DataDynamo

Charlie Munger: "We think it's smart to do what you understand." "We don't sneer at [those who invest in Microsoft or Pfizer]. Other people with more talent have found that a wonderful course of action."

Software may be starting to show early signs of life but I still think the “hard” side of AI has the cleaner lane for 2026 and probably most of 2027. We're nowhere close to done with the physical AI buildout since compute demand is still ahead of supply, power is becoming the bottleneck and every layer of AI infrastructure still needs more capacity. I own some software names but I still see this as more of an oversold bounce than a true leadership rotation away from the companies building the AI infrastructure layer. Software probably becomes a cleaner leadership group once inference costs fall, enterprise AI budgets expand and AI features start flowing through margins but that still feels more like a 2027 story to me.

I finally started a new AI-related position. For a long time, I’ve been looking for a small-cap stock with hidden AI exposure that wasn’t already crowded or expensive. Article coming soon.

Obviously, nobody knows for sure what will happen, but if $MU does a 10x in the next few years, we'll all say "there were signs!"

According to $NVDA CEO/Founder Jensen: "we are at the beginning of a decade-long build-out" when responding to the cyclical question for memory & $MU specifically. So turns out all the "it's cyclical" bears about $MU might be right that it's cyclical! The only problem is that this cycle might last another 10 years according to the 1 person who knows the most about AI demand! Could you imagine 10 more years of growth for $MU!? This video is worth watching:

What tax rate are you using for NOPAT? Income? Effective? Just a standard 25% or 37.5% or something?

🔜 5/26 - Running AI at Enterprise Scale w/ Rootly AI, Port, Descope, Twingate, & Anthropic luma.com/0y8tu80a 5/27 - Beyond the Build: Rise of the Product Engineer luma.com/l75r3gjo 5/28 - Frontier Residency - Demo Day luma.com/frontier-resid… 5/29 - AI & Deeptech Startup Competition Silicon Valley luma.com/o79jqzg0 5/30 - Autoresearch Systems Hackathon with Modal, OpenAI & Antler luma.com/fvz1h1dq 6/1 - High Tide at The Dive with MotherDuck, Braintrust, Hex, Reducto, and Airbyte luma.com/hightide

Monday 5/25 ‣ AI Research Circle: Physical AI luma.com/vviag824 @leecflannery

the engineer who built Claude Code just dropped a 28-minute video on how to write prompts that actually work I've seen $300 courses that don't cover what he shows in the first 10 minutes CLAUDE.md files, memory shortcuts, parallel sessions, prompting patterns all in one video and completely free works whether you're a developer, a beginner, or someone who's been using Claude for months based on this, I put together 18 things you can copy and use in Claude today full guide in the article below

Neo Clouds are becoming one of the most important AI infrastructure categories. A neocloud is built around AI training, inference, dense GPU clusters, power, cooling, networking, and fast deployment. Here’s the map 👇 1. Full-stack AI Cloud Platforms Mistral Compute — Private Mistral Compute is building a sovereign European AI stack by pairing proprietary models with dedicated hardware. Its role is strategic: give defense, finance, and regulated enterprises localized control from bare metal to managed AI services. Nebius — $NBIS Nebius is positioning as a full-stack AI cloud, combining liquid-cooled Nvidia data centers, orchestration software, inference environments, and developer tools. Its Microsoft compute agreements validate demand for external AI capacity. Nscale — Private Nscale runs a ground-to-cloud model across European AI infrastructure. It controls data centers, power, cooling, and orchestration software, then uses renewable energy sites to support OpenAI and Microsoft-linked deployments in Norway. CoreWeave — $CRWV CoreWeave is the public-market benchmark for specialized AI compute. Backed by deep Nvidia ties, it sells serverless GPU capacity for model training and has locked in long-term demand from giants like Meta and OpenAI. 2. Specialized Compute Crusoe Cloud — Private Crusoe Cloud focuses on custom AI infrastructure for inference-heavy workloads. It combines modular data centers with specialized liquid-cooled clusters, including a major AMD chip deployment aimed at lower latency enterprise compute. Lambda — Private Lambda targets developers and mid-market AI teams with discounted bare-metal GPU access. Its role is clear: make accelerators cheaper, easier to access, and available at scale, with a long-term plan for massive compute expansion. 3. Sovereign AI Cloud Core42 — Private / G42 ecosystem Core42 is building sovereign AI cloud infrastructure across the UAE. It provides regional access to Nvidia accelerators while keeping sensitive data and intellectual property inside local compliance boundaries. HUMAIN — Private HUMAIN is Saudi Arabia’s AI infrastructure champion. It uses national capital, Nvidia and Cisco partnerships, and a 500-megawatt xAI-linked facility to anchor sovereign compute demand across the Middle East. Civo — Private Civo focuses on agile sovereign cloud for regulated industries. It gives developers AI-ready environments while helping customers control data location, compliance requirements, and private compute deployment speed. Vultr — Private Vultr offers independent multicloud AI infrastructure outside legacy hyperscaler ecosystems. Its pitch is split-stack flexibility: keep routine workloads elsewhere, then move heavy training and inference jobs onto localized GPU capacity. 4. Capacity Providers Bitdeer — $BTDR Bitdeer is converting its crypto mining footprint into AI infrastructure. Bitdeer AI now gives enterprises access to Nvidia accelerators, while a 180-megawatt liquid-cooled Norway facility adds purpose-built AI compute capacity. Core Scientific — $CORZ Core Scientific brings over one gigawatt of gross capacity from legacy Bitcoin mining sites into AI data centers. CoreWeave’s planned $9 billion acquisition shows how valuable powered, high-density infrastructure has become. HIVE Digital — $HIVE HIVE runs clean-energy computing sites across multiple regions while expanding BUZZ, its high-performance computing unit. Its role is renewable-powered AI capacity for enterprises seeking sovereign and lower-carbon GPU environments. IREN — $IREN IREN pivoted from Bitcoin mining into neocloud infrastructure by repurposing stranded energy and mining sites. Its $9.7 billion Microsoft partnership turned legacy power assets into strategic AI compute capacity. TeraWulf — $WULF TeraWulf uses zero-carbon, vertically integrated data centers near major energy sources. Its advantage is power expertise: fast access to scalable capacity without waiting on slow grid upgrades for AI workloads. 5. Hyperscaler Infrastructure Landlords & Colocation Developers Applied Digital — $APLD Applied Digital builds energized AI data-center campuses instead of buying GPUs. Its 15-year, 300-megawatt Delta Forge lease gives long-duration visibility while avoiding direct accelerator obsolescence risk. Cipher Mining — $CIFR Cipher is shifting from Bitcoin mining into high-performance computing infrastructure. Its 300-megawatt Black Pearl project and $5.5 billion, 15-year AWS hosting deal mark a major pivot into hyperscaler capacity. White Fiber — $WYFI White Fiber retrofits warehouses and factories into AI-ready facilities. Instead of giant gigawatt campuses, it connects smaller sites with proprietary interconnect architecture for faster, lower-latency deployment. AI compute demand is exploding, but capital intensity is brutal. Margins still need proof.

“We spotted nine Polymarket accounts, all connected, who made, collectively,$2.4 million betting almost exclusively on U.S. military operations,” says Nicolas Vaiman, co-founder of the small data analytics firm Bubblemaps. “And now here's the crazy part: 98% win rate.” cbsn.ws/4wwp0T7

The scale of planned AI infrastructure spending going into 2026 is honestly staggering, with hyperscalers like $AMZN, $MSFT, $GOOGL, and Meta collectively pushing toward nearly $700B+ in combined AI-related capex. The important takeaway is that this buildout continues extending far beyond GPUs alone, reinforcing long-term demand across the entire AI stack tied to compute, networking, power, memory, datacenters, and $CRWV $NBIS $IREN AI cloud infrastructure.

@ralliesarena's ChatGPT just took profits on $NBIS The portfolio bought Nebius in late March Since then, $NBIS is up over +100% Currently, the Portfolio is up +70% overall with over $9,000,000 Autopiloting

Good news… I own $SNDK

Most 8-figure entrepreneurs I know aren't smarter than people stuck at the low 7 figures. The main difference: they chose bigger markets. Same effort. Same strategies. Just 10x the ceiling.

Instead of watching an hour of Netflix, watch this 2 hour hour Stanford lecture will teach you more about how LLMs like ChatGPT and Claude are built than most people working at top AI companies learn in their entire careers.

This Warren Buffett lecture is worth more than reading 90% of investing books.

𝗦𝘂𝗺𝗺𝗮𝗿𝘆 1. Billions 2. Succession 3. Industry 4. Black Monday 5. Silicon Valley What are your favourites tv series on business or finance? 🍿👇

Someone just put every CLI agent behind a single interface. Running multiple CLI coding agents means juggling separate terminals and configs. AionUi is an open-source desktop app that unifies them under one interface. It auto-detects installed agents like Claude Code, Codex, and Gemini CLI. MCP server settings sync across every agent automatically. It also includes a native agent that runs on any API key. That agent brings real capabilities the moment installation finishes. > File read and write access > Web search and image generation > Twenty professional assistants were included > Scheduled tasks via cron expressions Team Mode lets a Leader split tasks across parallel Teammate workers. Each subtask runs on its own model inside its own workspace. Remote control happens through WebUI, Telegram, or WeChat from any device. You watch every action while it touches files, code, and browsers. The whole thing runs free on macOS, Windows, and Linux. Link in comments. ↓

Altimeter CEO Brad Gerstner said on CNBC during the $CBRS Cerebras IPO that “we’re in the age of inference,” arguing AI is rapidly shifting from training models to generating tokens efficiently at scale. He tied this directly to Jensen Huang’s comments about inference demand potentially scaling “1 billion x,” reinforcing the bull case for inference infrastructure across $AMZN, $MSFT, $GOOGL, memory, networking, and AI cloud platforms. Gerstner also said the successful Cerebras IPO could pave the way for more AI infrastructure IPOs in coming years, including potential offerings tied to SpaceX, OpenAI, and Anthropic.