Bloomberg: Solidigm may IPO in 2027, raising ~$10B at up to $100B valuation. Nothing finalized. An enterprise-SSD valuation anchor for $SNDK $MU $WDC or competition for capital. Latest reports cite $100B vs $150B earlier. $GS, $MS, $JPM, $C, $UBS : fees on what could be a record tech IPO.

$APLD revenue: $341.9M (+322% YoY), including $183.5M tenant fit-out. Base rent: $65.8M. Interest: $77.4M vs adj. EBITDA $64.4M. Debt: $6.4B. Demand is there; cash flow must catch up. Watch the 300MW year-end delivery target. Revenue growth still has an interest bill.

SpaceX reportedly plans $40B in debt to buy Nvidia chips: $10B loans, $30B IG bonds. Potential upside for $NVDA. Management's $100B ARR claim needs cash-flow proof. GPUs can be bought on credit. Interest needs cash. Watch compute pricing, GPU life and debt terms. $NVDA $VRT $ETN $ANET $CRDO $ALAB

FOMC minutes: unanimous 25bp hike to 3.75–4.00%; most saw another hike likely by year-end. 10Y hit 5.36%, 30Y ~5.72%. Higher yields pressure equity multiples and floating-rate borrowers. Earnings must do more work.

$AMD reportedly seeks more 2027 capacity from TSMC, HBM suppliers and Foxconn. No volumes disclosed. AI bottlenecks support upstream pricing power. $TSM $ASX $AMKR $AMAT $LRCX and $KLAC Booking ahead signals demand. Track firm orders and double-booking risk before counting the growth.

Anduril's "largest shipyard since WWII" is real. Arsenal-2, Sparrows Point near Baltimore: → $3.7B of Anduril's own capital → 2M+ sq ft, ~40 football fields → Navy contract worth $2.9B, paid on demonstrated production → Hiring in 2029, operations in 2030 Same site where Bethlehem Steel built Liberty ships in WWII. Hence the Baltimore clipper reference. Arsenal-2 builds Virginia-class components and modules. Torpedo tubes first, then bows and sterns. Final assembly stays with $GD Electric Boat and $HII Newport News. The Navy buys ~2 Virginia-class subs a year. Industry has been delivering ~1.1–1.2. The bottleneck is capacity, not demand. Anduril is adding a big new module supplier to take load off the two prime yards. ⚠️ Anduril has no heavy-industry track record. Nuclear-grade welding, certification and skilled labor are a different game from drones and software. ⚠️ 2030 is a long way out. Little help for sub deliveries this decade. ⚠️ "Largest since WWII" is about new build scale. WWII yards built whole ships. Who benefits (my read): $GD $HII: near term, relief. More outsourcing = faster deliveries. Long term, watch whether Anduril becomes a competitor for their module work. $BWXT: open up the hull bottleneck and the reactor becomes the next one. That's pricing power. Anduril is private, so you can't buy it. The bigger signal: defense tech is moving from drones and software into the primes' heavy-industry turf.

ISM Services: 54.9; prices: 74.0, highest since July 2022. The 10Y stays near 5.3%. Slower growth + hotter prices complicate Fed easing and pressure equity multiples. Watch Oct. 7 FOMC minutes, 2pm ET. Bonds offer competition. Stocks need earnings to make their case.

Google signs a 20-year power deal with $CEG for 890MW added via nuclear uprates. CEG +14.84%. Big Tech is funding new capacity. Watch $GEV $VST $BWXT. AI needs power. After a near-15% rally, follow-on contracts, approvals and execution need to deliver.

CXL Isn't Killing HBM. It's Handing Hyperscalers a Weapon. There's an excellent explainer making the rounds arguing that CXL — pronounced dead in 2024 — is back. I agree with the frame. The key line is that CXL adds a floor to the memory pyramid; it doesn't replace one memory with another. HBM sits on top (bandwidth), DRAM in the middle, SSD at the bottom — and CXL wedges into the gap where "lukewarm" data lives. Three things revived it, and I'd sign my name to all three: 1. Training → inference. Training wanted bandwidth (NVLink's world). Inference wants capacity — the KV cache alone eats 80–120GB per GPU. Capacity is exactly what CXL does cheaply. 2. Crisis economics. Memory went from ~8% of hyperscaler capex to ~30% in two years. At that price, pooling idle "stranded" memory and reviving DDR4 you were about to scrap (Meta's Vistara paper) suddenly pencils out. 3. PCIe finally got fast enough. A PCIe 6.0 x16 link now carries ~2.5 DDR5 channels' worth of bandwidth. The road to the distribution center is finally wide enough to use. And the tell that matters most: in 2023 CXL was vendors pushing. In 2026 it's buyers deploying — Microsoft (Azure M-series), Meta (full-stack, own ASIC). That shift, from a seller's pitch to a buyer's order, is the whole story. So who benefits most? Read it as a value chain, not a ticker: The pick-and-shovel layer wins first and cleanest: CXL controllers and memory-interface IP (Montage, Marvell, Astera Labs, Rambus, and the EDA/IP houses). Every CXL module — at every deployment stage — needs a controller. That's the layer that gets paid regardless of who wins. The biggest dollar winner is the buyer itself — the hyperscaler. The entire point of CXL is to cut the memory tax: recycle what you already own, pool what sits idle, decouple capacity from the CPU. The value CXL creates is captured largely by the people deploying it. That's the counterintuitive part everyone misses. DRAM makers are mixed. Near term, CXL modules sell more standard DRAM and open a new outlet. Long term, pooling and recycling suppress incremental DRAM demand. Cut both ways. HBM is largely insulated. CXL solves capacity, not HBM's bandwidth job. "Does HBM demand take a hit" is the wrong question — CXL is complementary, not a substitute. Personal notes for informational purposes only. Not investment advice, not a trade signal.

US utility giant NextEra Energy #NEE# is reportedly in talks to combine with rival Dominion Energy #D# in a deal that could create a $400 billion power company as surging electricity demand from AI data centers reshapes the U.S. energy landscape. According to a report by the Financial Times on Friday, the talks could result in one of the largest mergers in corporate history and may be announced as soon as next week. The discussions are ongoing and could still collapse, the report said, adding that the transaction is expected to be structured largely as a stock deal.

#NVDA# Q1 2026 13F： Intel — 214,776,632 shares Nebius — 1,190,476 shares Coherent — 7,788,161 shares CoreWeave — 47,213,353 shares Generate Biosciences — 833,325 shares Nokia — 166,389,351 shares Synopsys — 4,821,717 shares