[image: 1758614107382-screenshot-2025-09-23-at-08.53.57.png] This imo will pave the way for wider agreement on many resources including high end silicon (GPUs). If the deal gets done and I think it will, China get chips. I think Trump wants the win and Nvidia wants to be able to service all comers-not that China is particularly relevant today! They likely will be in the future. Today Nvidia can't meet demand so excluding certain markets makes no difference(today). In the future, the demand/supply imbalance will converge, and if relations are not restored 'soon', there may not be a future in China, because China will pave its own way and their build out/roadmap will take another course. Feeding their lead to Huawei for even 1-2 years could very well see future market share(China) lost for good.

I follow Dylan Patel, founder of Semi Analysis. Very knowledgeable on the tech side, not the 'stock' side but for me, it compliments the knowledge and allows me to separate the hype from the reality . I know he met Chuck Liang recently to discuss their plans-he thinks something important is about to drop. Why Supermicro (SMCI) Gets the Spotlight in Dylan’s Tease—And Not Dell, Foxconn, or Wiwynn SM as a key supporter of his imminent “huge” framework on AI chips, inference, and infrastructure, due to drop by the evening of 9 October. SMCI is listed alongside hyperscalers/cloud players (CoreWeave, Nebius) and hardware/infrastructure specialists (Crusoe, HPE, Tensorwave), highlighting its pivotal role in the rack/server layer for optimised inference stacks. Notably absent are Dell, Foxconn, and Wiwynn (Wistron’s AI-focused ODM arm), despite their prominence in AI server markets. This isn’t arbitrary; it reflects SMCI’s unique position as the agile, high-density leader for the “neo-cloud” era,(think IREN) tailored to inference’s bursty(yes bursty-data that can burst from 1X to 100X in a nano second,) power-intensive demands. Here’s why SMCI gets the call-out, grounded in Patel’s reports, posts, and industry context:1. SMCI’s Edge: Speed, Customisation, and Hyperscaler Fit for InferenceRapid Prototyping and Deployment: SMCI excels at “just-in-time” manufacturing, delivering 100,000+ servers in weeks rather than months. This is critical for inference, where hyperscalers like CoreWeave (an SMCI client) demand swift iterations on hybrid NVIDIA/AMD setups to manage variable query loads. Patel’s September 2025 SemiAnalysis report on rack architecture (co-authored by him) praises SMCI’s modular designs for disaggregated PDUs and liquid cooling, enabling 250kW+ racks with a 30% better total cost of ownership (TCO) compared to rigid ODM builds. Foxconn and Wiwynn, as pure ODMs, prioritise volume for branded OEMs (e.g., Dell’s enterprise kits) but lag in bespoke hyperscaler customisation. Direct Hyperscaler Relationships: SMCI sells directly to neo-clouds (CoreWeave, xAI’s Colossus—partially SMCI-supplied) and AI labs (OpenAI’s AMD pivot), bypassing intermediaries. In his August 2025 “No Priors” podcast, Patel highlights SMCI’s vertical integration (from motherboard design to cooling), giving them a two-year lead on liquid-cooled hybrids, essential for inference’s 80%+ energy draw. Dell shines in enterprise/sovereign AI (per Patel’s May 2025 “How Dell Is Beating Supermicro” report), but their slower cycles, optimised for HPC stability, don’t match neo-cloud urgency. Inference-Specific Advantages: Backed by vLLM/SGLang (inference engines-sorry tech heavy!), the framework likely benchmarks rack-level metrics like tokens/second per watt. SMCI’s 8U/10U GPU trays (e.g., SYS-821GE with 8x Blackwell) blend NVIDIA prefill compute with AMD decode efficiency, reducing latency by 20-50%. Foxconn/Wiwynn handle high-volume NVIDIA HGX for cloud giants (e.g., Foxconn’s Oracle Stargate supply), but Patel’s critiques (e.g., 2023 posts on Foxconn’s public “reveals” being overhyped) underscore their ODM commoditisation—cheaper but less innovative for multi-vendor inference. Historical Context from Patel: SMCI as the “Crusher” in AI RacksPatel’s analyses consistently position SMCI as a leader in frontier AI infrastructure. A 12 September 2025 X post details his Supermicro factory tour with CEO Charles Liang, showcasing GB300/B300 and MI355X racks—directly tying to the tease’s NVIDIA/AMD backers. In contrast, his May 2024 report praises Dell for enterprise wins (e.g., Tesla, CoreWeave orders) but notes SMCI’s resurgence in neo-clouds via cheaper, denser cooling (e.g., April 2023 post: “I’m such an idiot for not going turbo long SuperMicro... they crush Dell and crew while being much cheaper”). Final thought- SMCI’s Unique PositionDylan’s call-out of SMCI reflects their role as the inference infrastructure “backbone” for collaborative, multi-vendor stacks, validated by factory tours, reports, and backers like HPE (also listed). Dell, Foxconn, and Wiwynn play critical roles in enterprise or ODM volume but lack SMCI’s hyperscaler agility and hybrid rack innovation for inference. If the full drop (expected ~evening 9 October BST) includes rack BOMs or benchmarks, SMCI’s prominence will likely grow.

HSBC Research today issued a report on Cowos capacity, stating, they see evidence of Nvidia wafer allocation at 800k units through 2027. This is the first sign of wafer accelerated growth since June 2025. A recap Here is my May 25 estimate based on TSM public comments-this proved accurate QoQ. The important column being Nvidia wafers, circa 105k by year end. I would think e are ahead of this. Remember, there is a direct correlation between wafers and revenue. And HSBC makes this very point. If we extrapolated these numbers using the existing expansion cadence we would arrive at circa 570k wafer through 2027. HSNC claims the number is now up to 800K and have modelled GPU(DC) segment revenues of up to $400B. Accordingly they have modelled just under $10 EPS! Accordingly they have issued a new price target of $320. Exciting times ahead [image: 1760598841437-screenshot-2025-10-16-at-08.11.35.png]

Hi Mike, These 'headline grabbing' numbers are spending over many years. For example the Open AI/ORCL deal, $300B over 5 years. The Stargate deal, the UAE deal. Once these projects get started, some allocate quickly and some stall. In many cases the physical infrastructure needs to be build. This could be the power source, the data centre itself or the compute hardware. I would think anyone announcing a new project today, would be waiting a long time for silicon. Case in point: Xai +800K GPU expansion of Colossus, which is being installed now. The chips were ordered a year ago and paid for at least 9 months ago. We must be at a $trillion YTD in planned projects, but as we have discussed, companies like Nvidia, AMD, TSM can only supply so much compute, today(CoWoS-L/packaging constraint), why, because the tools of production take > 12 months to build. It's a fallacy- give me 1m chips more and TSM can magic them up. And TSM isn't going to 10X their production because who will pay for it. There is a demand supply imbalance of anywhere from 5-10X:1. CC Wei has said he is on a 50% CAGR trajectory. Very strong but planned growth trajectory over 5 years to 2030. What I am saying here is regardless of the 'spend' on AI Nvidia is locked in to a 50-60% growth rate pretty much over the next 5 years because they are inextricably linked to TSM capacity. I can live with that. Personally, that is what I think it will be-in that range. In terms of valuation. If you look at the market you see hype for sure, some warranted, some not(imo). AI mania has caused almost everything in the space to appreciate materially. The Tesla effect if you will. Remember every kind of crypto imaginable, NFTs. Nothing but Tulips! Greed makes otherwise sane people do very silly things I have no idea what the stock price of palantir will be next year but I can look at its valuation today, its past growth, and take a view. It's interesting, their 5 year avg growth rate is 31%. It's good but many companies can match that. Nvidia's number is 65%. So in my mind, here is a company with half the growth rate(historical) and if anyone says that 31% is going to accelerate, id say show me the evidence(there isn't any). Faith. Sorry I need more than that. So Palantir has a PEG of 8 vs Nvida < 1. Palantir also gives almost all its earnings to the directors of the company. As we have all seen, at least for a while anyway, stock prices can diverge from all reality-that is until reality bites. I personally think this stock is grossly overvalued-good luck to anyone holding it. They will need a bull market, perfect execution and to smash estimates every quarter like clock work. Even if growth does accelerate-I see some 'experts' are saying their growth rate is actually 65%-double. OK so it's now only 4X more expensive than Nvidia. Why would I buy Palantir for $1 when I can buy Nvidia for 25c?.There is no margin of safety-none whatsoever. I would be buying 2030 earnings, which actually means the stock will flatline for the next 4-5 years as the business grows into its extreme valuation. I could be wrong, but as said the metrics are so off the charts it's just not worth taking the chance. And by that I mean what is the future upside. It might add another 50% but that would take its multiple to 400 and as we have seen there are alternatives with far better 'numbers'-this is what I mean about risk vs reward. It's literally gambling and we aren't in that business. I would bet if you asked the avg shareholder why they hold it, they would say 'to the moon' but could not articulate anything rational about their decision and the actual business fundamentals. In other words the-greater-fool-theory in play. So to answer your question. No I am not worried about a crash. We don't hold anything with a materially stretched valuation. Multiples in the 70-200 range. Some of our holdings are probably fully priced. Oracle might be short term. It's not easy to value because of their $300B DC deal with OpenAI. Lots of moving parts and contingent elements. But that is the basis of the stock market(auction based). We are not traders so if I have an opinion a stock is over/under on any particular day, we would not act on that. However we did act on Colgate when for no good reason(a flight to safety?) the stock moved up materially to an ATH. Range bound for 5 years 70-80 then pops to $105 in a very short period. The investment committee took a big chunk of weight out of that one and it was absolutely the right thing to do. The stock fell back and is now back in its multi year range @ $81. In conclusion, one shouldn't take a view 'when is the next crash'. Why would the broader market crash as opposed to be choppy- the US economy is ok. I wouldn't say great but it's doing well enough. But more specifically look at big techs earnings. It's at a record and healthy and they have had many quarters of growth with continued higher guides. GOOG still trades below a 22-23 multiple and its growth is still very strong, for example. There are pockets of extreme valuation for sure-best leave that to Cathy Wood

Over 50% of the value of the company is AWS-maybe 65% AWS generates approx 28.7b and earns 10.5(quarter) Azure generates about 24.5b(less than AWS) but also earns 10.5b Similar money makers similar growth. Msft earns significantly more money overall-circa 50%+ more.

Cheers Ronski That's saved taxing what little of the grey matter I have left

I don't mind disclosing the following with the caveat, everyone's situation is different by virtue of their: age appetite for volatility time to retirement total liquid assets lifestyle and cash requirements For example, we do have clients of advanced years(80+) with 100% of their assets with us in Tech. I happen to know that number represents less than 10% of their liquid funds. Conversely we have JISA accounts where their trustee has chosen MM or bonds. The point being, investment selection is based on many factors which you must weigh. Investment decisions should not be made based on a poll. If anyone wants to discuss their investment choices myself or Nik are happy to discuss. Our guidance service is FOC. Total assets under management % split and 3 year rank(1-5) performance: Lifestyle 40% (2) Tech 18%(1) IMOP 12%(4) PHE 8.5%(5) Index 8%(3) Other

some monthly memory spot price info: [image: 1761398953755-screenshot-2025-10-25-at-14.28.36.png] Critical weekly and monthly price changes AI HBM [image: 1761399574367-screenshot-2025-10-25-at-14.39.07.png] Setting up another beat and raise.

Thank you Nik. Perhaps you could propose giving the client the option to abandon the small change remaining in funds since it costs the company more to withdraw than the amounts themselves. This would help expedite the process. Frank

Work on that one Cappo.... it's almost funny!

Meta, Blue Owl to Finalise Record $30 Billion Private Financing Deal for Louisiana AI Data Centre Meta is set to complete an approximately $30 billion private capital financing deal — the largest of its kind on record — to fund Meta’s advanced AI data infrastructure project in Louisiana, Bloomberg News reported. Meta and alternative asset manager Blue Owl Capital will share ownership of the Hyperion data centre site in Richland Parish, Louisiana, with the technology giant retaining 20% of the asset. Hyperion is the largest of Meta’s 29 global data centres — a 4-million-square-foot complex located in rural Louisiana. Morgan Stanley arranged more than $27 billion in debt and roughly $2.5 billion in equity through a special purpose vehicle (SPV) to finance the development, according to the report. The transaction has been structured through this SPV to support large-scale infrastructure investment. The site is 2,250 acres. It's gigantic [image: 1760683491611-screenshot-2025-10-17-at-07.44.07.png]

Oracle is on the move again(up) based on rumours it will run the US operations of Tik-Tok. According to DT (Trump) 'a deal has been done'. The parties to said deal have yet to comment.

Quick update: with thanks to Nik, I've effected all the movements suggested and am hoping that the new investments in IM will grow as bountifully as the last ones have done! Thanks chaps.

Thanks. I did yesterday. Will wait or try to call. Thanks

Great to hear Ron. That portfolio has been very productive. I’m very proud of it and very pleased for anyone who put their hard earned money into it.

Breaking news-more to follow. OpenAI sign a deal today with AVGO/Broadcomm for the design of a new Aspics accelerator. The size of the deal is circa 10GW which would be at least $150B. AVGO is up about $40

ExIM, No problem, no intention of engaging but im not about to let intentional misinformation go unchecked.

Ok sure, send it over

@Adam-Kay said in Dashboard performance charts.: J, I wish you’d mentioned this a while ago. [image: 1740153593731-screen-shot-2025-02-21-at-15.56.38.png] Crumbs. That's not shabby.

Cheers Nik No urgency at all, I have another question, I shall ping off an email to you in the morning again no urgency on it