- Autowist
- Posts
- GPT Paradigm Shift in Automotive Design, Foxconn, Google Antitrust
GPT Paradigm Shift in Automotive Design, Foxconn, Google Antitrust
GPT -> Paradigm Shift in Automotive Design
Generative Artificial Intelligence (AI) has emerged as a transformative tool in the field of design. By combining optimization techniques with a sense of aesthetics, generative AI has the potential to revolutionize various industries, including automotive design. While, optimization alone can lead to pretty horrendous looking design.
Future Insight: There are layers of constraints that go into a car design starting from aesthetics, legal bounds for wheelbase, tonnage etc., drag coefficient, design for manufacturing and assembly. Now that's an interesting one. Because then it is not only an optimization problem + aesthetic considerations. But also a practical problem. To generate the design together with it's assembly and manufacturing instructions. Moreover, by integrating assembly instructions into the generative AI process, designers can create designs that are not only visually appealing but also optimized for efficient and cost-effective manufacturing.
Same goes for carbon footprint, recyclability, repairability, generating assembly instructions. Imagine if there is a design suite that inputs all kinds of parameters, constraints, including both engineering and practical ones. And the suite outputs designs, instructions for manufacturing, carbon footprint assessment, assembly instructions for workshops, material suggestions, sourcing options and a business case all in one workflow. Now that is a true "unlock" of productivity.
For example, a designer can input through text for the sketch to look more sleek or SUV-like and modern. The team focused on how the AI tool can be used to help improve aerodynamics, a critical factor in EV range.
TRI researchers released two papers describing how the technique incorporates precise engineering constraints into the design process. Constraints like drag (which affects fuel efficiency) and chassis dimensions like ride height and cabin dimensions (which affect handling, ergonomics, and safety) can now be implicitly incorporated into the generative AI process. The team tied principles from optimization theory, used extensively for computer-aided engineering, to text-to-image-based generative AI. The resulting algorithm allows the designer to optimize engineering constraints while maintaining their text-based stylistic prompts to the generative AI process.
Interpreting and Improving Diffusion Models Using the Euclidean Distance Function, F. Permenter, C. Yuan, 2023.
Drag-guided diffusion models for vehicle image generation, N. Arechiga, F. Permenter, B. Song, C. Yuan, 2023.
Toyota will use AI to maximize future EV range (Link)
Stellantis x Foxconn
Vertical Integration in the "new" automotive landscape is a trend that I have written about before. In this new joint venture between Stellantis and Foxconn, called SiliconAuto. The OEM aims to design purpose built semiconductors for their ECUs. This fits very well into their STLA Brain EE Architecture that they unveiled a while ago.
Insight: Stellantis intends to control the full stack. They are done being at the mercy of suppliers and the tiered supply structure in the automotive industry. Definitely a bold move. The trigger for such a bold move could be all the problems OEMs were in due to semiconductor shortage. And analysts all around the world propping up that shortages are here to stay. In addition to all that, the geopolitics angle, to reduce dependence from China.
Enter Foxconn.
Foxconn (also known as Hon Hai Technology Group) is famous for being suppliers to Apple. They produce iPads, iPhones and even gaming consoles like Xbox. Foxconn has a global footprint even outside of China, in India, Indonesia, Brazil. All markets where Stellantis have presence in. They are veterans in large scale operations. Although volumes might not match that of consumer market, Stellantis is a volume player with their 16 brands in total.
Stellantis and Foxconn announce joint venture to produce next-generation EV semiconductors (Link)
Tesla Manufacturing Everywhere
2 headlines last week were about Tesla manufacturing and here is a quick breakdown.
Gigacasting: Tesla uses house-sized die-casting machine to produce aluminium parts larger than any other OEM in the industry.
But why?: Well, larger parts means fewer parts. And that means lower overall costs.
Insight: There are a few companies that produce these machines. Italy-based IDRA, Buhler Group in Europe, Ube and Shibaura Machine in Japan, and Yizumi and Haitian in China. Other OEMs are planning to introduce this for their own production plants. One reason why Tesla has managed to to use this effectively is they have a rather small product line. That is unlike all other major OEMs who have multiple models, updated every other year with a facelift. This could in turn make OEMs boil down their existing product lines and make cars that look even more similar than they were before.
Giga Laboratory Concept: is a retail concept Tesla unveiled last week. The retail store is hsown with a large robotic arm right in the middle. In the announcement Tesla says:
The country’s first “Tesla GIGA Laboratory” made a stunning debut. Tesla’s “Most Beautiful Store in Chengdu” unveiled the mystery: An immersive experience of the magic of “building a car in 45 seconds”. Meet up with friends “Chat and take a break”.
Now does this mean a VR experience, or an actual assembly/mounting sequence is not clear. But one thing is clear, retail stores are key for Tesla sales. Although the OEM has said they will focus on online sales in 2019. The number of retail stores around the world has seen a steady growth.
Why are other automakers chasing Tesla's 'Gigacasting'? (Link)Tesla unveils new Giga Laboratory concept (Link)
Google Antitrust (Automotive edition)
What?: The EU has made a formal antitrust complaint against Google and its Ads Service
Why?: Google maintains itself as a popular choice among OEMs as a platform for their in-car infotainment services. Google Automotive Services(GAS) includes the play store, voice assistant, navigation etc. that OEMs can integrate into their infotainment systems. EU claims that being in this dominant position in the market allows Google to favour it's own Advertising intermediary AdX (Ad Exchange) over others.
Source: EU Commision Twitter
What could this mean for Google?: Well the EU has suggested that if the litigation stands as is, then there is no other option but for Google to sell of some parts of it's business.
Google's in-car tech sales face German antitrust crackdown (Link)EU suggests breaking up Google’s ad business in preliminary antitrust ruling (Link)
Strides in Software Defined Vehicle
Polestar forms a JV with Xingji Meizu, a tech company that is also part of Geely.Strategy: They will cater primarily for the Chinese market. To keep up with the Chinese in-car functions that is becoming more deeply integrated with Super Apps such as WeChat al the way to in-car Karaoke.
The plan is to use the Flyme Auto platform from Xingji Meizu, and adapt it to meet the requirements of Polestar.
Source: Flyme OS
Nardò Technical Center (NTC) establishes a new division focused on software development for intelligent and connected vehicles. With the new division, the proving ground in Salento (Southern Italy), which is owned by Porsche and operated by Porsche Engineering, further underlines its positioning as technology partner for the integrated development and validation of future vehicles.
Strategy: Apart from the obvious that Porsche is doubling down on software features, there is another interesting aspect. In the latest VW Presentation they announced that VW, Skoda, Audi and Porsche will merge into their common SSP BEV Architecture.
Now the question is whether Porsche is actually hedging their bets, given the turmoil CARIAD has been in the past few months. Or is the plan with the new technical center to further augment and enhance deliveries from CARIAD and to tailor-make them to Porsche.