ReflexFunds
Active Member
Wright’s Law/Learning Curves/Experience Curves are extremely important to understand when considering technological progress and are particularly key to what Elon’s companies are trying to achieve. Wright's Law is often misunderstood however so I think it’s worth explaining some of my thoughts about it.
We have found pretty consistently over the past 100 years that as the cumulative historic production of a product doubles, the production cost reduces by a roughly fixed percentage. This is Wright’s Law and it is a rule of thumb which works over long time periods over multiple product and factory generations, but is not likely to be very accurate when making short term predictions.
Wright’s Law works because of three different underlying mechanisms:
1) R&D spend roughly scales with an industry’s revenue, so cumulative R&D spend scales with cumulative production. If you have a larger industry, more resources are spent on technological progress and these breakthroughs drive down the cost. R&D appears to be the largest driver of experience curves but it doesn’t work on short time periods because it could be several years before new R&D is put into production. It can also be quite lumpy with cost breakthroughs coming in fits and starts when a new generation of technology is introduced.
2) For every product you produce the staff and company “learn” how to do their task better. This is a combination of more productive staff and better production methods introduced after fixing the bottlenecks and problems previously experienced. In other words, to some extent all Cost of Goods Sold of a product can double up as R&D. Possibly the key driver of success in Elon’s companies is his push to maximize the R&D value of all expenditure – whether that is a COGs cost, capex cost or R&D cost. For example much of SpaceX’s R&D was done via COGs as it tested new hardware on customer missions. Similarly, Tesla’s factory is set up to rapidly upgrade car designs once lessons are learnt from the production staff and the fleet (which are both expensed as COGs). He also experiments with multiple new production techniques when investing in capex for a new factory – if this works he will use it again, if it doesn’t work he can always return to traditional process for the next factory – either way he has learnt something valuable and capex has worked as R&D.
3) Increased economies of scale. These are actually a function of annual production rather than cumulative production. These scale advantages could be higher staff productivity, more fixed cost and depreciation leverage or better purchasing power from suppliers (plus suppliers passing on their own scale savings). Economies of scale can actually cap out fairly early – you get huge benefit from 1 production line producing 500k cars vs 100k cars, but two 500k factories only has limited increased economies of scale vs one 500k factory.
There is some nuance to these mechanisms though. Experience curves only really work on unique components – when looking at a system with multiple sub components you are really looking at an average experience curve over the whole product. You can’t expect an entire EV to follow the same experience curves – some of its components have already gone through 50 years of learning in ICE cars. Most of an EV Powertrain should follow steep experience curves to some extent - this includes Battery Cells, Packs, BMS, chargers, converters, motors, inverters, cooling etc. But only for costs which are not for off the shelf components or commodities (most battery “commodities” are actually highly value add products which can themselves follow experience curves as EVs increase their cumulative production – but this can also be cancelled out in the short term by supply demand pressures if raw material production expansion is lagging behind end product capacity/demand). Many of the components supplied to Tesla are actually designed by Tesla in-house with just manufacturing outsourced – so many supplier components can also follow experience curves the same as components manufactured in-house.
Another important note on Wright’s Law is that cost reductions at times can rapidly outpace the growth in cumulative production if R&D is suddenly scaled up significantly as a % of an industry’s revenues (or even just if R&D productivity per $ is suddenly increased). This means that Tesla can still achieve significant cost reductions in many of the car components common to both EVs and ICEs. R&D in real manufacturing innovation for car production has been very low for decades as the global auto industry settled into a comfortable oligarchy with no culture of progress or innovation. This means the industry has not been learning enough from the hundreds of millions of cars it has produced over recent decades and there is a lot of slack for Tesla to apply new technology to reenergize innovation in components which are already extremely high volume. This is similar to the Boring Company (almost no R&D for 50 years has left a lot of slack) and SpaceX (corrupt cost plus pricing to lobbyists had stifled innovation in rocket development). For this reason there is some argument that even non EV unique components in Tesla cars could follow quite steep learning curves in line with Tesla’s cumulative production (rather than tracking the much smaller increases in cumulative historical car volume from the past 100 years). But these components will still likely have a much lower learning rate than for the EV component alone.
Another important consideration on Experience Curves/Wright's Law - certain engineering philosophies and corporate cultures can allow you to achieve significantly higher learning rates than the competition - and Elon has fully embraced these advantages.
The three mechanisms driving Wright's Law (R&D scaling, Learning from experience and Economies of scale) are all somewhat interrelated and the lines can be blurred at times.
The main thing they all have in common is that delivering on each works by eliminating the bottlenecks to cost reduction or production increases.
I’ll call this the Bottleneck Theory of Progress and I think it is the key engineering philosophy behind all of Elon’s companies. Elon solves these Bottlenecks by breaking every large goal/problem down into Physics First Principles and then re-writing every challenge in terms of smaller steps. This gives his teams smaller, clear and achievable goals and allows Elon to focus his resources (human and financial) on the key bottlenecks where the most progress can be made for a given number of resources. This engineering philosophy allows Elon to accelerate the mechanisms which drive Wright’s Law and achieve far higher learning rates than previously shown in the industries/product.
This First Principles and Bottleneck approach is completely different to how R&D and innovation normally works in the economy. Generally R&D teams are tasked with iterating the current technology with a 1-5% improvement one step at a time. This leaves most products dependent on sometimes arbitrary design decisions taken years or decades ago, using outdated science and engineering tools and a huge reluctance to pivot from sunk costs. Elon instead looks at every problem from the bottom up, ignoring previous design decisions and pulling in all current knowledge from all branches of technology and engineering. He will rapidly pivot strategies if a newly considered approach solves more cost/volume bottlenecks than the current one, regardless of sunk costs.
One analogy to explain the differences between these approaches is to consider a person tasked with travelling from point A to point B without knowledge of the paths or terrain in between. On his journey he faces many decisions on which fork in the path to take. Some of these decisions are educated guesses while some are random. This person eventually finds themselves scaling a mountain with the route ahead getting gradually more and more difficult with progress getting slower and slower, but he keeps pushing on, one step at a time, refusing to ever turn back. This person is sadly the most common process of innovation in corporations today.
A second person on the other hand realises he can always turn backwards and attempt to find an easier route. He also realises many different people are trying to cross the same terrain (in the real world people pushing forward in different branches of technology/the economy) and that he can communicate with them to try to piece together a fuller picture of the terrain to better plan the route forward. Eventually he finds an alternative and far easier route around the side of the mountain. This person is Elon and this is the culture he has tried to instil in Tesla, SpaceX, Paypal, OpenAI, The Boring Company and Neuralink.