SpaceX President Gwynne Shotwell just revealed the enormous scale of Starlink: “As of June 30th, our constellation of operational Starlink broadband and mobile satellites in orbit grew to roughly 10,200, with our 9,600 broadband satellites delivering roughly 800 terabits per second of total downlink capacity Following a successful speed-run test of the V3 Starlink satellites on the most recent Starship launch...during which we connected every satellite through every laser link.....we intend to deploy V3 satellites into the constellation for operational use on the upcoming Starship missions”

NOD Glass Mousepad Batch 2 Pre-order in USA Store!​ Missed the first drop? Your second chance is here. NOD CR​ (Uncoated Control) & NOD SP​ (Nano-Coated Speed) 🗓️Estimated Shipping: August 30th​. 💎Pre-order now

Honestly spending my 30th with the fans who have made my life what it is at jingle ball, then throwing the most aggressive holiday party known to womankind... I just.. seriously feel so lucky and will spend forever trying to find ways to say thank you. *cries in Christmas tree*

Last wk, S&P/Nas/Mag7 +0.4%/+0.1%/-0.8%. Cooler inflation (CPI, PPI) & economic data (consumer sentiment, retail sales) but +5% oil steepened the yield curve but lowered odds of a rate hike. Looking forward, I continue to believe the impact of Agentic AI with the advent of OpenClaw on January 30th has at least a year to run: 1) Token production has gone up roughly ~7.5x from the end of January more than offsetting the nearly 50% token cost reduction seen since open-weight model usage started to take off in May. 2) Combined annualized run-rate revenues for OpenAI and Anthropic which ended last year at $29B seems to be around $100B currently with Anthropic getting profitable in Q2. 3) Capex from the Big6 hyperscalers accelerated from 84% y/y/ in CQ1 to 92% in CQ2 with forecasts for nearly 100% in Q3. But this is being supported by cloud revenue growth at the 3 Big Public cloud vendors of $AMZN $MSFT $GOOGL accelerating from 23% y/y in Q1:25 to 35% in Q1:26 to 43% in Q2:26. Arguable more important is public cloud operating margins expanded from 34% to 37% and 39% during those time periods. 4) The $500B financing deal backstopped by up to $125B from $NVDA adds even more lower cost money to fund AI capex spend for the non-hyperscaler players. Nvidia gained 0.5% last week. 5) The liquidation of Situational Awareness and retail accounts during July cleared out some of the frothiness in the AI related names In terms of negatives: 1) The cost of money (yields on government bonds) remain near the highest levels for the 30 yr tenor at 5.3% since 2007. 2) Given large scale offensive US military actions are seemingly off the tablein favor of financial sanctions, probably driven by current election polls, I now believe Iran is likely to hold the Strait of Hormuz hostage until past the US mid-terms. This would be akin to them releasing the US hostages in 1981 (they were held for 444 days) just hours after President Reagan was sworn in replacing Carter. There were severe financial sanctions then also. 3) Since 1990, which happens to be the Gulf War, from the end of July through November 9th, which covers the reaction to all mid-term results, the performance is worse than non mid-term years. For mid-term years the median S&P500 gain from 7/31-11/9 is 0.9% with gains 56% of the time but the median peak loss from 7/31 is 6.2% (intra-period median peak loss of 9.9%.) For non mid-term years the median gain is 2.7% from 7/31-11/9 with gains 59% of the time and the median peak loss from 7/31 is 3.5% (intra-period median peak loss of 5.2%.) This year with the momentum seen by the Socialists which are not big business friendly, I see more risk than normal. 4) The easy money on the AI technical rebound from oversold levels on 7/29 due to the forced sale by Situation Awareness is probably over. There were negative stock reactions to headline beat and raise earnings on both revs & EPS for AI infrastructure winners $CSCO (-8% for the week but still up +45% YTD), $AMAT (-6%/+97%) and $COHR (-14%/+77%). While negatives can always be found, their biggest crime was arguably their recent bounce from 7/29-8/7 of 8%, 24% and 71% respectively and their market beating YTD gains. In summary, I remain bullish. Even from the end of July through November 9th during mid-term years since 1990, the S&P has an additional median gain of 4.2% to its peak before giving some of that back closer to the election. Given some of the negatives, especially the reaction to solid earnings data, I would add some hedges back on further market gains and get more selective. Consumer discretionary hedges should also make sense if oil is higher for longer. I believe value should continue to accrue to the infrastructure layer which includes 1) the public cloud vendors such as Amazon, Microsoft, Google and 2) the semiconductor companies. $INTC, my favorite semi company, still gained 0.8% last week despite: 1) a $20B equity offering which causes ~5% dilution and 2) being up 178% YTD. This clears the funding overhang. All the best in the week ahead.

Last Sunday, I previewed caution on capex spenders & more optimism on their customers. This past wk, Mag7 (spenders) -5.8%, while Semi Index (customers) +1.2% with S&P -0.6%. This wk, we get earnings from four of the Mag7 for a more complete picture. Last week, WTI +8% and yields across the 2-30 yr curve increased by 9-15 bps, which was also a major issue for equity markets. De-escalation in Iran over the weekend should help both oil prices and yields during the start of this upcoming week. I continue to believe November 3rd mid-terms put a powerful impetus on the US administration to get oil prices down well before then. This is my current framework for AI investing: Positives: 1) The advent of Agentic AI on January 30th with the formalization of OpenClaw will drive 10-100x more token production. 2) Microprocessor vendors benefit in the move from 8 to 1 GPUs per CPU to ~unity in the move to Agentic AI. 3) The current philosophy of the hyperscalers that they cannot afford to lose the AI race keeps me more bullish on the recipients of that capex spend than the spenders. Negatives: 1) AI native revenues will be determined in the near-term by the trade-off between the top 1% of companies focusing on controlling their AI spend vs the ramp of the other 99% given the move from token maximization in March to token minimization by June. 2) Part of that control of AI spend will come at the expense of non-security software companies, IT services and headcount. 3) The cost of money is also more expensive which is a headwind to equity multiples. 11 of the major central banks have gone from cutting rates from roughly 2.7% in July of 2025 to 2.4% by early February to raising them to 2.6% today. At the same time, 10 year treasury yields across the 11 associated countries stayed flattish at roughly 3.2% from July of 2025 through February and has now risen to 3.8%. For megacap earnings this week, my belief is much like last week, ROIC (the interplay between capex and future EPS estimates) is likely to have an outsized impact on the near-term direction in stock prices. $META should have benefitted from increased engagement during the World Cup. But Meta revs grew 27% in Q3:25 vs 22% in Q2:25 creating tough comparisons for the Q3:26 guide. In addition, a potential launch of a public cloud and LLM API may embolden Meta to increase capex spend. Fortunately, valuation in the teens is low vs 22x CY26 PE for the S&P. $MSFT Azure has a high bar given Google GCP revs accelerated from +63% y/y in CQ1 to 82% in CQ2. But I wonder if the focus on AI cost control is an increasing tailwind for Co-Pilot. It operates natively within the Microsoft 365 ecosystem where enterprise work already happens. But Microsoft owns 27% of OpenAI which I remain negative on given they are caught between Google in consumer AI and Anthropic in enterprise. $AMZN AWS rev growth also has a high bar to clear due to GCP. In addition, higher oil prices are likely to be pressuring the logistics cost of their e-commerce business as well as consumer purchasing power. The shift of their 4 day Prime event to late June from early July should benefit Q2 but at the expense of Q3 guide. $AAPL is benefitting from the AI capex spend of others, especially their partner, Google. But I believe CQ3/CQ4 estimates are too high for both revenues and margins (due to rising semiconductor prices) and valuation at a 37x PE is expensive. From a longer-term perspective, I remain bullish on the potential upgrade cycle from a foldable phone with AI enabled Siri. Best of luck in the week ahead.

Every great industrial revolution has had overinvestment due to the potential for riches for the last companies standing. Eventually this ends in a bust. AI I believe is no different. But I believe the current “speedbump” is not the beginning of that ultimate bust. On Saturday June 20th, I laid out my near-term concerns of an AI “speedbump” due to: 1) Token minimization 2) Competitive low cost open-sourced LLM models 3) Rising semiconductor cost impact on Q3 guidance Monday June 22nd in hindsight turned out to be the short-term top for the semiconductor index and the momentum trade. Token maximation in March turned into token minimization by June with the most extreme example being the $COIN CEO on June 26th posting how they cut their token spend by nearly 50% by largely routing AI queries to cheaper models. In the near-term, the question becomes can token usage by the other 99% of firms go up fast enough to offset the top 1% of firms like Coinbase cutting their AI bills. Numerous cheaper LLMs have been introduced recently. Last week the introduction of Moonshot’s Kimi K3 (China-based) challenged the performance of the most advanced US models. Profitable AI native revenue growth that is ROIC positive is what is needed to keep the whole ecosystem functioning. $GOOGL reporting this Wednesday will provide the first major datapoint on the trade-offs between cheaper tokens and more token production. Google Cloud Platform has seen revenue growth year-over-year accelerate from 34% in Q3:24 to 48% in Q4 and 63% in Q1:25. This growth rate should accelerate further in Q2 due to token maximization. While I expect forward Rev/EPS to move higher post results due to their core business, comments on GCP growth are likely to drive the stock reaction. While $GOOGL is my favorite consumer AI play given they have the complete AI stack, I am not sure they are immune from the leading 1% of companies trying to cut their AI bills. Uber for example that blew their entire AI budget for the year in the first four months has GCP as their primary supplier. $INTC which reports on Thursday provides multiple ways to win at the AI infrastructure layer which I am more bullish on than the increasingly commoditizing LLM model layer: 1) Agentic AI is driving a surge in demand for server CPUs which are a new bottleneck, 2) their advanced packaging has already attracted several hyper-scalers and 3) their foundry business (our national champion with an investment by the US government) continues to improve and attract new customers. As for the current AI “speedbump” in stocks, the Morgan Stanley Momentum Index (MOMO) which subtracts the Long Index from the Short Index fell 28% since June 22nd through July 16th in just 24 days with a slight reprieve of 1.6% on Friday. Historical corrections of over 10% since 1996 have averaged 20% from peak to trough but have taken 48 days on average to bottom. But the rally of the MOMO index of 40% from the March 30th stock market bottom through June 22nd was also much sharper than historical precedents. In my opinion, investing is about the risk versus reward. On June 20th, it was not good with increasing examples of token minimization. From a technical basis, MOMO is still not oversold given the RSI only reached 35 on 7/16 and on average it bottoms at 31. But the risk vs reward is more favorable today with sentiment having fallen further following: 1) the negative pre-announcement by $IBM which declined 26% last week despite prior claims of being an AI beneficiary, 2) the 3% drop in $ASML and 8% decline in $TSM last week despite positive earnings and 3) the 10% decline in both the Semiconductor Index and MOMO last week. Finally, I believe the advent of Agentic AI which arguably started on January 30th of 2026 with the formalization of OpenClaw requires 10-100x more tokens vs Chat-based AI. As a result, I believe it is prudent to start adding back some exposure in the AI related infrastructure names. But I believe this needs to be balanced with prudence at the public cloud layer given the near-term focus on AI bills needing to be controlled by the top 1% of corporations.

Given a couple of my friends misunderstood my short tweet yesterday, let me clarify. My post on Hynix was to acknowledge a datapoint for those calling THE TOP of the AI cycle. I believe it is important as an investor to not just focus on what is supportive of your investment case but even more so to focus on what is directly against it, especially if it is a universally acknowledged loved idea like AI. Having said that, let me be clear. I still strongly believe this is a “speedbump” which is what I have been writing about since my original post on June 20th. But a “speedbump” can be ugly as I have pointed out before. Starting in both late 1995 and 1997 there were drawdowns of over 50% in the semiconductor index. But the semi index finished up roughly 850% from the end of 1994 to the peak of the internet buildout in March of 2000 despite this. Also back then, those drawdowns had fundamental drivers. In 1995, Windows 95 did not lead to the upgrade cycle expected and companies were sitting on a bunch of DRAM inventory. Intel wrote off about $1B in DRAM inventory as an example. There is no excess memory inventories on balance sheets today. In 1997, the Asian currency crisis forced a slowdown in end-demand. The Iran war and continuation of the Ukraine conflict are both disruptive but there is no general slowdown in demand being caused by them that I can see. From a technical standpoint, I believe forced liquidations and margin calls in both retail accounts and hedge funds that typically run with leverage over the past couple of weeks is leading to a technical bottom. From June 22nd, the peak of the momentum trade, the S&P is down only 2.1% and the Nasdaq is down 6.6%. But the SOX index (the tip of the spear in the AI trade) is down 28.6%, the Morgan Stanley Momentum Index is down 38.0% and their Momentum TMT (tech, media and telecom) Index is down 53.5%. The ferocity of these moves in roughly one month has never been seen before in some cases. I believe this has sped up the cleansing process with prime brokers not wanting another Archegos situation. From a fundamental standpoint, the advent of Agentic AI started arguably at the end of January 30th with the formalization of OpenClaw. This requires 10-100x more tokens than chat-based AI. No question that token minimization is going on at the top 1% of companies which was the main reason for my speedbump concerns but the other 99% are still ramping. In summary, my view is that we could have seen at least a short-term bottom today with a strong rally ahead of us in the sectors most caught in the latest speedbump. Time as always will tell.

Speed matters. And so does getting your money back. SN50 is built for premium inference, pairing RDU decode with existing GPUs for prefill. ROI in 6 months at a 70% billable utilization planning assumption. Read more in our blog ⬇️