So you want to build a battery factory

It’s harder than it sounds

We’re going to need a lot of batteries

These are exciting times for those of us cheering on the transition to an electrified global economy. In the automotive sector alone, OEMs have pledged well over $100Bn in investments to transform their product lines to battery power, and have announced plans to introduce somewhere between 450 and 500 new electric vehicle models in the next few years. The financial community is on board too, as each new public declaration of intent to electrify is greeted with a bump in the OEM’s stock price.

Amidst all this excitement, there is growing concern among those who follow these trends that there may not be enough battery cells to power all the electric vehicles that are supposed to hit the road by the middle of the decade. In a previous post (“The next chip shortage will be a battery shortage”), we crunched the numbers and predicted a significant shortfall in global battery production capacity relative to the anticipated demand from the transportation sector. That post also offered some thoughts as to how OEMs might avoid a repeat of the depressed production volumes, sales, and profits currently resulting from the semiconductor crunch.

We’ll make our own batteries. How hard can it be?

Clearly the auto OEMs have seen this shortage coming, and they understand that the future of their companies will depend on a steady supply of high-quality battery cells, as much as it does on rubber, steel, and aluminum (and microchips, for that matter). In response we’ve seen almost weekly announcements of brand new cell suppliers, and a growing list of car companies announcing plans to go vertical, either manufacturing their own battery cells or locking in dedicated production capacity through a closely-held joint venture. (An incomplete list of organizations declaring such intentions recently includes BMW, Ford, GM, Mercedes-Benz, Porsche, Stellantis, Tesla, Volvo, VW, and Apple!) Problem solved, right? What’s the catch?

The catch is that much as batteries are complicated, building and operating battery factories is really complicated — especially getting a new factory online and ramped up to full production volumes with the quality and yield needed for long-term viability. For evidence, we need look no further than the two titans of EV battery supply, “Tier 1” suppliers Panasonic and LG Chem (now LG Energy Solution). Panasonic, for years the sole cell supplier to Tesla, broke ground on the Nevada Gigafactory in 2014. This flagship facility didn’t begin producing cells until January 2017, and as of April 2019 was still only running at two-thirds of targeted production capacity, limiting Tesla’s output of the Model 3. Similarly, a large LG Chem battery plant in Poland was supposed to begin production in 2017. The start of cell production slipped well into 2018, and by early 2020 limited capacity led to follow-on delayed vehicle shipments from Jaguar, Audi, and Mercedes. Given the extensive experience and leadership positions of Panasonic and LG Chem, it should be sobering for OEMs planning to build their own batteries that these factories literally took so long to ramp to full production.

What’s so hard about making batteries?

Global automakers have decades of experience sourcing materials and components, and building them into complex, high-quality products. As such, one might assume that making batteries would be something they could master relatively quickly. However we’ve already seen that industry-leading Tier 1 suppliers still don’t have it fully figured out, so there must be some significant complexity under the hood. Here are a few of the primary challenges to making batteries at scale:

• Sensitivity: Battery performance depends on a number of complex factors, including the quality and purity of the materials that go into them. It’s fairly well known that tiny amounts of impurities can cause massive yield issues in semiconductor production, but battery production requires “clean-room” process standards as well (think bunny suits from those old “Intel Inside” commercials). Returning to the Panasonic Gigafactory example, as of April 2019 they were scrapping half a million cells every day, roughly 15% of total production at the time, reportedly due to lack of adherence to proper clean-room processes.
• Many, many variables: There are hundreds, if not thousands, of parameters to tune when ramping up a battery factory. Incoming materials properties, slurry viscosities, equipment setpoints, densities, temperatures, pressures, speeds, all the way through to the final “formation cycling” procedures during which a finished cell is carefully charged and discharged a few times to form critical layers and structures inside the battery. It’s simply a lot to keep track of, and any one of these steps can cause a yield issue.
• “Voodoo” optimization: Through our partners in the industry, we hear that much of this production optimization work is performed using “eyeball” estimates, gut feel, and an overall lack of systematic analysis. The data analysis done during this process often amounts to manually pulling data from the production line using thumb drives and analyzing it in spreadsheets. Given the urgency and time constraints in play, the industry is badly in need of modern, automated analytical tools to surface the key correlations that drive finished battery quality and yield.
• Talent vacuum: As we discussed in our previous post, global battery production needs to grow by an order of magnitude or more in the next few years. There simply aren’t enough experienced engineers to hire away from the existing Tier 1 battery suppliers to fill all the factories that the auto OEMs plan to build. They’ll need to look to automation and purpose-built tools to address this shortage as well.
• Continuous and constant change: All of this activity is taking place against a backdrop of rapid development and advancements in the underlying battery technology itself. New chemistries, larger form factors, new manufacturing techniques (see Tesla’s dry electrode processing showcased during last September’s Battery Day), and more. A battery factory built today won’t be the same as one built just a few years ago, so there is a lot of ad hoc learning to be done.
• Looooong feedback cycles: Fundamentally, it may take months to know if you made a good batch of batteries. Formation and aging (the last steps in the manufacturing process) take two to six weeks, and it will take a customer several more months to qualify a battery design for use in their product. As such, it’s critical to catch any quality issues as early as possible in production as you dial in the process.

Enterprise Battery Intelligence can accelerate production ramp

A new category of software, Enterprise Battery Intelligence (EBI), is designed to use large-scale data analysis to optimize every part of the battery lifecycle, including production. An EBI solution automatically aggregates data from a variety of sources including materials properties, production parameters, quality control measurements, and formation cycling and aging at the end of production, synthesizing correlations and insights that help to rapidly dial in the critical-to-quality production parameters. These types of tools are already common in industries like semiconductor and pharmaceutical production, and the advent of EBI arrives just as the battery industry needs to seriously up its game. With these modern, industry-specific tools, battery manufacturers will be able to adapt to rapid change and onboard staff more quickly as well, as there will be less need for people with “eyeball optimization”-type experience. Bottom line, the industry can’t afford to spend three to five years scaling up each new battery factory if we’re going to meet EV sales projections, so rapid analytics will be key to accelerate ramp times.

Where is this all going?

While there are major challenges ahead, we do expect the industry to rise to the moment and ultimately supply enough batteries to power many millions of electric vehicles. At Voltaiq our mission is to accelerate the transition to a battery-powered world, so suffice it to say we’re here to help. Whether this massively increased production capacity comes from existing Tier 1 battery suppliers, new entrants, or verticalization efforts from carmakers remains to be seen, but here are a few things we expect:

• Terafactories, not Gigafactories. The battery industry needs to scale rapidly, and battery production lends itself to scale. In the coming decade we’ll see new battery factories coming online with orders of magnitude greater production capacity than today’s largest factories.
• Building near cheap energy. One of the primary, and most expensive, inputs to battery production is electric power. Look for the Terafactories of tomorrow to be sited near inexpensive sources of clean, plentiful energy.
• Building near the car factories. Battery prices continue to fall and are thus getting more expensive to ship relative to their value. It fundamentally makes sense to make the batteries close to where they’ll be assembled into vehicles. Panasonic and Tesla started this trend with the Nevada Gigafactory producing both battery cells and Model 3s, and we expect to see this practice adopted more broadly across the industry.

However this all plays out, it will be an exciting ride and there is a lot of great work to be done to get there. Let’s get after it!