News

BMW Is Teaching AI How to Crash Cars Faster—and Smarter

Leave a comment

Artificial intelligence is coming for the auto industry in ways that go far beyond chatbots and touchscreen voice assistants. BMW’s latest move proves the next big AI battleground may actually be hidden deep inside the virtual crash lab.

The BMW Group has announced a partnership with French AI startup Mistral AI aimed at transforming the automaker’s crash-simulation process using highly specialized artificial intelligence models. While that might sound like another vague Silicon Valley buzzword exercise, BMW’s plan is rooted in something much more tangible: an absolutely staggering mountain of engineering data.

Every week, BMW runs thousands of virtual crash simulations as it develops new vehicles. Over the years, those digital impacts have accumulated into more than a petabyte of crash data—a library of structural deformation, material behavior, and safety-performance information massive enough to make even the biggest consumer AI datasets look quaint. Now BMW wants to turn that archive into an engineering brain.

The collaboration with Mistral AI centers around what BMW calls “Large Industry Models,” or LIMs. Think of them as the industrial equivalent of large language models, except instead of learning how humans write emails or generate memes, these systems are being trained to understand how a car’s chassis twists during a side impact or how different alloys behave in a high-speed frontal collision.

BMW says the goal is to improve the speed, accuracy, and overall quality of complex engineering work. In practical terms, that could mean engineers identifying weaknesses earlier in development, reducing costly physical prototypes, and accelerating the timeline between concept and production. In an industry where safety validation can consume enormous amounts of time and money, shaving even small percentages off the process matters.

“For the BMW Group, the use of industrial data is a key factor in translating artificial intelligence into value creation,” said Dr. Franz Decker, the company’s CIO and Senior Vice President. Translation: BMW believes its real competitive advantage isn’t just building cars anymore—it’s owning decades of highly specific engineering knowledge that AI systems can learn from.

That’s where Mistral AI enters the picture. The Paris-based startup has quickly become one of Europe’s most prominent AI companies, positioning itself as an alternative to American AI heavyweights. According to Mistral Chief Revenue Officer Marjorie Janiewicz, industrial AI represents “the new frontier” for artificial intelligence, particularly in engineering-heavy applications like crash simulation.

Unlike general-purpose AI tools, BMW’s LIM strategy focuses on domain-specific intelligence. The company isn’t asking AI to do everything. It’s asking AI to become exceptionally good at understanding one thing: vehicle development. That distinction matters. Generic AI may know what a crash test is, but BMW wants a system that understands precisely how a front subframe behaves under load at 40 mph.

The move also highlights a broader shift happening across the automotive world. Carmakers are no longer treating AI as a futuristic feature for infotainment systems—they’re embedding it directly into the engineering pipeline itself. The race now isn’t just about who builds the best EV or the fastest software-defined vehicle. It’s about who can turn decades of proprietary industrial data into a competitive weapon.

And if BMW’s AI can learn how to crash cars more efficiently before humans ever build them, the next generation of safer vehicles may arrive faster than anyone expected.

Source: BMW

News

2026 Opel Astra Expands Its Powertrain Empire

Leave a comment

The compact hatchback may not be glamorous anymore, but Opel clearly didn’t get the memo. Because while the rest of the segment is busy chasing crossover trends, the new Opel Astra has returned with something unexpectedly compelling: choice.

Real choice.

Electric? Plug-in hybrid? Self-charging hybrid? Diesel? Opel now offers the Astra in every flavor short of hydrogen, turning its long-running compact into one of Europe’s most versatile daily drivers. More importantly, every version feels sharpened with a clear purpose rather than engineered as a compromise.

At the center of the update is the improved Astra Electric, which now stretches its WLTP-rated range to 454 kilometers—roughly 35 kilometers farther than before—thanks to aerodynamic tweaks and drivetrain optimization. In an EV market obsessed with giant batteries and even bigger curb weights, Opel’s approach feels refreshingly disciplined.

The recipe remains simple: a 156-hp front-mounted electric motor, 270 lb-ft of instant torque, and a relatively modest 58-kWh battery pack. The result isn’t neck-snapping acceleration, but a genuinely usable electric hatch that still remembers how to be light enough to feel agile. Opel claims a 0–100 km/h sprint in 9.3 seconds, while top speed is capped at 170 km/h. That may not trouble a Tesla owner, but in the real world of European commuting, it’s more than enough.

More interesting is how thoughtfully Opel has refined the experience around the numbers. Regenerative braking can now be adjusted through three levels using steering-wheel paddles, allowing drivers to tailor the car’s coasting and energy recovery behavior. DC fast charging tops out at 100 kW, replenishing the battery from 20 to 80 percent in about 32 minutes, while an 11-kW onboard charger comes standard.

Then there’s the unexpectedly useful tech. Vehicle-to-Load capability means the Astra Electric can power external devices—from e-bikes to camping equipment—while battery preconditioning helps optimize charging performance before arriving at a fast charger. These aren’t headline-grabbing gimmicks; they’re the kind of practical details that make EV ownership easier.

For buyers not ready to fully commit to electrons, Opel’s revised plug-in hybrid may hit the sweet spot. Combining a 150-hp turbocharged 1.6-liter four-cylinder with a stronger electric motor, the setup now produces a combined 196 horsepower and 266 lb-ft of torque. More importantly, the battery grows to 17 kWh, boosting electric-only range to 84 kilometers on the WLTP cycle—or more than 100 kilometers in urban driving.

That’s enough to cover most daily commutes without touching gasoline, while still preserving the flexibility of a combustion engine for long-distance travel. Performance doesn’t suffer either. Opel says the hatch reaches 100 km/h in 7.6 seconds and tops out at 225 km/h, making it comfortably the quickest Astra in the range.

But perhaps the most intriguing version is also the least flashy.

The Astra Hybrid skips plug-in capability altogether, pairing a 136-hp turbocharged gasoline engine with a small electric motor and a six-speed electrified dual-clutch transmission. It’s designed for drivers who want better efficiency without changing habits—no charging cables, no wall boxes, no range anxiety. Around town, the system can drive electrically for short distances and spends up to half its urban operating time with the gasoline engine switched off.

In other words, it behaves like the hybrid solution many mainstream buyers actually want.

And then, almost defiantly, Opel still offers a diesel.

The 1.5-liter four-cylinder makes 130 horsepower and 221 lb-ft of torque, paired exclusively with an eight-speed automatic transmission. It’s not glamorous, but for high-mileage drivers covering serious autobahn distances, the diesel Astra remains deeply sensible. Opel claims a 209-km/h top speed and respectable 10.6-second acceleration to 100 km/h.

What makes the Astra lineup stand out isn’t any single powertrain. It’s the fact that Opel refuses to force buyers into one technological path. In an industry increasingly dominated by all-or-nothing electrification strategies, the Astra feels unusually pragmatic.

The EV is more efficient. The plug-in hybrid is more capable. The hybrid is more approachable. The diesel still exists for the people who genuinely need it.

That flexibility may not generate the loudest headlines, but it makes the Astra something arguably more important: one of the most intelligently engineered compact cars in Europe today.

Source: Stellantis

News

Volvo’s AI-Powered EX60 Can Read the Road Like a Human

Leave a comment

For years, automakers have promised cars that could “understand” the world around them. Mostly, that has meant lane-keeping systems that ping-pong between road markings or parking sensors that scream at trash cans. But Volvo and Google may have just shown a glimpse of something genuinely different—and, for once, it doesn’t sound like marketing fluff.

At Google I/O, Volvo used its upcoming EX60 electric SUV as the stage for what the company calls a world-first integration between Google Gemini and a vehicle’s external cameras. In plain English: Volvo is teaching its cars to actually see.

Not just detect. Understand.

That distinction matters.

With the driver’s permission, Gemini can interpret the world from the car’s point of view in real time. Parking signs, lane markings, landmarks, restaurants—suddenly the car isn’t simply recognizing objects, it’s contextualizing them. Think less “advanced cruise control” and more “rolling AI co-pilot.”

And honestly? Parking might be the killer app.

Anyone who has circled a downtown block trying to decipher a parking sign that looks like a legal contract written in hieroglyphics will immediately understand the appeal. Volvo says the system can read and interpret restrictions, permit rules, charging regulations, and time limits as you approach a space. Instead of gambling on whether your car will still be there after lunch, the EX60 could simply tell you whether the spot is valid.

That sounds mundane until you realize how useful it could become.

The bigger story, though, is what this says about where in-car tech is headed. Volvo’s latest demonstration suggests the next frontier won’t be screens, horsepower, or even autonomy—it’ll be contextual awareness. Cars that understand what’s happening around them and respond naturally.

Volvo says the technology relies on Gemini’s multimodal AI capabilities paired with the EX60’s neural-processing hardware and software-defined architecture. Translation: the EX60 has enough computing muscle to process visual data in real time without feeling like a science-fair experiment bolted onto the dashboard.

And Volvo isn’t stopping there.

The Swedish automaker also announced that Google Maps’ new Immersive Navigation system is headed first to the EX60, along with the larger EX90 and ES90 EVs. The feature overlays a more detailed 3D visualization of the road ahead, complete with rendered buildings, tunnels, intersections, and overpasses designed to make dense urban driving less confusing.

If you’ve ever missed a turn because your navigation screen looked like a 2009 smartphone app dropped into a sea of skyscrapers, you’ll understand why this matters.

The system also upgrades voice guidance to sound more human and less robotic GPS relic. Instead of “Turn left in 500 feet,” the car might say, “Go past this light and take the next left after the library.” It’s a small change, but one that aligns navigation with how humans actually give directions.

Of course, the automotive industry has a habit of overpromising futuristic AI experiences that end up feeling half-baked. But Volvo and Google have something many rivals don’t: a long-standing software partnership that already underpins some of the best infotainment systems in the business.

That gives this announcement more credibility than the usual CES vaporware.

The EX60 itself is shaping up to be one of Volvo’s most important vehicles yet—a midsize electric SUV that will likely sit at the heart of the brand’s lineup. Now it also appears poised to become a rolling laboratory for the next generation of AI-assisted driving.

Not self-driving. Not autonomous. Just smarter.

And for once, that may be exactly what drivers actually want.

Source: Volvo

Cars and catalogues