Work with us
Talks, briefings and research on the money behind AI
Briefings, talks and research for boards, investment committees, conferences and companies that need to know where the money in AI is going, who keeps it, and what the filings say about whether it is paying.
Start an inquiryWhy us
Most commentary on AI investing is either hype or alarm, and both usually come from press coverage and earnings-call enthusiasm. Our work starts from what companies actually report to the SEC: capital spending, segment revenue, margins, backlog, lease commitments, useful lives, and the financing links between suppliers and their customers, each quoted from its filing. That is what this site publishes, and it is what we bring into a room.
We combine two vantage points. Rob May has invested only in AI since 2015, through HalfCourt Capital, and wrote the book Investing in AI on how to tell AI moats from traps. Vik Chadha has spent his career building and scaling companies, co-founding UnifyCX and growing it to more than 6,000 employees, and now runs Scalable Ventures, a venture lab building AI-native B2B software, and invests as a general partner at Unbridled Ventures. Together we built and sold Backupify. We read AI companies the way an operator and an investor both would: what does it cost, who pays, and does it come back?
See the kind of work we do: the AI stack map, the AI money loop and following the capex money.
Ways to work with us
Keynotes and talks
Conferences, investor events, executive offsites, industry associations.
A talk built on the latest filings, with the charts updated for your date. Typically 30 to 60 minutes with questions, in person or virtual. The talks below are starting points; we tailor the emphasis to the audience.
Board and investment-committee briefings
Boards, investment committees, family offices, CFOs and strategy teams.
A focused session on what the AI build-out means for your company or portfolio: where you sit in the stack, how exposed you are to the capex cycle and its financing, which signals to monitor, and the questions to put to management or to the companies you own. Pre-reading and a short written summary are included.
Custom research and data
Investors, corporate development teams, publishers and platforms.
Filings-based analysis of a company, a peer group or a portfolio, using the same methods as this site: where each company sits in the stack, how much of what it is paid it keeps, who it depends on, and what its AI claims rest on. The underlying datasets (capex, segments, financing links and disclosures, refreshed each quarter) can also be licensed.
Advisory and board roles
Companies building or deploying AI, and funds investing in it.
Ongoing advice on AI strategy, product economics and capital allocation, from people who have built companies, invested in AI for a decade and read the filings of the companies that set the market. We take on a small number of advisory and board roles where we can be genuinely useful.
Who we work with
- Boards and investment committees that need an independent view of AI spending and its risks before a strategy or allocation decision.
- Asset managers, family offices and wealth teams whose clients ask what the AI build-out means for their portfolios.
- Companies selling into the AI build-out, from components to power, that want to understand their customers’ economics and the cycle they depend on.
- Software companies and their boards deciding how to price, build and talk about AI so that it shows up in the numbers investors watch.
- Conference organizers and publishers looking for a data-driven session on AI investing that holds up to a skeptical audience.
What a board briefing covers
A typical 90-minute session, adapted to your business and sent with pre-reading a few days ahead. Every chart and figure is current to the latest filings and traceable to its source. Our board guide sets out the twelve questions we start from.
- 01
Where you sit. Your company or portfolio placed on the AI stack: which layers your revenue and costs depend on, and which of the buyers’ decisions move them.
- 02
The cycle in numbers. The latest quarter’s spending, cloud revenue, supplier margins and financing commitments from the filings, and what changed since the last one.
- 03
Exposure. How your business or holdings would be affected if capex keeps rising, flattens or falls, and which of your counterparties depend on vendor financing.
- 04
Signals to monitor. The handful of disclosures worth tracking each quarter for your situation, and who on your team should own them.
- 05
Questions to ask. For management, portfolio companies or vendors: the questions that separate AI that is earning its cost from AI that is mostly talk.
Talks
Each talk is rebuilt from the latest filings for your date, so the numbers on the slides are current when you hear them.
Who gets paid when AI gets built
The largest five buyers spent $566B on capital projects in four quarters, and 97% of their operating cash flow in Q2 2026. This talk follows that money through the eight layers of the AI stack, from the clouds to chips, memory, networking, power and foundries, shows which layers keep the profit (chips and memory took 83% of the suppliers’ extra gross profit over two years) and asks whether the revenue is coming back.
The AI money loop
Suppliers now invest in, lend to, guarantee and buy capacity back from their own customers. We map the 25 links the filings disclose, from Nvidia’s ecosystem investments and lease guarantees to Amazon’s credit line for Anthropic and AMD’s warrants, and explain what vendor financing means for revenue quality, concentration and hidden leverage, with lessons from the telecom build-out.
Telling AI moats from traps
The six mental models from Rob May’s book Investing in AI, applied to today’s market: why the habits formed in the dot-com and SaaS eras mislead AI investors, how adoption and abandonment move faster, and how inference costs can eat margins that look durable.
Is your AI earning its cost?
A practical framework for boards and investors: the evidence in a company’s own filings that its AI is paying (margins, revenue growth, backlog, the specificity of its AI claims) and the warning signs that it is mostly talk. Based on the AI Signal Gauge we built and the backtests behind it.
The depreciation bill nobody is talking about
Capex arrives on the income statement slowly, over the useful lives companies choose for their servers. We show how much depreciation is already locked in, what a one-year change in server life is worth, and the signals in the filings that hardware is aging faster than planned.
How it works
Step 1
Tell us what you need
Use the form below. A few lines on the audience or decision, the timing and anything we should read is enough.
Step 2
A short call
We will suggest a 20-minute call to understand the question and agree what a useful outcome looks like. If we are not the right people, we will say so and, where we can, point you to someone who is.
Step 3
A clear proposal
Scope, format, timing and fee in writing before any work starts. Fees depend on the engagement, so we quote after the call rather than publishing a rate card.
Step 4
Delivery from the filings
Every figure we use is sourced from SEC filings or other public documents and can be traced. We update our numbers for your date, so a talk or briefing reflects the latest quarter.
What you get
Every engagement ends with something you can keep. For a talk, the slides with sources, updated to the date of the event. For a briefing, the pre-read, the charts and a short written summary of the questions and signals we agreed matter for you. For research, a written report with every figure traced to the filing it came from, and the data behind it in a form your analysts can use.
Because our numbers come from a pipeline we run every quarter, a briefing can be refreshed after the next earnings season at little extra effort, so a board can see the same frame, updated, at its next meeting rather than starting again. Many of the questions a board asks one quarter can be answered from the next quarter’s filings.
Independence
We don’t give personalized investment advice or tell anyone to buy or sell a security, in talks, briefings or research. Our work explains how the industry’s money moves and what the evidence shows; decisions stay with you.
We don’t accept payment to cover, feature or rate a company on this site. If we have a material interest that bears on an engagement (a holding, an advisory role or a commercial relationship), we disclose it before we start.
Start an inquiry
Tell us about the audience or decision, the timing, and anything we should read first.
Questions about working with us
Do Rob May and Vik Chadha speak at conferences?
Yes. We give keynotes and talks on AI investing, the economics of the AI build-out and how to read AI companies from their filings, in person or virtually. Use the form on this page with your date and audience.
What does a board briefing on AI cover?
Where your company or portfolio sits in the AI stack, how exposed it is to the capital spending cycle and its financing, which signals to monitor, and the questions to put to management. It is built from SEC filings and tailored to your business.
How much do you charge?
It depends on the format, scope and timing, so we quote in writing after a short call rather than publishing a rate card.
Will you recommend stocks?
No. We don’t give personalized investment advice or buy and sell recommendations. We explain where the money in AI goes, who keeps it and what the filings show, so you can make better-informed decisions.
Can we license your data?
Yes. The datasets behind the AI stack map, including quarterly capex, segment revenue, financing links quoted from filings and lease disclosures, can be licensed. Tell us how you would use them.
Can a briefing cover a specific company we own or compete with?
Yes. We can place any SEC-filing company on the AI stack and walk through what its filings show: its revenue and margins, its customers and suppliers, its commitments and any financing links. We don’t give buy or sell views.
How far ahead should we book?
For a keynote, a few weeks is usually enough; briefings can often be arranged sooner. If your date falls soon after a Big Tech earnings season, tell us, because we can build the session around the newest filings.
Do you give talks virtually?
Yes. Talks and briefings can be in person or virtual; tell us the format and time zone in the form.
Do you take on advisory or board roles?
Selectively, where our mix of operating, investing and AI research experience is genuinely useful. Use the form to tell us about the company or fund.