Comment by Animats | Hacker News Reader

Comment by Animats | Hacker News Reader

That's bill of materials cost. What's killing Tesla is labor cost in assembly. Tesla has about 10,000 employees in Fremont, to produce about 3500 cars a week., per Bloomberg.[1] (No, not 5000 cars; they got up to 4500 once, though.)

Modern auto assembly plants typically run about 15-30 labor hours per car.[2] Tesla's Fremont plant is somewhere over 110, even in good weeks. In bad weeks, the ratio is much worse.

The point of the teardown is that the design isn't inherently expensive. If Tesla ever gets their assembly plant working properly, they could make money.

Musk commented "go to Ford, it looks like a morgue", after apparently visiting the Ford Rouge plant. Ford's PR person replied "No doubt the vibe is funky in that 'makeshift tent,' but it's not bad either across the street at the #FordRouge plant where a high quality, high-tech F-150 rolls off the line every 53 seconds like clockwork." That's what's supposed to happen. Cars coming off the line at a steady rate, union workers not being worked to death.

[1] https://www.bloomberg.com/graphics/2018-tesla-tracker/ [2] https://www.motortrend.com/news/toyota-chrysler-have-north-a...

Not true, Munro does labor of cost assessments as well, not just BoM. If Munro is saying car is profitable then I am pretty sure he factored that in.

Does it factor in the warranty maintenance? Tesla seems to have a high rate of warranty work from noisy drive units to retrofitting titanium armor to all the battery packs.

You're referring to changes made to Model S vehicle manufacturing over half a decade ago. The lubrication of the drive units has been automated, and battery pack armoring is built into the current cost model for all production vehicles.

Other reviews have predicted that the Model 3, itself, is profitable. These say Tesla's bigger financial issue is the Supercharger network. Ideally, they would pay for Supercharger through perpetuities - or perpetual annuities - from interest on part of the profit from the car; but the invested profits are sequestered from operations' use.

Is your figure of 10,000 employees just for the Model 3 lines, or is it for the entire factory? Because they’re making about 2,000 Model S and X per week, so your numbers need to account for that if it’s the entire factory.

Animats has been repeatedly told that mixing the labor numbers for low-volume high-end aluminum-bodied cars with some parts insourced with a new high-volume line with kinks to be worked out is meaningless. But don't worry, he will probably be making the same comment next month, and the month after, in rotation with his other usual negative Tesla comments he makes in every Tesla thread.

Your tone is definitely negative, but, in this case, I'm afraid that it might be warranted. Its hard to imagine being this far off on production numbers by accident.

I'm all for having a conversation about Tesla's staffing numbers. Seeing the same comments repeated with absolutely no interaction, post after post, that's not conversing.

I cannot comment either way on Animats' thoughts (because I do not have enough information to do any concrete analysis).

However, for what it is worth, Bloomberg's latest article on the trials and tribulations of Model 3 production (https://www.bloomberg.com/news/features/2018-07-12/how-tesla...) indicates that "compared with Toyota...even with all the robots, Tesla spends more than three times that number of hours on each car".

Overall, I thought the Bloomberg article cited was pretty interesting/revealing and, crucially, it seemed pretty balanced.

Where I think that Tesla might have some issues is on Model 3 quality. The "burst periods" may be contributing negatively to that. But that remains an open question in terms of how it compares with competitors' first model year problems and industry standards overall (information that I do not have). There is some anecdotal evidence in the "Tesla forums" of strange quality issues. How representative that is, again, an open question in my view.

Tesla does more than final assembly, so it's an apples-to-oranges comparison. Plus, aluminum-bodied cars are famous for needing more human labor. Bloomberg fails to go into that, which makes for bad journalism. Great to have an actual discussion about it... for a change.

> Tesla does more than final assembly, so it's an apples-to-oranges comparison.

True, but, this estimate of labor was provided by Michelle Hill, a manufacturing expert at management consulting firm Oliver Wyman who is likely aware of that and, I would think, would add that fact into their analysis. Naturally, I cannot be sure that Michelle did this (because I do not have access to the report), but it is probable I suppose given their time in this industry.

> Plus, aluminum-bodied cars are famous for needing more human labor.

As someone who has worked in automative assembly (I design automation equipment and industrial systems), the "body line" where most of the "aluminum handling/assembly/welding" is typically highly automated (> 95% by time). This is true of Tesla now I think (https://electrek.co/2018/06/08/tesla-model-3-body-line/) and very much true of several of the big name automotive manufacturers that I have had the occasion to work with.

Again, it is difficult for me to make broad generalizations of Tesla's manufacturing situation. Mr. Musk has indicated some "over-automation" issues (a pretty classic amateur manufacturing mistake). The "tent" thing was strange, but I think it was a necessary evil likely in terms of the lack of capacity planning at the get-go. So Tesla did what it needed to.

If I were Tesla, I would take a serious look at quality over production volume (not implying that they are neglecting quality). Mandatory recalls and serious/safety quality issues in terms of consumer perception can bite much harder than just meeting a 5,000/week production volume.

F-150s are fully aluminum and they do pretty well turning those around.

Yep, but do they do the labor-intensive stuff in the plant that does final assembly? Given that it converted from steel to aluminum in a couple of steps (initially, just the hood was aluminum) and stayed at the same volume, I'd guess "no way".

It's much-discussed in the industry.[1] Tesla Fremont is running 3 shifts, 7 days, to get about 3500 cars a week out the door.

[1] https://www.forbes.com/sites/joannmuller/2018/05/01/no-way-t...

Except that their current production rate, according to your previous Bloomberg link and the well-known and little-changing S and X numbers, is 5,500/week.

Just one day later, and the Bloomberg estimate is 4,330 Model 3's per week, plus the usual 2,000 S/X. Good timing for getting a pessimistic comment in.

It’s easy for me to imagine. All attention is on the Model 3. You see a figure like “3,500 cars/week” and it’s easy to assume that’s everything they’re making, and forget they have two other models. Of course, this becomes less believable if it persists after correction....

It's always worth noting the disclaimer that Bloomberg has on their tracker:

> Given the exponential trajectory of a production ramp up, there may be sudden increases in output that this model is slow to catch.

When you're looking at Tesla's numbers, the tracker becomes more reflective of progress several weeks after a production peak. Given that the 5,000 peak was hit relatively recently, this means that we should expect the tracker to reflect this substantially later.

It's triple the labor hours, so you can happily assume a higher production ceteris paribus and it still looks bad.

I'm not sure where you got the idea that it is "triple the labor hours"?

"At a typical plant run by Toyota Motor Corp., widely seen as the most capable carmaker, a new car requires about 30 hours of labor. Even with all the robots, Tesla spends more than three times that number of hours on each car, says Michelle Hill, a manufacturing expert at management consulting firm Oliver Wyman."

https://www.bloomberg.com/news/features/2018-07-12/how-tesla...

If Tesla makes more components in-house than other automakers, that could shift costs from parts to labor without affecting overall costs. So I don't think having more employees or hours per car is evidence they're dysfunctional on the face of it.

Tesla produced 53,339 vehicles in the most recent quarter. That works out to ~4,444 vehicles per week. Also, the vehicles per week figure is increasing.

To be fair, they would only produce ~231k/year at this run rate, which would be disappointing for a factory that is capable of ~430k/year. Hopefully their production rate will improve considerably by the end of the year.

The core flaw to most negative comments on HN since the Model S was first announced or produced is: the skeptics pretend there will be zero volume improvement and Tesla is going bankrupt soon. If you remove those two absurdities, their arguments always implode.

With the Model S it was routinely claimed that because Tesla got off to a slow production start, that they'd therefore never manufacture the car in any volume. The same ridiculous argument was then wheeled out again for the Model 3. It will be wheeled out for the next model that has a slow production start or production problems. Which points to the obvious conclusion: it's about attacking Tesla, it's not about logic, reason, facts.

The Model 3 starts slow on manufacturing. The skeptic says: see, they've failed, it's over, they'll never produce it in volume. The Model 3 weekly production moves up to 500 or 1,000 units. The skeptic says: see, they have no idea what they're doing, they can't produce it any real volume, and besides they're going bankrupt in 1Q19 or sooner anyway. The Model 3 weekly production moves up to 2,500 or 3,000. The skeptic says: yeah but the car sucks, they're building them in tents, and besides they'll never manufacture 5,000 per week, and they're going bankrupt anyway. The Model 3 production moves up to 5,000+ units per week. The skeptic says: yeah, but they'll never make 10,000 per week, and they're still going bankrupt, they can't make money on the car.

That's how the Tesla attack, goal post moving has worked thus far. And it'll perpetually continue to work that way, they'll just move on to the next thing to focus on as an attack.

> The Model 3 production moves up to 5,000+ units per week

Do you have evidence that this has happened ?

Because from all accounts the 5,000 number was a once-off.

53,339 vehicles, but only 28,578 Model 3 sedans or ~2.3k per week. The 5k per-week goal is for the Model 3, not all vehicles combined.

‘Hi-tech’ F150? A cheap truck that doesn’t have to comply with the same safety or emissions standards as cars? It’s a very different beast to a BMW, let alone a Tesla..

Cheap? You clearly haven't looked at a dealership lately. F150s start around $30k for the most basic two-wheel drive v6s. Models people actually buy are closer to 40k, 45k new.

There's a reason Ford quit making sedans.

Reply on news.ycombinator.com