Advanced Analytics and Machine Learning
Every city needs foundations, power, roads, buildings, and people using them. AI is the same , just faster and more expensive.
Layer 1: The Sand - Chips & Semiconductors
Before any AI can run, someone has to build specialised hardware. Normal chips can't handle it. AI needs to do millions of calculations simultaneously, not sequentially — and that requires an entirely different class of processor.
- NVIDIA makes the gold-standard AI chip (the GPU). Basically everyone uses them. They're the arms dealer of the AI war.
- AMD makes competing chips. Cheaper than NVIDIA and catching up fast.
- Intel is trying to get back in the game after being late to AI. Turnaround story.
- TSMC actually manufactures the chips that NVIDIA and AMD design. Nobody builds at their scale. No TSMC, no chips.
- ASML makes the machines that make the chips. One company, total monopoly, Dutch. No ASML means no modern chips. Full stop.
- ARM Holdings designs the chip architecture blueprint that almost every device uses. They license the design — they don't build anything themselves.
- Broadcom, Marvell, Credo make the networking chips that connect all those AI chips together at speed. The glue between the GPUs.
- Micron, SanDisk make the memory chips. AI needs to store and retrieve data at extreme speed.
- That's what these companies do.
Layer 2: The Power Stations - Energy Infrastructure
AI data centres consume obscene amounts of electricity. A single large facility can use as much power as a small city. Someone has to generate that power reliably, around the clock.
- Bloom Energy makes fuel cells that can power data centres without relying on the grid. Always-on, no outages.
- Babcock & Wilcox, T1 Energy are nuclear and alternative energy plays. AI needs baseload power that solar alone can't guarantee.
Layer 3: The Land - Data Centres & Cloud
You need physical buildings stuffed with chips, cooled constantly, and connected to the internet at enormous speed. This is the real estate of the AI economy.
- Amazon (AWS), Microsoft (Azure), Alphabet (Google Cloud) own the biggest data centres on earth. They rent compute power to everyone else. The landlords.
- CoreWeave is a newer data centre company built specifically for AI workloads. Rents NVIDIA GPUs at scale to anyone who needs them.
- Core Scientific, IREN, Applied Digital, Cipher Mining started as crypto miners. They already own the buildings, power contracts, and cooling systems. Now pivoting to rent that infrastructure to AI companies. Smart repositioning.
Layer 4: The Roads - Networking & Connectivity
Data has to move between chips, servers, and data centres at ludicrous speed. These companies build the pipes and roads that make that movement possible.
- Arista Networks builds the switches that move data around inside data centres. The internal road network.
- Coherent, Lumentum, Lightwave Logic make photonics components — fibre optic connections that move data at the speed of light between data centres.
- EchoStar is satellite connectivity infrastructure. The long-range road that connects remote areas and emerging markets.
Layer 5: The Buildings - AI Platforms & Models
This is where the actual AI gets built and run. The companies building the models that everyone else uses, and the platforms that host those models.
- Alphabet (Google) built Gemini and owns DeepMind. Has the best data in the world via Search. The incumbent with the most to lose and the most to gain.
- Meta built Llama (open source). Owns Instagram and WhatsApp, which generate enormous training data. Playing a different game to everyone else.
- Microsoft owns a large stake in OpenAI (ChatGPT). Azure is how most businesses access AI. The enterprise distribution channel.
- Nebius Group is a lesser-known European AI cloud company rebuilding from Russian tech origins. Early stage but interesting positioning.
- Snowflake, MongoDB, Oracle are data platforms. AI is useless without clean, accessible data. These companies store, manage, and serve it.
Layer 6: The Shops - Software Built on AI
Companies building useful products on top of the AI models. These are the applications people and businesses actually interact with every day.
- Palantir does AI analytics for governments and large enterprises. Heavy defence and intelligence contracts. Controversial but deeply embedded.
- ServiceNow is enterprise workflow software baking AI into everything. The boring-but-essential layer of corporate IT.
- Tesla is as much an AI company as a car company. Full Self Driving and the Dojo supercomputer are massive AI bets hidden inside a car manufacturer.
- Pega Systems is enterprise software with AI automation. Slower moving but deeply embedded in large organisations.
- Robinhood is a fintech that benefits from retail enthusiasm around AI stocks. A meta-play on the whole trend.
- ImmunityBio, Lemonade are AI applications in biotech and insurance respectively. Niche bets on AI transforming specific industries.
Layer 7: The Guards - Cybersecurity
More AI means more attack surface. More data means more to steal. AI-powered attacks require AI-powered defence. The security layer grows in direct proportion to everything else in the stack.
- CrowdStrike uses AI to detect and stop threats in real time. The endpoint security leader. Every laptop in a big company is probably running this.
- Palo Alto Networks covers networks, cloud, and endpoints. One of the most comprehensive security platforms available.
- Zscaler does cloud-native security — protecting companies where the perimeter no longer exists, when everyone works from anywhere.
- SentinelOne is an autonomous AI security platform that detects and responds to threats without needing a human in the loop.
TL;DR
Someone makes the chips. Someone powers them. Someone houses them. Someone connects them. Someone builds the models. Someone builds apps on those models. Someone secures all of it.
That's the AI ecosystem. Seven layers, hundreds of companies, and an enormous amount of capital flowing through all of it. The infrastructure layers — chips, power, data centres, connectivity — tend to win regardless of which AI model or application eventually dominates. The picks and shovels tend to win regardless of who finds the gold.