← All Dossiers
Clearwood PBC – Pre-Seed
In Conversation Pre-Seed  ·  StoryHouse Fund II  ·  Enterprise / AI  ·  Portland, OR
Dossier generated 2026-07-09 by /deal-dossier  ·  Deal record: recvmsuT5RcklZTPd  ·  Source: Airtable appjxAR3LPe3fkHOp

One-Liner & Thesis

Clearwood is building white-box monitoring infrastructure for enterprise AI agent deployments, running cheap classifier probes on a model's internal activations instead of an always-on LLM judge to deliver as-good-or-better safety monitoring at 50x to 10,000x lower compute cost. The wedge is a genuine cost/latency arbitrage: today enterprises choose between fast-but-shallow classic NLP monitoring or high-quality-but-expensive LLM-as-judge (2x+ latency, prohibitively costly at scale). Clearwood pulls activations mid-forward-pass and classifies them, working directly on open-weight models and via a proxy model for closed-weight ones. The claimed moat is not the probe technique itself (which comes from the AI-safety research community) but the tooling around it: an automated probe-development engine, a labeled failure-mode library, and simulator environments to re-tune probes per client, because probes are sensitive to distribution shifts and a centralized lab library will not generalize to enterprise deployments (Internal).

Validation is early but real. The founders reproduced and extended Apollo Research's deception-detection work via proxy models and built a code-reuse probe with a worst-case AUC of 0.82 across model families, with signal firing 3,000 to 4,000 tokens before rollout end (early-cutoff potential) (Internal). Commercially they have four active BD conversations — a pharma company automating Phase 3 adverse-event reporting (most advanced, wants regulatory attestation), a consumer kids-athletics app expanding agent scope, plus earlier-stage finance KYC/AML and HR bias-screening use cases — but no signed design partners yet (Internal). This is a first StoryHouse conversation on a ~$1M pre-seed with no lead and no set valuation; the deal is early, Fund II Enterprise, and the founders are the draw (Pomona CEO, prior Yofi/NoFraud exit on the co-founder). SH context: check size and ownership are undefined pending the deck and DD form (Internal).

Investment Score & Recommendation

61/ 100
HOLD

A high-caliber technical team attacking a fast-growing AI-monitoring market with a credible cost-arbitrage wedge — the biggest driver is founder quality plus a real secular tailwind. The biggest drag is stage risk: ~6 weeks in, zero signed design partners, no set valuation, and no lead, so the deal is not yet underwritable.

Momentum: Accelerating Red flags: 3 / 9 Confidence: Medium
Market & TAM8/10
25% weight
Team & Founder7/10
25% weight
Product & Traction5/10
20% weight
Deal Terms & Return5/10
20% weight
VC Syndicate3/10
10% weight

Deal

Funding Round
Pre-Seed
Round Size
~$1.0M
Valuation / Cap
Not set
Lead Investor
None yet (a few angels interested)
SH Check
TBD
Fund
StoryHouse Fund II
Vehicle
Enterprise theme
Runway (raise)
18 months planned

Company Snapshot

Sector
AI  ·  Enterprise AI governance / security
Location
Portland, Oregon
Year Founded
2026
Structure
Public Benefit Corporation (PBC)
Website
clearwood.ai
Status
Stealth / Private

Market Size

$7.1B
AI Agent Observability TAMWeb
by 2035 (from $0.4B in 2025)
33.3%
CAGRWeb
2026–2035
$20B+
AI Security MarketWeb
by 2028 (from ~$5B 2025)
Early
Timing

Clearwood sits at the intersection of two of the fastest-growing enterprise-software categories: AI-agent observability (est. $0.4B in 2025 growing to ~$7.1B by 2035 at a 33.3% CAGR) and AI security (~$5B growing to $20B+ by 2028) (Web). The tailwind is the production deployment of autonomous agents, which forces buyers to fund monitoring, tracing, and governance to contain incident cost and satisfy regulators. The timing thesis is that white-box / mechanistic-interpretability techniques have lived in the AI-safety research community and have almost no enterprise footprint yet, so the category leader has not been minted (Internal).

Competition

PlayerPositioningFunding / StageEdge vs. them
ClearwoodWhite-box activation probes for agent monitoring; open-weight direct, closed-weight via proxy; probe-dev engine + failure-mode libraryPre-seed, ~$1M
GoodfireInterpretability primitives to monitor/steer models in productionSeries B, $150M @ $1.25B (Feb 2026)Clearwood is enterprise-deployment-first with a per-client customization engine vs. research-platform posture (Web)
White CircleReal-time monitoring/control of deployed models~$11M raised (2026)Clearwood's white-box probes claim far lower compute cost than black-box guardrail approaches (Web)
LLM-as-judge platformsAlways-on generative judge / eval (incl. forked "Control Tower")Category incumbents50x–10,000x lower compute; no 2x latency penalty (Internal)
Black-box AI securityInput/output and chain-of-thought monitoring, zero-trust/sandboxingWell-funded (Protect AI, Robust Intelligence)Clearwood reads internal activations, catching signal black-box tools cannot (Internal)

Moat: defensibility is claimed to come from the automated probe-development engine, a labeled failure-mode library, and per-client simulator environments that re-tune probes to each deployment's distribution — not the probe method itself, which is public research; this remains the central unproven question (Internal).

Traction

4
Active BD ConversationsInternal
pharma + consumer furthest along
0
Signed Design PartnersInternal
demos in prep
0.82
Worst-case Probe AUCInternal
code-reuse probe, cross-model
~6 wk
Company AgeInternal
at time of deck

Exit Potential

Strategic M&A
Likely Path
4–7 yr
Time to Liquidity
Active
Sector ConsolidationWeb

AI-security and observability is one of the most acquisitive corners of software right now: 2025 cybersecurity M&A hit ~$96B across 400 deals, and recent comps directly rhyme with Clearwood's category — Palo Alto Networks acquired Protect AI (~$500–700M) and Chronosphere ($3.35B for AI-era observability), Cisco bought Robust Intelligence (~$400M) and Galileo (AI observability, Apr 2026) (Web). Likely acquirers are the platform consolidators (Palo Alto, Cisco, Datadog, ServiceNow) plus model labs needing enterprise-grade monitoring. At a ~$1M pre-seed entry with no set price, the return math is undefined until terms exist, but the exit-comp density in the category is a genuine positive (Web).

Founders

Matthew Levinson
Co-Founder & CEO  ·  Pomona College (CS), UCLA Ph.D. Statistics
Spent 9 years as Senior Director of Machine Learning at Nike, with 20+ years of technical leadership across ML, statistics, and AI systems. A former AI Safety Fellow (Coefficient Giving, Simplex AI Safety) who brought mechanistic-interpretability research into the commercial founding thesis. Pomona grad (Claremont angle); based in Portland, OR. First-time founder.
Michael Klear
Co-Founder  ·  Machine Learning Engineer
Worked under Levinson in Nike's data-science division. Co-founded Yofi (return-fraud / reseller-detection), which was acquired by NoFraud in October 2025 (Yofi backed by Nyca Partners, Point72 Ventures, and eBay Ventures). Took a sabbatical to onboard on mechanistic interpretability alongside Levinson, then reconnected to build Clearwood. First-time founder as CEO but has a founding-team exit (Web/Internal).

Open Questions & Risks

Next Steps

Latest Meeting Notes

2026-07-09 First call Cold inbound intro on white-box agent monitoring

Matthew Levinson reached out cold via LinkedIn Sales Navigator and walked through Clearwood: white-box activation-probe monitoring for enterprise AI agents, positioned as a cheaper, faster alternative to always-on LLM-judge monitoring. Early but technically credible; both founders first-timers.

Source: Meeting Notes recuRPnGl8cm7C9va

Deal Timeline