Valuing an AI company is really a bet on how AI plays out. These are the pieces we think give you the clearest view, picked for quality, not because they agree with each other. We earn nothing from these links.
Fact Titles, authors and dates. Links checked on 2026-10-30. Our view Our picks and the one-line notes.
Essays that change how you see AI
Big pictureFree
The Commanding Heights
Will Manidis and Joe Roberts · Oct 2026
Book-length essay arguing that cheap intelligence won't leave the profits with the model makers: like railroads and oil, the fortunes go to whoever owns the rails it travels on. Read it, then ask which of our companies owns rails.
OptimistFree
Situational Awareness
Leopold Aschenbrenner · Jun 2024
The case that AI scales to superintelligence around 2030, with trillion-dollar compute clusters, power plants and a US-China race. Much of the 2025-2026 data centre spending reads like it.
OptimistFree
Machines of Loving Grace
Dario Amodei · Oct 2024
Anthropic's CEO on what powerful AI could do for health, biology and economies within 5-10 years. Read it knowing the author runs a company we value.
RisksFree
The Adolescence of Technology
Dario Amodei · Jan 2026
The sequel: the risks of powerful AI, from misuse and loss of control to concentrated power and wealth, and what could be done about them.
OptimistFree
The Gentle Singularity
Sam Altman · Jun 2025
OpenAI's CEO argues the take-off has started and the cost of intelligence will fall towards the cost of electricity. Same caveat: the author runs OpenAI.
RisksFree
AI 2027
Daniel Kokotajlo, Scott Alexander and others · Apr 2025
A month-by-month scenario of AI automating AI research, with two endings. Concrete enough to check against the news as it happens.
SkepticFree
AI as Normal Technology
Arvind Narayanan and Sayash Kapoor · Apr 2025
The counterweight: AI will spread slowly, like electricity did, held back by adoption, rules and the real world rather than by how good the models are.
SkepticFree
AI's $600B Question
David Cahn, Sequoia Capital · Jun 2024
The original "revenue gap" maths: how much AI revenue the chip spending needs to pay off. Our payback calculator on the AI boom page builds on it.
Big pictureFree
The Bitter Lesson
Rich Sutton · Mar 2019
Two pages explaining why more computing power keeps beating clever ideas, and so why AI companies spend hundreds of billions on chips.
RisksFree
The Intelligence Curse
Luke Drago and Rudolf Laine · Apr 2025
If AI does the work, what happens when states and companies no longer need people? A "resource curse" argument, with ideas to avoid it.
The economics of AI
EconomicsFree PDF
AI and Our Economic Future
Charles I. Jones, Journal of Economic Perspectives · Summer 2026
A leading growth economist on whether AI can lift growth above 2% a year, and why "weak links" (tasks only people still do) slow it down.
SkepticFree
The Simple Macroeconomics of AI
Daron Acemoglu, NBER · 2024
A Nobel laureate estimates AI adds only about 1% to US GDP over ten years: the most cited low estimate, worth knowing before you value trillion-dollar AI plans.
EconomicsFree
Economic Policy for AGI
Julian Jacobs and Alex Imas, Google DeepMind · Sep 2026
Scores eleven policies, from more unemployment insurance to a negative income tax and "universal basic capital", for an economy where AI replaces many jobs.
Big yearly reports
AI Index Report
Stanford HAI · Yearly, in spring
The reference yearbook: model performance, training costs, investment, adoption and policy, with free charts and data.
State of AI Report
Nathan Benaich, Air Street Capital · Yearly, in October
Hundreds of slides on research, industry, politics and safety, ending with predictions that it grades a year later.
Trends: Artificial Intelligence
Mary Meeker, BOND · May 2025
About 340 slides on how fast AI is being adopted compared with the internet, and what it costs to build and use.
Benedict Evans presentations
Benedict Evans · Twice a year
A tech analyst's big-picture deck on what AI changes for tech companies and who is likely to win.
Live trackers: watch AI as it happens
Dashboards that update as new models ship. Useful when you set growth assumptions in the lab: are capabilities, usage and spending still rising?
How capable is AI?
- Task time horizons METR
How long a task (in human working time) AI agents can finish. METR found the length doubled about every 7 months from 2019 to 2025. - Benchmarking Hub Epoch AI
Independent scores for the newest models on hard maths, science and coding tests, run by Epoch itself. - Model leaderboard Artificial Analysis
Compares models on intelligence, speed and price per token. Good for watching AI prices fall. - LMArena LMArena
People vote blind between two answers; the leaderboard is the crowd's ranking of chatbots. - ARC Prize leaderboard ARC Prize Foundation
Puzzles that are easy for people and hard for AI, plotted against cost per task. - SWE-bench Princeton and Stanford researchers
Can AI fix real software bugs from GitHub? The standard test for AI coding agents.
Who uses it, who pays?
- OpenRouter rankings OpenRouter
Which models developers actually pay for, by tokens used each week. A live clue to market share between labs. - Ramp AI Index Ramp
The share of US businesses paying for AI, and for which provider, from Ramp's card and bill-payment data. - Anthropic Economic Index Anthropic
Which jobs and tasks people use Claude for, by country. Company data from Anthropic, which we value on this site. - Frontier data centres Epoch AI
The largest AI data centres tracked from satellite images and permits: power, chips, cost and owner. - AI models database Epoch AI
Training compute, cost and release date of notable AI models, back to the 1950s.
Risks and forecasts
- AI Incident Database Responsible AI Collaborative
Reported real-world harms and failures of AI systems, searchable. - Metaculus Metaculus
Crowd forecasts on when AI milestones will arrive, scored over time. - Artificial intelligence charts Our World in Data
Free charts on AI compute, investment, jobs and public opinion, with downloadable data.
Newsletters and podcasts to follow
Free
Dwarkesh Podcast
Dwarkesh Patel
Long interviews with lab leaders, researchers and economists. The best place to hear the arguments first-hand.
Partly paid
Stratechery
Ben Thompson
Strategy of the big tech companies: who captures the value from AI and why.
Partly paid
SemiAnalysis
Dylan Patel and team
Chips, data centres and supply chains in depth. Much quoted by investors.
Partly paid
Exponential View
Azeem Azhar
The AI economy, and a regular look at whether the boom is a bubble.
Partly paid
Understanding AI
Timothy B. Lee
Plain-English explainers by a journalist who checks the hype.
Know a piece that belongs here? Tell us at [email protected].
Not investment advice. Stonkometer is an educational calculator, not investment advice. Valuations depend entirely on the assumptions you choose. We are not a broker or adviser, and nothing here is a recommendation to buy or sell any security. Do your own research and consider a licensed adviser. Full disclaimer