Counterparty and solvency (OLP)
OLP is the sole counterparty to a net-long book of about $270M. About $43M is visible on-chain; its venue capital, pool margin and hedges cannot be checked.
Section 5 →Partner due diligence · prepared for a Layer-1 foundation
Variational
Omni, a request-for-quote perpetuals exchange settling in USDC on Arbitrum One, with a single in-house market maker (OLP) on the other side of every trade. Pre-token: the $VAR TGE is planned for Q4 2026.
Demo report. This is a worked example of DINQ's agent-driven partner diligence. It is not for distribution, and Variational has not been asked to comment. It uses public information and on-chain data only. Nothing here is legal, investment or financial advice.
Recommendation
Proceed with conditions
Where Variational can be tested, it holds up. The collateral is real and matches what the company reports, treasury revenue reconciles to the cent, every sampled flow of user funds went where it should, and the founders, investors and screens check out.
The partnership risk sits where the chain cannot see. OLP's solvency cannot be verified. Sixteen single-signer keys can move user and OLP funds with no on-chain limit, and users have no way out on their own. By its own terms the company holds no licence for the leveraged derivatives it offers. None of this is evidence of misconduct, but each point depends on controls a foundation should see evidenced before it attaches its name.
Would move to “Do not proceed yet” if Variational cannot evidence conditions 1 or 2, if the hedge-venue accounts turn out not to be held for OLP, or if a regulatory action, incident or sanctions link surfaces.
Would move to “Proceed” once conditions 1–4 are met, a multisig or timelock is visible on-chain, and OLP attestations recur.
Conditions before signing
- OLP solvency evidenceA dated statement of OLP equity across every venue, backed by third-party statements or an attestation, with the hedge policy by asset class and confirmation of who controls the hedge-venue addresses.
- Key-control evidenceThe custody and approval policies for the admin address, the 16 provider keys, the treasury and the vault, and a dated plan for on-chain limits: multisig and timelock, capped fee collection, a pool-only destination check and a user exit path.
- Legal postureCounsel's view on licensing for the products offered, evidence of restricted-jurisdiction controls beyond an IP banner, and partnership terms that keep the foundation out of marketing to restricted jurisdictions.
- Corporate chartAll five entities with directors, owners and UBOs, which entity holds OLP's capital, and which entity issues $VAR.
$239.1M
USDC held in 203,417 settlement pools. Reported TVL matches within 0.15%.
On-chain · block 515,041,730
$272.5M vs $43.1M
Users' net long book against the OLP capital visible on-chain (about 16%).
Company API · on-chain · 15 Oct 13:28Z
16 keys
Single-signer provider keys that can move pool and vault USDC. No quorum, no timelock.
Verified source and roles · block 515,074,626
20% / 85.7%
Share of accounts holding at least $1, and the share of collateral held by the largest 1%.
On-chain · block 515,041,730
$61.8M
Raised in three rounds led by Bain Capital Crypto, Peak XV and Dragonfly.
Primary press and investor sites
0 licences
Found in the four registers checked. More than half of volume is on non-crypto underlyings.
ToS, ESMA, FinCEN, MAS, CFTC
Scorecard
Eleven dimensions, rated after verification
Red means the evidence a partner needs is missing or the design concentrates risk. It is not a finding of wrongdoing.
Smart-contract and admin risk
One admin address and 16 provider keys can move pool and vault USDC with no cap, quorum or timelock, and users have no exit of their own.
Section 6 →Legal and regulatory
Unlicensed by its own terms, offering leveraged derivatives and CFDs with more than half of volume on traditional-finance underlyings.
Section 7 →Custody and fund flows
Every sampled flow of user funds was clean, but the vault is a key-controlled wallet and OLP capital goes to venue accounts of unconfirmed ownership.
Section 5 →Traction and data authenticity
Collateral and revenue reconcile. Volume and OI are self-reported, 80% of accounts hold under $1, and 150 pools hold half the collateral.
Section 4 →Security programme
The only contract review began 14 days after launch and published 2 of 46 findings. The router has no public review. No incident found.
Section 6 →Corporate transparency
Five entities in four jurisdictions trace to official records, but no officers, owners or group chart are public for any of them.
Section 1 →Team
The founders' records check out against an SEC filing and Genesis's own release. No compliance, legal or risk function is visible.
Section 2 →Reputation and community
A large, mostly real audience with almost no funds-safety complaints, but coverage is largely paid or referral-driven.
Section 8 →Token design
Standard insider lock-up, but a fully liquid 32% airdrop, an undisclosed team/investor split and a buyback promise reworded twice.
Section 9 →Funding and backers
About $61.8M from established crypto funds; 8 of 10 named investors list Variational on their own sites.
Section 3 →What holds up
- The money in the pools is real. On-chain USDC matches reported TVL within 0.15% at three separate blocks.
- Revenue reconciles exactly. 69 hours of treasury inflows equal its balance change to the cent, all from user pools.
- User funds went where they should. In a 200-pool, 30-day sample, every dollar went to the OLP vault, the pool's own trader, the treasury or a fee key.
- The founders are who they say. An SEC Form D and Genesis's own release confirm the Qu Capital and Genesis history.
- Clean screens. No sanctions, litigation, enforcement or security incident was found.
What we could not verify
- OLP's total capital and hedges. Most of it, if it exists, sits off-chain at venues.
- How the keys are controlled. Fireblocks MPC policies, if used, are invisible on-chain.
- Trading volume. Revenue is consistent with it, but a 2× overstatement cannot be ruled out.
- Who owns the entities. No registry we could reach names officers we can tie to the team.
- Whether restricted users can trade. We saw a banner, not a block, and did not place trades.
Timeline
From a 2018 quant fund to a 2026 token launch
Dates come from filings, chain records, archived documents and primary press. Filled markers are the events this report leans on most.
- Qu Capital fund files Form D naming both future co-founders as executive officers. SEC EDGAR
- Genesis buys Qu Capital's assets; the founders join Genesis Global Trading until 2021. Genesis release
- Variational Research, Inc. is formed in Delaware. It still holds the variational.io domain. NY DOS · RDAP
- Seed round closes, priced equity with 1:1 token warrants per the CEO. It is announced three years later. The Block
- Applied Variational Research SEZC (Cayman) is named data controller. Archived Privacy Policy
- $10.3M seed announced, co-led by Bain Capital Crypto and Peak XV. Primary press
- MCMC Research Corporation, S.A. (Panama), the future operator, opens its folio. Registry aggregator · lead
- Core contracts go live on Arbitrum: factory, oracle and pool template, all immutable. Chain
- The only contract review starts, 14 days after deployment (Spearbit via Cantina, to 9 Mar). Cantina report
- $1.5M strategic round from Mirana, Caladan and Zoku. Wire release only
- Router deployed. It moves vault USDC into pools; its source is unpublished and unreviewed. Chain
- Points programme launches; account creation peaks three weeks later. Docs · chain
- Immunefi bounty goes live, capped at $100k. Immunefi
- Taiwan is dropped from the Restricted Persons list. Wayback
- ~$50M Series A led by Dragonfly with Bain Capital Crypto and Coinbase Ventures. Fortune · CoinDesk
- Taiwan is added back; the page's “last updated” date stays April 20. Wayback
- Variational Token Holding Ltd (BVI) obtains an LEI. GLEIF
- $VAR TGE announced for Q4 2026; the next day the buyback wording becomes “intend 100%”. X · docs
Section 1
Company and entities
Bottom lineVariational runs through at least five entities in four jurisdictions and publishes no group chart. Each link to the brand traces to an official document, but none of the registries we could reach names an officer or owner we can tie to the team.
| Entity | Where | Role | Status |
|---|---|---|---|
| MCMC Research Corporation, S.A. | Panama | Operator and data controller under the current Terms of Use (folio opened about 7 weeks before the first ToS) | Verified role · Lead registry details |
| Variational Research, Inc. | Delaware, US | Registrant of variational.io; NY biennial statement past due since Nov 2023 | Verified |
| Applied Variational Research SEZC | Cayman Islands | Data controller May 2024 – Jan 2025; “Sponsor” of the 2026 trader campaign | Verified |
| Variational Token Holding Limited | BVI | LEI issued 21 Sep 2026, two days before the TGE announcement; no recorded link to Variational | Lead role |
| Variational Foundation | Not disclosed | Named by the company as holder of the 18% ecosystem reserve | Company claim |
The operator is a Panamanian company with professional officers
The Terms of Use and Privacy Policy name MCMC Research Corporation, S.A. as operator. Two Panamanian aggregators tie it to Mercantil folio 155760399. One cached registry copy shows the folio opened on 11 Dec 2024, about seven weeks before the first ToS, and lists officers who hold many other directorships, as is typical of law-firm incorporation services. No Variational founder appears. We could not read the official Registro Público, which needs a personal account and a CAPTCHA.
A US company holds the domain
The .io registry lists the registrant of variational.io as Variational Research, Inc. (New York, US). New York lists it as an active Delaware corporation formed on 7 Oct 2021, with its biennial statement overdue since November 2023 and no CEO on file. Its current role is not disclosed. This matters in section 7, because US persons are restricted from the product.
Ask VariationalQ8 corporate chartQ15 token issuer
Section 2
Team
Bottom lineThe founders are credible and their history checks out against primary records. The concern is what is not there: a derivatives venue with a single counterparty shows no compliance, legal or risk function, and the trading and risk work appears to rest on the two founders.
21
LinkedIn profiles; about 17 core staff
10
of the 21 joined in 2026
0
compliance, legal or risk titles
0
profiles from Virtu or Jane Street, as the docs claim
The founders' history holds up against filings
An SEC Form D signed on 10 Jul 2018 lists co-founders Lucas Schuermann and Edward Yu as executive officers and directors of Qu Capital US Fund, LP. Genesis's own release of 19 Sep 2019 announced it was acquiring Qu Capital's assets, and both founders' profiles show Genesis Global Trading from 2019 to 2021. One small difference: Edward Yu's own profile and website give his Genesis title as “Head of Quantitative Research”, where Variational's docs say “VP of Quant Trading”.
No link to the Genesis and DCG litigation
Neither founder, Qu Capital nor Variational is named as a party in the 25 Genesis and DCG federal dockets we reviewed. OpenSanctions found no sanctions or PEP match for the founders or five senior staff, and we found no adverse media. The search did not cover New York state courts, the Genesis claims register, or PACER beyond what RECAP holds.
A small, recent team with no visible control functions
LinkedIn shows 21 profiles, about 17 once community and affiliate roles are removed, and 10 of them joined in 2026. None holds a compliance, legal, counsel, risk, AML or KYC title, and the docs name no CCO, general counsel or CRO. Besides the founders, no quant or trading staff are visible. The docs' description of staff from “Virtu … Jane Street” is supported by none of the 21 profiles.
Section 3
Funding and backers
Bottom lineAbout $61.8M from established crypto investors, with the two main rounds reported independently. The terms of the later rounds and the investors' token share are undisclosed.
| Round | Announced | Amount | Investors | Status |
|---|---|---|---|---|
| Seed | 23 Oct 2024 (closed Dec 2021) | $10.3M | Co-led by Bain Capital Crypto and Peak XV; priced equity with 1:1 token warrants, per the CEO | Verified |
| Strategic | 4 Jun 2025 | $1.5M | Mirana Ventures, Caladan, Zoku Ventures. Caladan is a market maker; see Q18 | Company claim |
| Series A | 20 May 2026 | ~$50M | Led by Dragonfly, with Bain Capital Crypto and Coinbase Ventures | Verified |
Eight of ten investors list Variational on their own sites
All three announcements went out as paid wire releases, but the seed and Series A were also reported independently, by The Block, Fortune and CoinDesk. Eight of the ten named investors list Variational in their own portfolios. Peak XV's page dates the relationship to 2026, which does not fit a seed co-lead whose round closed in 2021. No SEC Form D names any Variational entity or founder after 2018.
Ask VariationalQ21 round termsQ18 investors as counterparties
Section 4
Traction and data authenticity
Bottom lineThe collateral is real and the treasury's revenue reconciles with what Variational reports. Volume cannot be independently verified, most accounts are dormant, and a few hundred accounts hold most of the money.
Reported TVL matches on-chain collateral within 0.15%
We listed all 203,417 settlement pools from the factory's events and read each pool's USDC balance at a pinned block. The total matched the stats API's tvl within 0.15% at three separate blocks. The API leaves out the $26–44M held at the address Arbiscan labels “Core OLP Vault”, although the docs define TVL as pools plus the vault, so the reported figure understates rather than inflates. Pool balances also include some OLP margin, so this is collateral in pools, not user deposits alone.
Reported TVL against on-chain collateral
USD millions at two full scans. The docs' own definition of TVL (pools plus the OLP vault) would be higher than what the company reports.
tvlShow data
Revenue reconciles to the cent; volume stays a company figure
Over 69.4 hours from 12 to 15 Oct, the treasury received $333,747.31 in 619,657 USDC transfers, all from 19,406 settlement pools, and sent nothing. That equals its balance change exactly, a run rate of about $42M a year. Over the same window the API's cumulative volume rose by $10.21B, so revenue is 0.327 bps of reported volume. At the documented 20% treasury share, that implies about 1.63 bps paid per dollar traded, inside the half-spreads Variational quotes (about 0.65 to 3.2 bps). The check rules out gross fabrication but cannot exclude volume overstated by up to about 2×.
Most accounts are dormant, and the incentive-era cohorts the most
Of 203,417 accounts, 41,148 (20.2%) hold at least $1, and 19,406 paid the treasury in the 69-hour window. Accounts opened from October 2025 to February 2026 make up 59% of all accounts, but only 14.8% of them still hold $1 or more, against 26.7% for the March–September 2026 cohorts. That fits incentive-driven sign-ups around the points programme. It does not prove them: we did not run sybil clustering.
New accounts per week, and how many still hold money today
Settlement pools created per week (one per account), split by whether they held at least $1 on 15 Oct 2026. Markers show dated events.
PoolCreated events and balances at block 515,041,730; event dates from docs, chain and pressShow data
A few hundred accounts hold most of the collateral
The 2,034 largest pools, 1% of all pools and 5% of funded ones, hold 85.7% of the $239.1M. Just 150 pools hold half. Revenue is less concentrated but still steep: the 100 largest payers produced about half of the treasury's inflow in the window.
Share held by the largest accounts
Cumulative share of collateral (all pools) and of treasury revenue (paying pools, 69 h), against the number of largest accounts counted. Log scale.
Show data
Section 5
OLP counterparty risk and fund flows
Bottom lineOLP, Variational's own market maker, takes the other side of every trade. Users are net long about $270M; OLP's capital visible on-chain is about $43M. The rest of its capital and all its hedges sit at venues the chain cannot see. Red here means solvency cannot be verified, not that OLP is insolvent.
Visible OLP capital covers about 16% of users' net position
Variational's own API showed traders' longs exceeding their shorts by $272.5M at 13:28 UTC on 15 Oct across 572 markets; OLP carries the opposite position before hedges. At the same moment the Core OLP Vault held $43.09M of USDC and the named hedging wallets about $0.04M. Before hedges, a 10% rise across every market would cost OLP about $27M, equal to 63% of that visible balance. This is a sensitivity, not a loss estimate: OLP's margin inside pools, its venue capital and its hedges cannot be seen. The docs state that if OLP were insolvent, users' accrued profits “would be considered bad debt”.
Users' net long book against OLP capital visible on-chain
USD millions, 15 Oct 2026, 13:28 UTC. The net book is the company's figure; OLP's balances are read from chain. A small net short in “other” markets (−$0.3M) is left out.
Show data
Every sampled flow of user funds went where it should
We traced 30 days of outflows from 200 pools, the 100 largest plus 100 at random: $436.8M in all. 87.0% went to the OLP vault, 12.8% to each pool's own registered trader, 0.13% to the treasury, 0.03% to a Variational fee key, and nothing anywhere else. In the 24 hours to 15 Oct 13:00 UTC the router sent 147,882 transfers ($47.06M), all to pools. The treasury has never sent USDC. Pools emptied during the window could not be sampled.
Where the Core OLP Vault's USDC went in seven days
8–15 Oct 2026. Band width is proportional to USD. Most of the money cycles between user pools and the vault; $10M left for a hedging venue.
Show data
OLP's off-chain capital goes to exchange-linked accounts we cannot attribute
Since May 2025 the vault has sent $25.91M to its EVM hedging wallet, including $10.0M on 14 Oct. Since May 2026 that wallet has bridged $24.31M to Ethereum through Circle's CCTP, passing each tranche on within 21–99 minutes. On Ethereum the money went to three addresses, none labelled as Variational's: $10.93M to an address with the deposit pattern of a Coinbase Prime account (per Etherscan and Open Labels tags); $11.01M to an address first funded with gas by a wallet labelled as a proprietary trading firm; and $2.37M to a 2022 Safe multisig that has received funds from Bybit- and MEXC-labelled wallets. One address, funded by Variational's own hedging wallet, connects all three, so they may be Variational's own venue accounts. Who controls them is not established.
The “vault” is an ordinary key-controlled wallet
The docs call the OLP vault “a smart contract where the capital that powers OLP (USDC) is secured”. The address the docs give for it has no code and signs its own transactions: it is an externally owned account. Its safeguards are therefore operational rather than enforced by code, and whether it is a single key or MPC cannot be seen from the chain.
Section 6
Smart contracts, keys and security
Bottom lineThe code gives Variational's keys complete control of user funds, with no on-chain limit and no exit for users. We saw no misuse: every observed movement was routine. But nothing in the contracts prevents misuse, so a partner is relying on how the keys are run.
Any one of 16 provider keys can move user and OLP funds
Each of 16 externally owned “provider” addresses can, with one signature, make any settlement pool pay any amount up to its balance to that same address, through Oracle.collectFees. It can also move any amount from the vault to any non-zero address through transferFromOLPToPool, because that function never checks that the destination is a pool and the vault has given the Oracle an effectively unlimited approval. No quorum exists anywhere in the deployed code. One further address, 0x8e4d…eabf, owns the factory, administers the Oracle and the router, and has never changed on-chain. We saw collectFees used only for small routine fee sweeps, and every one of 1,649 sampled vault-to-pool transfers landed in a real pool.
Who can move what
On-chain roles and powers at block 515,074,626. Rounded boxes are key-controlled addresses; square boxes are contracts. Every arrow is a single-signature power.
Users cannot withdraw on their own
Every way out of a pool needs a transaction from the contract the factory names as its oracle, which Variational's provider keys drive. If Variational stopped processing, or paused the Oracle, pool funds could not leave until processing resumed. The factory owner can replace the oracle for all pools in one transaction with no delay; this has been used once, at setup. The docs' statements that “uninvolved parties do not have access” and that a user “may withdraw the unused balance” describe how Variational operates, not what the code enforces. In practice withdrawals are processed routinely, and the Oracle has never been paused. The code also allows a few admin mistakes that would be hard to undo; we have shared the detail in question 9 rather than here.
The only contract review came after launch, and the router has none
The core contracts went live on 30 Jan 2025 and cannot be upgraded. The only public smart-contract review, by Spearbit via Cantina, ran from 13 Feb to 9 Mar 2025. Its public report shows only the 2 High findings of 46, both acknowledged and not fixed; one is present in the deployed Oracle but only triggers if the admin changes the Oracle's factory, which has never happened. Zellic's public report covers the web application only. The router, which moves about $40M of vault USDC a day, has unpublished source and no public review. The Immunefi bounty, live since March 2026, caps critical payouts at $100,000, about 0.04% of the money in pools; critical theft bugs in the unlisted contracts still appear to fall under the platform's primacy-of-impact rule. No exploit, hack or outage appears in DefiLlama, SlowMist, DeFiHackLabs, rekt.news or a web sweep.
Review coverage against what was deployed
Reviews (bars) and deployments and programme dates (markers), Nov 2024 – Oct 2026.
Section 7
Legal and regulatory
Bottom lineVariational says it holds no licence, and most of its volume is on traditional-finance underlyings, including single stocks and pre-IPO companies. Its restriction of US users is an IP banner we could see past. No enforcement action exists today. This section is context, not legal advice.
No licence, by its own terms and in the registers we checked
The Terms of Use say the company, the platform and the services “are not registered or licensed by any governmental or regulatory authority or agency”. We found no listing in the four registers we could query with a working control: ESMA's MiCA CASP register, FinCEN's MSB list, the MAS directory and the CFTC's list of designated contract markets. We could not complete checks of the NFA, the FCA, CIMA, the BVI FSC or Panama's SMV.
More than half of volume is on non-crypto underlyings
On 12 Oct, Variational's API attributed 51.7% of $3.06B in 24-hour volume to non-crypto markets, mostly index swaps; one US-100 index swap alone was 25.5%. It lists 94 equity-class perpetuals, including OPENAI and ANTHROPIC, which the docs call pre-IPO perpetuals. By 15 Oct the non-crypto share was 63.3%. The current ToS describes the products as “derivatives transactions (including perpetual futures contracts and contracts for difference)”.
24-hour volume by underlying
Share of $3.06B, 12 Oct 2026, by the asset class Variational's API assigns.
Show data
Restricted countries see a banner, and the trading page still loads
On 15 Oct we loaded omni.variational.io once per country through datacenter proxies. From US, Canadian and Taiwanese addresses it showed “Service access is restricted in your region” but still served the trading page; we did not test whether a wallet can connect or trade. The UK, where retail crypto derivatives have been banned since January 2021, showed no banner. For context, in a 2023 settled order against Opyn the CFTC found that steps “such as blocking users with U.S. internet protocol addresses” were “not sufficient to actually block U.S. users”; that was a finding on Opyn's facts, not a general rule or a ruling on Variational. Two more US touchpoints: a US company holds the domain (section 1), and X's “About this account” panel places the official account in the United States, which X infers from login IP addresses and which therefore says nothing about domicile.
Wide unilateral powers, and a restricted list that changed without a new date
The ToS let the company amend the terms with no notice period, suspend access to deposited funds over source-of-funds concerns, change fees at any time and retract points; liability is capped at the lower of six months' fees or US$100. Taiwan was on the Restricted Persons list until at least April 2026, was dropped in the version dated 20 April 2026, and was added back in mid-September 2026 while the page kept the 20 April date.
Clean sanctions and litigation screens
No court docket, regulator action or warning names any of the five entities. Seven protocol wallets are not listed by TRM, the Chainalysis sanctions oracle, OpenSanctions or OFAC, each checked with a working positive control. Indirect exposure through intermediate wallets was not assessed.
Section 8
Reputation and community
Bottom lineA large audience that looks mostly real, and almost no complaints about the safety of funds. Most of what has been written about Variational was paid for, curated or referral-driven, and nobody independent has examined OLP or custody.
73.9k
X followers; 12% show three or more low-quality signals
71.3k
Discord members, approximate
927
Telegram announcement subscribers
4 of 436
sampled posts were complaints; none about withdrawals
A big audience with few complaints
In a uniform random sample of 1,000 X followers, 12.2% showed three or more low-quality signals and only 1% were created in the last 90 days. Complaints were rare: 4 of 436 sampled posts, about downtime or the TGE and points timeline. Targeted searches found none about withdrawal delays, unfair liquidations or auto-deleveraging. Our search covered X in English; Reddit was not covered.
Most coverage is paid, curated or referral-driven
A Xangle report carries a “paid” flag, all three funding announcements were paid wire releases, and 35% of replies to the official X account in a 13-day sample contain referral codes. The one substantive independent critique, by a Blockworks Research analyst on 25 Sep 2026, called pre-market pricing “overly optimistic”: the ~$1.5B FDV implied by prediction markets is 57× annualised treasury revenue against a 6.3× median for recent perp-exchange launches. The same analyst said they remained optimistic on the project.
Ask VariationalQ24 paid coverage and referral campaigns
Section 9
Token and TGE
Bottom lineInsider vesting is standard. The rest is soft: a third of supply unlocks to an incentive-driven user base at launch, the team and investor split is undisclosed, and the buyback promise has been reworded twice and binds no one.
$VAR allocation
Share of total supply, per the current docs. The token has not launched.
The buyback promise has changed twice in four months
From June 2025 to May 2026 the token page promised buybacks of “a minimum of 30% of protocol revenue”. By 1 Sep 2026 it said allocations “may be modified or discontinued at any time”. Since 24 Sep it says “we intend to use 100% of revenue directed to the treasury”. The base has shifted from protocol revenue to the treasury's share, which is 20% of spreads today and “still being tested and is subject to change”. The points end date has also moved from the end of Q3 to the end of Q4 2026.
Ask VariationalQ15 issuer and supplyQ16 buyback mechanics
Section 10
Questions for Variational
Ranked by how much each answer would move the recommendation. A foundation would normally send these before any term sheet.
- Would change the verdict
- OLP equity. What was OLP's total equity across all venues at a recent dated time, with independent evidence (venue or prime-broker statements, proof of reserves, or an attestation)? What hedge ratio does OLP keep by asset class, and how are index, single-stock, commodity and pre-IPO exposures hedged?
- Key custody. How are the admin address, the 16 provider keys, the treasury and the vault custodied? What approval policies, quorums and limits apply, and who can approve? Why does the treasury also hold the provider role, and why do four never-used provider keys hold full power?
- On-chain limits. Will you put the factory owner and Oracle admin behind a multisig and timelock, cap fee collection per pool, check destinations in
transferFromOLPToPool, shrink the provider set and add a user withdrawal path that does not depend on the operator? On what timeline? - Hedge-venue addresses. Who owns or controls the three Ethereum addresses that received bridged OLP funds, and the address that links them? Are they segregated accounts of a Variational entity or a hedging agent? What did the $10.0M sent on 14 Oct fund?
- Loss backstop. If OLP's losses exceeded its capital, what would happen to users beyond the docs' “bad debt” clause? Is there insurance, a backstop, or a limit on net exposure relative to OLP equity?
- Licensing. Do you hold any licence or registration, or a legal opinion, for leveraged derivatives and CFDs on index, single-stock, commodity and pre-IPO underlyings in Panama, the UK, the EU and other unblocked jurisdictions? Do you have consent or a licence for the OPENAI and ANTHROPIC references?
- Restricted users. Which controls apply beyond the IP banner: wallet screening, KYC at any threshold, VPN detection? What share of volume comes from restricted jurisdictions? What does Variational Research, Inc. do, and where are the teams that run the official channels based?
- Corporate chart. Please provide the chart for all five entities: agents, directors, shareholders, UBOs and intra-group agreements. Which entity owns OLP's capital? Why is the New York biennial statement overdue?
- Needed for a complete file
- Runbooks. What is the procedure for a pause, a compromised key or an operator outage, given that a pause also blocks provider revocation? How are pool outflows, fee collections and the router's settings monitored?
- Router. Will you publish and verify the router's source? What can its operators do with it, and has it been reviewed?
- Audits. Which commit matches the deployed bytecode? Please share the 44 unpublished Cantina findings and their status.
- Control functions. Who owns compliance, AML and sanctions screening, legal and risk, in-house or outsourced? What is total headcount including contractors, and what cover exists if a key person in trading or risk is unavailable?
- Volume. Does reported volume count one side of each trade or both? When did the 10% and 20% treasury shares take effect? Can volume be independently attested?
- TVL. The API's
tvlexcludes the OLP vault although the docs include it. Which is intended? Will you publish the split between OLP and user margin in pools? - Token issuer. Which entity issues $VAR, and what are the Foundation's form and governance? Is there a token legal opinion or a MiCA white-paper plan? What are the reserve's release schedule, the team/investor split and the expected circulating supply at TGE?
- Buybacks. How will buybacks work: venue, schedule, verifiability, decision-maker? Will the treasury share be fixed, and why did the wording move from a 30% floor to discretionary to “intend 100%”?
- Litigation. Please represent whether there is any pending litigation, regulatory inquiry, subpoena or Wells notice for any entity or officer, including any role in the Genesis and DCG matters, and provide a KYT report on indirect sanctions exposure for the vault and hedging wallets.
- Investors as counterparties. Do any investors act as OLP hedge counterparties, liquidity partners or $VAR market makers, and on what terms? Who are the traditional-finance liquidity partners behind the swaps?
- Policy changes. What is the change-notice policy for the ToS and the Restricted Persons list? Why was Taiwan removed and re-added without a change of date?
- Accounts and points. How many accounts were excluded for inorganic behaviour, what share of points does the top 1% hold, and which sybil filters apply at TGE? What explains the account-creation spikes?
- Round terms. What instrument, warrant terms and valuation applied to the strategic round and the Series A, and which entity issued the equity?
- Fee key. What is the fee key that receives about 18% of swept fees, and who benefits from what it pays back into pools?
- Personnel exposure. Have any key personnel held roles in, or given evidence about, sanctioned or mixer-related protocols, and what is your policy on such exposure?
- Paid coverage. What commercial relationships do you have with research outlets, referral campaigns and community managers who publish about Variational?
Appendix
Method and evidence
DINQ built this report with research agents working on a tested source harness. Each source was first measured for success rate, speed and cost. Six agents then researched one area each, and four separate agents tried to refute every headline claim against the raw evidence, the chain and primary sources. None of the headline claims was refuted; 43 corrections were applied after verification. The most important change: we first counted 17 keys that could drain pools; the verifier showed one of them, the router, cannot, so the report says 16.
130
data sources catalogued
92
sources tested; 24 in the working set
1,430
logged API and page calls
≈ $2
metered API cost for the whole research
How to read the labels
VerifiedRead from the chain or verified source code, or confirmed by a primary record (a registry, a filing, a court, an official document) or by two independent reputable sources.
Company claimWhat Variational says or reports, including its own API figures.
LeadAn inference, or a fact from a single secondary source. Worth asking about, not established.
ContradictsA primary record disagrees with what the company says.
Not foundWe searched the named sources and found nothing. It is scoped to those sources.
What we could not check
| Gap | Why | Effect |
|---|---|---|
| Panama, Cayman, BVI and Delaware registries | Personal accounts, CAPTCHAs or paid portals | No officers or owners for any entity |
| NFA, FCA, CIMA, BVI FSC, Panama SMV | Blocked to automated queries | The licence check covers four registers |
| Historical on-chain balances | The public RPC keeps no archive state | TVL is checked at three points, not as a series |
| Trading volume | No per-trade data on-chain; the company runs both the feed and the sweep | Volume and OI stay company figures |
| OLP venue balances and hedges | Off-chain; the vault-performance API refuses every location we tried | Solvency cannot be assessed |
| Key-management policies | Off-chain (Fireblocks MPC, if used) | The admin rating rests on on-chain powers only |
| Router source and audit commits | Unverified source; private repositories | Router behaviour is read from bytecode |
| Trading from a restricted country | Would require connecting a wallet and accepting the ToS | Geo-blocking effectiveness is untested |
| Indirect sanctions exposure | Needs a paid transaction-monitoring service | Screens cover direct listing only |
| Sybil clustering; Reddit and non-English media | Out of scope for this version; missing keys | Incentive-driven activity is suggested, not measured |
Evidence index
Every finding ID above maps to raw files in the project's evidence store. Each file is listed with its source, fetch time and SHA-256 hash in a manifest.