Thank you

Foxconn: AI demand continues to surge — November 2026 Foxconn, one of Nvidia's principal AI server manufacturing partners, has reported exceptional growth and expects AI-related operations to expand further in Q4. October revenue: $39.7 billion, up 46% year on year, setting an all-time monthly record. First 10 months of 2026: Approximately $280 billion in revenue, up 40.4% year on year. AI servers: Strong demand for Nvidia systems continues to support manufacturing activity and expansion. Q4 outlook: Foxconn expects further growth in AI-related operations, alongside seasonal strength in consumer electronics. US expansion: Foxconn is expanding its AI server manufacturing footprint in Houston, Texas. Further management comments Foxconn's chairman, Young Liu, also made several notably bullish statements on 30 October: No AI bubble: Liu explicitly rejected concerns that AI infrastructure investment is a bubble. 2027 growth: Foxconn expects another year of double-digit growth. AI server shipments: Management expects AI rack shipments to more than double in 2026. Longer-term demand: Foxconn previously indicated that AI production capacity demand should remain very strong in 2027. OpenAI: A further collaboration is being developed, potentially extending its role in AI infrastructure. Reuters: Foxconn is reporting record revenue, forecasting continued AI growth into 2027 and expanding production capacity. For Micron, the read-through is strongly positive for memory volumes, although Foxconn has not explicitly forecast DRAM shortages or higher memory prices.

Yes-Berkshire hold almost $40B in GOOG/L-all acquired between Q2 and Q3 2026. It represents about 13% of their entire Equity portfolio

@Adam-Kay said in Performance of the various Cobens portfolios: Google’s talks to lock memory supply through 2035 (plus HP/Dell deals to 2031) are strongly bullish for Micron. Customers only commit a decade out when they fear prolonged shortage, not a quick glut. This adds further credibility to a structural AI-driven cycle lasting far longer than traditional 2-4 year booms. Micron already has multi-year agreements covering ~35% of revenue through 2030 with price protections. Extended visibility reduces classic cycle risk, supports sustained high margins, and favors $MU’s earnings power for years ahead. There's also the Buckethead track Have you seen the length of his fingers?! Thanks Adam. I'll take that as 'all-on-track' and zero to worry about

Thanks Nik

not at the moment, just drop me a message at nik.burrows@cobensdirect.co.uk and I can sort it for you Cheers

Amazon delivered a solid result-the .1% miss is down to FX conversion and shifting Prime Day extra quarter. AWS is growing at its fastest clip ever-management see a path to $1T in AWS revenue. Capex was gigantic and guided to grow materially! [image: 1793347747224-screenshot-2026-10-30-at-08.07.08.png] [image: 1793347759398-screenshot-2026-10-30-at-08.07.35.png] [image: 1793347769519-screenshot-2026-10-30-at-08.08.04.png] CEO Andy Jassy — six particularly important comments Capacity is being consumed almost immediately AWS is installing an AI accelerator every 14 seconds. Customers are taking up that capacity almost as quickly as it arrives. Jassy sees strong demand extending several years ahead. AI stimulates conventional cloud growth AI training, inference, reinforcement learning and agent activity all drive additional CPU, storage and networking consumption. AWS's traditional compute and storage businesses recorded their fastest growth in several years. Inference is becoming the larger AI opportunity Inference already accounts for well over half of AWS AI revenue, and its share is growing. Jassy believes enterprise production deployment is still at an early stage. Bedrock adoption is exploding Bedrock token volumes increased nearly 17-fold year on year. Fortune 500 spending on Bedrock increased almost sevenfold. Jassy called it AWS's fastest-growing multibillion-dollar service. Cloud migration has decades to run Jassy estimates approximately 85% of global IT spending remains on-premises, with a potential 10–20-year transition towards cloud infrastructure. AWS backlog compared with Microsoft and Google Company Cloud backlog / RPO Amazon AWS $578bn Triple-digit growth Microsoft ~$625bn Commercial RPO, broader than Azure Google Cloud ~$561bn Cloud backlog The headline comparison is striking, but these are not directly equivalent measures. Microsoft's figure includes commercial obligations beyond Azure, and the companies disclose different contract mixes and recognition schedules. Amazon's $578bn figure is particularly encouraging because of its rapid growth and the fact that AWS revenue is already accelerating. Q4 guidance Metric Q4 2026 guidance Revenue $240–247bn Revenue growth +12–16% Operating income $28.5–33.5bn Revenue FX headwind ~60 basis points Advertising growth High teens

Yesterday Oracle updated the market at their investor conference. What an update-def worth a read. Look at teh growth(guidance). Almost 4X from now until 2030! Management have guided EPS to rise from 7.60 in 2026 to $21 by 2030! [image: 1793268451762-screenshot-2026-10-29-at-10.04.36.png] [image: 1793268459212-screenshot-2026-10-29-at-10.04.23.png] Management also revealed: $664bn contracted backlog, over 90% associated with infrastructure. 7 GW of additional capacity planned over approximately two years. Approximately 10 GW total capacity by the end of FY2028. Cloud database growth expected to exceed 50%. Existing four-year-old GPUs have been renewed or resold at prices 20% above their original contract prices. Right there is hard evidence there is money in GPUs The real surprise wasn't the $225b revenue target, which was broadly already known. It was the financing commitment. Oracle said it needed only another $20b of capital-market funding through FY2030, comprising $10bn of debt and $10bn of equity, on top of the $20b equity raise already completed. This contradicts all the FUD about their inability to fund their expansion. The mechanisms include supplier financing, customers supplying their own GPUs and customer prepayments. These reduce Oracle's immediate cash requirements, although vendor financing and leases still create financial obligations.

[image: 1793259784949-screenshot-2026-10-29-at-07.41.48.png] Microsoft's Earnings: AI Spending Is Delivering Exceptional Growth Microsoft delivered another outstanding quarter, comfortably beating expectations on revenue, earnings and Azure growth, demonstrating that the AI infrastructure investment cycle is translating into substantial commercial returns. Q1 FY2027 revenue increased 19% to $92.3 billion, while operating profit rose 21% to $45.9 billion, representing an exceptional operating margin of nearly 50%. Adjusted EPS reached $4.89, up 18%. Azure was the standout performer, with revenue surging 48% year-on-year to $33.2 billion. Microsoft Cloud revenue reached $63.9 billion, growing 30%. Perhaps most impressive is Microsoft's $677 billion commercial backlog, up 73%, providing extraordinary forward revenue visibility as contracted demand converts into recognised revenue. Microsoft generated $58 billion in quarterly operating cash flow and $18.6 billion in free cash flow, despite investing heavily in AI infrastructure. Capital expenditure, including leases, reached $45.2 billion, with spending expected to exceed $50 billion next quarter. Azure growth is forecast to accelerate further to 49–50%. The key takeaway: Microsoft is simultaneously delivering accelerating AI demand, exceptional profitability and enormous contracted future revenue. For NVIDIA, Micron and the wider AI infrastructure supply chain, Microsoft's continued investment represents another powerful confirmation of sustained demand.

NVIDIA and Foxconn: Monetising AI NVIDIA's partnership with Foxconn provides a compelling example of AI moving beyond hype and delivering measurable commercial benefits. Foxconn, the world's largest electronics contract manufacturer, is deploying NVIDIA's AI technology across its manufacturing operations, reporting remarkable initial results: over 90% improvement in defect-related processes, 50% less injection-moulding changeover work and 25% fewer labour hours spent on production scheduling. With more than 200 manufacturing facilities globally, the potential savings are enormous. Estimates suggest that a mature deployment could generate approximately $3 billion in annual productivity savings, against $1.5 billion in ongoing costs and an initial investment of $3–5 billion. The crucial point? NVIDIA isn't simply selling expensive chips. It's selling the infrastructure that enables customers to generate substantial financial returns. Now multiply that opportunity across automotive, aerospace, pharmaceuticals, logistics and manufacturing worldwide. This is how AI becomes an economic revolution.

That didn't really end very well, currently suspended as they manage a wind-down.