The Token Engineering Commons was the flagship deployment of the augmented bonding curve, and it is now the DAO industry's best-documented deliberate wind-down. Between July and December 2025 its remaining members argued for closing it, voted to close it, and then closed it on-chain - publishing the reasoning, the treasury arithmetic, the redemption formula and the claim contract as they went. Almost every DAO that dies leaves an unresolved treasury and a forum that simply stops. The TEC left a receipt.
The organisation described itself as the world's first explicit economic commons on a blockchain, built on Elinor Ostrom's design principles for governing a commons and funded through a curve that minted $TEC to anyone who deposited into its reserve. It funded cadCAD, the Token Engineering Academy and the Bonding Curve Research Group, among others. cadCAD itself has since gone quiet in the way the TEC did not: cadcad.org, community.cadcad.org and docs.cadcad.org all stopped serving in 2026 (the first two answer Cloudflare error 1033, an origin that no longer answers behind live DNS; the Internet Archive's last capture of the site is 20 May 2026), and the repository has had no commit since April 2024. The MIT-licensed code and the PyPI package are what remain.
On 23 July 2025 the steward rex published “Sunsetting the $TEC: Proposal & Treasury Distribution Framework”, and its argument is unusually blunt for a document written by the people whose roles it eliminates. The numbers it put on the table:
| Measure | July 2025 |
|---|---|
| Treasury and reserve, combined | ~$300,000 |
| Common Pool | $225,000 |
| Reserve backing the curve | $54,000 |
| Dormant funds in inactive multisigs | ~$25,000 |
| Operating cost | ~$10,000 per month |
| $TEC supply | 1.209M tokens |
| Spot price / fully diluted valuation | $0.18 / $217,000 |
| Treasury value per token | $0.25 |
The decisive line is the last two rows together: the token traded at roughly a 33% discount to the treasury standing behind it, in a favourable market. A token below its own redeemable backing is a market saying it expects the organisation to consume the difference. The proposal read that as a mandate rather than an insult, noting that the TEC had briefly broken even in 2024 before grants funding resumed without a sourcing plan, that governance participation was low, and that signers were going inactive faster than the treasury was shrinking. Its framing was a software end-of-life: deprecation notice, scheduled shutdown, archival and handoff.
A fuller shutdown proposal followed on 17 October 2025 from natesuits, extending the plan from the token to the organisation and arguing the case on legacy grounds - that a commons which taught the field how to launch should also demonstrate how to close.
“Proposal: Initiate the Sunsetting Process for the $TEC Token” ran on Snapshot from 30 July to 4 August 2025 and closed 264,496.79 $TEC for, 6,773.42 against, 0 abstaining - a 97.5% majority cast by eleven addresses. Queried through Snapshot's GraphQL hub on 28 September 2026 it is still the newest of the space's 28 proposals; nothing has been put to the space since.
Eleven voters is not an anomaly in this dataset. The space's preceding proposal, a coordination-team funding request, drew ten. A DAO whose participation has collapsed to a double-digit voter set can still execute a decision perfectly well - what it can no longer do is claim the decision represents its membership. This is the endgame of voter apathy: not a failed vote, but a passed one nobody contested.
The mechanics were published as they happened, in “TEC Token Sunset” (2 December 2025) and its replies. On 10 December 2025 the vote cleared the DAO's delay app and executed. Per the implementing contributor's own account, that single execution burned roughly 24,000 $TEC from the Common Pool, approved the transfer of the reserve's rETH and DAI to the claim contract, and closed the bonding curve - buy and sell calls against it now revert with APP_AUTH_FAILED. The $TEC snapshot was taken at Optimism block 144895034.
Both reference blocks resolve first-party against an Optimism RPC node:
| Block | Timestamp (UTC) | What it marks |
|---|---|---|
138850000 | 2025-07-23 19:26:17 | The day the wind-down was announced |
144895034 | 2025-12-10 17:47:25 | Dissolution; the curve stops trading |
Forty-one addresses held different balances at the two blocks. Redemption runs through TECClaim in CommonsBuild/tec-shutdown, and the design is worth reading if you ever have to build one: a UUPS-upgradeable contract that creates a non-transferable clone of the $TEC token using MiniMe's createCloneToken, frozen at the snapshot block, so the claim ledger cannot itself be traded. Holders burn snapshot tokens to claim a proportional share of the redeemable assets, which prevents double claims; the owner can adjust balances before activation, block addresses, and reclaim whatever is unclaimed after a deadline.
The transferable idea in the whole exercise is one line of the final plan. Each address's claim was capped at the lower of two balances:
Take the two cases in turn. Someone who bought $TEC after the announcement - while it traded below treasury value - is capped at their July balance, so the purchase buys no extra claim. Someone who sold after the announcement is capped at their December balance, so the sale is not undone. The cap therefore removes the entire trade: for the four and a half months between announcement and execution, no market action could change what anyone was owed. What is left is a claim on the treasury fixed by who was a member when the news broke.
This is the wind-down analogue of a snapshot block, and it solves a problem snapshots do not. A single snapshot at dissolution rewards whoever front-runs the announcement; a single snapshot at announcement traps sellers into a claim they no longer want. The minimum of the two is the only choice under which the announcement is not itself a trading opportunity - and it costs nothing but a second balance query. Any DAO facing a redemption event can copy it.
The cap did what it was designed to do. What it could not do was make anyone show up. The claim window opened after the 10 December 2025 dissolution and ran for the six months the shutdown proposal had committed to; three days after it closed, the steward rex published the turnout (a TEC forum post, offline since October 2026 and never archived) (22 June 2026): of 724 eligible wallets, 59 had claimed - about 8% - leaving roughly 450,000 $TEC, some 40% of the snapshot supply, unredeemed, almost all of it in ordinary externally-owned accounts rather than lost contracts. His own reading is the honest one - a 92% no-show by wallet count reads like people missing the window, not people declining their share - and he argued for an extension on exactly that ground. The wallet-by-wallet data is published on Dune.
The window was extended by a week past the original deadline (rex's status update dates that deadline 19 June 2026, the final-results post 20 June), and the final numbers (a TEC forum post, offline since October 2026 and never archived) followed on 11 July 2026:
| Measure | Final result |
|---|---|
| Eligible wallet addresses | 724 |
| Wallets that claimed | 70 (9.67%) |
| $TEC eligible / used for claims | ~1,107,831 / ~720,865 (65.07%) |
| DAI claimed / remaining | 77,370.98 of 118,908.57 / 41,537.59 |
| rETH claimed / remaining | 21.4107 of 32.9053 / 11.4946 |
| Approximate value claimed | ~$146,500, priced as each claim settled |
The extension bought eleven more wallets and about six percentage points of the treasury. Two true statements come out of that table pointing in opposite directions, and the TEC published both: nine in ten eligible wallets never claimed, and two thirds of the money went home. Roughly 90% of the claimable allocation sat in the top 100 wallets, so the value-weighted figure is the fairer measure of the wind-down's execution and the wallet-count figure is the fairer measure of what a claim window asks of an ordinary member. A distribution concentrated enough to make the first number respectable is the same distribution that makes the second one inevitable.
The unclaimed remainder did not sit. It was swept out of the claim contract on 2 July 2026 and, per the TEC's own announcement, sent to Giveth. That transfer resolves first-party against an Optimism RPC node: transaction 0x445a2114…b636a7a98 in block 153675651 at 06:54:55 UTC, a Safe execution moving 41,537.590536 DAI and 11.494624 rETH out of the TECClaim contract - the published remainders to the last decimal - into 0x4d9339dd…04d1c2cd, a thirteen-owner Gnosis Safe holding 2,699,249 GIV. Three days later the shutdown had its permanent archive, and the forum has been quiet since.
The transferable finding is the uncomfortable half. The TEC did the mechanism design well - the cap is genuinely worth copying - and still ended with nine in ten members not collecting. A redemption that has to be claimed inside a window is a redemption most of the membership will not claim, however long the window and however hard anyone chases it, because the constraint is attention rather than mechanism. Any DAO planning a wind-down should budget for a majority no-show and decide in advance, in public, where the residue goes.
As of 13 August 2026 - eight months after the curve stopped trading - tecommons.org answers HTTP 200 with a front page inviting readers to Build With Us and Get $TEC. There is no sunset notice anywhere on it. The token it points at cannot be minted; the curve behind that button is closed.
No link checker can see this, because nothing is broken: the site resolves, the pages load, the links work. The organisation's real status is legible in three other places - its Snapshot space (silent since August 2025), its forum (where the shutdown threads were the most recent substantial activity), and its GitHub, whose four most recently pushed repositories are tec-shutdown, TECredeem, tec-dao and TECsunsetcomingsoon, all created in the closing weeks of 2025. Its X account @tecmns is dormant. The gap has since widened rather than closed: the TEC published a permanent archive of the whole six-year project on 14 July 2026, and as of 28 September 2026 the front page still does not link it, still invites readers to Build With Us, and still offers Get $TEC.
Update, 8 October 2026: tecommons.org lapsed. The registration reached its expiry on 6 October 2026 and entered the registry's auto-renew period. Two days later tecommons.org served a domain-parking page and forum.tecommons.org refused secure connections, so the front page that never mentioned the sunset is gone, and so is the forum where the sunset was decided. The links on this page now go to the Internet Archive's last captures. Two threads were never captured: rex's claims status update and the final-results post, whose figures survive here and on the Dune dashboard cited above. The TEC Archive on GitHub Pages, published in July, is still up and is now the project's only first-party home.
For anyone auditing a directory of DAOs, that is a usable rule: a project's repository names change before its marketing does. The corollary matters too - the same gap in reverse is a live project with a frozen website.
The TEC needed a proposal, a vote and a purpose-built claim contract to give members their share of the treasury, and eleven addresses decided it for everyone. In a caper, leaving is a standing right rather than an event someone has to organise: a member can exit at any time for their fair share of the treasury, in one transaction, and nothing has to pass first.
The TEC's stewards were reacting to a token trading below the treasury behind it while the organisation spent the difference. In a caper that gap is not left to a vote to close: a member who thinks the treasury is worth more than the token can leave with their share on their own. A caper can still end, and no member waits on a vote to leave it.
The claim window is the other difference. Nine in ten of the TEC's eligible wallets missed a six-month window that was extended and chased by hand. A caper's exit has no window, so a member who notices in 2029 leaves in 2029, on the same terms, with nobody needing to process it.