# Largest Hedge-Fund Buys in Q2 2026 (by Dollars)

> Stocks that took in the most estimated institutional buying dollars in Q2 2026, from new and added 13F positions. Sourced to SEC EDGAR.
> Period: Q2 2026. Teaser: top 25 of 100 ranked names.
> Board recomputed 2026-09-06 from the latest effective filings.
> Berkshire Hathaway Class A leads this board at $265.0B.

## Ranking (teaser)

| # | Ticker | Issuer | bought |
|---|--------|--------|------||
| 1 | BRK/A | Berkshire Hathaway Class A | $265.0B |
| 2 | SPCX | Space Exploration Techn-CL A | $98.1B |
| 3 | AZN | AstraZeneca | $43.0B |
| 4 | NVDA | Nvidia | $37.7B |
| 5 | MSFT | Microsoft | $36.2B |
| 6 | AMZN | Amazon.com | $33.6B |
| 7 | GOOGL | Alphabet Class A | $33.2B |
| 8 | KLAC | KLA | $32.3B |
| 9 | AAPL | Apple | $29.7B |
| 10 | MU | Micron Technology | $27.1B |
| 11 | GOOG | Alphabet Class C | $23.8B |
| 12 | TTE | Totalenergies SE | $22.7B |
| 13 | AVGO | Broadcom | $21.4B |
| 14 | META | Meta Platforms Class A | $17.9B |
| 15 | BKNG | Booking Holdings | $15.3B |
| 16 | TSM | Taiwan Semiconductor ADR | $13.5B |
| 17 | SNDK | Sandisk | $12.9B |
| 18 | AMD | Advanced Micro Devices | $12.6B |
| 19 | LRCX | Lam Research | $12.0B |
| 20 | INTC | Intel | $11.4B |
| 21 | AMAT | Applied Materials | $10.6B |
| 22 | V | Visa Class A | $10.4B |
| 23 | NFLX | Netflix | $10.3B |
| 24 | LLY | Eli Lilly & Co | $10.0B |
| 25 | CBRS | Cerebras Systems Class A | $9.8B |

ETFs and index wrappers (9 in the top 100) are listed separately on the human page, not in this stock teaser.

## How this is computed

Computed from the latest non-superseded 13F filing of every tracked fund as of Q2 2026, grouped by security. A 13F shows long positions only, on a quarterly lag. No look-ahead. Counts reflect the 13F filers and securities Arkolith tracks, not the entire institutional universe.

## Full data via the API and MCP

This page is a teaser. The complete book, full quarter-over-quarter history, and per-datapoint
provenance are available through the metered REST API and MCP server.

- Endpoint: https://arkolith.com/api/mcp (transport: streamable-http)
- Auth: Bearer API key. Mint one and copy a ready-to-paste command at https://arkolith.com/connect
- Claude Code: `claude mcp add --transport http arkolith https://arkolith.com/api/mcp --header "Authorization: Bearer YOUR_API_KEY"`
- Discovery: `tools/list` costs no credits but needs the API key (the server answers 401 without one); `tools/call` is metered by credits.

### Machine-readable
- Tool catalog JSON: https://arkolith.com/api/mcp/catalog
- MCP descriptor: https://arkolith.com/.well-known/mcp.json
- OpenAPI 3.1 REST spec: https://arkolith.com/openapi.json
- Full LLM guide: https://arkolith.com/llms-full.txt
- Human docs: https://arkolith.com/docs

## Source and method

- Source: SEC EDGAR 13F (public domain).
- Human page: https://arkolith.com/rank/largest-buys
- Per-datapoint lineage: call the `provenance.get` MCP tool for the source URL and fetch history behind any figure.

Paid product. $1 trial at signup: 300 research credits over three days, then $29/mo, cancel any time. Most calls cost 1 credit (~$0.01); heavier point-in-time / cross-source tools cost more.

_Latest disclosed public filings. Data, not investment advice._