Market Overview

AI infrastructure remains the dominant market narrative as we enter June. Today’s action is bifurcated: Marvell surges 25% on Jensen Huang’s explicit endorsement at Computex, HPE rips 24% on a blowout AI server quarter with raised guidance, while the broader Dow drifts lower on unresolved U.S.-Iran tensions. The semiconductor and optical networking complex is on fire — the SPDR S&P Semiconductor ETF (XSD) has tripled in a year — but geopolitical uncertainty is keeping a lid on the cyclical/value rotation.

Top Movers

MRVL (+24.8%) — $273.80 → $300 (+9.6% upside) Thesis: Jensen Huang publicly blessed Marvell’s custom silicon business at Computex, calling out the company by name in the context of a “$1 trillion” custom chip TAM. This is the single strongest catalyst a fabless chip company can receive — an Nvidia endorsement effectively validates the design pipeline and signals hyperscaler wins are real. RSI at 50 despite today’s move suggests the stock wasn’t overbought coming into this, so there’s room to run before exhaustion. Levels: Exit at $300 (psychological round number, likely where profit-taking intensifies). Support at $240 (pre-gap level; any fill below $250 invalidates the breakout thesis).

HPE (+24.3%) — $58.42 → $62 (+6.1% upside) Thesis: Fiscal Q2 was a genuine blowout — record backlog, AI server revenue inflecting sharply higher, raised FY26 guidance, AND they issued a FY27 outlook of 8-12% sales growth. Multiple analysts raised price targets and one upgraded. This isn’t a one-quarter wonder; the backlog gives multi-quarter visibility. However, RSI at 91.8 screams near-term overkill — this is a name to buy on a pullback, not chase at the open. Levels: Exit at $62 (where stretched valuations likely cap near-term upside for a hardware name). Support at $52 (prior resistance now support; 53-week high before this gap).

FLNC (+6.4%) — $28.90 → $34 (+17.6% upside) Thesis: Fluence jumped 42% yesterday on an AI data center partnership with Nvidia, Siemens, and nVent for a 136MW facility. Today’s follow-through suggests the market is re-rating this from “energy storage company” to “AI infrastructure picks-and-shovels play.” The 35% weekly gain tells you this is early in the re-rating cycle — the thesis shift from grid storage to AI power enablement meaningfully expands the TAM narrative. Levels: Exit at $34 (measured move target from the breakout base). Support at $24 (pre-announcement consolidation zone).

AAOI (+8.7%) — $201.74 → $225 (+11.5% upside) Thesis: Applied Optoelectronics is up 439% YTD as the optical networking space gets aggressively bid on AI interconnect demand. The Mediacom contract win gives it a cable/fiber revenue floor, while analyst bullishness on optical stocks broadly provides a tailwind. RSI at 59.7 is neutral — not overheated — and the stock bounced hard off Friday’s 9% decline, showing aggressive dip-buying. Levels: Exit at $225 (prior intraday high from the recent range). Support at $175 (Friday’s low; break below means momentum is broken).

CLS (+7.7%) — $459.26 → $500 (+8.9% upside) Thesis: Celestica continues its run as the premier AI infrastructure contract manufacturer. RSI at 70 is borderline overbought but not extreme, and the 24% weekly gain reflects genuine institutional accumulation. The company has a pristine 99 EPS Rating and is benefiting from the exact same hyperscaler capex wave driving MRVL and HPE. This is a “picks and shovels” play on AI data center buildouts. Levels: Exit at $500 (round number resistance, likely options gamma magnet). Support at $420 (prior breakout level from last week’s gap).

KSS (+6.9%) — $15.92 → $18 (+13.1% upside) Thesis: Kohl’s posted its best comp sales in 4+ years with digital up 4%, and management emphasized early turnaround traction. RSI at 12.9 — extremely oversold on a longer timeframe — means this is bouncing off a severely depressed base. The UBS note flagging softline retailers benefiting from data center construction boom employment is an unusual but interesting angle. This is a contrarian long in a deeply hated name showing green shoots. Levels: Exit at $18 (prior resistance and round number). Support at $14.50 (recent 52-week low area; break below means turnaround is failing).

VNET (+6.5%) — $11.29 → $13 (+15.2% upside) Thesis: Chinese data center operator reporting 19.8% revenue growth driven by AI wholesale demand, reaffirmed 2026 guidance. RSI at 16.7 is absurdly oversold despite the stock being up 268% over 3 years — suggesting a vicious recent pullback that’s now reversing. The AI data center demand in China is real and VNET is a direct play. Levels: Exit at $13 (prior consolidation zone from recent weeks). Support at $10.50 (where buyers stepped in during the pullback).

LAC (+13.9%) — $6.28 → $7.00 (+11.5% upside) Thesis: Lithium Americas turned profitable for the first time as Thacker Pass construction advances and Argentina operations generate cash flow. This is a real fundamental inflection — going from cash-burning explorer to profitable producer changes the investor base entirely. The potential ASX listing could unlock new capital pools. Lithium sentiment has been in the gutter, making this a contrarian play with actual earnings support now. Levels: Exit at $7.00 (round number and prior resistance). Support at $5.50 (the base from which this rally launched).

Headlines to Watch

  • Jensen Huang endorses Marvell at Computex, frames custom silicon as “$1T” opportunity — This is the kind of partner validation that creates multi-month momentum; watch MRVL, AVGO, and custom ASIC plays.
  • Alphabet plans to raise $80 billion for AI infrastructure — Confirms hyperscaler capex isn’t slowing; direct read-through to AVGO, MRVL, CLS, and networking names.
  • HPE raises FY26 guidance and issues FY27 outlook of 8-12% growth — First enterprise server company to give ‘27 visibility; validates the multi-year AI infrastructure thesis.
  • U.S.-Iran talks “continuing at a rapid pace” per Trump, but Iran reportedly hitting brakes — The ambiguity is keeping oil elevated and creating a headwind for the broader market; resolution either way unlocks the next leg.
  • Fluence Energy joins Nvidia/Siemens on 136MW AI data center design — Energy storage just became an AI adjacency; watch for more partnerships as power becomes the bottleneck.
  • SpaceX, OpenAI, and Anthropic IPOs reportedly coming — Yardeni says won’t “suck oxygen” from markets — If anything, these IPOs may create a wealth effect that flows back into public AI names.
  • Broadcom earnings tomorrow — The biggest event of the week for semiconductors; a beat/raise could send the entire AI chip complex higher, a miss would be devastating to the names running today.

Claude’s Top Picks

MRVL (+24.8% today, +31.5% week) — $273.80 → $300 (+9.6% upside) Valuation: Trading at a premium to AVGO on forward P/E, but the custom silicon TAM is expanding faster than legacy chip businesses — the growth rate justifies the multiple if hyperscaler wins materialize as guided. Upside: Jensen Huang’s endorsement is the most powerful catalyst in semis; institutional money will chase this name for weeks as sell-side raises targets. Risk: Broadcom earnings tomorrow — if AVGO disappoints, it drags the entire custom silicon narrative down with it, including MRVL.

FLNC (+6.4% today, +35.5% week) — $28.90 → $34 (+17.6% upside) Valuation: Still well below its 2022 IPO-era highs despite a fundamentally stronger business; the AI data center partnership reframes the story entirely. Upside: The thesis shift from “grid storage” to “AI power infrastructure” is only 48 hours old — most institutional investors haven’t had time to update models yet. Risk: Energy storage remains a lumpy, project-based revenue business; one delayed contract and the stock gives back 20%.

KSS (+6.9% today, +19.8% week) — $15.92 → $18 (+13.1% upside) Valuation: Trading at roughly 5x forward earnings — absurdly cheap even for a struggling retailer, and now showing first evidence of operational improvement. Upside: RSI of 12.9 means this was washed out and hated; best comps in 4+ years changes the narrative from “death spiral” to “turnaround in progress.” Risk: Retail turnarounds have a high failure rate; one bad quarter erases all goodwill and sends it back to $12.

VNET (+6.5% today, +13.8% week) — $11.29 → $13 (+15.2% upside) Valuation: China data center peers trade at higher EV/Revenue multiples; VNET is discounted despite 20% revenue growth and AI tailwinds. Upside: RSI of 16.7 on a stock with reaffirmed guidance and AI-driven wholesale growth is a textbook mean-reversion setup. Risk: China regulatory risk and wider net losses could keep a lid on multiple expansion; geopolitical headlines can crush ADRs overnight.

Avoid

HPE ($58.42, +24.3% today) — RSI at 91.8 is extreme overkill for a hardware company. The quarter was legitimately great, but chasing a 24% gap on a $58 stock with a hardware multiple is asking to be the exit liquidity. Wait for a pullback to $52-54.

PENG ($69.92, +17.1% today) — Semiconductor sympathy move with RSI at 81.4. The Zacks upgrade and Nokia board appointment are real but modest catalysts that don’t justify a 27% weekly gain. This is riding MRVL’s coattails.

HSAI ($22.07, +10.2% today) — RSI at 0.74 is a data anomaly, but the stock was down 16% after its Mercedes-Benz announcement — today’s bounce looks like a dead cat. Lidar companies have repeatedly disappointed on commercialization timelines. The Mercedes deal sounds great but the market already faded it once.

WSB Sentiment Check

SPCE — WSB says: BULLISH (80% bullish) Claude says: DISAGREE — 2,622 mentions with 30K upvotes screams retail FOMO frenzy. Virgin Galactic has been a serial retail trap; without seeing this on the top gainers list today, the enthusiasm is likely narrative-driven (SpaceX IPO halo effect?) rather than fundamentals-driven. When WSB is THIS unanimously bullish with this many mentions, you’re usually buying the top.

HPE — WSB says: BULLISH (80% bullish) Claude says: PARTIALLY AGREE — The earnings are genuinely strong and the AI server thesis is real. But WSB is discovering this AFTER a 24% gap and 53% weekly gain with RSI at 92. They’re right on direction, wrong on timing. The smart play was owning it before earnings, not chasing it today.

MU — WSB says: MIXED (55% bullish) Claude says: AGREE with the mixed read — Micron sparked a semiconductor rally last week (per PENG news), but memory is cyclical and we’re likely mid-cycle on pricing. The 55% split accurately reflects the uncertainty. Wait for the next earnings report for clarity.

NVDA — WSB says: BULLISH (80% bullish) Claude says: AGREE — Nvidia’s Computex presence, new PC chip, and Jensen’s active endorsement of partners all confirm the ecosystem is thriving. NVDA near buy points with the AI infrastructure thesis intact is one of the few times WSB consensus aligns with institutional positioning.

MSFT — WSB says: BULLISH (80% bullish) Claude says: AGREE — Microsoft just got initiated with fresh coverage, Azure AI growth remains the core bull case, and the stock is a core holding for any AI portfolio. Not exciting, but WSB being bullish on mega-cap quality during a bull market isn’t delusional — it’s just consensus.

Earnings Scorecard

AVGO — BEAT (implied) | Stock: +5.1% | Reported: After Close (last session) The modest 5% move on what prediction markets priced as a “near-certain beat” suggests the bar was already high. Broadcom is setting up for tomorrow’s official report — if this was a preliminary leak or pre-announcement, the real fireworks come Wednesday. The stock at $460 near all-time highs means the beat was priced in; any guidance miss tomorrow would be severely punished.

VEEV — Likely MISS or in-line | Stock: -3.6% | Reported: After Close Veeva selling off 3.5% post-earnings in a bull market suggests either a miss or weak guidance. For a healthcare IT name trading at premium multiples, any deceleration gets punished. Not a buy-the-dip unless you see acceleration in the next quarter’s guide.

DOCU — Likely MISS or soft guide | Stock: -3.3% | Reported: After Close DocuSign’s persistent inability to re-accelerate growth continues to weigh on the stock. The 3% decline suggests in-line results with disappointing forward commentary. Avoid — this is a value trap masquerading as a SaaS name.

PANW — Likely soft guide | Stock: -1.7% | Reported: After Close Palo Alto’s modest decline suggests the quarter was fine but the market wanted more. Cybersecurity has been an underperformer relative to AI infrastructure names, and PANW’s platformization strategy needs another quarter to prove out. Hold, don’t add.

RBRK — Stock: -2.8% | Reported: After Close Rubrik selling off modestly after reporting is notable given how hot cybersecurity/data protection narratives have been. Likely a “sell the news” on a stock that ran into earnings. Wait for stabilization before adding.