PROJECT RESEARCH DOSSIER — Nous Research / Psyche Network

Pre-TGE, no token issued, no public sale/ICO — primary sources only Stage: Pre-TGE | Raise: Private only (seed + Series A, no public round) | Prepared: July 7, 2026

1️⃣ What it builds and why it matters

Nous Research (founded 2022, SF) builds open-source LLMs (Hermes family) and Psyche — infra that coordinates large-scale neural network training across distributed, heterogeneous GPUs using Solana smart contracts as the coordination layer. Core bet: DisTrO/DeMo optimizers compress inter-node data exchange by orders of magnitude, making internet-bandwidth training competitive with centralized data centers. Not just narrative — two production runs shipped as open weights: Consilience-40B (~20T tokens, reported as the largest distributed pre-training run at the time) and Hermes 4.3-36B (144k tokens/sec across 24 nodes).

2️⃣ Tokenomics — FDV vs circulating, vesting, allocations

⚠️ No token has launched. No allocation table or vesting schedule has been published. That absence is itself the key tokenomics finding. What IS verifiable:

Seed: ~$20M total ($5.2M Jan 2024 from Distributed Global/OSS Capital + ~$15M follow-on) — source: The Block Series A: $50M, led almost entirely by Paradigm, closed April 2025 — source: The Block, Fortune/Yahoo Implied FDV: $1B token valuation per the SAFT terms signed by Series A investors Public sale/ICO: none conducted or announced As of the raise, team hadn't decided between a proprietary token or using Solana's native SOL — Fortune interview, Karan Malhotra, Apr 2025 Vesting/cliffs/emissions: not disclosed anywhere

3️⃣ Unlock & dilution risk (modeled, not disclosed)

Since Nous hasn't published real numbers, here's an illustrative scenario built on 2025-26 SAFT-to-token industry norms (NOT project-specific fact):

Team/advisors (~15-25%): ~12mo cliff, 36-48mo linear vest → High risk, classic post-TGE overhang Investors (~10-20%, likely near top given $65M raised): ~6-12mo cliff, 18-36mo vest → High risk, Paradigm's stake alone is a large single-holder unlock event Ecosystem/community (~20-30%): ongoing usage-tied emissions → Medium risk Public allocation: unknown, no sale format announced Base rate: cliff-heavy (vs linear) unlocks correlate with ~25% avg price drop on unlock day (Bitget Academy TGE guide, 2026).

4️⃣ Team verification — verifiable vs marketing

✅ Verifiable: Founding team named consistently across independent press (Fortune, The Block) — CEO Jeffrey Quesnelle, Karan Malhotra, "Teknium," Shivani Mitra. CEO's GitHub handle (jquesnelle) is the top committer on PsycheFoundation/psyche (Apache-2.0) — rare case where the named founder is the primary engineer. Research lineage (Hermes, DisTrO/DeMo) predates funding and is public on GitHub/HF. ⚠️ Marketing/unverified here: live node counts and "how decentralized" claims aren't independently auditable (dashboard is JS-rendered per third-party trackers); download/star counts weren't re-verified by me — spot-check via HF/GitHub API before citing.

5️⃣ Key risks

Technical: DisTrO's sub-linear communication-growth claim only validated on 2 disclosed runs, not at frontier scale. Execution: Docs reportedly avoid the word "mainnet," framing current phase as interim — no committed graduation date. Governance: No DAO, no on-chain governance. 4-person founding team controls all decisions; no published transition plan. Incentive/tokenomics: ~$65M in insider capital priced at $1B FDV before any public holder — largest unresolved risk in this dossier. Legal/regulatory: Series A structured as a SAFT (has drawn SEC scrutiny in prior cycles); single-jurisdiction US entity.

6️⃣ Disclosure

No token exists, so no position is possible. No consulting/advisory/compensation relationship with Nous Research, Paradigm, or any named investor. Prepared for research/educational purposes — not investment advice. Sources: nousresearch.com (Psyche architecture post, "Next Phase of Psyche"), github.com/PsycheFoundation/psyche, The Block (4/25/25), Fortune via Yahoo Finance (4/25/25), Bitget Academy TGE guide 2026 (benchmark only). Constraint check: No public raise, no active ICO/TGE as of writing — funding is private-only (seed + Series A/SAFT). Full PDF version attached below. Pushback welcome — especially on whether the illustrative dilution model in §3 is too generous or too harsh given the SAFT structure.