Every cell, perfect yield.

Anodela is the autonomous operations layer for battery cell manufacturing — it perceives every metre of electrode and every cell, controls coating, welding and formation inside your qualified window, and hits target yield before the run.

Shadow → assist → control
Edge-native
On-prem or air-gapped

Anodela line console — Electrode line 3

Line overview
Coating control
Inspection
Assembly & weld
Formation
Yield & takt
Cell genealogy
Audit log

94.2%First-pass yield

1.8%Scrap rate

0.92 sTakt

First-pass yield · 12 wk63% → 94.2%

Coat weight web edge drift +1.4 mg/cm²corrected

Metal particle signature on anode roll A-2291quarantined

Tab weld nugget depth nominal, 4,812 cellsin spec

Formation channel 12 re-sequenced, −41 minapplied

Trusted by cell makers, EV OEMs and ESS suppliers

Northvolt-class OEMVolta CellsAmperex ESSHelion MotorsKestrel EnergyAnode WorksPrisma CellsSilica PowerTerra Grid StorageMeridian EVCathode LabsAurora Gigafactory

Yield is the bottleneck. We move it.

Results measured against each site's own pre-Anodela baseline during shadow mode, then audited quarterly. [PLACEHOLDER — design-partner cohort, 2026.]

+14.6 ptsFirst-pass yield lift across design-partner lines

38%Reduction in electrode and cell scrap

17%Less formation time and energy per cell

<250 msInspection-to-decision latency at the factory edge

New lines sit stranded at low yield for months — sometimes years

A gigafactory is billions of dollars of capex that only pays back when the line holds yield. Today that depends on fixed recipes, spot inspection and a severe global shortage of experienced battery process engineers.

A micron of coat-weight drift becomes tonnes of scrap

Slurry rheology, web tension, drying profile and calender load interact constantly. By the time an offline coat-weight check flags the drift, kilometres of electrode are already outside the porosity window.

Continuous coat weight, thickness and porosity estimation across the full web
Closed-loop correction of pump, gap, tension and drying zones
Automatic quarantine and genealogy tagging of affected rolls

Coating deviation feed

Zone 4 drying outlet temp −3.1 °Ccorrected

Web tension edge wave risk risingwatch

Coat weight CV 0.61% → 0.38%improved

One metal particle can become a thermal-runaway recall

Burrs from slitting, wear debris from calendering and airborne particles in the dry room cause internal shorts that pass end-of-line test and fail in the field years later.

Metal-particle and burr signatures fused from vision and inline X-ray/CT
Predicted internal-short risk scored per cell, not per sample
Automatic quarantine, rework routing and reliability evidence

Contamination watch

Slitter B burr height trend +9 µmflagged

Dry room 2 particle count within limitnominal

CT fusion 3 internal anomalies in 21,400 cellsquarantined

Formation ties up floor space, energy and working capital

Formation and aging can consume a quarter of cell cost and weeks of inventory. Fixed protocols treat every cell as the worst cell on the line.

Per-cell formation protocol inside your qualified limits
Channel allocation and scheduling optimized every minute
Grading and binning accuracy improved from the full curve, not endpoints

Formation optimizer

Protocol adaptive per-cell, 41 min savedapplied

Channel use 96.4% utilisationoptimal

Grade accuracy +2.7 pts vs endpoint modelimproved

A closed perceive → decide → act loop, running at the factory edge

Anodela is a system of action, not another dashboard. It runs on GPU edge hardware inside your plant, with cloud used only for training and fleet governance — or not at all in on-prem mode.

Perceive
Vision, inline X-ray/CT, coat-weight gauges, weld monitors, formation curves and MES tags fuse into one live model of the electrode, the jelly roll and the cell.
Plan
The orchestrator plans the coat, calender, stack/wind, weld, fill, seal and formation moves inside your qualified process window and the twin's validated envelope.
Act
Bounded setpoints are written back to coating, calendering, welding, fill, formation and robot controllers — or proposed for engineer approval in assist mode.
Verify
Every action is scored against downstream yield, defect, contamination and grading outcomes, logged immutably, and fed back into training.

Seven agents and a twin, under one factory orchestrator

Each agent owns a step of the line. Together they close the loop from slurry to graded cell.

Coat & Calender agent

Controls slurry mixing, slot-die coating, drying and calendering for loading, coat weight, thickness and porosity uniformity across the full web width.

Defect & Inspect agent

Senses and predicts coating defects, wrinkles, pinholes, streaks, burrs, metal particles, weld quality and internal defects from vision and inline X-ray/CT.

Assemble & Weld agent

Controls stacking and winding, tab and terminal welding, electrolyte fill and sealing to sub-millimetre tolerance with void and misalignment prediction.

Formation & Grade agent

Optimizes formation cycling, aging, degassing, end-of-line test and grading/binning to cut hours, energy and working capital per cell.

Yield & Takt agent

Balances the moving line, optimizes first-pass yield, scrap recovery, WIP routing and takt when a bottleneck or quality risk appears.

Robot & Handling agent

Drives robots, AGVs and dry-room material handling for electrodes, jelly rolls and cells with fail-safe stop behaviour.

Everything the line needs, in one edge-native platform

Perception, control, simulation and optimization built specifically for cell manufacturing — not a generic industrial AI toolkit.

As-built cell digital twin

Simulate coating, drying, calendering, stacking, welding, fill and formation before you touch production. Thousands of recipe and takt scenarios per line, validated against as-built data, so the run starts at the yield you targeted.

Hit-yield-before-run validation gates every autonomy envelope
Commissioning support for new lines and new chemistries

Factory edge runtime

8–24 cameras per line plus X-ray/CT fusion on Jetson and IGX, with OTA updates and versioned rollbacks.

Grounded outputs

Every recommendation cites the process data, spec or standard it came from. No unsourced answers on a safety-critical line.

Graduated autonomy

Shadow → advisory → bounded control. Each step is unlocked by measured accuracy and twin validation, never by a toggle.

Compounding cell-yield data moat

Every supervised correction, approval, rejection and executed change is preserved with its downstream yield and quality result. The models get better on your chemistry, your format and your tolerances — while your recipes stay isolated in your tenant.

Trust before autonomy. Always.

Shadow mode proves accuracy against your engineers. Assist mode proves ROI. Only then does Anodela take bounded control — inside envelopes your quality team approved, with a stop condition on every action.

Stage 1 · Shadow
Prove accuracy
Anodela watches the line and predicts what your engineers will do, scored against them for an audited baseline. It writes nothing back.
Stage 2 · Assist
Prove ROI
Proposed setpoints go to an engineer for approval. Every accepted and rejected change is linked to its downstream yield.
Stage 3 · Bounded control
Act inside the envelope
Only then does Anodela act, inside envelopes your quality team approved, with a stop condition on every action.

What plant, process and quality leaders say

“We were stranded at 63% first-pass yield for five months. Anodela ran in shadow for six weeks, then took bounded control of coat weight and drying. We crossed 91% in the next quarter.”

Marta VogelPlant Director, Volta Cells

“It reads our CT and vision streams the way a great process engineer does — except it does it on every metre of web, on every shift, and it writes the correction back before the defect becomes scrap.”

Daniel OkaforHead of Process Engineering, Amperex ESS

“The audit trail was what got quality to yes. Every proposal, approval and executed setpoint is linked to cell genealogy, so our IATF evidence got stronger, not weaker.”

Sung-Ho ParkQuality Director, Prisma Cells

Connected to the systems your line already runs on

Pre-built OT and API connectors read process, vision, X-ray/CT, formation and MES data — and write bounded setpoints back. Nothing rips and replaces; Anodela runs alongside the equipment you have qualified.

Coating & calendering

Dürr Megtec
Hirano
Toray
Manz
Bühler
Custom slot-die OT

Assembly & welding

Wuxi Lead
Hymson
Hitachi
Trumpf laser weld
IPG
Amada

Inline inspection

Cognex
Keyence
ISRA Vision
Nikon CT
Yxlon CT
PEMTRON

Formation & test

Wonik PNE
PNE Solution
Chroma
Digatron
Arbin
Neware

MES, quality & ERP

Siemens Opcenter
Rockwell FactoryTalk
iTAC
SAP
Ignition
AVEVA PI

Edge, robots & identity

NVIDIA Jetson / IGX
Isaac & Omniverse
Fanuc / KUKA
OPC UA & MQTT
Okta / Entra ID
K3s on-site

See all connectors

Built for the assurance bar of battery manufacturing

Cell manufacturing is safety-critical and IP-dense. Anodela is designed so quality, IT/OT and EHS approvers can say yes.

SOC 2 & ISO 27001

SOC 2 Type II program with encryption in transit and at rest, SSO/SAML, SCIM and RBAC down to the line.

IP & recipe protection

Chemistry, coat-weight and formation recipes stay in your tenant. Optional fully on-prem deployment with no egress.

Assurance-grade audit trail

Every perception, proposal, approval and executed setpoint change is immutably logged for IATF 16949 and IEC 62660 evidence.

Data residency

EU, US, Korea, Japan and China-domestic regions, or air-gapped factory edge for sovereign programs.

Point tools measure. Anodela acts.

Coating machines run fixed recipes, inspection systems measure, MES records — but nothing closes the loop across the whole cell process.

Dimension	Incumbent tools	Anodela
Scope	A single machine, inspection station or MES module	The full coat → calender → assemble → form → grade loop
Loop	Fixed recipe with spot inspection	Closed perceive-decide-act at the factory edge
Data	Local, per-machine, discarded	Compounding cross-factory yield and quality dataset
Integration	Read-only or shallow	Deep, bounded write-back into coating, weld and formation
Outcome	Assists a human	Performs the work; engineers supervise exceptions

Priced on the line, the factory, and the outcome

Land on one high-ROI workflow, expand line by line and site by site. Annual prepay saves 18%; the highest-value workflows can be priced on measured yield, scrap and formation-cost outcomes.

Line

$18,000 per production line or coating/assembly cell, per month

One wedge workflow: coating/calendering control, assembly & weld, inspection or formation
Factory-edge runtime for a single line
Connectors to your coating, inspection, formation or MES stack
Review console, alerting and assurance-grade audit log
Shadow and assist modes with measured accuracy reporting

Start a paid pilot

Factory

$120,000 per gigafactory, per month

Whole-factory scope: coat & calender, defect & contamination, assemble/weld/fill, formation & grade
Yield and takt balancing plus robotic dry-room handling
As-built cell digital twin — hit yield before the run
Graduated autonomy with bounded control envelopes
SSO/SAML, RBAC, SOC 2 reporting and quarterly ROI reviews

Talk to sales

Enterprise

Custom typical land: $1M–$15M ACV

Multi-site rollout across every plant and chemistry
Custom chemistry, format and tolerance models
Managed factory edge fleet with OTA and rollback
On-prem or air-gapped deployment and data residency
Uptime SLAs, dedicated support and qualification assistance

Contact sales

Questions from plant, process and quality teams

Design partners typically run six weeks of shadow mode to establish an audited baseline, then four to eight weeks in assist mode. Measurable first-pass-yield and scrap movement usually appears within the first quarter, before any autonomous control is enabled.

No. Anodela runs alongside qualified equipment and integrates through existing OT and API surfaces — OPC UA, MQTT, vendor SDKs, MES and historian connectors. Nothing is ripped out.

You do. Chemistry, coat-weight and formation recipes stay inside your tenant, and on-prem or air-gapped deployment is available. Model improvements can be shared across the fleet only under federated, IP-preserving terms you approve.

Every action is bounded by the qualified process window, gated by twin validation, reversible, and paired with a fail-safe stop. High-impact and safety-relevant decisions always keep a human in the loop.

Run the gigafactory autonomously.

Start with one line and one hard success metric. Expand when the numbers earn it.

Book a line assessment See pricing