Hacker News Reader

Comment by cs702 | Hacker News Reader

So far, only Tesla is using giga-presses to reduce the number of parts that must be welded together in each vehicle and lower costs. Other car manufacturers have barely started exploring the possible use of these giga-presses, and seem reluctant to transition away from time-tested approaches to manufacturing. Quoting from the OP:

> After initially considering die casting for its upcoming Trinity model, Volkswagen (VOWG_p.DE) has backtracked, while BMW (BMWG.DE) has never expressed an interest. Ferrario said the auto industry tended to be conservative and that no one liked upending established processes, but he rejected idea that die casting posed a risk to jobs at carmakers, noting body-making was already highly automated.

Thanks in part to its use of these giga-presses, Tesla currently has the highest profit margins of any mainstream car company. It's not a coincidence.

View on HN

Is it the gigapress that allows Tesla to have high margins?

Could it also be that EV cars are subsidized around the world by governments? Maybe Tesla's "self driving beta" addon adds to the profit more?

I genuinely don't know.

https://english.hani.co.kr/arti/english_edition/e_business/1...

> The secret behind Tesla’s 30% gross margin

> The company's Q3 sales rose 58% year-on-year despite a 6% decrease in average selling price, allowing for high margins for the electric vehicle manufacturer

> Analysts point to another factor behind Tesla's high profit margin. Park Hyung-keun, a senior researcher at POSCO Research Institute, said, "Through vertical integration by directly being involved from floor design to parts supply and demand, production and service, Tesla has helped reduce costs by raising the degree of its parts integration and cutting overlapping costs."

> Tesla's unique structure of vertical integration, ranging from the development of semiconductor chips, software and batteries for electric vehicles, to charging, unmanned driving and insurance services, helps lower costs. Its “do-it-all” approach simplifies the automotive production process in a manner resembling that of electronic products. In contrast, other automakers actively utilize production outsourcing to diversify vehicle quality risks and raise output efficiency.

> A leading example is Tesla's “giga” aluminum die-casting process. A Giga Press weighing more than 1 giga pound (400 tonnes, or around 900,000 pounds) stamps the entire rear chassis of a car with a large aluminum alloy. About 70 metal plates can be welded to the chassis, but giga casting can simplify the process and slash production costs by about 40%. This is why Tesla electric vehicles have recently reduced panel gaps issues — defects caused by misaligned steel plate seams.

So instead of figuring out "precision" welding (maybe SpaceX could have helped out?), Musk used pounds instead of kg or tons in order to implement "Giga"presses? Well, at least the manufacturing inexperienced Tesla crowd has something new to rave about... Still no < 35k Model 3 so.

They found a cheaper, faster, more efficient way to make structural components. SpaceX already shared with Tesla how to friction stir weld [1] (and that technique is still used currently [2]), but that is not as fast as stamping out body components. Simplicity is the ultimate sophistication. There is no extra credit for making your manufacturing process more complex than necessary (if you want to show off, fire your EV into space). Tesla's manufacturing ramp rate is primarily a function of how fast the org can improve the speed and process at which atoms travel through the manufacturing process, feedstock to final product.

[1] https://electrek.co/2015/05/24/spacex-transferred-novel-weld...

[2] https://www.teslarati.com/tesla-model-y-spacex-welding-techn...

Tesla tried full automation, which didn't work. Now they use pressed parts limiting the use of platform startegies, which enable all other manufacturers to produce a multitude of models, including loe cost ones, using the same tooling, parts and R&D. I guess we will see were that ends.

Why would, or should a business decide to make less money selling a cheaper version of a product when demand for the more expense (and hence more profitable) product shows no sign of waning? Don't get me wrong, I'd love to buy a $35,000 Model 3, but it makes no business sense, especially considering the fact that the average new car cost is something like $44k these days.

Just food for thought… $35k in Feb 2019 when Tesla announced the “$35k Model 3” is now just under $42k simply adjusted for inflation.

All EVs are subsidized and pretty much all of them (except Tesla) are loosing money while Tesla has industry leading gross and operating profit margins.

So it's definitely not subsidies.

On a definitional level, gross profit is difference between price and cost.

Tesla has biggest difference which implies that it's a combination of charging higher prices and having lower production cost that other.

Only EVs assembled in the US can receive the full subsidy now, ever since the Inflation Reduction Act came into force.

All auto manufacturers are subsidized in some sense. Remember that GM was straight up insolvent in 2009 and had to be bailed out with enormous federally-guaranteed loans.

But none of that discounts the value of talking about engineering activities involved in reducing the costs of production, because that stuff matters too. And yes, single-piece chassis are absolutely part of that, as is the minimalist interior (the cockpit BOM for a Tesla is a tiny fraction of what you see on competing EVs), the ongoing sensor fusion architecture (yes, everyone loves to scream about it here, but the upshot is that Tesla doesn't pay for the radar units everyone else uses), etc...

They're actually extremely trim vehicles from an assembly perspective. There's a Munroe video out somewhere where he estimates production costs for a Y vs. a Mach-E and figures there's something like a $10k advantage for Tesla.

Telsa has been relentless in simplifying the part count of a tesla. Each part needs to designed, tracked, inventoried, installed, verified, made available for repairs, increases the size of the factory, and decreases the rate of production. Added parts also require space to install, clearance, sometimes custom tools, testing, might create vibration/wear/noise that requires mitigation, etc.

The front and rear chassis used to require 100s of robots and parts, and then alignment would have to be checked, more material used, more time, more robots, etc. Replacing the front and rear chassis removed 300 robots ... each, and of course the factory space for the robots, and the time for the assembly line to run past those robots.

The glass roof, which IMO isn't a big feature, can be installed after the chassis is built. Enabling robots to install the dashboard, central console, seats, etc before the roof is installed.

On most cars the dashboard is complex, multilayered, complex set of sensors, displays, spinning needles (speed, rpm, temp, fuel levels, etc). On a Tesla the dash is built on a big square straight piece of aluminum, a simple slot for airflow (no fancy/fiddly air flow controls), and a 15" display. Likely reducing 100s of parts.

The octovalve and related cooling system is a marvel as well, nicely integrating heating/cooling of the cabin, motors, and battery. Dramatically simpler than competing solutions, in particular the Ford Mach E and Lightning.

Most cars have an extremely complex set of CPUs, sensors, and control distributed all around the car. Chips in doors for locks/windows, a separate system for ABS (usually from bosche), networks of temperature/pressure/movement sensors, torque controls/monitoring for window motors, windshield wipers, airflow motors, etc. In the model 3 it's largely integrated into a single board at a level of integration Munroe claims they have seen in no other car, or even any product ... outside of a satellite.

This might sound like hand waving and marketing, but one metric that supports this is the cars products per square foot of factory per hour. Tesla is way ahead. It's also supported by high profit margins, even when compared to companies that have as high or higher prices.

So yes I'd say that the gigapress is a key part of reducing the part count, assembly time, and number of robots required to build a car. The result is a Tesla factory builds more cars than the competitions factory of the same size.

The underlying trick to this, in part, is the fact that Tesla's making exactly 4 models (S3XY). The models are long-lived. This lets them invest in specialized parts that reduce the cost (significantly!) of their vehicle. I read a stat somewhere that the Y takes 10 hours to build.

Disclaimer: I respect Tesla's manufacturing prowess, I despise the self driving claims, and think the vehicles are hideous.

Indeed. Much like Apple. Apple has 4 ish laptops, that generally don't change much year to year. Compared to Dell or HP who have MANY more models, often seems like 4 models or so per market (edu, business, consumer, gaming, etc.) and are refreshed often.

Agreed on the self driving. Not a big fan of the Tesla look, but after living with one I'm a fan. Then again I've had some ugly cars, early Acura GSR (with the tiny headlights), early Subaru WRX, and early Forester Turbo. None would win any beauty contests. Model 3 generally seems like the look is determined by a wind tunnel. At least the model S (at least in some trims) is pleasantly curvy and has some style.

Thats just standard manufacturing

> Most cars have an extremely complex set of CPUs, sensors, and control distributed all around the car. Chips in doors for locks/windows, a separate system for ABS (usually from bosche), networks of temperature/pressure/movement sensors, torque controls/monitoring for window motors, windshield wipers, airflow motors, etc. In the model 3 it's largely integrated into a single board

yeahnah, thats deffo not true.

tesla has a similar number of sensors, actuators and other junk. It even has an ABS pump too. How do they think they measure torque if there aren't any torque sensors?

Having everything wired directly to a single board makes the wireloom really really unwieldy. Telsa use the same automotive busses everyone else does. (it probably uses ALL THE BUSSES, because why not.)

the reason the reason tesla make a profit is that they've been value engineering the same three cars for close to ten years. however, they have only been making a profit since 2020.

The other key is producing a car to a standard of something retailing for less than half the price. The top end Kia EV 4 feels much more "posh" than a model y, which is much more expensive. The killer feature of the EV 4 is that its waterproof.

I think the biggest reason for Tesla's gross margins is that millions of people want EVs for various reasons (gas prices, environmental concerns, fun, status) and Tesla is one of the only companies making them in large quantities. They don't have to be the best (even though they probably are on many metrics). They just have to be available and they can kind of set their price.

It's one of the reasons, but not the only one.

That's why I wrote "Thanks in part to..."

Everything is subsidized.

I think it's vertical integration in general that drives their profit margin. It's why Teslas are good but not great, and will rarely be the best at any particular thing. Brewing up your own rain-sensing wipers saves on component costs, but is inferior. Building your own seats saves on costs, but is inferior. They're playing for the middle. Not a bad strategy at all if you're going for profits, so it makes sense.

It'll be interesting to see how it all settles out compared to the legacy manufacturing style of subbing out components. Some people will pay extra for Recaros (me!) but a lot of people couldn't care less.

Never seen a Tesla review claiming the seats were inferior. In a few they were claimed they claimed they were among the best, but most reviews don't mention the seats in particular.

The seats in my X and 3 are the only car seats I've tried that does not give me back pain after long trips.

They are one of the biggest reasons i drive tesla instead of the etron.

They're suspected of underallocating funds for warranty repairs. Those funds would normally eat into the profit margin, but at Tesla they're supposedly accounted as good will service, so it comes out of a different post.

Yes and no. Automakers reluctance to new design methodologies is coming from a place of safety - the "traditional" engineering mindset. In civil and mechanical engineering, it takes a long time to be fully confident that you've found all the edge cases that might cause catastrophic failure - and catastrophic failure in certain industries (Automotive industry included) generally means death.

Tesla is coming from a "software" engineering mindset. Move fast, break things, etc, etc. This is far easier when dealing with bits and with code. The likelihood of death and dismemberment is also far lower if you make mistakes.

> Other car manufacturers have barely started exploring the possible use of these giga-presses

Lets be clear here, tesla are doing this for the rear subframe: https://electrek.co/2021/01/11/tesla-starts-production-model...

And they want to scale it up, from what I see to the whole bottom of the car. This seems like a big reaction from how they used to do things with a billion different fasteners.

Now, I think why the other companies don't immediately jump on this is that making a die that big is really expensive, and I imagine high maintenance. Tesla will probably get away with it because they don't really care all that much about tolerance, so will run the die a lot longer than a safe manufacturer.

Also, tesla don't actually make that many models so they only need a limited number of presses and dies to be effective.

They also do the front subframe for the Model Y at some factories.

https://www.notateslaapp.com/news/1037/a-peek-at-tesla-s-sin...

Reply on news.ycombinator.com