Topic momentum
How much each topic moved, day by day: the score from each daily deep analysis, combining fresh coverage, independent sources and how central the topic is to what we track. 30 analyses, October 18, 2026 to November 16, 2026. How it's scored
Energy ›
The binding constraint: how the power for AI gets generated, moved and paid for.
Compute ›
The substrate: the chips and the factories that make them, and the datacenters on Earth that run them.
Semiconductors & chip supply chain
GPUs, TPUs and custom accelerators; foundries, lithography, HBM memory, packaging and EDA; chip export controls; fab projects such as Terafab.
Datacenters & cloud
Hyperscale and neocloud datacenter builds on Earth: capacity, cooling, networking, power procurement, training versus inference capacity.
Intelligence ›
The models, the labs that build them, and the agents built on top.
Frontier labs & models
OpenAI, Anthropic, Google DeepMind, xAI, Meta and peers: model releases, evaluations, fundraises, valuations and talent moves.
Agents & applied AI
Agentic frameworks, coding agents, enterprise copilots and workflows, the software harness around models, inference economics.
Open-weights models
Openly released model weights, fine-tunes and licensing fights; open-model share of usage; China's open-weights strategy.
AI safety & security
Alignment and frontier-safety research, dangerous-capability evaluations, AI-enabled cyber offense and defense, model security.
Physical AI ›
Atoms, not bits: AI that moves through the world.
Humanoid & general robotics
Optimus, Figure, Agility, 1X, Unitree and peers: humanoid and general-purpose robots, dexterity, robot foundation models, deployments.
Autonomous vehicles
Robotaxis, FSD and driver-assist, autonomous trucking, delivery drones, eVTOL; their fleet data, safety records and regulators.
Industrial & materials
Factory automation and manufacturing methods (gigacasting, composites), EV and battery supply chains, critical minerals.
Orbit ›
Where the build-out is heading: launch capacity, AI compute and power in orbit, and the networks that tie it together - unlimited solar energy and radiative cooling, with no land to buy.
Launch & the space economy
Starship, Falcon and competing launchers; launch cadence and cost to orbit; lunar and Mars programs, in-space manufacturing, debris and space regulation.
Space-based compute & power
AI compute satellites and orbital datacenters (SpaceX, NVIDIA, Google and others), space solar power, radiative cooling, inter-satellite laser links for data.
Global connectivity
Satellite broadband (Starlink as an ISP, direct-to-cell), subsea cables, 6G and non-terrestrial networks.
Symbiosis ›
Humans and AI co-evolving: how people, science and work change as intelligence becomes abundant.
Human-AI co-evolution
Personal AI and agents in daily life, brain-computer interfaces, how individuals and organizations learn to work alongside AI, human-AI partnership.
AI for science
AI accelerating discovery: biology and drug discovery, materials, fusion, climate and weather modeling, digital twins.
Labor & abundance
Employment and productivity data, jobs displaced and created, the abundance and universal-income debate.
Capital & Power ›
Who pays for the build-out and who governs it.
Capex & corporate spending
Hyperscaler and corporate AI capital expenditure: quarterly spend, guidance, GPU order books, what share of announced capex is deployed.
Venture, sovereign & project finance
Startup rounds, datacenter project finance and GPU-backed debt, sovereign AI funds and national compute programs.
Public markets & the AI trade
AI-exposed public companies as investments, earnings, IPOs, the bubble-or-not debate. Informational only, never advice.
Policy, geopolitics & China
AI regulation and executive actions, antitrust, export controls, US-China competition (labs, chips, robots, reactors), Taiwan risk.