LoopEmerging trends in the
AI industry
The key developments that are shaping AI right now,
curated by a Senior AI Engineer.
| Technology | Time horizon | Key stat | Summary |
|---|---|---|---|
| AI coding agents | Here now | 90% enterprise adoption | AI programming tools are now standard across engineering teams. |
| AI video generation | Here now | Veo 3, Sora 2, Seedance | Production-quality video is now possible from text prompts. |
| Enterprise AI platforms | Here now | $13B market | Full-stack AI platforms are dominating enterprise procurement. |
| Robotaxis | Here now | 14M Waymo trips/year | Autonomous ride-hailing is operating at scale in major US cities, with Waymo expanding rapidly. |
| AR glasses | Here now | 7M Meta sold in 2025 | Consumer AR glasses have crossed into mainstream adoption. |
| Autonomous trucks | 1–2 years | Highway corridors expanding | Long-haul autonomous trucking is scaling on fixed highway routes. |
| Humanoid robots | 1–2 years | Factory pilots underway | Humanoid robots are being deployed in factories and used for narrow tasks. |
| World models | 1–2 years | Next frontier for AI | Models that can understand physics and spatial reasoning are advancing rapidly. |
| Custom AI chips | 1–2 years | Apple, Google, Amazon | Big tech is designing proprietary silicon to reduce its dependency on NVIDIA. |
| AI hardware (Jony Ive) | 1–2 years | Stealth → reveal soon | AI-first consumer hardware is being developed by Apple's former designer. |
| Quantum computing | 3–5+ years | Commercial in ~5 years | Fault-tolerant quantum computers for commercial use remain years away. |
| Fusion power | 3–5+ years | First plants by ~2030 | Fusion energy could be used to power data centres, but it's still pre-commercial. |
| Space data centres | 3–5+ years | Early R&D phase | This area is being actively explored by startups, but it's incredibly early days. |
| Full global autonomous driving | 3–5+ years | Regulatory + infra gaps | Worldwide autonomous driving depends on regulation and sufficient training data, which isn't there yet. |
The big picture
The infrastructure arms race is intensifying
Companies are spending more on AI data centres than oil infrastructure for the first time in history. Nvidia is the clear winner, but everyone from chip startups to fusion energy companies are racing to meet the insatiable demand for AI.
AI agents are finally going mainstream
Coding assistants have proven that AI agents can deliver real value, with enterprises in the early stages of deploying them across their business. The big tech companies are working to standardise how agents talk to each other and access external tools.
Governments can’t agree on how to regulate AI
China is adding new safeguards around chatbots and mental health, while Europe is loosening its AI Act to stay competitive with the US. Safety incidents and copyright lawsuits are forcing regulators to act, but there’s no global consensus emerging.
Defence tech is having its moment
Autonomous drones and AI-powered fighter jets are no longer science fiction. Defence startups are attracting record investment as governments realise that they need to modernise their militaries fast, especially with war raging in Europe.
- •OpenAI finalises a $100 billion deal, now valued at over $850 billion
0 new trend added · 1 trends updated
Anthropic is now the top enterprise provider for AI, beating OpenAI
Enterprise LLM market share, 2025
Data centre spending overtakes oil investment for the first time
Global investment in 2025, $B
You might also like
AI use cases for everyone
Practical ways that you can use AI. At work, at home, and everywhere in between.
Strategies for using Claude Code
Practical guide for getting the most out of Claude Code — from writing clear prompts to using plan mode, MCP servers, and CLAUDE.md.
Trend report for Q2 2025
The top 10 trends that stand out from Q2. If you want to know where tech is genuinely heading, start here.