John Petersen
On Wednesday a reader sent me a new Goldman Sachs research report on the advanced battery sector titled “Advanced Batteries: Light, but the tunnel is long; Buy ENS, HEV Neutral, AONE Neutral.” If the essence of legal argument is “the plausible boldly asserted,” then I’d suggest that the essence of sell-side analysis must be “the implausible accepted without question.” While I agree with Goldman’s conclusions that Enersys (ENS) is a bargain at the current price and caution is best when it comes to A123 Systems (AONE) and Ener1 (HEV), I have major problems with their assumptions about lithium-ion battery costs, safety and consumer acceptance.
Since we’re starting a long holiday weekend, I’ll keep it brief.
Cost Assumptions
There is no issue in the advanced battery sector that’s more contentious and wildly speculative than the future cost of lithium-ion battery packs. While reliable numbers are scarce, there are a couple of clear reference points. In May, Nissan’s chief product planner for North America told the Wall Street Journal that their manufacturing cost for the 24 kWh battery pack used in the Leaf is roughly $18,000, or $750 per kWh. By the time you include a normal markup, the OEM price from an independent battery manufacturer would be about $1,000 per kWh. In its most recent Form 10-Q, A123 reported $1,375 in cost of goods sold for each kWh shipped in the first quarter. While these two examples are not conclusive and rumors of low prices are common as weeds, I tend to believe a pack level cost of $1,000 per kWh is a safe current estimate.
According to Goldman Sachs, the United States Advanced Battery Consortium is targeting pack level costs of $500 per kWh by 2012 and $300 per kWh at volume production by 2014. The long-term goal is $250 per kWh, the current cost of the least expensive consumer batteries. Faced with ambitious auto industry and government expectations, battery developers are caught between a rock and a hard place. They can either tell the automakers, the financial markets and the government “your goals are unattainable” and reject the funding opportunities, or they can say “we think we may be able to reach that goal,” trusting that the money will flow and forgiveness will be easier to come by than permission.
In its manufacturing cost discussion, the Goldman report breaks down the current cost of a typical automotive lithium-ion battery pack into the following broad classes of economic inputs.
| 5% | Lithium |
| 7% | Other active materials |
| 32% | Purchased parts |
| 11% | Labor |
| 14% | Depreciation |
| 14% | R&D |
| 4% | Scrap |
| 13% | Markup |
Later on, in the appendices, Goldman identifies the principal manufacturers of binders, foils, coating equipment, separators and electrolytes, and manufacturing systems, which are with very few exceptions Asian.
I’m not a trained industrial engineer, but I am a trained cost accountant. When I take time to consider the classes of economic inputs and their relationships to total product costs, I can’t imagine where savings of 50% to 70% will come from. We live on a resource constrained planet and commodity prices have been increasing for decades. With escalating global demand for all the necessities of life, it seems patently absurd to believe that the prices for lithium and other active materials will collapse or that prices for materials-based foils, separators, electrolytes, cases and other purchased components can fall significantly. It’s possible to reduce labor costs and jobs through increased automation, but the trade-off is increased depreciation. When you factor in the reality that over 60% of the economic inputs are directly or indirectly attributable to imports, it’s hard to see how medium term forecasts of $500 per kWh and long term forecasts of $250 per kWh are reasonable.
Economies of scale are easy to discuss as glittering generalities. They are extremely hard to achieve in volume manufacturing of a mature product. The Japanese have been making lithium-ion batteries for almost 20 years and it’s highly unlikely that Sony, Panasonic, NEC and Toshiba have overlooked easy economies of scale while building billion dollar businesses. I won’t argue that the widely hyped cost savings and performance gains can’t happen, but I haven’t seen anything that leads be to believe they will happen.
Safety Assumptions
In the lithium-ion battery industry the standard penetration safety test is conducted by piercing a single cell with a 3.5 mm steel rod. This is a wonderful test of the damage that might be inflicted if a carpenter drops his tool belt on a battery powered screwdriver, but it proves absolutely nothing about battery pack safety in the class of catastrophic penetration events that can and do occur in automobile accidents.
When automakers conducted safety testing of NaS battery packs in the mid-90s, they took fully charged battery packs and drove a foot long four-bladed wedge into them with a hydraulic ram because it was the only way to learn what the pack level interactions would be in a catastrophic penetration event.
The only event I know of that involved a cascade of battery failures was a spontaneous fire in a storage bunker at the Toxco recycling facility in Trail, B.C. last November. While the newspaper reports of the Toxco fire are sobering, they don’t have the same impact as this YouTube Video that shows why pack level studies of cell interactions and failure cascades are an essential prerequisite to battery pack safety claims.
I understand that Tesla Motors (TSLA) is the only automaker that uses or plans to use lithium-cobalt-oxide cells, the most dangerous lithium-ion battery chemistry, but even the benign chemistries use organic electrolytes that are inherently flammable and react v
iolently when penetrated. Single cell safety tests are interesting, but until lithium-ion battery developers conduct pack level tests of catastrophic penetration, there will be no basis to make any safety claims or assumptions.
Consumer Acceptance Assumptions
The nice thing about consumer acceptance forecasts is that they’re only wrong in 20/20 hindsight. For most consumers, an automobile is the second most expensive asset they own. Car buying decisions are not made lightly, and while most buyers carefully consider their capital, operating and maintenance costs, they make a buying decision based on a subjective assessment of how well a vehicle suits their needs and driving habits.
All of the consumer acceptance forecasts I’ve seen assume that battery prices will plummet and large numbers of consumers will be willing to overlook the quirky limitations of EVs to make a political or ideological statement. That may be enough motivation for a few diehards, but it won’t cut it for the average consumer who believes the green in his wallet is more important than the green in his cocktail party conversation.
In the final analysis, I believe forecasts about consumer acceptance rates are only as reliable as the future cost assumptions that underlie the forecast. Since there’s good reason to believe the cost assumptions are garbage in, there’s also good reason to believe the consumer acceptance assumptions are garbage out.
Investment Recommendation
The Goldman report was neutral on A123 and Ener1, but included a buy recommendation for Enersys with a six-month target price of $29. It also reiterated a buy recommendation on France’s Saft Groupe (SGPEF.PK). While Goldman does not currently cover Exide Technologies (XIDE), which generates more annual revenue than Enersys and Saft combined, I expect that to change within the next 12 months if the earnings recovery Exide reported for its most recent fiscal quarter continues.
Disclosure: No positions.
