Good find, but that contradicts many of their earlier numbers.
More importantly, Tesla bought the old NUMMI plant from the GM/Toyota consortium to use as their manufacturing facility. The peak production under NUMMI was only 400k cars/year, so they'd have to get an additional 25% utilization out of a facility that was running 3 shifts/day in many areas. I'm fairly skeptical they can do that, especially by 2020.
Going from 60 cars/day (current production rate) to 1,500/day is no small feat.
[citation needed] on the 400k, but at least on wikipedia the average number is 6000/week. At 3 shifts/day * 5 days gives the really too nice and round 400 / shift (or almost 1 a minute off all lines give a 7 hour production). And this is with Toyota's lean changes so making significant improvements would be hard. For example if you watch UltimateFactory you can determining the takt time in your head of other auto manufacturing facilities, some including GM are able do do 450+ in their plants (btw for those of you in the EU and you ever get the chance see Porche's factory, it is fantastic). But the fact is that the cars being made are not the same and so you might be able to get that efficiency gains simply from the smaller number of value added steps required. Really it all depends upon the amount of waste (non-value add) that is in the system and how easy it is to make improvements. Given the opportunity I would love to checkout Tesla's facilities and would probably geek out more on it than the car. 25% might be doable, but without more real information I guess this is like saying sports team X could win Y without looking at the current stats.
One of the more interesting things about factories is that a facility that has a robots doesn't guarantee high throughput. If the robot can work 10 times faster than the rest of the line does that mean it is mostly idle or does it work on multiple lines or do you have a wip queue somewhere and then you have more work moving stuff from the robot to inventory etc and lastly the great thing about humans over robots is that we are much more flexible and can try new things sooner to see if they work better. I wouldn't be surprised if there are stories of factories that got robots and their metrics went down.
Peak production at the 370 acre facility on the outskirts of
Silicon Valley topped 400,000 cars and trucks in 2005.
I've been on tours of several plants, include GM's Delta Township plant which at the time was the most automated in the world as well as the factory we're discussing (NUMMI both Pre and Post Tesla ownership).
The thing that makes me skeptical with Tesla is the amount of custom work they need to do. Some of the GM / Toyota plants have higher throughput, but that's because they're largely just assembly buildings. Tesla is custom forming many parts on site, far more than a typical assembly building. It would be hard to scale many of those areas without huge investment.
Very interesting, even with a big investment to try to scale their tooling that doesn't happen overnight. At the very least it could be a difficult/fun challenge for someone to take on.
If they maxed out the "Tesla Factory" (formerly NUMMI), they'd still need ~100k vehicles/year from somewhere else. GM just announced they're spending ~$1.5B on a 150k vehicle/year plant in China. Assuming the new Tesla factory would be more expensive, since labor cost in the US demands more automation, they'd be looking at ~$2B to build a new facility. Plus the costs from the upgrades at the Tesla Factory.
I have a hard time seeing Tesla deploying $5B in CapEx in the next 5 years when they only have $700M in PPE now and $800M cash in the bank.
When they retool for the Gen III 'Model E', I'd expect about $1B in capital investment to satisfy a production level of ~150k cars/year.
More importantly, Tesla bought the old NUMMI plant from the GM/Toyota consortium to use as their manufacturing facility. The peak production under NUMMI was only 400k cars/year, so they'd have to get an additional 25% utilization out of a facility that was running 3 shifts/day in many areas. I'm fairly skeptical they can do that, especially by 2020.
Going from 60 cars/day (current production rate) to 1,500/day is no small feat.