One model for each stock in your portfolio.
CNW trains a dedicated AI model for each equity and delivers the signal without changing how you operate. More alpha. No operational cost.
NASDAQ-6 active portfolio.
NASDAQ-6 · Igual peso 16.7% · r_atr_ha_guard · 6/6 modelos activos · May 2024 – May 2026
| Strategy | Total return | Volatility | Sharpe | Max. drawdown | Calmar |
|---|---|---|---|---|---|
| Optimized | +28.44% | 33.00% | 0.862 | −30.09% | 0.945 |
| Buy & Hold | +15.10% | 37.62% | 0.401 | −38.62% | 0.391 |
| Difference | +13.34pp | −4.62pp | +0.461 | +8.53pp | +0.554 |
Integrates into your process. Doesn't replace it.
CNW acts as a signal layer on top of your existing process. You receive a daily signal via API for each position, with a confidence score and suggested sizing. Your execution doesn't change.
You define your universe
You provide your equity universe. CNW maps each position to a dedicated model instance. The S&P 500 set is available immediately.
We train one model per asset
Each security has its own model trained exclusively on its own historical data: price microstructure, fundamentals, macro context and derived signals.
You receive the daily signal
Post-market delivery via REST API. Directional (long / flat / short) with confidence score and suggested sizing. 67% of signals are automatically filtered before reaching you.
Validated out-of-sample results.
Three independent lines of research. Walk-forward throughout: no look-ahead bias.
View all papers →A global model learns what is true on average.
What is true on average is precisely wrong for each individual stock. CNW solves this with infrastructure that until recently didn't exist.
The signal layer added 2.41× Sharpe while reducing maximum drawdown by nearly a third — without a single change in how we managed the book.
Signal before it reaches the price.
We aggregate and normalize 28 alternative data sources in real time — options flow, institutional sentiment, market microstructure and macro regime — to extract signal before it is reflected in the price.
View Alt Data product →Frequently asked questions.
Everything you need to know before requesting access.
A daily signal per asset: directional (long / flat / short) with a 0–1 confidence score and suggested sizing as a percentage of the book. Delivered post-market via REST API, ready before the next day's open.
No. CNW acts as a signal layer on top of your existing process. Your OMS, broker and execution don't change. You simply receive an additional signal that you can incorporate with whatever weight you decide.
Between 2 and 4 weeks from signing. Includes asset universe setup, API integration testing and a paper trading validation period before go-live.
We don't set a fixed AUM threshold. Access is evaluated case by case: hedge funds, prop desks, family offices and managers with equity universes are the typical profiles.
All published results use strict walk-forward: the model only sees data prior to the prediction point at each moment. We use point-in-time constituent lists for the S&P 500, eliminating survivorship bias.
The full S&P 500 set is available immediately. We also cover NASDAQ-100, Russell 1000 and custom universes from 50 positions. Models are trained specifically for each asset in the requested universe.
Fill in the access request form. We respond within 48h with an initial proposal and a diagnostic call to assess the fit with your investment process.