Coming out of stealth, a startup is out to overturn the CAE industry with a combination of AWS-based HPC capabilities and what it says is superior multi-physics design simulation software and a new, TCO-cutting “as-a-service” pricing model.

According to Cupertino, Calif.-based OnScale, its CAE platform is the industry’s first Solver-as-a-Service offering, giving engineering teams pay-as-you-go pricing while avoiding investments in on-premises HPC capabilities. Co-founder and CEO Ian Campbell said that for a team of 10 engineers consuming 100,000 CAE core hours over a three-year-period, OnScale delivers 50 percent savings over traditional CAE solutions, such as those from Ansys, Comsol and ABaCUS.

OnScale was founded as a spin-off from Thornton Tomasetti, a nearly 70-year-old science and engineering consulting services company, which also led the $3 million seed financing round, with participation from Silicon Valley investor Michael Lehman (former CFO of Arista Networks, Palo Alto Networks, and Sun Microsystems) and CampbellKlein, an early stage high-tech venture fund.

“The CAE/HPC space is ripe for disruption,” Campbell said, “because yesterday’s systems are too expensive and cumbersome to solve tomorrow’s engineering problems. Our platform eliminates the risk associated with sequential trial-and-error physical prototyping, reduces design cycles from months to weeks or even days, and delivers unequaled computing performance that easily scales to meet ever-changing CAE workloads.”

Source: OnScale

OnScale will charge customers on a per-core-hour basis, with the software and compute power accessed in the AWS public cloud service. Campbell said this delivery model means engineers no longer have to pay for CAE software licenses, maintenance and support or HPC hardware. Nor, he said do they have to wait for IT to deploy and maintain CAE systems. He added that the traditional CAE pricing model can mean that too few licenses are purchased, leading to struggles among engineers seeking access to the limited number of seats.

OnScale solvers have been developed, used and maintained by Thornton Tomasetti software engineers over the last three decades, Campbell said. A key OnScale performance advantage, he said, is that its design modules are tightly coupled and integrated, enabling the software to take on large-scale design challenges that need to be viewed in 3D, such as RF (radio frequency) filters and switches for 5G smartphones and base stations (other target verticals include IoT and IIoT sensors, biomedical and autonomous vehicles).

Campbell said it’s the loosely coupled nature of traditional CAE software that results in overburdened HPC systems used by engineering firms. “To solve a problem like a 5G RF filter,” he said, “they couldn’t solve that model in three dimensions because they’d exceed the amount of RAM available to them – a problem like that requires much more than the 2TB of RAM they typically have available to them.”

But OnScale solvers (such as mechanical and electrical) share the same memory space, which reduces the amount of CPU power required while also giving engineers access to what Campbell called the “virtually limitless” HPC capabilities in AWS. “They can solve these problems in 3D doing massive parametric sweeps or optimization studies, so they’re not just looking at one design, they’re looking at thousands of designs to find the most optimal for their system.”

Campbell said OnScale will be available in Q2 of this year, and that its solvers are in use at roughly 200 companies, most of them in the biomedical field.

“Our engineering design cycles come in bursts throughout the year,” said Mo Zarkan, USM development engineering manager, at Emerson Automation Solutions. “OnScale’s SaaS model helps us to best utilize our R&D expenses as demand for CAE and HPC fluctuates.”

“OnScale’s SaaS pricing model gives us access to CAE and HPC resources that would traditionally only be available to large companies with massive R&D budgets,” said Dr. Michael Weaver, CEO and founder  of Cenofex Innovations. “OnScale gives our startup a competitive edge against giants in our industry.”

This article originally appeared in our sister publication, EnterpriseTech.