Loops HouseNEAR Thesis: an investment case you can re-run

NEAR HACKATHON
Links
Team
1 member- AROwner
Arpit Singh Bhatia
Overview
A long-term investment thesis for NEAR that anyone can re-run. Every number in THESIS.md is computed from public, keyless data by the scripts in the repo, and a check fails if the thesis and the data disagree.
The thesis in five lines:
- NEAR's long-term case is NEAR Intents, not the L1: Intents did $32.9B of swaps in the last 365 days, 7.3x the year before, #18 of 1,157 DEX protocols on DefiLlama.
- Its fees are real ($5.99M in 30 days) and since 2026-02-23 NEAR's share buys NEAR on the open market: a $21.9M-a-year run rate.
- NEAR halved its maximum inflation from 5% to 2.5% onchain (read from RPC before and after; supply growth fell from 4.9% to 2.5% a year).
- Against it: the L1 has shrunk (62,933 daily active accounts, 3.1% of peak), DeFi TVL is flat, Intents is ~93% of NEAR's DEX volume, and the buyback covers only 19.8% of new NEAR while the token trades at 205x it.
- So holding NEAR is an option on Intents growing about 5x at today's take rate; six falsification tripwires say each quarter whether that is getting closer.
What's in the repo: a standard-library Python pipeline (49 public sources fetched, rebuilt into CSVs and metrics.json), THESIS.md rendered from the data, a dashboard of 8 headline tiles and 11 charts in plain HTML/SVG with no build step, and make check: unit tests plus 7 groups of checks and the six tripwires, offline on the committed snapshot. make all refetches everything in about 2 minutes.
Figures are from the snapshot fetched 2026-11-12.
Also entered by the same author: Glyf × NEAR, NEAR Intents in an AI keyboard, https://github.com/The-x-35/glyf/tree/near-hackathon