
GRIDACTA
Brains for Power‑Hungry Machines
AI Co‑Pilot for Data Centers
One agentic platform orchestrating three specialized engines: real‑time physics control (P1), GCLE contracts‑as‑code (P2), and guardrailed market bidding/VPP (P3).
Agentic AI Service-as-Software
Problem
Data centers burn megawatts and money. Energy, thermal, and market ops are split across tools (EMS/ETRM/CLM), contracts are opaque, and control loops aren't coordinated with market and legal constraints—leaving savings, resilience, and revenue on the table.
Solution
An agentic co‑pilot that plans and acts across physics, contracts, and markets. The core orchestrates three engines: (P1) RL/MPC control via a digital twin; (P2) GCLE contracts‑as‑code for enforceable limits and auditability; (P3) bidding/VPP with guardrails for DR/ancillaries, FERC 2222‑ready.
Key features & benefits
Agentic Core
Goal‑seeking planner coordinating engines, tools, and data under policy guardrails.
Digital‑Twin Control (P1)
IC‑LSTM forecasting + RL/MPC for HVAC/chillers/gensets/storage; optimize cost, PUE, and risk.
GCLE Contracts (P2)
Machine‑readable clauses → enforceable limits, automated compliance, audit logs.
Market Bidding/VPP (P3)
Guardrailed bids for DR/ancillary markets; FERC 2222‑ready aggregation.
Deep Integrations
BMS/SCADA, meters, weather, ISO/RTO APIs; unify EMS/ETRM/CLM workflows.
Security & Compliance
Least‑privilege (RBAC), audit logs, SOC 2 Type I path.
Traction / proof
- PhD‑validated methods: IC‑LSTM forecasting, Q‑Learning + MPC control (open‑source exemplars).
- Live site + early‑access pipeline (gridacta.com); customer discovery with DC operators.
- Prototype agentic core coordinating control ↔ contracts ↔ bidding; print‑ready audit logs.
- Pilot partnerships targeted for Oct–Dec 2025; goal: 3–5 pilots within 12–18 months; pre‑seed raise in motion.
Business model
- Service-as-Software (SaaS) platform fee per site or per MW‑month.
- Performance fee on verified savings and market revenue.
- One‑time integration & onboarding; optional managed service.
Ideal customer profile (ICP)
- Hyperscale campuses >50 MW; regional colos 10–50 MW.
- Mission‑critical enterprises with DER (PV, batteries, gensets, thermal storage).
- Energy retailers/aggregators seeking AI‑native VPP partners.
Team
Senior ML/AI researcher (Generac). Built IC‑LSTM forecasting and RL‑MPC control for DER/HVAC; full‑stack data + controls. Focus: agentic systems for energy‑aware data centers.
Now hiring: Commercial Co-Founder (Enterprise/Revenue lead).
- Advisors (non-operating): add names/titles here to mirror the deck.