This is Engineering Paper, and here’s the latest design and simulation software news.
There’s a herd of engineers in Houston, Texas this week, and they’re branded by backpacks emblazoned with 3DExperience World, Dassault Systèmes’ annual user conference.
Like last year’s conference—also held in Houston—3DExperience World 2026 has a Texas-sized focus on one technology in particular.
“A lot has happened since we met last, but one topic has stayed firmly in the spotlight: AI,” said Manish Kumar, Solidworks CEO, in his opening address.
Indeed, Kumar and other Dassault executives have been shining that spotlight throughout the event. We’ve seen a mix of AI past, present and future on the 3DExperience platform: some of it intriguing, some of it confusing, and all of it steeped in the usual blend of hype and asterisks. Here are some high-level highlights.
The AI model maniac
One of the traditions of 3DExperience World, dating back to when it was still called Solidworks World, is a Solidworks speed modeling competition called Model Mania. Contestants are given a 2D drawing and must make it into a 3D model as quickly as they can.
This year, human modelers will have a new foe: AI.
In a live demo, Kumar showed off an AI feature in xDesign that can convert a PDF of a drawing into a series of sketches, and then turn those sketches into a parametric 3D model—exactly what human model maniacs do.
“You all love the Model Mania contest,” said Pascal Daloz, Dassault Systèmes’ CEO, who was on stage with Kumar. “This year, Leo will not just be watching. Leo will compete against you.”
Leo? This is where things start to get a bit head-scratching. Leo is one of three AI agents, or as Dassault prefers to call them, virtual companions.
Aura, Leo and Marie
Aura, the virtual companion that Dassault introduced last year, has two new friends: Leo and Marie. The three companions are complementary but specialized, according to Daloz.
“Aura, she is orchestrating the knowledge and the context from requirements to project to changes. Leo is bringing engineering results: mechanics, structure, motions, simulations, manufacturing. And Marie, she is bringing scientific disciplines: materials, chemistry, formulations, regulations,” Daloz explained in his Day 1 keynote.
More succinctly, Daloz likened Aura to having an MBA, Leo an MSc, and Marie a PhD. Business bot, engineer bot, and scientist bot.
In a press conference, Kumar elaborated on the three companions: “Aura is the virtual companion that connects with everything, whether it is internal knowledge or external knowledge… core competencies are business strategy, interrelationship, compliance, management.
“Leo, the engineer, the builder, he’s more grounded… Leo wants to confirm that you can build it, you can manufacture it.
“And Marie is essentially the scientist which is going to go into the deep scientific knowledge. Typical competencies are molecular science, microbiology and so on.”
Apparently the three companions even have distinct personalities, per Kumar: Aura is agreeable, Leo is assertive, and Marie is even more assertive.
Why have three different AI companions? “We do not believe that one single intelligence is going to provide all possible capabilities,” Kumar said.
I don’t quite understand that explanation, and it’s not clear whether users are expected to choose between the three companions, but I’ll report back when I get more details.
Meanwhile, one more fun fact: apparently these virtual companions named themselves. Aura is an extremely forced acronym for “assisting you to realize your ambitions,” Leo is named after da Vinci, and Marie is named after Curie.
Nvidia and Dassault Systèmes team up on AI
The second-day keynote had a special guest. Jensen Huang, CEO of Nvidia, took the stage with Daloz to announce that Nvidia and Dassault are strategic partnering to “establish a shared industrial architecture for mission-critical artificial intelligence across industries.” (Sounds a bit familiar, doesn’t it?)
“We met 25 years ago [during] a transition from mainframes and workstations to personal computers, the last major platform shift,” Huang said in a press conference with Daloz. “This is the beginning of a new platform shift and the largest integration of our technologies in our history.”
That integration will involve Nvidia’s AI, CUDA-X, and Omniverse libraries and infrastructure and Dassault’s 3DExperience portfolio. Additionally, Nvidia announced it’s using Dassault’s model-based system engineering (MBSE) to design its AI factories.
For Daloz, the new partnership will help enable Industrial World Models—a new piece of Dassault jargon that fits somehow with 3D Univ+rses.
“Together with NVIDIA, we are building industry World Models that unite Virtual Twins and accelerated computing to help industry design, simulate and operate complex systems in biology, materials science, engineering and manufacturing with confidence,” Daloz said in the partnership press release.
AI in the works for Solidworks
Naturally we heard about several AI features coming to Solidworks, some as soon as the next functional delivery release. I’ll cover these in more detail soon, but here’s a quick preview:
- Automated drawings (called it), now with more user customization
- Automatic fastener recognition and assembly
- “What’s wrong” root cause analysis for modeling errors
- Assembly structure generation
- “Assembly doctor” recommendations to optimize performance
- Chat-driven modeling
- Picture-to-mesh generation
3DExperience gets 16% less confusing
Finally, a non-AI piece of news. You may remember an article from last year called 37 things that confuse me about 3DEXPERIENCE, written by Peter Brinkhuis of CAD Booster. Pascal Daloz clearly remembers it, because he brought it up at the end of his first keynote.
“We fixed six, and we will continue to fix them,” Daloz said. Among those fixes is a simplified compass that highlights apps users have access to and provides descriptions of what they do (with names like 3DSwym, who could be confused?).
Brinkhuis was at the conference, so I couldn’t help but ask his thoughts. Sporting a bright orange shirt with the number 37 on it, Brinkhuis was surprised the CEO was aware of his complaints—but he’s willing to let him write a new number on the orange shirt.
“I know they fixed a couple of things. I’m keeping score, but I need to update the score,” he said. “It was nice to hear that they know they have to simplify things.”
I’ll continue to hunt down details on these topics here in Houston and beyond. If there are any that piqued your interest, let me know at malba@wtwhmedia.com.
One last link
My colleague Martin Rowe from EE World Online discusses 6G, test equipment, Boston, and the blues in his guest appearance on The Amp Hour Podcast: The Measurement Blues.
Got news, tips, comments, or complaints? Send them my way: malba@wtwhmedia.com.
The post AI is the new model maniac appeared first on Engineering.com.
