QS: Path To Scale

From Raptor to Cobra and Beyond

I previously explored Quantumscape’s potential commercialization structures. There are a lot of steps between where they stand today, as a low volume producer, and becoming a mature company that’s capable of pumping out GWh’s of production.

Note that QS is now aided by PowerCo in this quest for scale. But whether it’s QS, alone, or a team effort, these growing pains will have to get worked through one way or another.

In this article, I want to review where Quantumscape currently stands today, and what will be needed to finally hit the fabled GWh number.

Film Start Production

First, let’s define a “film start”. A film is simply one ceramic separator layer in their cell stack-up. When Quantumscape discloses production numbers of 100,000 film starts per week for Cobra (as a projection), they simply mean that they anticipate being able to produce 100,000 separators per week.

Figure 1 shows separator production targets for Raptor and Cobra. We can see that Cobra will employ much higher throughput numbers while occupying a much smaller footprint than previous iterations. Quantumscape management has heralded this as a massive breakthrough.

Figure 1: Current Scaling Projections

Okay great. 100,000 film starts sounds like a lot. What does that mean in terms of cell production? …or how many vehicles will Cobra potentially be able to support per year?

Here’s the math:

Quantumscape discloses a cathode loading of 6.2 mAh / cm^2 for their B0 sample (QSE-5) cells. This is basically how much energy (in Ah) a layer in the cell contains per unit of area.

Nominal voltage for an NMC cell is 3.84 V.

And a Quantumscape cell is listed at an area between 6 cm x 7.5 cm - 7 cm x 8.5 cm.

Putting this all together:

\(energy = 6.2 \frac{mAh}{cm^2} \times 6.5 cm \times 7.5cm \times 3.84V = 1.16\ wh\ per\ film\)

So if Cobra is projected to produce 100,000 films per week, that means that they can produce 116 kWh of battery cells per week (or right around 6 MWh annually).

Assuming that a typical vehicle has a 100 kwh battery pack, Cobra is slated to produce enough cells to build 60 vehicles per year. This is a far cry from the millions of vehicles that the PowerCo announcement hopes to build with QS cells. Cobra will still be two orders of magnitude shy from GWh scale.

In fact, if we were to reproduce Figure 1 to show the film starts needed to reach 1 GWh scale, it would look as follows.

Figure 2: Cobra to Giga Scale

GWh scale requires over 16 million film starts per week. Cobra doesn’t even register in the grand scheme (as seen in Figure 2).

To reach Giga watt-hour scale, QS will either need to build close to 200 Cobra lines or make a Cobra machine 200 times larger or 200 times faster (or some combination of these).

Quantumscape and PowerCo have their work cut out for them. We’ll really need to start hearing more about plans beyond Cobra if we want to believe that there’s a chance for millions of vehicles with QSE-5 cells on the road before the end of this decade.

I want to spend the rest of this post exploring some of the patent filings to try and glean anything we can on Quantumscape’s progress, to date.

Raptor?

Patent: Rapid thermal processing methods and apparatus

Raptor is basically old news at this point, so I won’t cover it in depth here. What we do know is that Raptor is essentially the first iteration of an automated continuous manufacturing line with an estimated throughput of about 15k film starts per week.

On to Cobra…

Cobra?

Patent: Processing apparatuses and methods of using

Here’s an excerpt from the filing speaking to the potential throughput.

[0181] In some embodiments, including any of the foregoing, the process apparatus is capable of producing at least 200,000 yielded sintered bilayers per week. In some embodiments, including any of the foregoing, the process apparatus is capable of producing at least 1,000 m² of sintered bilayers per week.

In addition,

In some embodiments, the sintered article comprises a trilayer, with two ceramic layers sandwiching a metal layer.

As we can see, this image really highlights the small footprint claimed by Quantumscape.

Immediately, the most notable thing that jumps out from the patents is that Cobra seems capable of producing 200k film starts per week (fspw) in the bilayer configuration - or 400k fspw in the trilayer configuration. This already implies that the estimates given by Quantumscape management may be conservative. Instead of the 60 cars per year estimate given above, QS-0 may be able to supply hundreds of vehicles starting as early as this year. This is still a far cry from Giga scale, but will be a major accomplishment none the less.

King Cobra?

Patent: Rapid ceramic processing techniques and equipment

This may be the next step in terms of scaling, basically taking Cobra and super-sizing it.

We can calculate the max film start production rate based the information provided in the patent.

In some examples, the sintering film, or sintering bilayer, may move through the CML [continuous manufacturing line] at a rate between 2 centimeters/minute (cm/min) and 100 cm/min.

and,

The process of claim 33, wherein the bilayer has a width, after sintering, of between about 0.8 mm to about 5 m.

Here’s the calculation:

\(production = 100 \frac{cm}{min} \times 500 cm \times 0.0238 \frac{wh}{cm^2} \times 525,000 \frac{min}{year} = 0.624 \frac{GWh}{year}\)

So “King” Cobra may have the ability to get us most of the way to Giga scale. And if the trilayer configuration works (2 separator layers going through the heat treat step at the same time), that yields giga scale from just one machine.

Note that this assumes the machine is running 24/7 with no downtime, 100% yield, and in the largest configuration and fastest throughput covered in the patent. Naturally, this is an upper bound for production capacity based on this specific design.

Summary

It’s important to note that these are just “best guesses” on my part based on the guidance provided by Quantumscape management over the years and my interpretation of that guidance…along with the patent filings linked above.

I suspect that near-giga capacity per line is required to actually run a giga factory. If King Cobra were only at 100 MWh (still an order of magnitude larger than “baby” Cobra), that would still mean that PowerCo would need 800 production lines to meet the maximum capacity spec’d from the licensing agreement (80 GWh). And as Cobra stands today (100k fspw), PowerCo would need 10,000+ production lines to reach that 80 GWh target - in my opinion, this would just be uneconomical.

My Take

Based on the information provided above, I believe that Quantumscape does indeed have “line of sight” on Giga factory scale. But, they still have a long ways to go.

If I were to put an estimate on it, I would guess 2027 would be the earliest year that PowerCo could reasonably start cell production from their first Giga factory using Quantumscape technology. This assumes that Cobra production out of QS facilities goes off without a hitch in 2025 and equipment ordering, installation, and qualification for “King Cobra” can be accomplished in 2026 at PowerCo facilities.

We’ll know more in a couple of weeks when QS reveals their 2025 milestones at their next earnings call.