Welcome back to The Battery Insider.
In Episode 01 of “Battery Gigafactory Building Practice” series, we covered the strategic, compliance, supply chain, delivery and talents foundations of a global battery gigafactory. Last episode, we discussed why the cell blueprint is the first financial model of your gigafactory—and why product definition must come before site selection and equipment/material procurement. This edition moves one step downstream: once the blueprint is defined, how do you select the right process route, and how do you architect yield targets that are ambitious but achievable?
Table of Contents
1. The Process Route Is the Factory's Operating System
If project teams treat process route selection as a technical formality—a box to tick before moving on to equipment procurement. This is a costly mistake.
The process route determines:
Which equipment is required, and in what configuration?
How many process steps exist, and where the bottlenecks will be?
What micro-environment each step requires?
How materials flow through the factory?
Where yield losses will occur, and how they accumulate?
How much capital is locked into each stage?
How long the ramp-up will take?
In other words, the process route is not just a sequence of steps. It is the factory's operating system. Once frozen, it dictates everything downstream. Changing it after equipment is ordered is not a redesign—it is a rebuild. That's why we discuss with our clients before: Don't just find an equipment manufacturer to provide turnkey solutions.
2. The Three Segments Process, Three Different Logics
A battery cell factory is typically divided into three process segments: electrode manufacturing (front-end), cell assembly (mid-end), and formation, aging, and testing (back-end). Each segment has a fundamentally different logic, and each must be designed with its own yield architecture.
2.1 Front-End: Electrode Manufacturing
This segment mainly covers mixing, coating, calendering, and slitting. The core challenge is managing rheology, adhesion, and dimensional consistency. Yield losses here are often invisible at the time they occur but propagate through the entire downstream process. A coating thickness variation that passes visual inspection may become a separator puncture several steps later. The front-end yield architecture must therefore focus on process capability, not just defect detection.
2.2 Mid-End: Cell Assembly
This segment mainly covers winding or stacking, jelly-roll encapsulation and filling etc.. The core challenge is precision and contamination control. A metal particle generated during laser cutting can travel through multiple process steps before it causes a micro-short. The mid-end yield architecture must therefore focus on contamination control, equipment stability, and traceability. A defect that cannot be traced to its source cannot be fixed.
2.3 Back-End: Formation, Aging, and Testing
This segment mainly covers formation, aging, grading, and packaging. The core challenge is electrochemical consistency and safety. Cells that pass electrical testing may still fail in the field if the formation process was not properly controlled. The back-end yield architecture must therefore focus on process window control and statistical process control, not just pass/fail testing.
Above we talked about yield integrated into process, then let take it into deeper.
3. Yield Architecture Is Not a Single Number
One of the most common mistakes in factory planning is treating yield as a single number—typically a steady-state target applied uniformly across the financial model. This is not a yield architecture. It is a wish, which will cost financial imbalance.
A proper yield architecture defines:
Step-level yield targets: What is the expected yield at each individual process step?
Cumulative yield: What is the combined yield across the full process route?
Ramp curve: How does yield evolve from first production to steady state?
Bottleneck yield: Which step has the lowest yield, and how does it constrain the entire line?
Sensitivity: How does yield at each step affect material consumption, equipment utilization, and cost per GWh?
Without this architecture, the financial model is disconnected from the factory's actual performance. A project that assumes 95% yield from day one will run out of cash before it reaches 80%. We should have seen similar projects before.
So process route and yield architecture are not sequential decisions. They are simultaneous ones, should be designed together. The process route determines where yield losses can occur. The yield architecture determines whether the process route is economically viable.
This is why process route selection cannot be delegated to equipment vendors. Vendors optimize for their own equipment. They do not optimize for the factory's overall yield architecture. That responsibility belongs to the project team.
4. The Bottleneck Is Not Always Where You Think
In most factory designs, the bottleneck is assumed to be the slowest or most complex step. In practice, the bottleneck is often the step with the highest yield loss, not the longest cycle time.
A step with a 99% yield and a 10-second cycle time may process fewer good units per hour than a step with a 95% yield and a 5-second cycle time. The bottleneck is determined by good-unit throughput, not raw throughput. This distinction is critical for capacity planning and for setting realistic ramp targets.
5. Common Pitfalls
Equipment-first thinking: Selecting equipment before the process route is frozen, then forcing the process to fit the equipment.
Yield as an afterthought: Treating yield as a performance metric to be improved later, rather than a design input to be architected upfront.
Uniform yield assumptions: Applying a single yield number across all process steps, ignoring step-level differences and cumulative effects.
Ignoring the ramp curve: Using steady-state yield in the financial model, which understates early-stage losses and overstates early-stage cash flow.
Vendor-driven process design: Letting equipment vendors define the process route, which optimizes for their equipment rather than for the factory's overall economics.
Confusing cycle time with throughput: Assuming the slowest step is the bottleneck, when the real constraint is good-unit throughput.
Before freezing the process route and yield architecture, the project team should summary out related connections and constrains.
Here is a question worth discussing with your team:
If you would explain your yield architecture to your board tomorrow, could you show how each process step contributes to—or constrains—the factory's overall economics?
Feel free to reach out when you need compare notes on how other projects are approaching process route and yield architecture decisions. We are always interested in exchanging perspectives on what makes a process route truly bankable.
- Future Topics in “The Battery Insider”
Topics: Battery Gigafactory Building Practice, Tech Innovation, Application Breakthroughs, On-site Factory Tours etc.