Wattness

The verifiable intelligence layer for the energy machine economy.

Learn More

We build the verifiable intelligence layer that lets flexible energy assets act with precision in wholesale power markets.

Physics-grounded agents. Deterministic strategies. Auditable by design.

Learn MoreOur PlatformWattness is a member of NVIDIA Inception.

The engine

One corridor, five layers, and the record they leave.

The physical edge, the agents that read it, the proofs that make their work checkable, the interface that turns all of it into dispatch a control room already knows how to accept, and the coordinated region that shape becomes once it is repeated.

Layer 1.0 · Physics Core
A deterministic model evaluated at the asset: thermal state, degradation, state of charge. Measured directly from physical equipment rather than estimated from seasonal averages.
Layer 2.0 · Agent-as-a-Service
The physics core packaged as an edge service, exporting standard schemas that an energy manager and a Qualified Scheduling Entity (QSE) can ingest without workflow changes.
Layer 3.0 · Verifiable Witness
Reconstructible proof that an edge optimisation ran as declared, anchored off-chain. No power market clears on a ledger; verification sits alongside dispatch.
Layer 4.0 · ISO Interface
Agent decisions leave as native offer curves, telemetry, and settlement files that a control room already accepts. Immediate deployment without tariff restructuring.
Layer 5.0 · Unified Machine Economy
Every asset executes its own physics and proves its own state. Coordination forms as an open mesh between edge nodes rather than a single central controller.
BESS Facility
· Deterministic Physics Core Active · Cell Degradation Tracked · Thermal Margin Monitored in Real Time
AI Data Center
· Edge Agent Active · Deferrable Workloads Isolated · Dynamic Interconnection Ramp Ready
Substation POI
· Dynamic Line Rating Active · Local Constraint Evaluated · Verifiable Dispatch Standby
Wattness Verifiable Decisions
What the witness field produced: an optimisation with a proof attached, arriving as something a control room can already accept.
Grid 2 GCAP Access Classes
Tiered access, one band per class. A curtailable megawatt and a firm one are not the same product, and an interface that cannot say which is which is not an interface.
Grid 1 Firm Baseline
The interconnection's own firm capacity, as the grid already states it. Nothing here is ours.
Interconnection
The point of interconnection. Everything standing on this site is scheduled against what this bus can physically accept under real-time thermal limits.
Storage
The only asset that operates on both sides of the market. Deciding when to charge, hold, or discharge requires accounting for electrochemical wear.
Flexible Compute
A workload where execution can pause or shift. Deferrability transforms a raw electrical load into a responsive balancing resource.
Delivery, Measured
Delivered performance scored against declared capacity. A continuous measurement record disciplines declarations without introducing penalties.
No Central Bottleneck
Edge intelligence scales without a singular operator. Assets optimize independently against observable physics while remaining auditable.

Our Businesses

We build deterministic intelligence to bridge real-time grid physics with wholesale power markets.

Physics-grounded by design, our architecture operates natively within existing ISO tariffs while unlocking the next generation of flexible, connect-and-manage infrastructure.

Storage Intelligence

1.0

Storage Intelligence

We develop deterministic, degradation-aware dispatch models for utility-scale BESS and hybrid assets. By factoring real-time state of charge, electrochemical fade, and thermal headroom directly into marginal cost curves, we co-optimize participation across energy arbitrage and ancillary services without eroding asset warranties or long-term operational life.

Storage Intelligence

Flexible Load Orchestration

2.0

Flexible Load Orchestration

We enable hyperscale data centers, AI training clusters, and industrial facilities to participate as dynamic grid resources. By mapping internal workload deferrability and thermal inertia to real-time nodal congestion signals, we allow large loads to secure rapid, non-firm grid access under connect-and-manage tariffs without risking operational downtime.

Flexible Load Orchestration

Verifiable Market Dispatch

3.0

Verifiable Market Dispatch

We interface autonomous edge decisions with wholesale ISO clearing engines through auditable, sub-minute execution. By pairing selective cryptographic proofs with standard QSE workflows, we deliver native nodal offer curves and transparent delivery scoring—requiring zero regulatory overhauls or changes to existing control-room settlement practices.

Verifiable Market Dispatch

Coordinating the Machine Economy

Our Ambition

Our ambition is to build the verifiable intelligence layer that coordinates the next era of energy and compute—enabling high-density loads, autonomous storage, and dynamic grid assets to unlock non-firm capacity at scale without compromising physical reliability.

Featured Research & Insights

Wattness Response to Grid 2.0 RFC #1

Every Place That Ever Was

The Grid Is an Exchange: A Crypto Trader's Guide to US Wholesale Power Markets