Would it have been possible to predict the arrival and increased popularity of electric vehicles? If yes, how early could you have known?
With Focus, the automotive industry could have known that lithium-ion batteries were getting cheaper much faster than combustion engines as early as 1992.
The entire automotive industry got taken by surprise. Lithium-ion batteries got affordable years before they and their advisors thought they would.
“The cost target of $250 per kWh is unlikely to be achieved at the cell or pack level by 2020”
Boston consulting group 2010
Wrong estimate on Li-ion price development
Electric vehicles are becoming more common by the day and are expected to completely replace internal combustion engine vehicles. However, until recently, it was very common to hear industry participants speak about the possibility of profitably producing and selling an electric vehicle due to the high costs of lithium-ion battery packs.
To get a definitive answer on when battery packs would become cheap enough to use in mass production vehicles, the American car industry hired BCG to write an extensive report in 2010 (Boston Consulting Group, 2010).
In its report, BCG concluded that: “The cost target of $250 per kWh is unlikely to be achieved at either the cell level or the battery pack level by 2020”. In reality, lithium-ion battery packs cost $137 per kWh in 2020. BCG was almost 100% off with their estimate. This may sound extreme, but it is common for expert predictions to be this far off.
In this example, we investigate how the car industry could have used Focus to predict how quickly lithium-ion battery packs would get cheaper.
Could the industry have predicted electrification?
By training only 2 AI classifiers, any actor within the car industry could have predicted that lithium-ion batteries were going to get cheaper much faster than what BCG was predicting. With these classifiers, the industry could have identified all relevant inventions, and Focus would have been able to calculate the improvement speeds for each technology.
To start off, lets investigate the overall patenting trend.
As you can see, the overall patenting trend holds no predictive power whatsoever. To this day, combustion engines are being patented more heavily than lithium-ion battery patents, even though the disruption of the internal combustion engine vehicle is in full swing.
Next, let us look at Focus’ technology improvement speed estimate, which is calculated using the same patent data, but then while looking at Focus’ cycle time and knowledge flow metrics rather than patent counts.
The first commercial lithium-ion battery was invented in the late 80’s. As you can see, from the early 90’s onwards it was already clear that lithium-ion batteries were improving much faster than combustion engines were.
With this insight it would have been clear that the disruption of internal combustion engines was inevitable in 2010. It would also have been clear that BCG’s forecast was off by a lot. The price forecast of battery packs would have been much more accurate if they used Focus. By using the estimated improvement speed of 26% for lithium-ion batteries in 2010, and by extrapolating into the future, Focus would have predicted a price per kWh of $120 in 2020. This is a little below the actual price of $137, but much more accurate than BCG’s forecast.
The car industry could have predicted that the electric vehicle revolution would happen much faster than they had anticipated. This would have allowed them to start planning for the shift faster, and to avoid being disrupted by Tesla as much as they have.
With Focus, you can see technology disruption coming in any technology domain up to decades in advance. After training AI classifiers, which takes only a few hours, you can monitor the improvement speed of any technology. Among competing technologies, the fastest improving one ultimately always wins.
The car industry could have known that electrification was inevitable as early as 1992.
Using Focus, the car industry could have also gotten a much more accurate forecast of 2020 lithium-ion battery pack costs in 2010.
Without Focus, it is impossible to get a quantitative measurement of a technology’s improvement speed.
All the insights shown in this example were generated with data that was publicly available at the time, there us 0 hindsight.