Warren, Mich.-based Atomic Industries Inc. is using artificial intelligence (AI) to build injection molds and, according to company officials, improve the quality of molded parts.
Founded in 2019, the venture-capital backed Atomic has been focused on creating AI-powered manufacturing products.
The Michigan startup also recently expanded its scope to provide finished parts and is seeking production space.
Atomic says it is leveraging AI-designed molds to "commoditize" tool building and optimize parts production in the process.
To that end, the company invested in four specially-designed presses from LS Mtron Injection Molding Machine USA that feature 46 input/output (I/O) ports each — triple the number of most other presses — to gather and leverage a wealth of process data.
"Our long-term vision is to use this technology to change the way parts are made, because ultimately the most important part of widget production is the mold," Atomic co-founder Lou Young said in a news release. "Our goal is to commoditize the mold-building process, so when we quote a project, we'll only be quoting a part price. The cost of the mold will be rolled into that cost."
More than a dozen processors from the automotive, medical, electronics and packaging markets have turned to Atomic seeking AI-designed molds that have features optimized for each part, the release says.
Company officials are eager to show other processors what Young describes as purpose-built tools that can produce better parts faster and more affordably.
"The only way to get an industry to shift is to show them that it's shifting," Young said. "It's going to be important for us to show how this technology is impacting production of plastic parts."
A paradigm shift in the automotive industry will be especially notable, Young added.
"In the industry now, when an automaker is kicking off an A pillar, they have 20 or 30 people around the room going, 'Here's where we want the gates' for one little plastic part. And the car has thousands of plastic parts in it," Young said. "The AI-designed injection mold we're building is going to have the best gate location possible for that part, the best waterline design. You won't need 30 people sitting around a table to make sure it runs right. It's just going to run right. That mindset will start to shift."
Atomic purchased four LS Mtron machines — 240- and 500-ton all-electric presses and 240- and 1,000-ton machines — and all new machine tools for mold testing.
"The machines Atomic bought from us came with 46 I/Os, which provides tremendous opportunities to monitor our machines and gather the data they require to optimize the performance of their molds. They're really taking this to the next level — from building the mold to running it," said Peter Gardner, president of LS Mtron Injection Molding Machine USA.
Real-time inputs quickly convey issues, Young said, like "we thought this ejector pin was going to be pushing this hard, but it didn't push anything; or, we thought the cooling of the part in this area would be affected more by this waterline than it was. As that data comes in, our software team will tweak the models."
AI can gather knowledge from reams of moldmakers' insights, combing the internet for data on best practices, said Paul Caprio, president of sales for LS Mtron Injection Molding Machine USA.
"Atomic is a tech company that intends to turn the mold-making industry on its ear," Caprio said, noting it employs many software engineers.
"A mold shop engineer will look at a part design today and say, 'I know where the water line should be because of my experience,'" Caprio said. "Atomic is going to the cloud, where countless mold makers have documented where the water line should be — and that's a much better, faster way of determining the best water line location."
What's next?
"We need a work cell to drive the cost of tooling to where it's like a commodity. We're driving the pain and heartache out of the whole process, so you don't even think about it anymore, you just order the widget," Young said.