Open-Source Intelligence · Federal Spending
Machine Age Money
A16z just raised $1.1 billion to build the physical layer of AI. The federal government has been paying for that layer for years. This resolves the public award record into entities and funding linkages — every dollar tagged to artificial intelligence, compute, and semiconductors, mapped to the agency that moved it and the entity that received it.
§01Entity Topology
Fifty-five billion dollars, resolved to entities.
Each large node is a funding agency; each node inside it is a recipient entity, sized by total dollars received and lit on a cold thermal scale — the brightest nodes took the most. Select an agency to open it. Select an entity to pull its file: every individual disbursement, linked to its official record.
§02Temporal Analysis
Where the money went, and when.
The same records across time. Left: obligations by fiscal year, split into grants and contracts. Right: every award placed by date and size on a log scale — the megadeals stand apart from the long tail. Rendered with Plottable, Palantir's charting library.
§03Funding Linkages
Which agency funds which entity.
The same capital as directed flow. Left: the agencies that moved the most money. Right: the entities that received it. Ribbon width is dollars. Trace a band to isolate one agency-to-entity linkage.
§04Top Entities
Highest-value recipients.
Total obligated to each entity across every award in the set. Select a name to pull its file into the topology above.
§05Record Search
Query the full record.
All 4,818 awards, indexed and searched in your browser — no query leaves the page. Every result links to its official USASpending.gov record.
§06Collection Notes
What this is, and is not.
Read before concluding
- A slice, not the whole federal budget. These are award records whose entity, agency, or description matched keywords like
artificial intelligence,machine learning,semiconductor,supercomputing,quantum,GPU,microelectronics,data center,neural network. - Every node is verifiable. The record link opens the official USASpending.gov page for that award. Nothing here is computed or invented; amounts and names are as the government reports them.
- Keyword matching is coarse. A contract for a building's data center and a grant for a machine-learning model both match. Read the description before drawing a conclusion.
- Award amount is the current total obligation at pull time, which can revise. Topology, linkages, search, and index all run client-side.
Sources: USASpending.gov · form after DataRepublican · framing after a16z's Machine Age Fund. Topology & linkages in D3; temporal charts in Palantir Plottable; UI in Palantir Blueprint (bp5-dark).