Rand Protocol

Private computation, hash-based proofs

Programmable and private: a layer-1 chain with a shielded note pool and a RISC-V zkVM, proved by a hash-based STARK with no trusted setup. For venture and strategic investors.

Investor overview·Version 6·October 2026

Records of 2 October 2026: a live public test network (chain 20, cut on v0.6.8 on 1 October), not a main network. Draft for counsel's review before distribution; not an offer or solicitation of any security, token or financial product. Test-network RAND has no monetary value.

The transparency tax

A public ledger taxes everyone with something to protect, and the largest pools of capital decline to pay it

Value lost on transparent ledgers, and the buyers it keeps away. Figures are third-party, each with its publisher and date.

What a public chain publishes

Who reads it

Who stays away: the veto

Orders, before they fill

every swap readable in the mempool

Sandwich and MEV bots

trade ahead of the order, then behind it

Funds and market makers

will not trade a strategy in public

Positions and their triggers

every loan down to its liquidation price

Liquidation hunters

push the price to where the line is

Large borrowers

stay off-chain or under-borrow

Payroll and supplier terms

every salary, every vendor's price

Competitors, counterparties

read margins, runway and terms

Corporate treasurers

settle on bank rails instead

Bids, quotes and votes

every auction bid and every ballot

Last bidders, vote buyers

win on latency; pay against a ballot

Issuers, dealers, DAOs

accept worse prices and bought votes

$60M

Lost to sandwich attacks

by Ethereum traders, 12 months to Oct 2025 (EigenPhi via Cointelegraph, Dec 2025)

$226B

Stablecoin supplier payments

by businesses in 2025, every one public (McKinsey, Artemis, Feb 2026)

$3.2B

Of listed equities on chain

of $152 trillion listed worldwide (rwa.xyz, 2 Oct 2026; WFE, end-2025)

$5.6T

Hedge-fund capital

the buyers of a dark pool, absent from public chains (HFR, Q2 2026)

Our solution

Until now a market could be verifiable or confidential, not both: Rand proves the rule ran without publishing the inputs

Banks, dark pools, ATSs

private, if you trust the operator who sees everything

Permissioned ledgers

Canton: private among approved members

Rand

private by default; programs on private inputs in RISC-V, the instruction set proposed for Ethereum's ZK future; a viewing key for the auditor; a dollar from four chains

Monero, Zcash

private payments; no programs

Railgun

shielded tokens on Ethereum

Aleo

private programs in Leo

Aztec, Midnight

privacy optional

Ethereum, Solana

every input public

Quantus

post-quantum; public

Positions are the project's reading of each system's design as of mid-2026, not independently checked. No uniqueness claim: other private chains are live; Rand's combination is the claim.

Market structure

The operator leaves the value chain

A dark pool is dark because a trusted operator keeps the book; a public chain runs anyone's code but keeps nothing dark. On Rand the matching rule is public code run on private inputs, so the trust an operator sold is replaced by a proof anyone can check.

Compliance

Private to the market, open to the auditor

A viewing key reads and cannot spend. The venue's compliance officer, the accountant or a regulator holds one; competitors do not. Keys scoped to one batch or counterparty are designed, not shipped.

Execution

Nothing in the mempool to trade ahead of

A transfer or call arrives as a proof: amounts, parties and inputs were never there. The block leader still chooses the order; sealed ordering is not built.

How it works

Keys and proofs stay on the user's device; the network checks a proof and never sees the payment

proof

commit

shares a key

bridge

public view

Your device

keys, wallet and prover: the payment is built and proved here

The network

26 validators check the proof, never the payment

The ledger

records only: proof valid, gas paid in RAND

Your auditor

a viewing key: reads your history, cannot spend

Dollars in

USDT or USDC from Ethereum, BNB Chain, Tron or Solana become zUSD

Everyone else

sees a proof and a fee; not who, what or how much

On chain 20 today, 2 October 2026

Private by default. Value is held only as shielded notes: no accounts, balances or transparent mode. Fees, timing and bridge amounts stay public.
Programs in Rust or C run in a RISC-V zkVM; the chain stores a proof and output words, not the inputs.
A dollar from four chains. USDT or USDC locked on Ethereum, BNB Chain, Tron or Solana becomes one shielded dollar, zUSD. Core-team guardians attest (6 of 8); caps are small.
Wallets for six platforms, an explorer and a public RPC. A viewing key reads and cannot spend. All 26 validators are run by the core team.

2.9 MB

Proof per transfer

hash-based STARK, no trusted setup

15 s

To prove a transfer

16 cores of one laptop; earlier build

32 ms

To verify a proof

warm key, one laptop core

Transfers per block

bounded by proof bytes, not compute

Use cases

Twenty industries, one property: the first to ship are few, valuable transfers that today's chain already carries

Payroll and treasury

zUSD; the accountant holds a viewing key

Sealed auctions, RFQ, OTC

a small rule; escrow in a program vault

Private ballots

a vote that cannot be sold

Credentials

prove one fact, keep the rest

Private lending

needs an oracle and blind keepers

Prediction markets

needs batching

Dark pool, tokenized equities

needs batching, custody and securities law

Confidential HFT perp DEX

like Hyperliquid; needs a far larger matching engine

Trade finance, bank analytics

joint checks on books no one opens

Private AI, clinical data

needs production-size proofs

Cases from randprotocol.org/usecases; placement is the founders' judgement from the chain 20 maturity table. Pink: buildable today. Dashed: blocked on work not yet built.

Why this order

Today's chain suits few, valuable transfers. About seven fit a block and a proof takes about 15 s: negligible on a payroll run or an OTC settlement, too slow for a retail swap.
Calls are capped at 16,383 cycles. A payroll split or a sealed-bid rule fits.
Program vaults (v0.6.8) let a program hold and pay out value, which escrow needs; it is the newest part, with the least run time.
The largest pools come last. Equities and lending need an oracle the program can read, batching rules and a custodian; a confidential perp DEX in the style of Hyperliquid needs a matching engine beyond the cycle cap and throughput far above seven transfers a block; AI needs proofs at production scale.

Why now

Why now: the dollars are on chain, and regulators now ask how privacy fits

Supply (left) and transactions a year (right), $ trillions

Pink: supply (DefiLlama); dashed, Standard Chartered ($2T by end-2028) and Citi ($1.9T base, $4T bull by 2030). Grey: transactions, Visa Onchain Analytics adjusted (raw about $33T in 2025); dashed, Citi's ~$100T a year by 2030 at fiat-like velocity, a scenario. Estimates drawn straight from today. Pools behind the use cases: cross-border B2B $31.6T a year (FXC, 2024); hedge funds $5.6T (HFR, Q2 2026).

Aug 2024

NIST

NIST post-quantum standards FIPS 203–205; Rand uses ML-KEM-768 and Dilithium2 (pre-standard ML-DSA).

Dec 2024

MiCA fully applies Licensed stablecoins and service providers across the EU.

Mar 2025

OFAC

Tornado Cash delisted Sanctions on the privacy software lifted.

Jul 2025

US law

GENIUS Act A federal regime for payment stablecoins.

Dec 2025

SEC

Privacy on the agenda Roundtable on financial surveillance and privacy, 15 Dec.

Dec 2025

DTC

Tokenized securities No-action relief for DTC to tokenize custodied assets; rollout from H2 2026.

Sep 2026

Senate

CLARITY Act stalls Cloture failed 15 Sep: market-structure law still open.

Jul 2027

AMLR

Anonymity-enhancing coins barred At service providers; disclosure by design is the test.

2030–35

NIST

RSA and ECC retired Deprecated after 2030, disallowed after 2035 (IR 8547, draft).

Why us

Why us: private by default, programs in Rust, post-quantum and no trusted setup; Aleo, Railgun and Quantus each lack at least two

	Status	Private by default	Programs, and how you write them	Proofs and setup	Post-quantum
Rand	Test network (chain 20)	Yes, always	Rust or C on RISC-V; EVM and Solana bytecode translators (partial)	Hash-based STARK, no setup; MB proofs	Notes, signatures, proofs
Aleo	Mainnet 2024	Records private; much state public	Leo, its own language, or Aleo instructions	Pairing-based SNARK; universal setup	No
Railgun	Live on Ethereum, Arbitrum, Polygon and BSC	Opt-in, per token	No private programs; calls public DeFi	Groth16 SNARKs; trusted setup	No
Quantus	Mainnet Sep 2026	Optional privacy layer	None	Not a programmable chain	Yes, ML-DSA signatures
Aztec, Midnight	L2; mainnet Mar 2026	Optional, per function	Noir and Compact, their own languages	SNARKs with a universal setup	No
Zcash, Monero	Mainnets since 2016, 2014	Zcash opt-in; Monero yes	None	Halo 2; Bulletproofs	No

Why developers choose Rand

No new language, and the ZK industry's instruction set

Write Rust or C, build with the standard compiler, deploy the binary; the chain proves it ran. RISC-V is what the leading zkVMs prove (RISC Zero, SP1) and what Vitalik Buterin proposed for Ethereum's execution layer in April 2025, so tools and talent carry over. Aleo, Aztec and Midnight need their own languages. Cost: megabyte proofs.

Where others are ahead

Users, liquidity and smaller proofs

Aleo, Zcash and Monero are mainnets with users; Railgun sits where Ethereum's liquidity already is; their proofs are kilobytes where Rand's are megabytes. Rand's claim is the combination. Others' positions: our reading of their public documents, mid-2026, not independently checked.

Progress and plan

A year of research, then a live network in three weeks; the seed buys the steps to a main network

Research papers

21 December 2025, then revised in git

Node releases

v0.1 (16 Sep) to v0.7.1 (2 Oct 2026)

Test chains

chains 14 to 20, 20 Sep to 1 Oct

2,172

Commits

233k lines of Rust, 2 Oct 2026

Research, Dec 2025 to Sep 2026

Dec 2025

Papers

Four papers Protocol, post-quantum security, stablecoin design, investor summary.

Jan 2026

Proto

First prototype An SVM node with confidential tokens; the RandScan explorer.

Jun 2026

Design

EVM execution Execution moved from SVM to EVM; private gas settlement.

Aug 2026

Design

Private computation Confidential programs; shielded bridged dollars.

Sep 2026

Rebuild

RISC-V zkVM A hash-based STARK, no trusted setup; the node from 9 Sep.

Shipped, Sep to Oct 2026

16 Sep

v0.1

First release A week after the first commit.

20 Sep

chain 14

Test network First public test chain.

28 Sep

v0.6

Delegated proving A prover that can read, never spend.

29 Sep

chain 17

Split authorisation The spend key's proof stays on the device.

1 Oct

chain 20

Programs with state Program vaults, a consensus fix, a bridge round trip.

2 Oct

v0.7.1

Signed releases Built by CI, attested and signed.

Next, months from the seed close

0–3

Operate

Keys leave the team Three operators by 1 Nov (target).

3–6

Scale

Founding round Gated on the test-profile exit.

6–9

Secure

Independent audit Slashing on; scoped viewing keys; a custodian.

9–12

Open

Outside validators An open validator programme; larger programs.

12–16

Sell

Validator sale To operators who ran a stage.

16–18

Launch

Main network Genesis frozen; no calendar date published.

Valuation

The seed is $0.03 a token, $30 million on RAND's 1 billion maximum supply: a third of the next round's price

Round prices from the seed materials of 29 September 2026 (tokenomics Draft 2). Proposed, not set in code: test-network RAND has no monetary value.

$30M

Seed valuation

$0.030 a token, open now

$100M

Founding round

$0.100; gated on a measurement

$150M

Public auction

$0.150, a week after genesis

16–18

Months to genesis

lock-ups count from genesis

Round	Price	Valuation	Supply
Seed, open now	$0.030	$30M	11%
Founding	$0.100	$100M	9%
Validator and prover	$0.120	$120M	4%
Strategic	$0.150	$150M	2%
Public auction	$0.150	$150M	4%

Why $30 million

Our own number for a two-founder, pre-revenue, pre-mainnet project: 1.5× Solana's $20M seed, and a fraction of what privacy projects raised at in 2025–26: Zama $550M, Aztec about $557M, Octra and Arcium $200M.

Where value accrues to RAND

Every fee is paid in RAND, whatever asset moves; validators stake at least 1,000 RAND each.
The bridge earns 10 bps of zUSD in and out: at 1% of 2025's $226B in B2B stablecoin payments, about $2.3M a year.

No return is implied or assured, and a token price can fall to zero. The company has no revenue today. Valuations are on the 1 billion maximum supply; the comparables in full and downside scenarios are for counsel's review.

The round

The round: $3.3 million for 11% of RAND's 1 billion maximum supply, sold as SAFTs; about 18 months of runway

Seed terms as set out on 29 September 2026; counsel is to review them before this deck is distributed. Nothing here is an offer.

Use of funds	Share	What it pays for
Engineering	45%	Four hires: node and standing review, bridge and infrastructure, wallet and product, proof systems
Independent audit	15%	$350–750K programme: circuits, recursion VM, bridge, consensus, wallets
Custody decentralisation	12%	Independent guardian and validator operators, RPC diversity, timelocked multisigs
Legal and compliance	10%	Counsel's opinions per role, the sale structure, custodian onboarding
Test network and go-to-market	10%	Operator rewards, design partners, viewing-key tooling
Reserve	8%	Contingency

Burn about $190K a month on the base plan: the seed alone covers about 18 months.

The ask

$3.3M across at least four investors (maximum ticket $0.9M under the 3% cap). Instrument: a SAFT (simple agreement for future tokens) for 110M RAND, 11% of a maximum supply of 1 billion RAND, at $0.03 a token ($30M fully diluted), issued by Randprotocol Development LLC (Wyoming) and delivered at genesis. Lock-up: 12-month cliff after genesis, then 24 months linear. Rights: pro rata in the founding round, information rights, an observer.

Milestones before the founding round: the first independent audit under way; guardians and validators at operators other than the core team (three by 1 November 2026, a target); design partners signed. Already done: the CON-4 consensus fix, live on chain 20 in v0.6.8.
Main network: genesis 16–18 months after the seed closes, gated on measurements rather than dates; no calendar date is published.

What the SAFT buys: the right to 110M RAND at genesis, not equity in the LLC; supply is capped at 1 billion. The company has no revenue today.
Diligence: the papers, the live test network, the measured figures with their sources, the internal review reports, the tokenomics model; a term sheet within four weeks of a first meeting.

Team

The team: two founders who designed the protocol and are building it themselves, from the papers to the wallets

Biographies as on randprotocol.org/team, 6 October 2026, from the founders' LinkedIn profiles and Google Scholar.

Founder · protocol and code

Dendi Suhubdy

Machine-learning research at Mila, Université de Montréal, with Yoshua Bengio: the MILABot chatbot, ChatPainter and Fraternal Dropout, and a contributor to Theano (671 citations). Brings market experience as co-founder of Bitwyre, a crypto spot and derivatives exchange whose trading engine and architecture he built (since 2019), and of Systematic Trading Group, a proprietary trading firm (2025). Co-founded Grokvideo, a text-to-video search engine in Montreal (2018–22, since acquired), sold two further companies in 2025, and has invested in KAST and Parts Order. At Rand he writes the code, every commit to date. LinkedIn · Google Scholar

Founder · security and design

Anish Mohammed

Cryptographer and security researcher with 20+ years of experience in cryptography, distributed systems and blockchain. An early member of the Ethereum Swarm team and reviewer of the Swarm Orange Paper, he later co-founded Panther Protocol, serving as CTO and Chief Scientist, and was principal architect at Boson Protocol. Anish has advised Ripple and Ocean Protocol and has published and lectured extensively on zero-knowledge proofs, privacy, blockchain security and cryptographic protocols, including research in the IACR Cryptology ePrint Archive and academic work on privacy and security. LinkedIn

Hiring with the seed

Four roles

A senior node engineer and standing reviewer; bridge and infrastructure; wallet and product; proof systems. Key-person risk is one engineer today. The first independent audit is a condition of the founding round.

Entity

Randprotocol Development LLC

Wyoming; the seed's issuer. No foundation is formed yet, so the LLC holds every foundation role. Cap table: founders 60:40; no other SAFTs, grants or options. Rand's IP to be assigned to it before close.

In print, open source

Five papers; GPL-3.0-only

randprotocol.org/papers. The full node (zkVM and recursion VM), the wallets, the bridge and the circuits are on github.com/randprotocol; the wallet core is WebAssembly built from the open-source clients.

Thank you.

randprotocol.org

github.com/randprotocol

Dendi Suhubdy · Anish Mohammed

Randprotocol Development LLC, Wyoming

Not an offer or solicitation of any security, token or financial product. Test-network RAND has no monetary value. Draft for counsel's review before distribution.