BYD - Presselinks WKN: A0M4W9 ISIN: CNE100000296 Forum: Aktien User: Morpheus33

Automakers' share of China NEV exports in July: BYD leads with 32.2%, Tesla rises to No. 3 with 12.3% BYD continued to dominate China's NEV export market in July with 173,721 vehicles, while Tesla China exported 66,330 units. http://cnev.co/BZn5VAI

Nearly 90% y/y growth for the world's biggest car exporter is a bit crazy. A stat I just dug up. From 1970 to 1980 Japan's auto exports increased from 1m cars to 4.5m cars, so +3.5m over ten years. China's car exports are poised to jump 3.5m in 2026 alone!

#BYD has published six additional solid‑state battery patents. 🔋 These focus on maintaining stable electrolyte‑electrode contact during cycling, and also cover material development, manufacturing, and quality control. Among the six, four are materials‑chemistry‑related: ▶️CN122494747A – Dual‑electrolyte approach: an electrochemically stable ion‑conducting layer between halide and sulfide electrolytes to prevent unwanted reactions and extend battery life. ▶️CN122494554A – Solid‑state electrolyte buffer layer between cathode active material and sulfide electrolyte to reduce thermal runaway risk. ▶️CN122494567A – Dual‑layer cathode composite: inner layer uses single‑crystal cathode material combined with solid electrolyte; outer layer uses polycrystalline material with different electrolyte composition. Single‑crystal resists cracking; polycrystalline offers better rate capability but is prone to intergranular cracking. The design combines the strengths of both while mitigating their drawbacks. ▶️CN122494558A – Three‑material approach for active cathode: metal oxides and sulfides are first lithiated, then mixed with solid‑state electrolyte. This reduces interfacial resistance, improves capacity, rate capability, and cycle life. ▶️CN122494552A – Ionic liquid wetting agent in the active electrode layer, distributed in a gradient (higher near the surface). Improves solid‑solid contact, ion transport, and battery life. ▶️CN122494553A – Quality control parameter “Re” measuring cathode‑electrolyte contact. Requires at least 60% of cathode particle perimeter to be in contact with electrolyte to improve performance, capacity, and lifespan.

Automakers' share of China's NEV market in July: BYD leads with 23.5%, Tesla falls out of top 10 Nio's retail sales surged 70.5% year-on-year in July, keeping it at No. 8 in China's NEV market. cnev.co/sszIg1e 👇

Facing sales pressure, Fangchengbao added a new entry-level trim for the Ti 3 at 143.8K RMB ($20.1K). The RWD base model retains second-gen Blade Battery flash charging while adjusting range to 510 km, explicitly designed to lower the buy-in threshold. When the Flash Charge edition arrived in March, Fangchengbao trimmed the lineup to just two variants, which pushed starting prices up. Monthly sales haven't matched earlier levels since, settling around 5,400 units—a direct reflection of that higher price tag.

BYD's large Shenshan plant sits in its special industrial zone which used > 1 TWh in 1st half of the yr. 48% (490 GWh) were used to produce NEVs! It sure requires a lot of electricity to do anything. Shenshan local govt provided the full infrastructure to attract BYD's major investment in the area.

BYD sales in Indonesia hit 6569 in July + another 159 w/ Denza brand BYD Indonesia continues to increase localization of its Subang plant & expects to hit the 60% TKDN target in 2027. BYD is now #3 in Indonesia. Toyota is the undisputed leader there.

https://de.motor1.com/news/804665/byd-seal-u-dmi-meisterverkaufter-plug-in-hybrid/

https://x.com/i/status/2087551150371311694

Hier ist die einfache und genaue Erklärung, worum es in dieser Arbeit geht:Der Titel und das ZielDas Paper stellt ein neues KI-Modell namens HyWorldVLA vor. Es handelt sich um ein sogenanntes Vision-Language-Action (VLA) Modell. Das bedeutet, die KI verarbeitet Bilder (Vision), versteht Sprache (Language) und steuert daraufhin das Auto (Action).Das Problem, das gelöst wirdBisherige Systeme versuchen vorherzusagen, wie sich die Umgebung des Autos in der Zukunft verändert. Dafür nutzen Forscher meist zwei verschiedene Ansätze, die beide Vor- und Nachteile haben:Pixel-Ebene (Pixel-level): Die KI versucht, die Zukunft als echtes Bild (Pixel für Pixel) vorherzusagen. Das ist sehr präzise, geht aber bei schlechtem Wetter (Regen, Nebel, Dunkelheit) schnell kaputt.Latente Ebene (Latent-level): Die KI wandelt das Bild in mathematische, abstrakte Codes (Zustände) um. Das ist robuster gegen Bildstörungen, aber für Menschen schwer nachzuvollziehen und verliert oft wichtige Details.Die Lösung: Ein Hybrid-AnsatzDas Modell HyWorldVLA kombiniert nun beide Welten (daher „Hybrid“) in zwei Schritten:Vortraining (Pre-training): Die KI lernt, wie ein Video aufgebaut ist. Sie sagt abstrakte Video-Codes voraus, rekonstruiert das Video aber gleichzeitig auch als echtes Bild, damit sie die genaue Umgebung versteht.Feinabstimmung (Fine-tuning): Die KI nutzt die robusten, abstrakten Merkmale, um sie an ein „Aktions-Modul“ weiterzuleiten. Dieses berechnet dann die sicherste Fahrkurve (Trajektorie) für das Auto.Das ErgebnisDie Entwickler haben das System auf bekannten Teststrecken für autonomes Fahren (NAVSIM v1 und v2) geprüft. Das Ergebnis zeigt, dass dieser kombinierte Ansatz deutlich besser und sicherer fährt als die bisherigen Methoden – besonders dann, wenn „Rauschen“ (wie schlechtes Wetter oder Bildstörungen) ins Spiel kommt. Zudem haben die Forscher einen neuen Maßstab (Benchmark) aufgestellt, um zu messen, wie gut solche Systeme mit Bildstörungen umgehen können.

Last month, the AI team at #BYD's New Technology Research Institute published its first paper, proposing a hybrid world model called HyWorldVLA for autonomous driving. The model combines pixel-level and latent-space world modeling and adopts a Vision-Language-Action (VLA) architecture. On the public autonomous driving benchmark NAVSIM v1, the model achieved a PDMS score of 90.59, setting a new record. On NAVSIM v2, which features a more comprehensive set of metrics, HyWorldVLA also achieved a leading score of 89.71.

Chengdu Auto Show opens next week with nearly 120 brands, 1,600 vehicles The 2026 Chengdu Auto Show will cover 220,000 square meters, making it China's first A-level show of the second half of the year. http://cnev.co/TiInL7j

According to Schmidt Automotive Research, #Porsche has officially exited the Volkswagen Group's CO₂ emissions pool. To meet European CO₂ emissions targets, the German sports car brand has concluded a carbon credit agreement (an “open-pool”) with Chinese automaker #XPeng for 2026 and 2027.

In Jul, China's PV exports reached 877.9k units, +4.7% MoM and +86.4% YoY.🇨🇳 • #BYD ranked 1st with 161.4k units, up 130.9% YoY, more than doubling its exports. • #Chery ranked 2nd with 121.9k units, up 37.3% YoY, maintaining its strong leading position. Tesla ranked 3rd with 66.3k units, up 143.2% YoY. • Several brands recorded remarkable growth, including Zeekr (+391.4% YoY), Leapmotor (+281.5% YoY), and #Xpeng (+223.3% YoY), all accelerating their export expansion. https://x.com/thinkercar/status/2087472946038366453?s=20

In Jul 2026, China's PV wholesale volume reached 2.3M units, +4.5% YoY, -4.2% MoM. 🇨🇳 • #BYD led the ranking with 349.7k units, up 11.0% YoY, maintaining steady growth on an already massive base. • Chery secured 2nd place with 159.8k units, up 6.2% YoY, showing stable performance. #Geely Galaxy took 3rd with 107.8k units, up 14.8% YoY. • #Leapmotor followed closely in 4th with 101.3k units, soaring 102.0% YoY, and is now just behind the 3rd‑placed Galaxy, narrowing the gap significantly. • Toyota ranked 5th with 93.7k units, but dropped 22.5% YoY. https://x.com/thinkercar/status/2087464145298018380?s=20

China's NEV sales rise 23.7% in July, with exports surging 1.5 times, CAAM data shows NEV sales were 1,561,000 units in July, accounting for 60.4% of total new vehicle sales, topping 60% for the first time, according to CAAM data. http://cnev.co/8C4S4Cx 👇