Discussion of New Platform for Model Y

[FrozenCanuck]

Hey gang. So last night on the Tesla Q1 F2017 conference call Elon said that there would be a "step change" in manufacturing efficiency when the Model Y comes out.

For context, and for those who haven't heard the call yet, this was brought up during a discussion of Model 3. Elon and JB were discussing how Model S is nothing amazing in terms of automated assembly, but the Model 3 line will be approximately as good as the world's best auto manufacturing lines with 5x the volume but the same labour, so WAY more efficient per vehicle. Then Elon said Model Y would be a massive difference and a new platform.

This has the usual suspects up in arms claiming that Tesla is throwing away it's prior plan and changing its mind all over again. I STRONGLY DISAGREE, and would love to have a dialogue about it here.

Let's go back to Elon's original "machine that builds the machine" commentary for a second. He said we'd have version 0.5 when M3 launches, then every 18 months or so we'd see an iteration. So version 1.0, then version 2.0 in about 3 years time.

So Model 3 was designed from the ground up, but they are still re-using a lot of concepts from the traditional car industry. Cable harnesses, 12V systems, etc. Sounds like they're chucking it all out the window for Model Y. No more 12V, and the use of multiple flex assemblies (just like in electronics) to reduce assembly complexity. Gotta go after 100% automated assembly, right?

Here's what I think: Model Y will be the first car Tesla makes that approaches FULL automated assembly, and, in parallel with that work, Tesla will be redesigning everything about the Model S, Model X and Model 3 to make use of the same innovations. The Model S and X should hopefully migrate to a single platform, and the Model 3 and Y will also be on a shared platform.

To all the folks arguing about how Tesla is throwing away capital by doing one-off platforms (S, X, 3), I argue it's still the best LONG TERM plan. No point making a Model Y on the M3 platform if you know that the M3 platform isn't as good (automation wise) as it can be. Better to invest in pushing things forward.

Thoughts?

 

[Krugerrand]

The answer is easy. They simply didn't have enough time to do what they wanted with the machine (and design) for M3. Let's not forget that the timeline for M3 got pulled wayyyyy ahead, which has never happened in the history of Tesla. Timelines have always slipped.

They now have time to do the machine (and design) right for the MY expected to come out in 2020 (or possibly late 2019).

 

[ecarfan]

I agree with both you guys. While Model 3 design is a big improvement over S/X in terms of ease of production, Tesla is still a small company with signficant engineering resource limitations and they have to make incremental steps towards Elon's goal of hyper-automated (my term) vehicle assembly.

It is impressive to see how Elon relentlessly drives the company forward. I was just rewatching the 2012 National Geographic "Mega Factories" video about Tesla. It was going on and on about how the Model S assembly line is so advanced and automated. But that was Tesla's first assembly line, and only two years later it was dismantled, improved, and moved to a different part of the factory to make way for the new S/X line which was much larger, the line currently running. Now the Model 3 line, even more advanced, is just a few months from starting up. All that in less than 5 years. Imagine what is going to happen over the next 5 years with the Model Y line becoming operational and a Tesla Semi line and who knows what else.

In just 9 years Tesla has gone from producing several hundred cars a year to now producing cars at the rate of 100,000 per year and by the end of 2018 looks very likely to be producing at the rate of 500,00 per year. So in a decade going from ZERO to a half million cars a year as well as launching a stationary storage business with an industry-leading kW/h low price, all due to the decision to build the largest factory on Earth.

So many investors and analysts seem to miss all that. Instead they obsess over quarterly earnings reports and small changes in delivered vehicles.

 

[jaguar36]

Aren't the S and X already on the same platform?

 

[RobStark]

Aren't the S and X already on the same platform?

That was the original plan.

The requirements of the FWD changed that.

S/X platform have 30% commonality.

 

[SwTslaGrl]

Perhaps it'll use PLC to reduce wiring Power-line communication - Wikipedia ?

 

[electracity]

Aren't the S and X already on the same platform?

Probably more correct to say that they can be built on the same line. From the model Y forward it will probably be correct to say that multiple vehicle types share a platform.

The model 3 platform will be orphaned if the model Y works out.

The assembly idea for the model Y are probably pretty wild. Likely a highly modular design. The risk here is delay, leaving Tesla without the most purchased vehicle type.

 

[FrozenCanuck]

The answer is easy. They simply didn't have enough time to do what they wanted with the machine (and design) for M3. Let's not forget that the timeline for M3 got pulled wayyyyy ahead, which has never happened in the history of Tesla. Timelines have always slipped.

They now have time to do the machine (and design) right for the MY expected to come out in 2020 (or possibly late 2019).

This comment makes a ton of sense. I completely forgot to acknowledge that the Model 3 got pulled ahead in my original post.

 

[GoTslaGo]

Don't forget that Elon mentioned that they will be using a lot of Model 3 components for the Semi, which I found super interesting. So if Semi becomes a hit, they will likely be already maxing out Model 3 components for both the model 3 and the Semi. (So in a sense the Semi could be to Model 3 like X is to S...and we all thought Y would be to 3)

So in that sense it does make sense to me that they would consider a whole new set up for Model Y. Production constrained again...

 

[BillO]

Well if it all works out, it will be great. But didn't the risk of going broke just go up? Surely a new platform and factory is going to cost billions.

 

[EinSV]

This was a very interesting development. I think it must have been a very difficult decision to make this pivot but as others have noted Tesla was committed to a very fast timeline on Model 3 and sticking to that timeline has been a major challenge and is obviously very important to customers and the company.

The timing could end up working very well as Model 3 should allow Tesla to scale into a mass production company and generate significant cash flow to develop the Model Y Alien Dreadnaught production.

Since the CUV market is bigger than the sedan market, in the long run (after 2020) the Model Y seems likely to be the higher volume vehicle so it should work out well to have the next generation manufacturing methods used there first as they will get the most bang for the buck, while getting more value out of the initial Model 3 factory by using the lines and equipment longer.

Obviously a ton of engineering and software development is needed to do this right so the somewhat longer than expected timeline for the Model Y launch makes sense. Tesla will have its hands full with Model 3 production, Tesla Semi etc. in the meantime.

 

[RobStark]

Well if it all works out, it will be great. But didn't the risk of going broke just go up? Surely a new platform and factory is going to cost billions.

Marginally.

More likely Model 3 and TE provide billions in profits. Tesla is just about break even with Gen II profits.

IF Tesla wants 4 new GF-AF ASAP in addition to new product will that will cause dilutive capital raises.

But IF Tesla goes just insanely fast instead of maximum plaid then it can probably funded organically.

 

[omarsultan]

It surprised me when I read this, and, to be honest, I have to question this a bit. They have spent a few billion dollars developing the Model 3 platform and building out the production line. They are currently developing platforms for the semi and the pickup as well as needing to build out production lines for both of those vehicles, each of which is a multi-billion dollar investment. On top of this, they want to build yet another platform? Oh, and somewhere in where, the Model S is due for a refresh.

At some point, I gotta wonder where all this money is going to come from. From a purely engineering perspective, probably there is an oppty to further optimize the Model 3 platform for the Model Y, but will it be in a way that an owner will notice/care? Can you really not leverage the Model 3 platform and get more than a couple of years out of that investment and is there really another better you can be dong with that money?

 

[RobStark]

Can you really not leverage the Model 3 platform and get more than a couple of years out of that investment and is there really another better you can be dong with that money?

If you don't disrupt yourself because you keep looking at sunk cost and can't move on someone else will disrupt you.

 

[MitchJi]

At some point, I gotta wonder where all this money is going to come from. From a purely engineering perspective, probably there is an oppty to further optimize the Model 3 platform for the Model Y, but will it be in a way that an owner will notice/care? Can you really not leverage the Model 3 platform and get more than a couple of years out of that investment and is there really another better you can be dong with that money?

It's about production efficiency. If you are patient, you will see the production line for the MY spitting out MY's between one car per second and one car per 2.5 seconds! Elon said multiple times that people probably won't believe that but he is very confident that they will be able to do that. If you don't believe him I suggest that watching the video of the reused rocket landing on the barge, and thinking about the implications of that might be useful.

If you think about the time and money that Tesla spent getting the MS-MX production line up to 25k per week, and consider the financial benefits of doing that, how can you believe that something that will revolutionize production efficiency will be a poor investment? It's a total no-brainer!

Elon said he was sees that as part of the reason that Tesla will grow to be bigger than apple. It's really impressive to me how quickly his theory of alien dreadnaught technology has progressed. When he first talked about his Production Epiphany he didn't mention modifying the cars or the robots. Now he's saying that the robots will all be controlled by the same software. It's going to work like an orchestra conductor!

You can definitely infer from his comments that he believes that this technology is an unassailable moat. IMO the unassailable moat is Elon!

 

[electracity]

At some point, I gotta wonder where all this money is going to come from. From a purely engineering perspective, probably there is an oppty to further optimize the Model 3 platform for the Model Y, but will it be in a way that an owner will notice/care? Can you really not leverage the Model 3 platform and get more than a couple of years out of that investment and is there really another better you can be dong with that money?

The way I see it, at this point, is that "alien dreadnaught" is Tesla's shot at becoming a substantial car company. The four largest non-Chinese automakers each make the number of cars in a day that Tesla takes a quarter to build. Making a half million cars per year in Fremont doesn't change that math much. Tesla is still tiny.

Musk needs the next exciting thing after the model 3 launch to continue to raise capital. Aiming at next gen auto production will keep supporters happy into the next decade. Just waiting for the model Y would probably get boring quickly. The affordable CUV is the world's choice of car today. Tesla will not have one until probably 2021. In that time frame many good EVs will arrive.

All this news now is likely timed to be sure investors are taking the long view. Model S sales tanking for a couple of quarters would in reality be just a small bump in the road. Musk obviously fears or expects this to happen, and that it will be used by the shorts and other critics at a time when he wants to commit to massive capital expenditures.

 

[cdub]

They really need to do an update to the S soon... at least on the interior... to make the extra money more worth it. I also think that once the federal tax credit goes away the Model S will be able to be $10,000 cheaper due to economies of scale with the 2170s once they use them in the Model S and X.

 

[ecarfan]

If you don't disrupt yourself because you keep looking at sunk cost and can't move on someone else will disrupt you.

Correct.

Walter Isaacson quoted Steve Jobs as saying, "If you don't cannibalize yourself, someone else will". I think Elon agrees.

 

[MitchJi]

They really need to do an update to the S soon... at least on the interior... to make the extra money more worth it. I also think that once the federal tax credit goes away the Model S will be able to be $10,000 cheaper due to economies of scale with the 2170s once they use them in the Model S and X.

And the AP hardware will be much cheaper and the software costs will be amortized over a much bigger volume of cars.

Eventually they'll replace the MS-MX platforms to be compatible with alien dreadnaught production technology which will allow them to reduce their costs drastically. The question is how much of the alien dreadnaught technology that they can add incrementally to the existing production lineup (I'm sure it's substantial), and how cost effective that is vs replacing the entire MS-MX platforms.

 

[jbcarioca]

I largely agree with FrozenCanuck and most of the rest of this thread. I think, though, there is an easier way to look at Tesla model evolution:

1. Platforms- in traditional vehicle building context are indeed changing 1. Roadster, 2. S/X, 3. Model 3, 4. Model Y. Roadster could be considered a technology demonstrator, but let'd call it platform 1 anyway. That really means the battery casing and chassis points, mostly, with some electrical componentry. For practical purposes one might extend platform to mean things purchased from third parties that are essentially generic, so 12V electrics, for example.
2. Manufacturing- in this context 'platform' means anything that can run on the same assembly line with mostly same techniques. Thus S&X are the same platform but with ~30% parts commonality.
3. Model 3, described as .5 in factory automation, could not be built without 12V electrics because the volumes had not yet been credible enough to justify suppliers building custom components at cheap cost. It also was designed before the entire notion of automated construction had been fully defined. Thus it is an interim solution, but at volumes that make it exponential in Tesla terms and global impact.
4. Model Y will arrive simultaneously with widespread premium BEV production, so suppliers will be willing to produce for everyone electrics compatible with main battery power, probably at a pretty standard approach, with Tesla, BD, BMW, VW will agree on, so others will follow. The platform difference in Model Y probably will include major simplifications (imagine, say, regenerative braking with zero friction brakes) that will form a core physical base that will in fact be a new 'platform for upcoming pickup, S and X replacements and probably even new Roadster. The 1.0 in factory automation certainly includes the basis for continuing rapid evolution without needing completely new lines, so 'platform's the correct word.

The larger point is that 'platform' means only compatibility with a given factory automation solution, that is, run on the same line, but nothing more or less than that.

The other elements can mostly be platform independent. These can be giant advances in efficiency of electrical components, vehicle automation and other topics. Tesla has redesigned a huge number of components that can be retrofitted to any given platform while others really cannot be retrofitted. My speculation is that move away from 12V is so far reaching that it needs to be in new platforms. I also think the 2170's can be retrofitted so S and X will get them just as soon as production volumes permit it, Panasonic will be happy to accommodate the change. JB and Elon have mentioned component improvements many times.

Imagine, say, the motor manufacturing process for Model 3 yields much lower cost, higher durability and better performance. It is quite probable that S, X and everything else would transition to new motors to spread that benefit. They might not even mention the change.
That is similar to the switch from titanium to carbon fiber radiator mounts in the S. That was never mentioned, but it saved ~$150 per car, was lighter, stronger and cheaper. I suspect we'll see continuous improvement like tat in every model Tesla. Unless a change necessitates a new assembly line it will not be called a 'platform' change. If it does not make a user visible change they might not even announce the change.

Software changes will almost never be traded as physical component changes. But, just as with other computers, there will be the odd CPU upgrade, memory upgrade or peripheral upgrade. Elon hinted that when discussion level 5, "if we need another processor, no problem, there's room for another one..."

Tesla discusses manufacturing and technological change much more than do most vehicle makers and terminology is an odd mixture of automotive and Silicon Valley terminology. I think that is why people are constantly a little confused about terms. Why was there not "production prototype" instead or in addition to "release candidate"? What does "platform" mean? Why can AP1, AP2 be released and have functionality done later. It's bizarre in automotive terms because Tesla releases when they have hardware, traditional automakers do Hardware and software and never improve. Then Tesla continually tweaks software to do more and even gives us "more cowbell".

I wonder how Model 3 owners will adapt to this approach, beloved to many of us, detested by some?