works if you believe the robotaxi/VW catalyst actually lands. Right now, it’s a waiting game.
Wolfe upgrades to Outperform from Peerperform. The bear case from last year — rising churn, pricing pressure, Starlink IPO overhang — is fading. Wolfe now sees a clearer path to EBITDA upside in 2026, and they're leaning into the buyback acceleration.
Q2 print was the catalyst: postpaid phone churn FELL 1BP Y/Y TO 0.86% (vs Street 0.90%) despite a price hike on existing subs. Upgrade rate also dropped 10bps to 3.2% (est 3.5%) — ending that five-quarter streak of rising upgrade activity. Less churn, less handset subsidy drag, more FCF.
The EPS beat was real: $0.65 vs $0.59 consensus. Revenue light ($31.6B vs $31.8B) but nobody cares when cash flow is compounding and buybacks are accelerating. Management reiterated long-term guidance.
Wolfe’s $29 PT excludes the 4.8% dividend yield. They argue double-digit EPS and FCF yields deserve more attention given the "pessimistic buy-side consensus."
"The company’s double-digit EPS and free cash flow yields warrant greater attention amid what it described as a pessimistic buy-side consensus."
That’s the setup. If churn stays low and buyback pace holds, T starts looking like a yield + optionality story (not just a yield trap). Biggest risk still the macro and potential Starlink disruption, but for now the operational tide is turning.
Verdict: This is a liquidation story, not a turnaround. Cantor just downgraded to Neutral with a PT of $4.84 – essentially the cash-out value post asset sale. Stock at $4.07, down 80% in a year. The math is simple: Progress Software buys the operating biz for $400M, Domo becomes a shell with $246M net cash ($4.84/share) and $900M+ NOLs. Founder Josh James stays. No shareholder vote needed (majority consent already done). Close by Nov 30, no financing condition.
The upside lever? The NOL monetization. Board will "identify opportunities" and maybe return capital. But they're also under a 3-year global non-compete covering BI, data integration, AI. So no quick pivot into something adjacent. The $13.5M termination fee if Domo walks suggests Progress wants this locked.
The bear case: You're buying a cash shell with a founder who still controls it, limited operating expenses, and a vague mandate to find something smart to do with the tax assets. The 81% premium to VWAP is already priced in. If you want a special situations vehicle for NOL harvesting, fine. But don't confuse this with a software re-rate.
No strong analyst quote to pull – Cantor's move is purely mechanical. The real signal is the rate of change: asset sale locked, volatility compressed, upside capped by the cash balance. Not a PM's dream unless you're heavy into tax-advantaged arbitrage.
Guggenheim drops a buy with a $68 PT (54% upside) on AI-as-inflection-point thesis. Construction software's least-digitized industry finally getting its AI moment, and they think newly appointed CEO Ajei Gopal is the right guy to shepherd it. PCOR checks four of their five "AI moat" criteria: context-rich system of record, heavily regulated end market, multi-stakeholder complexity, and outcome-based pricing. The stock's down 39% YTD – that's the hook. Valuation at ~4x EV/NTM recurring revenue, well below their no-growth intrinsic estimate of 5.4x. 80% gross margins, profitability expected this year. The short-term noise? Q1 revenue beat but EPS miss ($0.34 vs $0.36) – not ideal, but the street seems to be looking through it.
"Procore meets four of five criteria we've identified for protection against AI obsolescence... context rich systems of record, a highly regulated industry, multi stakeholder focus, and outcome based pricing."
The broader analyst landscape is a mixed bag but collectively constructive. D.A. Davidson stuck at Neutral/$55 (revenue beat below 8-qtr average, fair), Stifel Buy/$63 (post-jobsite tour), Citizens Market Outperform/$71 (ahead of Q2). No one's screaming bear – just a range of conviction levels. The Guggenheim initiation is the most aggressive call on both thesis and upside. Construction spend is soft, but customer backlogs are elevated and labor markets tight – that's exactly the environment where PCOR's value prop shines. If AI actually starts driving adoption acceleration here, the multiple compression narrative reverses hard. Risk: EPS trajectory needs to catch up to the revenue story.
Needham cuts PT to $14 from $18 but stays Buy — they're leaning into the 11% drop to $10.45. The Q2 pre-announce miss ($280-282mm rev vs prior guide) gets pinned on a single European customer's project timing and elevated memory/freight costs. Not ideal, but the firm explicitly says this doesn't bleed into 2027. They're more bullish after the pullback.
"does not expect the second-quarter challenges to impact results into calendar year 2027"
The rest of the news flow is noise — Wi-Fi 7 gateway expansion, a patent litigation dismissal (all claims with prejudice), and a board appointment. None of it moves the needle near-term. The setup here is a one-off customer glitch + cost spike vs a stock that's already <1x forward sales. Decent r/r if you trust the timing issue is isolated.
Bernstein cuts PT to $96 from $102, keeps Outperform. That’s a 24% implied upside from here ($77.40) — but the stock has been cut in half over two years, so the market is pricing in a lot of skepticism. The cut is tactical, not strategic: they still see 2026 as the organic growth trough, and the fiber/small cell divestitures (completed two months early) are the real story this quarter.
The Q2 beat was modest ($0.69 adj. FFO vs $0.39 consensus — though that comp is distorted by the old fiber footprint). The real driver: $35M lower interest expense and $14M more interest income from the sale proceeds. Pure-play tower now, no more drag. But the broader tower environment “has not meaningfully improved and likely will not for some time.” Management calling 2026 the low point — we’ve heard that before, but the 600MHz re-farming, upcoming auctions, and densification are concrete catalysts for ’27+.
“The long-term opportunity for Crown Castle remains underappreciated.”
5.48% dividend yield provides a nice floor while you wait, but the 26% Y/Y decline screams the market doesn’t trust the inflection yet. I’d want to see organic leasing activity turn before adding size — this is a patience trade, not a momentum one.
Benchmark cut PT to $650 from $695 (still Buy) — but the stock’s a full 30% below that even after the revision. The call is less about near-term bearishness and more about timing: they need proof that the current OpEx ramp is temporary, that ad revenue accelerates in 2H, and that sub adds aren’t front-loaded. EV/NTM EBITDA multiple has halved from the 2025 peak. That’s the setup — cheap if the story holds, dead money if it doesn’t.
“India’s Premium Platinum tiering, which has achieved approximately 7% new user attachment at roughly twice the price, as evidence that tiering can scale ARPU independent of AI-product monetization.”
Other shops are circling the same narrative. Rosenblatt went to $531 (Neutral) on FX drag, Citizens reiterated $625 (Outperform) on 2027 product tailwinds, Cantor bumped to $520 on the AI roadmap. Bottom line: the bull case rests on tiering + AI monetization showing up in margins by 2027 — and Benchmark isn’t buying the consensus on OpEx just yet.
Charter is trading like it’s going out of business. Benchmark says that’s the trade, not the thesis. The stock’s down 68% in a year, P/E of 3.45x, FCF yield of 23% — levels you see in distressed credit, not a cable duopoly fighting for share. The firm is sticking with its Buy and $380 target, arguing the market is pricing in a near-total collapse of the converged broadband/mobile model, which is wildly inconsistent with even the low-end consensus through 2030.
“Charter’s stock price action elicits more discomfort than for probably any stock the firm has ever followed.”
Benchmark’s 2027 FCF estimate of $47.84 per share implies a ~37% yield off that target. The asymmetry is absurd — but timing is everything. 2026 EBITDA likely flat (tens of bps variance either way). The bull case rests on broadband share stabilizing via bundle value and network evolution (multi-gig). The market doesn’t believe it.
Other analysts are less convicted: Wells Fargo cut PT to $160 (from $170) on expected 140k residential broadband net losses in Q2; UBS stays Neutral, forecasting a 2.4% revenue decline and 2.7% EBITDA decline for 2Q. The Comcast spinoff-driven 14% pop was a one-off, not a trend.
Verdict: The valuation is screaming, but near-term broadband erosion is the wall. Benchmark’s call is a deep-value contrarian bet that requires patience PMs may not have. Worth watching the earnings print for any narrative shift — options imply 12% movement.
BofA is leaning into the Q2 print. They bumped revenue and EBIT estimates above consensus — $198.8B vs $196.8B Street, $24.1B EBIT vs $23.6B. The real juice is AWS: now modeling 33% y/y growth (from 31%), a 5ppt acceleration vs Q1, driven by Anthropic revenue scaling and OpenAI workloads hitting Bedrock. Margin math gets messy quarter-over-quarter (SBC step-up) but BofA sees AWS margins expanding y/y to 34% on utilization and pricing.
"Amazon could raise its 2026 capital expenditure outlook to $210 billion on higher memory costs."
That’s the headline risk — not demand, but cost. Higher memory costs (HBM, DRAM) squeezing the CapEx line. BofA thinks it’s manageable but worth watching on the call. Telsey and Truist echoing the same bullish AWS acceleration thesis — 30%+ growth, strong online spend, operating income up ~24%. Consensus is already pricing in good news (PEG 0.81, 29x P/E). The question: is $210B CapEx a signal that margins peak sooner, or just a pass-through cost? Verdict: BofA stays Buy, $310 PT. The setup is clean if AWS beats and Q3 guide holds the $200-205B range.
Guggenheim starts Buy, $232 target — that's ~22% upside and 9.8x NTM recurring revenue. They see the Salesforce migration as a moat-builder, not a risk. The big picture: VEEV is a vertical software compounder serving a $2T life sciences TAM, growing revenue mid-teens with FCF margins in the high-30s/low-40s. The $6B run-rate by CY30 target is achievable, says the analyst.
"The company prioritizes customer value over monetization in its AI approach."
That's the right call in a market where AI hype is getting sticky. Stifel and Oppenheimer also chimed in recently (both Buy/Outperform, $230/$225 targets) post the Copli acquisition — that Falcon MLR product should automate pharma review workflows. New EHS app launching this summer is a small bolt-on but shows the platform expanding beyond core CRM/CLM.
Verdict: VEEV is a grind-it-out compounder in a solid niche, not a momentum name. If you want software with visibility and cash generation, this fits. Multiple is fair given growth durability.
Verdict: Buy. AI traffic fears are real but overblown. Reddit has the leverage.
The WSJ story spooked the street — logged-out user growth at risk if Google cuts referral traffic from AI Overviews. DA Davidson keeps $200 PT (stock at $168, -6% on the week) and makes the right call: Reddit’s human-first content is irreplaceable for LLMs, and the company is already forcing mobile web users into the app to build direct DAU. That’s a hedge, not a headwind.
"Reddit’s content represents a unique source of human-first content that has avoided getting overwhelmed with AI-generated material... putting Reddit in a position of leverage." — Wyatt Swanson, DA Davidson
Multiple firms are in the same zone: Cantor Fitzgerald to $200 (Neutral), Jefferies flagging strong ad momentum (mid-40% rev growth Q2), Citizens at $240 (Market Outperform). The numbers back it up — 91% GROSS MARGIN, 71% REVENUE GROWTH, more cash than debt. The AI access risk is a negotiation, not an existential threat. Reddit will cut a deal with Google that keeps the traffic flowing while protecting the moat. Buying the dip here makes sense.
Cantor Fitzgerald just threw $294 at ADP – up from $244, Overweight maintained – and they're not sweating the AI narrative. The call: stable employment backdrop, durable pricing, continued margin execution (48.5% gross margins, 7% rev growth LTM). Multiple expanding from 18x to 22x on CY27 EPS of $12.75.
The real signal here is positioning. Cantor flags that renewed buying interest reflects investors getting comfortable that AI disruption is longer-duration and less severe than initially feared. Market's pricing a beat-and-raise cadence starting with FY27 guide (expected broadly in line).
"Investors are increasingly positioning for ADP’s typical beat-and-raise cadence, beginning with fiscal 2027 guidance that is broadly in line with Street expectations."
Stifel also bumped to $260 last week (Hold), which is fine – not a conviction call, but adds to the bid. The trade is simple: defensible moat, slow-but-stable growth, and a dividend aristocrat (27 years of raises, 2.76% yield) getting re-rated as AI fears fade. Not a high-octane name, but a PM comfort blanket into earnings on July 29.
Guggenheim initiates Buy with $180 PT (26% upside), calling Guidewire the P&C insurance industry's emerging system of record. But the Q3 print tells a more nuanced story — ARR of $1.147B (+19% YoY) missed whispers, even as EPS and revenue beat. The Street's reaction was a quick round of PT cuts: Wells Fargo to $190 (Underweight), Stifel to $200 (Buy), RBC to $215 (Outperform). Net/net: the secular cloud migration and AI up-sell thesis is intact, but the cadence of ARR acceleration is the key near-term bogey.
"Guggenheim believes Guidewire will meet or exceed near-term and long-term targets and prove itself to be a durable high-teens grower for the next few years."
Bull case: Dominant position across claims, policy, and billing. Two prior tech shifts navigated successfully (mainframe→client-server, on-prem→cloud). Now they're the logical partner for AI deployment in a $200B+ P&C market. High-teens revenue growth + 11x EV/NTM recurring multiple leaves room.
Bear case: ARR deceleration is real — 19% YoY is below the "20%+" narrative. Competition from point solutions and in-house builds remains a risk. At $11B market cap, the premium already prices in a lot of success. The $180 PT from Guggenheim is actually below the Street median (~$200), so even the new bull isn't stretching.
Verdict: Buy the dip. Guggenheim throws down the gauntlet with a $440 PT (47% upside) on the cloud migration narrative, and the rest of the Street is already parked in the same camp.
TYL is getting crushed — down ~47% in the last year, trading at $284 near its 52-week low of $270.71. The stock looks like a value trap until you realize this is a classic "busted growth" setup in a vertical that actually has a structural tailwind. Guggenheim initiates Buy, $440. DA Davidson and Truist already have Buy ratings with $460 and $440 targets respectively. Consensus is screaming "the migration is real."
The core thesis: TYL owns the system of record for critical local government functions (property taxes, courts, public safety, etc.) and is leading the shift from on-prem maintenance to SaaS. Management expects ~85% of customers to migrate by 2030 at a 1.7x revenue uplift. SaaS growth is already >20%. The transaction-based business adds a 10% CAGR layer on top. At ~6x EV/NTM recurring revenue, this is cheap for a name with that kind of visibility.
Guggenheim specifically flags that TYL meets four of five moat criteria against AI obsolescence — that's rare in this space. Street estimates look achievable, with upside depending on migration pace.
"Tyler Technologies serves as the system of record for critical functions in local governments and leads the cloud transition of the traditionally tech-laggard industry."
Also worth noting: a $150M buyback under Rule 10b5-1 runs through July 30, which provides some near-term support. The company has a perfect Piotroski Score of 9 — balance sheet is not a worry.
Risk: Transaction business could be lumpy (Truist noted Q2 headwinds). Migration pace could slip. But the r/r at these levels is compelling for a multi-manager book that can stomach a few quarters of chop.
Buy the sum of the parts discount. APPF is getting street respect but the stock still trades at ~5.1x NTM recurring revenue — a multiple that implies zero terminal growth and impaired FCF margins. Actual LTM revenue growth is 20%+, gross margins are 63.8%. Guggenheim just launched with a Buy, $232 PT (7.4x NTM rev). That's in the same zip code as Benchmark ($226), DA Davidson ($225), and KeyBanc (Overweight, insurance as 30%+ ARPU driver). The cluster is tight — the message is clear: the market is pricing in an extinction-level AI threat that isn't coming.
"We estimate the current valuation implies 0% revenue growth and a significantly impaired free cash flow margin profile in the terminal period." — Tamjid Chowdhury, Guggenheim
The AI angle is actually a tailwind, not a headwind. AppFolio's Realm-X integration with Anthropic's Claude is live, and insurance is becoming a real revenue escalator. Macro backdrop is mixed — high rates keep renters renting (good for property managers), but rising vacancy offsets some of that. Still, at these multiples, even a conservative 5% growth / 35% FCF margin DCF gets you to $232. If they hit consensus (and they beat Q1 by $4M), the risk/reward is heavily skewed to the upside into 2027.
VERDICT: CACI IS CHEAP FOR A REASON, BUT THE FLOOR IS HOLDING. Cantor reiterates Overweight with a $675 target as the stock sits 16% below YTD open (S&P +9%). They're betting on constructive FY27 guidance with at least 5% organic growth, but the entire set-up hinges on defense budget outcomes from mid-terms and the FY27 appropriations battle.
"CACI trades at a 3-4x discount vs its historical 30-45% premium to the S&P in 2002/2008... the overhang from budget negotiations is priced in, but execution on guidance is non-negotiable."
Bull case: Valuation is compelling. Cantor notes CACI has swung from a 30-45% premium to the S&P (2002/2008 defense cycles) to a discount below its 10-year average of ~15%. If FY27 budget lands at ~$1.1T (March timeline), the current price already embeds worst-case cuts. Recent contract wins (VA $308M ERP modernization, Navy $113M, SkyValor drone system at Southern Border, Army C5ISR +$140M ceiling increase) show the pipeline is active — not dead.
Bear case: Defense stocks are a crowded short on budget anxiety for a reason. Mid-term elections inject binary risk into FY27 appropriations. Even if guidance is "constructive," 5% organic in a potentially shrinking top-line environment is a low bar. The stock is cheap, but cheap can get cheaper if the macro overhang persists through Q4 print. Not wrong to be wary.
Verdict: Buy the print, but don’t chase. Davidison sticking to $82 (54% upside) is the bull case – QTWO is a cloud-migration story with margins inflecting and a PEG ratio of 0.1. But the recent JPM cut to $60 and MS to $63 tell you the AI-defensibility debate is real. Results on the 29th are a tightrope.
Davidson reiterates Buy at $82 ahead of Q2 (July 29). They expect a “meet-or-slightly-exceed” quarter, an affirm or modest fine-tune of 2026 guidance, and – most importantly – preliminary 2027 guidance. That’s the catalyst. The cloud migration completed in January has already widened margins, and subscription revenue grew 17% YoY in Q1.
“We anticipate the company will affirm or modestly fine-tune its 2026 guidance ranges with the results.” – D.A. Davidson
JPMorgan cut PT from $80 to $60 (still Overweight) citing valuation, but admitted concerns around AI-defensibility are “overstated.” Morgan Stanley went to $63 (Equalweight) on higher margins. The risk: new 2027 guidance needs to show the migration-led margin expansion is durable, not a one-time pop. If 2027 guide is soft, the 54% implied upside evaporates fast.
INTC — Q2 blowout is real: 15-year high growth, data center +59%, demand validation from hyperscaler buildout. But the $20B capex hike and negative $8.4B FCF means margin recovery is a multi-year story. Bulls see 48% gross margin heading toward 60%+ as 18A ramps; bears say the balance sheet can’t support the foundry ambition without a capital raise. Rate of change: ACCELERATING, but risk/reward is binary — you're betting on 14A being a real TSMC competitor by 2028.
NVDA — The GPU floor guarantee is a clever financing tool that turns speculative neocloud clusters into bankable projects (banks won't lend without that backstop). But it also shifts residual value risk onto NVDA's balance sheet — a sign natural demand alone isn't enough to absorb all production. Vera Rubin's cableless design fixes the Blackwell reliability nightmare, but the CUDA moat narrative is under attack from DeepSeek's TileLang and Huawei's 950 supernode. Near-term demand is supply-constrained; long-duration share risk is building.
AMKR — The $1.5B prepaid packaging deal with Nvidia is massive validation that advanced packaging is the critical bottleneck in the AI supply chain. This is a geopolitical shift: onshore packaging for AI, with Nvidia putting cash upfront like hyperscalers do for GPUs. Stock +15% AH is pricing in a multi-year strategic partnership, but OSAT margins are structurally low (high capex intensity). Key question: does this re-rate Amkor as a strategic bottleneck play, or is it just one deal?
MU — NAND shortage is structural, driven by AI storage needs (CMX racks can't ship fully populated). HBM4 pricing at $35-40/GB is narrowing the SRAM cost gap. Memory forward PEs are cheap (MU 6x, SKHY 5x) but cycle risk is the debate. Bulls see YMTC staying consumer-focused; bears worry about Samsung capacity additions. Rate of change: ACCELERATING tightness — this feels early-cycle, not peak.
TSM — 25 fabs in 5 years, capex raised to $60-64B — the most visible sign that AI semiconductor demand is structural. Insider cluster buying June 16-July 9. TSM benefits regardless of which chip designer wins (NVDA, AMD, Google, even Intel). But the debate is margins: can they maintain 68% gross margin with that capex load? Intel's 14A being competitive is a credible long-term threat, but TSM's scale advantage ($60B+ capex vs Intel's $20B) is enormous. Rate of change: STABLE at dominant position.