The Dimetrics MCP
Every HPC lease, the power behind it, and the debt that financed it — answered in Claude, with the filing cited.
Stop rebuilding the numbers from 8-Ks. Ask a question in Claude, Codex, or Excel and get a structured, EDGAR-cited answer on the companies building AI and bitcoin datacenter capacity.
7-day trial includes live Cipher Mining (CIFR) data · no card required
Sourced from SEC filings
Every figure resolves to an EDGAR anchor
Citation-backed
Verbatim quote behind each number
Verified, never inferred
Debt-to-lease links are source-cited only
In your stack
Claude · Claude Code · Codex · Excel
Built by an analyst who covered these names — for the desk that trades them
Wells Fargo SecuritiesInvestment Banking AnalystGalaxy DigitalVP of Mining · led mining researchSecond Gate AdvisoryCo-Founder · boutique BTC-mining advisory
Through Second Gate Advisory, advised Bitdeer, Compass Mining, and Sphere 3D. Speaker at The Bitcoin Conference, the MIT Bitcoin Expo, and the Nashville Bitcoin Mining & Energy Summit.
The fundamentals that move these stocks are buried in the filings.
Energized vs. secured megawatts. Contracted critical-IT load. Which senior note funds which lease, at what coupon. It is all in the 8-Ks and indentures — scattered, unstructured, and stale the day after you finish reading them. Generalist terminals do not carry it. Decks self-report it. So you rebuild it by hand, per name, every quarter.
Live example · on your trial
One question. The whole debt-to-lease chain, sourced.
You ask Claude
“Which debt instruments finance Cipher Mining’s HPC leases, and at what coupon?”
Dimetrics MCP returns · $4.54B across 4 senior notes
$2.0B6.125%due 2031Black Pearl· Amazon (AWS)
$1.4B7.125%due 2030Barber Lake· Fluidstack
$810M6.00%due 2031Stingray· Amazon (AWS)
$333M7.125%due 2030Barber Lake· Fluidstack (tack-on)
Each instrument links to its SEC filing with the verbatim source — e.g. “we raised $1.4 billion by selling five-year senior secured notes of 7.125% to fund the development at Barber Lake.”
Try this exact query on the free trial. CIFR is included; the Analyst plan opens the full universe.
What you can ask
Nine read-only tools. Plain-language questions in, structured + cited answers out.
Lease & deal economics
On trial“What’s Cipher’s contracted critical-IT MW, and the $/MW/yr on the Fluidstack lease?” — gross & critical-IT MW, contract value, term, NOI margin, $/MW/yr, $/kW-month.
Power capacity & pipeline
On trial“How much of Cipher’s power is available but unleased?” — the VanEck ladder: operating, energized, secured, prospective; available-unleased; per-site by state and grid operator.
Debt-per-lease linkage
On trial“Which note funds Barber Lake, at what rate and maturity?” — verified, EDGAR-cited instruments tied to the lease they finance.
Capital structure
Analyst plan“Roll up the issuer’s debt and tie it to the latest balance sheet.” — full tranche-level cap structure with per-field provenance and reconciliation.
Political & regulatory risk
Analyst plan“Score the regulatory risk on this operator’s sites.” — county-level political-risk scores for datacenter locations.
Moratorium intelligence
Analyst plan“Where are datacenter moratoriums spreading, and what correlates?” — dataset-wide US moratorium overview, by-state, recent, and correlates.
Built for the desk that owns these names
The analyst
Stop reading every 8-K. Get the contracted MW, the financing, and the $/MW in seconds — and cite the filing in your note without leaving Claude or Excel.
The family office
Defensible, auditable diligence on a fast, opaque sector. Every figure traces to a primary source you can click through.
The fund
Sector depth your generalist terminal does not carry — debt-per-lease, capacity ladders, and moratorium risk across the whole universe, queryable from your stack.
Why not the alternatives
A $24k generalist terminal
Broad coverage, no digital-infra depth — no debt-per-lease, no critical-IT capacity ladder, no moratorium data.
Reading the filings yourself
It is all there — across dozens of 8-Ks and indentures, unstructured, and stale the moment you finish. Hours per name, every quarter.
Mining newsletters & trackers
Commentary and hashrate, not structured, citation-backed lease, debt, and capacity data you can query.
The MCP is included with the Analyst plan
$1,000/year — all nine tools, the full Power Shells + Neo-Clouds universe, queryable from Claude, Codex, and Excel. That is roughly the cost of an hour of an analyst’s time per month, against one better-informed position.
Free 7-day trial on CIFR · then Terminal $29/mo or Analyst $1,000/yr
Questions
What is the Dimetrics MCP?
A Model Context Protocol server that connects Dimetrics data to Claude, Claude Code, Codex, and Excel. You ask a question in plain language; the model calls Dimetrics tools and returns structured, citation-backed answers on HPC lease economics, power capacity, debt financing, capital structure, and political risk for US bitcoin miners and neoclouds.
Where does the data come from?
Primary-source SEC filings. Every figure resolves to an EDGAR URL with the verbatim quote it was drawn from. Debt-to-lease links are verified and source-cited only — never inferred.
What does the free trial include?
Signing up starts a 7-day MCP trial scoped to Cipher Mining (CIFR) with the HPC-lease and lease-debt tools — list deals, deal economics, debt-for-lease, and available capacity. No card required.
What does the full plan unlock?
The Analyst plan ($1,000/yr) unlocks all nine tools — adding political risk, datacenter-moratorium intelligence, full issuer capital structure, instrument detail, and debt-to-balance-sheet reconciliation — across the entire Power Shells + Neo-Clouds universe.
How do I connect it?
In Claude, add a custom connector pointing at https://www.dimetrics.ai/api/mcp and sign in. In Claude Code: claude mcp add dimetrics --transport http https://www.dimetrics.ai/api/mcp. Full setup is in the docs.
Is this investment advice?
No. Dimetrics provides sourced data, not recommendations. There are no price targets or buy/sell signals — the edge is your judgment on better information.