The engineers are ready. The test is set. The button is pushed. Roar! After just over a year of engineering and creativity, a brand-new engine for a Formula 1 race car comes to life.
With the help of a cloud-based supercomputer, the team creating the new power unit for Oracle Red Bull Racing’s 2026 car have done something remarkable—designing and building a new engine prototype from scratch in just over a year. In fact, the team is creating an entire power unit: the engine, the turbocharger, and the hybrid systems.
The development organization, called Red Bull Ford Powertrains, understands the enormity of this endeavor. They’re designing and manufacturing the new race car power unit for use in F1, the pinnacle of racing performance—and doing it in less than four years. For comparison, most automotive engine development programs take almost twice as long to achieve the same end. What’s more, they started with no existing engine design team, no data center to run engineering simulations, and no factory to build it all.
The computing element has been critical to their development speed. Red Bull Ford Powertrains tapped cloud-based high performance computing (HPC) capacity from Oracle so they could start doing engineering simulations without building a major data center. Using cloud infrastructure to perform that level of computationally intensive work can save considerable time and money. As it did for Red Bull Ford Powertrains, cloud-based high performance computing will make it possible for many more people and organizations in other industries to take on dazzling feats of engineering and design.
Red Bull Ford Powertrains is designing and building the new power unit in time for the 2026 F1 season, which ushers in a new hybrid engine design and sustainable fuel requirements. While the team now has a working prototype of the 2026 engine, there are many iterations and further developments yet to come before its descendants hit the track, not to mention thousands of hours of rig testing to ensure reliability. And with Oracle Red Bull Racing, the defending champion team of the 2022 F1 Championship, there’s only one ultimate measure of success: the top spot on the driver and team rankings.
“Red Bull are going to tick every box that’s required in order to deliver a power unit that can win a championship,” says Matt Cadieux, who, as CIO for Oracle Red Bull Racing, leads the computing team behind the work. “And that’s exciting, challenging, and stressful.”
Oracle Red Bull Racing’s decision to build their own power unit started when the team’s former engine maker, Honda, announced their retirement from F1 racing back in 2021. After initial discussions with several manufacturers, Oracle Red Bull Racing decided to do something unheard of: design and build their own power unit rather than buy one from an established engine maker. They soon joined in a strategic partnership with Ford Motor Company, who will lean in with their substantial expertise in EV development, controls, and battery technology to support the development program.
With the decision made, the work began. Red Bull Ford Powertrains staffed up with expert engineers, built facilities to construct and test engines and other components, and put in place the computing horsepower required to run the sophisticated models and engineering simulations needed to optimize the power unit while meeting F1’s exacting requirements.
The Red Bull Ford Powertrains’ engineers and designers are using high performance computing clusters in the cloud to simulate possible designs for the new engine for both Red Bull–owned teams (Oracle Red Bull Racing and Scuderia AlphaTauri).
The workloads, which are hosted on Oracle Cloud Infrastructure (OCI) on a cluster of bare metal servers, focus on repeated iterations of computational fluid dynamics (CFD) models to inform some of the most intricate designs inside the internal combustion engine part of the power unit. This includes simulating fuel spray patterns, combustion performance, and how the shape of the inside of a cylinder could affect and optimize these to deliver maximum performance.
The team members construct detailed mathematical models of the car, set an initial set of values for the many variables, and then run a simulation. Based on analysis of that simulation, the engineers adjust the model or tweak some of the values and run the simulation again. Over and over.
Automakers and other engineering-intensive companies have their own high performance computing environments, but building and equipping a new on-premises data center would have taken many months—losing precious time and delaying the start of design work. Instead, using OCI, Red Bull Ford Powertrains was able to get started immediately. What’s more, in the early days, the optimal system configuration to maximize simulation performance wasn’t finalized. Using OCI gave them the flexibility to evaluate different computing shapes and configurations.
“We were able to experiment with hardware and software combinations and different HPC shapes without having to buy hardware, take it out of boxes, put it into racks, connect it to the network, load the software,” Cadieux says. “All that stuff is super time-consuming.”
Deciding to develop a power unit is a huge leap for a new organization like Red Bull Ford Powertrains. In doing so, they’re taking on rival manufacturers that have decades of F1 experience. Yet, while the engineering team feels the pressure to succeed, they also feel the support. The mission is clear, and the computing infrastructure and other tools are in place for them. Cadieux says this approach fits the Red Bull philosophy.
“If you give a very talented group of people the tools to do the job, you’re setting them up for success, and you’re giving wings to their imagination and their innovation and their ability to go and win,” he says.
Every few years, Formula 1’s governing body, the Fédération Internationale de l’Automobile (FIA), revises the specifications for the F1 Championship car. Those specs cover many factors, including the shape and weight of the vehicle, the safety equipment, and so on. The FIA can also change the regulations on the power unit, which happens less frequently due to the long and expensive development journey. The last major regulation change in this area took place in 2014, and the next-generation power unit will debut in the 2026 season.
A winning F1 engine needs to efficiently burn all the fuel that’s injected into the engine to get the maximum performance. Declan Ward, senior performance simulation engineer at Red Bull Ford Powertrains, explains how he runs detailed analysis and 3D modeling to make sure the fuel intake system vaporizes and mixes the fuel in just the right way and sprays the right amount of fuel in exactly the right spot for combustion.
“You want to ensure that all of that fuel is available when it comes to combustion for every single cycle in each cylinder,” he says. “Some fuel with a misguided injection event or spray pattern could be lost in the exhaust system.”
While the specs of an F1 race car are intensely detailed, there are still many, many decisions that go into the design of a power unit, including choice of materials (steel, titanium, and aluminum alloys are allowed in various places), rates of burn for the fuel, lubrication, exhaust system, and much more.
Red Bull Ford Powertrains engineers use OCI to run several computational fluid dynamics software packages for the design and simulation work to assess those variables. One is CONVERGE, a popular CFD package from Convergent Science that contains configurations for internal combustion engines, fuel injectors and sprays, exhausts, and other areas of interest to F1 designers. The team uses CONVERGE for thermodynamics, intake system combustion, exhaust systems, and fuel conversion simulations.
The team is also using Simcenter STAR-CCM+ from Siemens for CFD for turbocharger design and simulation.
When it comes to running math-intensive software such as CFD, speed matters—so having the latest hardware and choosing the right processors matters. When OCI moved the team onto the latest generation of processors in December 2022, it delivered a 10% performance gain. From a customer point of view, implementing hardware upgrades in OCI is a much easier project than doing so in an on-premises data center.
OCI also let Red Bull Ford Powertrains invest in only the computing capacity they needed, rather than make a big up-front investment in the early days of the engineering program when requirements were still evolving. When they kicked off this work in mid-2021, says Cadieux, the computation was small scale. But since moving into a new building on the Red Bull Technology Campus, the team undertook a major recruitment drive. “We started off using a relatively small number of OCI compute resources,” he says. “But as the team started growing, as the model sophistication started growing, we were able to easily and quickly provision more OCI resources in a phased manner to keep up with growing demands. OCI gave the team the agility it required as well as a financial model where we only paid for what was needed.”
“The Red Bull culture is to bring in the right people and enable them with the right tools. We approach the powertrain with that kind of open mindset—we found the best technology so we could give our people the best tools for the job. It’s the Red Bull way."
Creating a high performance computing cluster is more complex than following a standard recipe—the cluster might comprise hundreds of servers, each with multicore CPUs and GPUs, memory, cache, storage, and high-bandwidth interconnects between the chips and between the servers. The Red Bull Ford Powertrains team was able to leverage Oracle’s existing relationships with the simulation software providers to help determine the optimal processors, GPUs, and memory for specific workloads.
Building an HPC cluster on-premises requires a vast inventory of different types of hardware as well as the manual labor to reconfigure them and reinstall and tune the software. But by using the knowledge Oracle and the software providers already had about OCI, the Red Bull Ford Powertrains engineers “could quickly converge on the high performance sweet spot,” says Cadieux.
And once the engineers settled on that sweet spot, the team was running production-level HPC cluster configurations in about a week. “For us, that would have been a month of work with people unboxing a mountain of hardware, connecting it all, imaging it,” Cadieux says. “Instead, it just magically happened in Oracle Cloud, and we didn’t have to think about it.”
As of mid-2023, Red Bull Ford Powertrains is modeling, designing, simulating, and testing, over and over, to finish a power unit design and get it manufactured for the 2026 season.
“We’re David versus Goliath on the Formula 1 power units because we’re going up against the might of the OEMs,” says Cadieux, referring to the established engine manufacturers for race teams.
Equipped with the best engineering software running on OCI’s high performance computing, the Red Bull Ford Powertrains team relishes the challenge.
“The Red Bull culture is to bring in the right people and enable them with the right tools,” says Ward, the simulation engineer. “We approach the powertrain with that kind of open mindset—we found the best technology so we could give our people the best tools for the job. It’s the Red Bull way.”
See how to use a high performance computing cluster in the cloud with powerful CFD software and the very latest CPUs and GPUs.
Cloud-based high performance computing offers advanced compute, storage, networking, and software technologies at a predictable cost while delivering or exceeding on-premises levels of performance.