Original: @cnfinancewatch Primary IR: Iridium stockholders approve Rocket Lab acquisition (Sep 24, 2026) — ~99.6% of votes cast; close targeted mid-2027 subject to remaining approvals. Desk read, not a buy/sell.

Today we are announcing that S&P Global has entered an agreement to acquire OpenZeppelin. Onchain finance is growing from an emerging market into core financial infrastructure, and the standards and rails our team and community built are becoming the rails of global finance. OpenZeppelin smart contracts facilitated over $37 trillion in value transferred, with the vast majority of the largest DeFi protocols, blockchain networks, stablecoins and tokenized funds relying on them. With S&P Global, we expect to accelerate the impact of onchain finance, backed by more than a century of trust in global markets, benchmarks, and risk frameworks. To our clients and to all the users of OpenZeppelin open source tools: • OpenZeppelin Contracts and all our open source applications and tools remain open source, free, and publicly maintained on GitHub. Building open source standards stays a core priority. • Audits, engineering work, and ecosystem programs continue with the same team, brand, quality, and customer experience, with what will be the added benefit of S&P Global's research capacity, market data, and institutional reach. For the last decade, OpenZeppelin has set the security standard for onchain finance. Today begins a new chapter for that mission, together with one of the most trusted names in global markets. Read the full announcement:

Desk check on BTC ETF tape: Sep 15 −$450.4m Sep 16 −$295.9m Sep 17 +$159.5m (IBIT led the flip) ETH same day still −$39.3m. Fear & Greed: 56 Greed (was 50 Neutral). If a take still says “post-FOMC = ETF still draining,” the Sep 17 BTC print already disagrees. Hangover ≠ still outflowing. Not advice.

GOOGL +3.85% on the session we flagged — 3/3 tracked desk notes still constructive on the AI-infra read (@aleabitoreddit · @pequityresearch · @JohnsonZ91127). Price and multi-source commentary lined up. Views tracked here: Not advice.

BTC spot ETF tape (Farside): Sep 10: ~−$283M net (−282.7) Also red Sep 8–11 (multi-day withdrawals). Crypto Fear & Greed ~57 (Greed) the same desk day — mood vs ETF flow aren’t the same signal.

Two AI labs, two listing paths — same selloff tape: Anthropic still pushing a Nasdaq IPO track. OpenAI: not going public this year. SoftBank Group got hit hard in Asia — Nikkei ~−11%; Reuters saw as much as −13.2% in the session. AI upside ≠ a clean trade in the listed vehicles that fund it.

US Crypto Regulation Mapped 1️⃣ Payment Stablecoins → GENIUS Act (already law) - Sets the rules for payment stablecoins - Requires 1:1 reserves backing each coin - Guarantees redemption rights for holders 2️⃣ Token Fundraising / Securities Law → Regulation Crypto Assets (proposed SEC rule, not final) Core question: Is the token itself a security? YES → it falls under the existing SEC securities framework NO → next question: How was it sold? Does the offering satisfy the Howey test? ✅ Satisfies Howey → it is an investment contract → treated by SEC as a "covered investment contract" → Regulation Crypto Assets applies ❌ Does not satisfy Howey → no investment contract, so Regulation Crypto Assets is generally not needed (note: other laws can still apply) If Regulation Crypto Assets applies, there are three exemption paths: A. Startup exemption: raise up to $5M over 4 years + narrative disclosures B. Fundraising exemption: raise up to $75M per 12 months + narrative disclosures, financial statements, and ongoing reporting C. Investment contract safe harbor: an independent path where, once managerial efforts are completed or ceased and a transition report is filed, the investment contract is deemed to end and the token is no longer subject to it 3️⃣ Crypto Market Structure → CLARITY Act (bill, not yet law) Answers "who regulates what" by splitting jurisdiction: - SEC: digital securities, investment contracts, securities offerings - CFTC: digital commodities, spot market oversight, digital commodity exchanges, brokers, and dealers $BTC $ETH

bitcoin:native is up over 20% in a week, but funding has already cooled off. If the recent surge were driven mainly by fresh leveraged longs, funding would likely stay elevated instead of cooling down so quickly. This is also supported by Coinbase Premium data. The index has been deeply negative since the 10/10 crypto crash last year - clearest illustration of just how much damage that crash did to the crypto market. And since May, it has stayed in red without turning positive even once. The index is now back near zero recently. If Coinbase Premium Index turns positive from here, it indicates that the US spot buyers are back. This would be bullish.

HSBC: Memory-led price hikes could drive FY28 earnings upside for $NVDA HSBC expects Nvidia to be able to pass higher memory costs through via higher product pricing. The bank raised: - FY28 Data Centre revenue estimate from $528.1B to $608.8B, ~12% above consensus - FY28 EPS to $14.40, 13% above consensus With bunch of bullish headlines are being released: - AI server prices reportedly +15% or more - $6B deal to strengthen Nvidia's open-weight AI push - H200 shipments to China have started We're very bullish on $NVDA - tomorrow's market open should be fun to watch :> But expect volatility on earnings day itself. If Jensen confirms memory remains tight and customers are absorbing higher prices, it basically implies that memory inflation has not yet reached the point where it starts destroying AI demand. Overall, this is bullish for almost the ENTIRE Nvidia ecosystem. HSBC estimates Nvidia could consume around 63% of TSMC's CoWoS capacity in 2026 and 52% in 2027, which shows just how much Nvidia demand can propagate through the semiconductor supply chain. $TSM → wafers + advanced packaging $MU / $SKHY → HBM $ANET / $MRVL / $AVGO → networking + connectivity $COHR / $LITE → optical infrastructure $VRT → power + cooling $DELL / $SMCI → AI servers In terms of stock performance, $NVDA heads into earnings after 6 straight red closes, its longest losing streak since 2022. Nvidia's earnings are now one of the market's biggest remaining catalysts - effectively every US stocks investor needs to pay attention. If NVDA fails to hold up here, there may not be much left to support the market.

What we got right this week, and what we didn't ✅ Called $SNDK to pull back after its 40% run over the week ✅ Predicted $SKHY would surge on the buyback news, mirroring $SNDK's move after Investor Day ❌ Stayed bullish on Unitree until we realised physical AI might just be... a robot ❌ Missed how fast $BTC and the broader market would rip back on the White House conference and the new Regulation Crypto Assets Gotta do better next week 💪🏼

GOLDMAN SACHS: South Korea's chip exports rose 11.5% MoM seasonally adjusted in the first 20 days of August. Exports to China and the US also jumped 13.3% and 10.0% MoM, respectively. Which suggests that the semiconductor demand is still holding up despite all the concern around the cycle. Obviously bullish for SK Hynix, Samsung, $MU, and $SNDK. To investors, this is especially relevant because memory is unusually sensitive to supply-demand changes. Strong export momentum makes it harder to justify concerns that demand has suddenly collapsed.

Anthropic's potential IPO timeline has reportedly been brought forward to the end of August. Per Bloomberg, the company is targeting an IPO raise that matches or exceeds SpaceX’s record-setting offering. For context, SpaceX raised roughly $75 billion earlier in 2026. Anthropic’s funding rounds, based on publicly disclosed information: Series A - May 2021; raised $124M Series B - Apr 2022; raised $580M --> SBF invested Series C - May 2023; raised $450M Series D - Early 2024; raised ~$750M Series E - Mar 2025; raised $3.5B Series F - Sep 2025; raised $13B Series G - Feb 2026; raised $30B Series H - May 2026; raised $65B Anthropic went from a $550M valuation in 2021 to $965B in 2026, a roughly 1,750x increase in five years. And it may not stop there. Investors are already eyeing a ~$2T valuation for Anthropic’s IPO, which would make it 2026’s second trillion-dollar IPO after $SPCX. If so, the company would be worth roughly 3,600x more than in its Series A round.

JP MORGAN: Google is potentially turning Marvell into one of the key suppliers of the infrastructure surrounding its TPU ecosystem. And the numbers behind the deal are pretty insane. Potential $120B Google opportunity through FY33. The warrant structure implies more than $120B of cumulative Marvell custom-product revenue if fully vested, equivalent to roughly $19.2B per year over 6.25 years. For context, JPM's previous AI ASIC revenue expectation for $MRVL was only around $11.5-$16.8B for FY27/FY28. One important distinction though, is that JPM does NOT think Marvell won Google's core TPU. The opportunity is more about the chips sitting around the TPU and making the entire AI system work. Basically, Google is expanding its custom silicon strategy beyond just the accelerator itself. And $MRVL is positioning itself to supply more and more of that surrounding infrastructure. It validates the broader custom ASIC TAM thesis. Hyperscalers are increasingly designing custom silicon not only for compute, but also for networking, storage, memory and other adjacent workloads. JPM reiterates OVERWEIGHT on $MRVL with a $240 PT. $MRVL has also been one of the stronger AI semicons names recently, holding up relatively well despite volatility. People are now calling $MRVL the next $MU - do you agree with this 🤔 Fun facts we bet you didn't know about $MRVL: - Members of Congress (Rep. Byron Donalds & Rep. Maria Salazar) were buying before the recent run - Jensen Huang literally called Marvell the "next trillion-dollar company.

Chinese traders are BEARISH ON UNITREE 🫪 A lot of the criticism comes down to valuation. People are questioning how a company with just ~RMB 2B in revenue and ~RMB 300M in R&D can justify a RMB 400B valuation. Even Unitree CEO Wang Xingxing has been pretty cautious about the near-term outlook. He said the “ChatGPT moment” for embodied AI has not arrived yet. In the best case, it could take another 2-3 years. In a slower scenario, it may take 5 or even 10 years. One of the biggest problems is that robots still need to be retrained whenever they are given a new task, which makes deployment inefficient. That probably explains why he looked so stressed during the IPO ceremony.

Just how fast the night changes. Another $1,000 up for $BTC, and @saylor will be back in profit again. In 3 days times, $BTC +15% -> $MSTR +25% $ETH +22% -> $BMNR +19% Insane rally across the overall crypto market after BTC remained stable around $64,000 for two months! Markets are finally rewarding the patient HODLers.

FUN FACT: Morgan Stanley tracks IBM Consulting job postings as a leading indicator for Consulting revenue. As of Aug 10, daily active Consulting postings went down -53% vs end of Q2. This is a clear negative read-through for IBM's Q3 Consulting growth, contrasting with management's constructive tone at Q2 earnings.

TLDR: Stripe essentially believes that AI agents will become economic actors. That appears to be the main strategic rationale behind its acquisition of OpenRouter. Strikingly, 88% of the Forbes AI 50, including OpenAI and Anthropic, are building on Stripe, while Stripe says the share of revenue coming from AI companies and crypto is more than doubling YoY. Overall, payments and financial infrastructure are still the foundation of the business. Meanwhile, AI is becoming the main strategic direction. Its existing businesses (see table below) are still growing very fast. According to the letter, H1 2026: - Stripe net revenue grew 41% YoY - FCF grew 43% YoY - Stripe Billing grew 71% YoY - Stripe Capital exceeded 100,000 active loans - Link passed 300 million users - Stripe Treasury is described as one of Stripe's fastest-growing products ever This is where OpenRouter fits into the picture. Stripe manages the money flow, Metronome measures usage, while OpenRouter determines where AI/token demand gets routed across models. Also, don’t expect a Stripe IPO anytime soon. Despite now being valued at $159B, Stripe appears comfortable staying private.