US imposes 25% tariff on Chinese EVs
CN EV exports ↓ → MX contract manufacturing utilization ↑, US used-EV prices soften
AGENTIX · DECISION INFRASTRUCTURE
Agentix builds WOMO: the enterprise world model.
Every organization holds a model of the world and a model of itself, and the two never touch. WOMO holds both, and the causal path between them. Deployed inside your perimeter, where your data never leaves.
Every system you own is a system of record. It can tell you what happened. WOMO is a system of foresight: a live model of your entity and its world, run forward.
THE QUESTION NOTHING ELSE ANSWERS
Freight costs spike thirty percent. Every system in the building can tell you that happened, and several will give you a sentiment reading on it. None of them can tell you what it does to this company's cash position and covenant headroom over the next eighteen months. That path is the product.
WOMO, LIVE
SIMULATION REPLAY · 194K CANONICAL ENTITIES · 119K VALIDATED CAUSAL EDGES
GLOBAL IMPACT GRAPH · 6 ACTIVE
Live signal propagation
POSITIVE NEGATIVE NEUTRAL
SIGNALS / S347K
EVENTS / 24H128
COUNTRIES HIT47
AVG LATENCY3.2S
SENTIMENT−0.14
LIVE FEEDSTREAMING
SIMULATION REPLAY · 194K ENTITIES · 119K EDGES
OPEC+ announces surprise 1.2M bpd supply cut
Brent ↑ → IN refining margins compress, DE chemicals feedstock costs ↑
Sony shifts image-sensor production to Hanoi
VN electronics FDI ↑ → CN sensor makers cut prices, smartphone BOM deflation
Panama Canal draft restrictions reinstated
Daily transits ↓ → US Gulf LNG shipping costs ↑, Asian spot LNG firms
TSMC accelerates N2 capacity ramp in Arizona
US advanced-node supply ↑ → JP specialty-gas exports ↑, TW industrial power demand growth slows
SIGNALS / S347K
EVENTS / 24H128
COUNTRIES HIT47
AVG LATENCY3.2S
MIRAGE
MIRAGE: THE BENCHMARK FOR FORWARD SIMULATION
Existing benchmarks grade how convincing output looks; MIRAGE measures coherence: whether a system holds one consistent internal world. Protocol, anti-contamination design, and reference baselines published at Womo Labs; frontier results next.
SEE THE METHODOLOGY →Three systems. One model: WOMO.
SUBSTRATE
ONE REPRESENTATION OF YOUR MARKET, NOT A LIBRARY OF DOCUMENTS
119K+ VALIDATED CAUSAL EDGES · 194K+ CANONICAL ENTITIES
LEARNING
THE MODEL DOES NOT GET WORSE AT WHAT IT ALREADY KNEW
CONTINUAL LEARNING, NOT CONTINUAL RETRAINING: THE MODEL UPDATES IN PLACE, WITHOUT FULL RETRAINS.
BACKWARD TRANSFER ≈ 0%
REASONING
SHOWS THE PATH, NOT JUST THE ANSWER
GROUNDED IN DO-CALCULUS
DEPLOYMENT
Sovereign by architecture, not by policy.
Others ask you to send data to their models. We send the model to your data. Model, data, and inference stay within your perimeter, no call home, no shared tenancy, no data egress by construction. Autonomy is governed: every threshold, action boundary, and escalation path is set by you and auditable by you. Deployable in-region, air-gapped where required, with no hyperscaler dependency.
SOVEREIGN VPCZERO DATA EGRESSGOVERNED AUTONOMY12 PATENTS FILED
Who is behind the company.
Agentix is founded by the operator behind Microsoft's first enterprise Copilots, with five prior exits across the founding team, and researchers from NYU's deep learning group.
FIVE PRIOR EXITS ACROSS THE FOUNDING TEAM · 12 PATENTS FILED · MICROSOFT, GOOGLE, AND STANFORD/CMU/NYU/COLUMBIA ROOTS
FOUNDER & CEO
Ex-Microsoft
Led the team behind Microsoft's first enterprise Copilots and AI agents across a $50B portfolio. Member of the Microsoft AI Council. Prior data-center compute infrastructure startup backed by Cisco, Fidelity, and Nokia, exited to NTT Comm. Now building WOMO, the first true Enterprise World Model.
12 PATENTSMICROSOFT AI COUNCILEXIT → NTT COMM
CTO
Stanford · ML platforms
Three decades building machine-learning platforms and infrastructure, including engineering leadership at two industry-defining ML platform companies (Gartner Magic Quadrant Visionary). Stanford MS, computer architecture. Three prior exits.
STANFORD MS3 EXITSML PLATFORMS AT SCALE
FOUNDING TECHNOLOGIST
CMU PhD, ex-Google
Scaled a vertical-AI company from 3 to 200 engineers through a $700M+ acquisition. Led AI engineering at Google across ads forecasting and monetization. CMU PhD in statistical machine learning; 20+ research papers. First investor in the company.
CMU PHDEX-GOOGLE$700M+ EXIT3→200 ENGINEERS
FOUNDING RESEARCHER
NYU deep learning group
World-model researcher from NYU's deep learning group; prior exited founder. Research focus: world models, benchmark design, and model evaluation. 15 papers, ~100 citations.
NYUEXITED FOUNDER15 PAPERSWORLD MODELS
FOUNDING MTS
RL and simulation
Builds the reinforcement-learning training pipelines. Prior simulation-environment work at an a16z-backed AI company, since acquired.
RL TRAINING PIPELINESSIMULATION ENVIRONMENTSFOUNDING EXPERIENCE
CHIEF GROWTH OFFICER
Pre-revenue to unicorn
Scaled an enterprise AI company from pre-revenue to unicorn valuation as growth leader: 2,000+ customers acquired, user base grown from zero to 70K globally in under two years. Built the go-to-market machine, commercial ecosystem, and partner operations from scratch. Columbia Business School.
PRE-REVENUE → UNICORN2,000+ CUSTOMERS0 → 70K USERSCOLUMBIA MBA
INVESTORS AND ADVISORS
Operators, not tourists
Current and former executives and technologists from Microsoft, Google, Amazon, Salesforce, and Snowflake.
MICROSOFTGOOGLEAMAZONSALESFORCESNOWFLAKE
TEAM ALUMNI
INVESTOR & ADVISOR AFFILIATIONS
FULL TEAM BIOS, PATENT FILINGS, AND BENCHMARK METHODOLOGY AVAILABLE IN THE DATA ROOM UNDER NDA.