Archimedes Tech SPAC Partners
#44 of 11753/100 from 1 resolved vehicle (1 closed, 0 failed), 17% of the raw 68 after small-sample shrink, completion credit gated ×0.96 by the measured post-close record. Confidence: low.
Sponsor DNA
what has happened before, with its sample size- Completion rate—n=1 resolved vehiclederived
- Liquidation rate—n=1 resolved vehiclederived
- Median post-close return—n=1 priced completed deSPACderived
- Median redemption—n=0 redemption events with a stated ratederived
- Deals terminated0terminated dealscounted
- Extension votes on record0extension votescounted
2 of 6 statistics carry a figure for this sponsor. A rate is published from 3 resolved vehicles and a median from 3 observations: two points have no middle, and a rate over two can only be 0, 50 or 100. Counts have no threshold — a count is an observation, not an estimate.
Everything marked derived is arithmetic we did to rows we hold, not a figure any filing states.
What this panel will not tell you, and why (6)›
Score breakdown
every component, what it measured, and what it could not- Deal completion20% weightn=196/100
1/1 resolved vehicles closed a deal (100%); 0 liquidated, 0 terminated. Gated ×0.96 by measured post-close quality (46/100): closing deals that ended below trust value is not a completed job, so only 96% of the completion credit is earned. Full credit resumes at outcome quality 50/100 (the median deSPAC ending at trust value); the gate can never exceed 1×. Small sample — the shrink below keeps this near neutral.
- Liquidation / termination drag16% weightn=3100/100
0 liquidations and 0 terminations across 3 vehicles raised → 0% attrition (terminations 1.25×, stale shells 0.75×).
- Post-close outcome quality40% weightn=146/100
1 priced deSPAC vs trust value (prior vehicles against the $10.00 IPO baseline, in-DB vehicles against the trust they filed): median -26%, 1/1 still worth at least half of trust, 0 at under a tenth of it. Worst: SOUN -26%. n=1, pulled toward neutral.
- Redemption behaviour10% weightnot measurable
No redemption events extracted for this sponsor yet (coverage is partial) — held neutral; absence of rows is NOT evidence of zero redemptions.
Held at the neutral 50 across its full 10% weight — missing data is never scored as a failure, but it never earns credit either.
- Extension reliance8% weightnot measurable
No extension filings extracted for this sponsor — held neutral (partial coverage, not a clean record).
Held at the neutral 50 across its full 8% weight — missing data is never scored as a failure, but it never earns credit either.
- Live fleet vs trust6% weightn=2100/100
2/2 live vehicles trading at or above the trust value they filed.
- Measured weak recordflat penaltyn=1not measurable
Only 1 measured prior vehicle (median -26%) — one vehicle is an anecdote, not a record; the rule needs ≥2.
How the number is built: weighted mean of the six components above = 68, then pulled 83% of the way back to the neutral 50 for small sample size (1 resolved vehicle) = 53.
2 components are not measurable for this sponsor (redemption behaviour, extension reliance) — 18% of the weight is a neutral placeholder rather than evidence. That is why the confidence chip reads low.
How the Sponsor Score worksoutcome-first weighting
The score answers one question: did this sponsor make money for the people who held through the merger? Not “did they get a deal signed”. Those are different questions, and most sponsor rankings quietly answer the second one.
So post-close outcome quality carries 40% — the realised return of every prior vehicle we can price from a primary filing, measured against the $10.00 trust baseline. Deal completion carries 20%, and it is gated: closing deals that ended below trust value only earns part of the completion credit, because closing is a precondition for a return, not a return. Liquidation and termination drag takes 16%, redemption behaviour 10%, extension reliance 8%, and what the tape says about the live fleet just 6% — a quote is an opinion, not evidence.
A component with no data is never guessed. It is held at the neutral 50 across its full weight and labelled “not measurable”. Dropping it and re-weighting the rest would quietly reward a sponsor for having no verifiable record — exactly backwards. The consequence: a sponsor with no post-close evidence at all cannot read above 71, and cannot be labelled a strong operator no matter how many deals it closed.
Experience never inflates the score. There is no “years in business” component. A first-time sponsor sits at exactly 50 and reads “unproven” with low confidence — new is not bad. Sample size only pulls a score toward or away from that neutral 50, so nobody is called great or terrible on one vehicle.
Every input is a row already in the database, sourced from SEC primary filings: prior vehicles verified on EDGAR, redemption results read out of 8-Ks, prices from public feeds. The arithmetic is deterministic — no model, no LLM, no judgement call. Research tooling, not investment advice.
Prior vehicles
1 SEC-verified — what happened to holders who stayed in| Vehicle | Outcome | Became | vs $10.00 | Today | Source |
|---|---|---|---|---|---|
| Archimedes Tech SPAC Partners CoIPO 2021 | Completed | SoundHound AI, Inc.SOUN | -25.7% | Trading$7.43 · Aug 14, 2026 | 0001213900-22-023274 |
1 of 1 prior vehicle carry an honest post-close price, split-adjusted against the $10.00 trust baseline a holder gave up at the merger. Cash buyouts are read from the per-share consideration stated in the DEFM14A / SC 14D-9; a buyout no filing prices stays unpriced and stays out of the score. “Listing ended” means the quote stopped with no buyer — scored as a total loss because that is what the evidence says, but never printed as a percentage we cannot source.
Current fleet
the vehicles running todayResearch profile
synthesized from SEC filings + sourced researchOne sponsor-LLC series — Archimedes Tech SPAC Sponsors LLC, II LLC and III LLC — behind Archimedes Tech SPAC Partners Co (SoundHound), II (ATII) and III (ARCI). EDGAR full-text search on "Archimedes Tech SPAC" returns exactly those three filers and no other corporate entity.
Data provenance & audit trail1 internal entry
Written by SpacBrain’s data agents whenever a figure is captured, corrected or flagged, and kept verbatim so every number on this page can be traced back to the filing that states it. This is a running log, not the current record: an early entry may be superseded by a later correction — the panels above always hold the current values.
Family asserted from primary SEC data only (Form 3/4 reporting-owner XML, EDGAR submissions formerNames, sponsor-LLC registered addresses, and the vehicles' own prospectuses).
The Sponsor Score is a deterministic research heuristic over primary-sourced rows — never a recommendation, and never a prediction. It cannot tell you whether this sponsor’s next deal will work; it tells you, precisely and with its own uncertainty attached, what the last ones did.