The AI Value Chain

Today, the AI value chain is built on a complex ecosystem that spans hardware, cloud, software, AI models, and vertical and consumer applications.

This is an ecosystem in full development, where rather than demand bottlenecks, there are supply bottlenecks.

This means the ecosystem is still developing to cover for a potential embedding of AI into every possible commercial use case.

Yet, this process might take a couple of decades to develop fully to accommodate the mass adoption of AI and to make inference available to everyone.

In the last few weeks, I’ve been tackling a key topic in tech: competition from multiple perspectives.

There, I’ve started by defining what it means to build AI moats if you’re outside the frontier model stack.

Table of Contents

In This Issue!

The AI industry is structured in layers, starting from hardware to cloud infrastructure, AI models, vertical AI applications, and consumer apps/hardware.
Competitive moats can be built at any layer, either through branding, UX, or architectural/manufacturing advantages.
No single company can dominate the entire AI stack; instead, they focus on specific layers or intersections.

AI Chips: Market leaders include NVIDIA (GPUs), Google (TPUs), AMD (AI accelerators), Intel (Habana Gaudi), and startups like Cerebras, Graphcore, and Hailo.
Memory Systems: Samsung, SK Hynix, Micron dominate high-bandwidth memory (HBM) and AI-optimized storage.
Specialized Processors: Companies like IBM, Tenstorrent, and Mythic AI design AI-specific chips.
Interconnect & Data Fabric: Broadcom, Cisco, and NVIDIA (Mellanox) provide networking solutions for AI workloads.

Cloud AI Providers: AWS, Microsoft Azure, Google Cloud, IBM Cloud, and Oracle Cloud offer AI training, inference, and deployment services.
GPU Cloud Infrastructure: CoreWeave, Lambda Labs, Cerebras Cloud specialize in high-performance AI compute.
AI Storage & Data Management: VAST Data, WekaIO, and Pure Storage optimize data pipelines for AI.
Model Deployment & Orchestration: Databricks, Hugging Face, Run:ai, OctoML enable AI model hosting and scaling.
Networking & Interconnect Solutions: NVIDIA (Mellanox), Broadcom, Arista Networks, Cisco support AI data centers.

Foundation Model Developers: OpenAI, DeepMind, Microsoft, Anthropic, Meta AI, Mistral AI, Baidu, Alibaba Cloud.
Model Fine-Tuning & Customization: Hugging Face, Cohere, Databricks, OctoML.
Domain-Specific AI Models: BloombergGPT (finance), IBM Watson (enterprise), Character.AI (chatbots).
Model Deployment & Orchestration: Run:ai, CoreWeave, Lambda Labs.
AI Safety & Alignment: OpenAI, Anthropic, DeepMind focus on responsible AI.
Open-Source AI: Hugging Face, Mistral AI, Meta AI (Llama), BigScience (BLOOM).

Voice & Speech AI: ElevenLabs, Deepgram, Resemble AI.
AI for Robotics & Automation: Skild AI, Shield AI, Figure AI, Agility Robotics.
AI Video & Media: Synthesia, Runway ML, Pika Labs, HeyGen.
Enterprise AI & AI Integration: Distyl AI, Cohere, Glean Technologies, Adept AI.
AI in Defense & Aerospace: Shield AI, Anduril, Palantir.
AI in Healthcare & Biotech: Tempus AI, Insilico Medicine, Owkin.
AI in Finance & Legal: Kensho, Zest AI, Luminance, LawGeex.
AI in Retail & E-Commerce: Vue.ai, Focal Systems, Syte AI.
AI in Manufacturing: SparkCognition, Uptake, Bright Machines.
AI in Agriculture & Food Tech: iFarm, Prospera, Taranis.

AI-Powered Mobile & Web Apps: AI assistants, productivity tools, and AI-enhanced UX.
AI Consumer Devices: AI-native smartphones, AR glasses, smart displays.
AI-Specific Hardware Features: On-device AI processing, custom AI chips, neural engines.
Next-Gen AI Interfaces: Brain-computer interfaces, AIoT devices, ambient computing.

The AI industry is highly competitive but offers multiple entry points for companies to establish a moat.
Success depends on leveraging strengths in specific layers of the AI stack or integrating AI capabilities across multiple layers.
The market is rapidly evolving, with emerging players disrupting established incumbents in hardware, cloud, models, vertical AI, and consumer applications.