Massive M&A shakeup in AI networking after the bell. ​Bloomberg reports Marvell $MRVL is in advanced talks to acquire Lumentum $LITE , with an official deal potentially dropping as soon as next week. ​After-Hours Reaction: ▪️ $MRVL: $229.80 (-8.68% AH, after closing +3.32% at $251.65) ▪️ $LITE: $990.00 (+8.32% AH, after spiking past $1,000 from a $913.95 close) ​The biggest question on the desk right now: With Lumentum’s CEO confirming earlier this month that their laser capacity is completely sold out through early 2029, why sell the company now? ​Three practical reasons: 1. ​Peak Valuation Leverage: $LITE is sitting near an $83B market cap after a massive multi-month run. Management never has more negotiating leverage to command a top-dollar premium than when 70% of customer demand is going unfulfilled. 2. ​CapEx & Competitive Risk: Scarcity attracts supply. Rivals like Coherent ($COHR) are aggressively doubling Indium Phosphide laser output through 2028. Scaling fabs to meet 2029 backlog takes billions in CapEx and heavy execution risk. 3. ​Full-Stack Scale: Combining Lumentum’s photonics and Optical Circuit Switching (OCS) with Marvell’s DSPs and custom silicon creates a $300B+ vertically integrated giant capable of going toe-to-toe with $AVGO. ​One major wrinkle to watch: $NVDA put $2 billion into Lumentum back in March to secure its own optical supply chain. Watch how Jensen reacts to Marvell trying to buy his laser supplier outright.

$MSFT Q1 FY27: Clean beat across the board. Azure hits +48% YoY. ​Revenue: $92.3B (Est. $90.69B) | +19% YoY 🟢 ​Adj. EPS: $4.89 (Est. $4.73) | +18% YoY 🟢 ​Operating Income: $45.9B (Est. $44.06B) | +21% YoY 🟢 ​Azure: $33.2B (Est. $32.32B) | +48% YoY vs. +45.2% est. 🟢 ​Agents & Infra: $76.8B (Est. $75.6B) | $40.6B OpInc 🟢 ​Devices & Consumer: $15.5B (Est. $15.0B) | $5.4B OpInc 🟢 ​Cash CapEx: $39.4B (Est. $46.59B) 🟢 ​Commercial RPO: $677B (Est. $705.6B) Azure at $33.2B (+48% YoY) is now on a $132.8B annualized run rate, accelerating from 43% last quarter. Combine that with cash CapEx coming in $7.2B below Street estimates and a near-50% company-wide operating margin, and it’s clear why the stock is ripping past its $551 52-week high after hours.

Massive quarter from Vertiv this morning. $VRT just showed how strong physical data center demand actually is in Q3: ▪️​Revenue: $4.31B (+61% YoY, beating estimates by $550M) ▪️​Adj. EPS: $2.21 (+78% YoY, beating by $0.38) ▪️​Organic Orders: Up 100%+ YoY ▪️​Americas Revenue: $3.0B (+75% YoY) ▪️​Adj. Operating Margin: 25.2% (+290 bps YoY), guiding up to 25.3%–26.3% in Q4 ​Vertiv also raised its full-year 2026 outlook to $14.65B in revenue (at the midpoint) and $7.10–$7.20 in Adj. EPS, pulling in $2.6B–$2.8B in free cash flow. ​Pair these numbers with yesterday’s $1.45B acquisition of UtilityInnovation Group, which gives Vertiv behind-the-meter power and microgrid architecture and their strategy is clear. They are locking down the entire chain from grid interconnect straight to the rack. $VRT up 15% in pre-market.

OpenAI reportedly plans to buy $GOOGL TPUs, right as $ORCL signs a deal to become the first external cloud to rent them out. ​Google’s custom chips are no longer locked inside Google Cloud. ​Nobody is walking away from $NVDA, but frontier labs can't afford to rely on a single hardware vendor anymore. They need every gigawatt of compute they can get, and they need cheaper inference. ​Big validation for custom ASICs and a major tailwind for $AVGO as the multi-chip era scales up. This is exactly why I’ve been accumulating $AVGO on the recent dip ($320s–$330s). ​Simple logic: when every tech giant is racing to build custom chips to lower their compute costs, you buy the company building the custom chips. After defending support last week and reclaiming $340 today, the setup looks primed for a run back to $400 by EOY.

SK Hynix $SKHY just posted an 80% operating margin on $78.9B in quarterly revenue. ​Those are software margins on heavy semiconductor manufacturing. ​Q3’26 Numbers: • ​Revenue: $78.9B (+328% YoY, +32% QoQ) - beat by $4.1B • ​Oper. Profit: $62.8B (+632% YoY, +38% QoQ) - beat by $4.3B • ​Gross / Oper. Margin: 85.7% / 80.0% • ​Cash vs. Debt: $86.3B cash vs. $11.9B debt Market spent the last month worrying about a 2027–2028 memory cliff. Look at what management actually put on the table today: 1. ​5-Year LTAs: Signed with 14 major customers (incl. global CSPs), locking up over 50% of 2027–2028 capacity right now. 2.​HBM4 Transition: HBM4 passes HBM3E in both shipments and revenue share in Q4, with 2027 HBM contracts already finalized and conventional DRAM margins rapidly closing the gap. 3.​ Q4 Sequential Acceleration: Guiding DRAM bit growth up mid-20% QoQ after a high-20% ASP jump in Q3. 4. ​Massive Cash Return: $15.08/sh Q3 dividend (boosted by Kioxia gains) + $15.1B in Q3 share purchases, pushing full-year 2026 shareholder returns to ₩100T ($75B). ​ Net income came in at $30.0B due to a $24.6B non-operating hit ($17.3B investment mark-to-market loss and $8.8B FX translation from a stronger won). Operationally, cash flow hit $41.3B for the quarter and management expects memory demand to outrun supply through 2030.

$SOFI just dropped a monster Q3 beat-and-raise, plus their first look at 2027. ​Wall Street spent all month cutting price targets to $15 while the business was quietly putting up record scale. ​Q3’26 Numbers: • Revenue: $1.4B vs. $1.25B Est. (+44% YoY) • EPS: $0.18 vs. $0.16 Est. (+64% YoY) • Adj. EBITDA: $469.8M (+70% YoY — 34% margin) • GAAP Net Income: $241M • Members: 17.1M (+35% YoY | Record +1.2M added in Q3) • Products: 26.7M (+44% YoY | Record +2.2M added in Q3) ​FY26 Guidance Raised: • Adj. Revenue: ~$5.04B vs. $4.87B Est. (+40% YoY) • Adj. EPS: ~$0.61 vs. $0.59 Est. • Adj. EBITDA: ~$1.66B | Net Income: ~$840M ​Brand New FY27 Guidance: • Adj. Revenue: ~$6.4B vs. ~$5.9B–$6.0B Est. (+27% YoY) • Adj. EPS: ~$0.90 vs. $0.82 Est. • Adj. EBITDA: ~$2.24B | Net Income: ~$1.28B (~20% net margin) $SOFI is trading at ~17x 2027 EPS with $51.3B in deposits, a 6.14% NIM, and $570.6M in quarterly fee-based revenue. ​8 straight quarters of beating estimates. The fundamentals are completely detached from the YTD stock chart.

“Alea iacta est” The battle lines in telecom have officially been drawn. ​SpaceX’s 80,000 ground tower push to build a 4th carrier is the confirmation of ASTS’s core thesis: Even with 15,000 satellites, you still need massive terrestrial infrastructure to serve the retail market. ​By choosing to build its own 4th carrier, $SPCX forces AT&T and Verizon to accelerate $ASTS commercial rollout to defend their turf. Space-based cellular won’t kill mobile network operators, it is going to save them.

Mqrket spent the last week selling $RKLB down to $51 over the Neutron push to 2027. Meanwhile, the core business has a completely different problem: clearing a massive line of paying customers. ​Rocket Lab’s bottleneck right now is testing and pad throughput, not demand. ​Look at the actual numbers behind this new test facility: ▪️​100+ launches sitting in the backlog (boosted by the 20-launch ▪️Synspective deal locked in on Sept 30) ▪️​97 missions flown to date, with 18 already in the books for 2026 ▪️​$2.36B+ total backlog (+137% YoY) and $2.13B in cash on the balance sheet ​Expanding static fire capacity for Electron and HASTE means faster turnaround times, higher annual launch cadence, and quicker revenue conversion while we wait for Neutron Stage 1 testing in Q1. ​With Q3 earnings coming up on Nov 9 ($250M–$265M revenue guide), the core engine is running hot. https://t.co/dpatqsEUXy

$BE down 8.7% today and another ~3% AH to ~$210 after posting one of the strongest growth quarters in the energy sector: ​🔹 Rev: $1.24B (+138% YoY) vs. $1.06B est. 🟢 🔹 Adj. EPS: $0.94 (+527% YoY) vs. $0.68 est. 🟢 🔹 Adj. Gross Margin: 34.5% (+412 bps YoY) 🟢 🔹 FY26 Rev Guide: $4.4B–$4.6B vs. $4.1B est. (3rd raise this year) 🟢 🔹 FY26 EPS Guide: $3.10–$3.40 vs. $2.71 est. 🟢 ​The only blemish? $27.1M in operating cash flow because they are aggressively building out manufacturing capacity to keep up with hyperscaler data center orders. ​Triple-digit revenue growth, expanding margins, $2.49B in cash, and an undeniable bottleneck in grid power. If the market wants to sell this off on CapEx timing, it goes straight to the top of the dip-buy list.

SpaceX $SPCX is coming for the Big 3 telecoms. ​Per Bloomberg, SpaceX is in talks to lease at least 80,000 cell tower sites across the U.S. to build a 4th nationwide wireless carrier from scratch. ​The puzzle pieces are falling into place fast: ▪️​Tower Infrastructure: Negotiating multi-billion-dollar deals with major tower operators, including Crown Castle. On yesterday's Q3 call, $CCI CEO Chris Hillabrant called taking over abandoned DISH network sites the "industry cheat code" for speed to market. ▪️​Spectrum: Locked in a deal on Oct 8 to acquire nationwide 800 MHz low-band spectrum from Grain Management (on top of its $17B EchoStar mid-band deal). ▪️​Space Layer: Secured FCC approval on Oct 7 to deploy 15,000 VLEO satellites for direct-to-cell service. ​Elon Musk confirmed on this week's Q3 earnings call that SpaceX is building out terrestrial networks alongside Starlink to make Starlink Mobile a major U.S. carrier. ​The immediate after-hours market reaction tells the whole story: 🔹️​Tower REITs ripping: $CCI closed flat at $78.83 (+0.23%) and just spiked +10.33% to $86.97 after hours. $AMT is up +3.55% ($188.02) and $SBAC +1.46% ($181.00). 🔹️​Legacy carriers dumping: $T (-5.85%), $VZ (-4.62%), and $TMUS (-4.52%) all gave back today's regular-session gains in minutes.