At Badger Commerce, scaling smarter beats scaling harder. So this year we made a bold move: our infrastructure platform is now 100% ARM-based.
And the results have surpassed our expectations.
Why We Migrated
We’d run on mixed x86 and ARM-capable infrastructure for a while—familiar, reliable, but increasingly expensive and complex to manage.
As our product, traffic and internal workloads kept growing, we faced scaling out more machines rather than getting more from each machine. Meanwhile, we kept hearing about the industry shift: companies like Google are turning to custom ARM hardware and reporting approximately 65% better price-performance and 60% better energy efficiency.
So we asked ourselves: could Badger Commerce get more capacity, lower cost, and simpler operations—by going all-in on ARM?
The answer: yes.
The Hard Metrics
Here’s what we delivered:
~75% reduction in compute & infrastructure cost – thanks to better performance per watt, optimized right-sizing of instances and fewer idle resources.
2× effective capacity – workloads that previously required broad horizontal spreads now run denser and faster on ARM nodes.
Environment consolidation – with higher density and performance per node, we’ve been able to retire multiple duplicate or shadow environments, simplifying our deployment matrix and reducing operational overhead.
Simpler infrastructure architecture – fewer scaling edges, fewer instance-types, fewer special-cases. That means fewer surprises, fewer bugs, faster time to market.
How We Did It
Here are the key enablers that made this migration work—inspired in part by what we observed in the broader industry and adapted for our platform:
1. Benchmarking & pilot phase – We ran identical workloads on x86 vs ARM, measured performance, cost, memory and I/O behaviour, and found strong advantage for ARM-based instances in our workload patterns.
2. Refactoring build/test pipelines – As others have found (e.g., Google’s internal build/test tooling pivot) moving architecture means you must adjust CI/CD, build flags, test suites. We adopted automated validation to catch architecture-specific failures early.
3. Right-sizing and capacity planning – With better performance per core and per socket, we redesigned our node sizing model: fewer but larger ARM instances where it made sense, and eliminated many smaller legacy hosts.
4. Environment rationalisation – Because we unlocked more capacity per node, we consolidated dev/test/QA/staging clusters, removed duplication, retired older x86 clusters altogether.
5. Savings reinvested – Cost savings didn’t go into spreadsheets alone; we reinvested into product innovation, reliability engineering, monitoring, autoscaling improvements—effectively letting this migration fund future growth.
What It Means for Our Customers
You won’t see “We now run on ARM!” on your checkout page—but you will feel the difference:
Faster response-times and better headroom for growth (we’ve doubled capacity without doubling cost).
Fewer maintenance windows, fewer infrastructure surprises because we simplified the architecture.
More predictable pricing and lower cost-base for us means we can stay competitive and continue to invest in new features and reliability.
Lower infrastructure overhead means a lower environmental footprint (better power utilisation, fewer idle machines)—aligning with our commitment to sustainability.
Looking Ahead
This migration isn’t the end—it’s the foundation.
We will continue optimising and tuning workloads to take full advantage of ARM-specific features (e.g., vector-extensions, custom caching behaviours).
We will monitor real-world telemetry to push performance further.
We will explore hybrid multi-arch strategies (e.g., ARM + GPU + specialised instances) where new workloads require it, but with ARM as our core.
We will keep translating infrastructure savings into platform innovation for our customers—more features, better stability, more value.
Why This Matters
If you’re evaluating an e-commerce platform, you’re not just buying software—you’re buying reliability, performance and headroom for growth.
With this move, Badger Commerce has taken a foundational step: cutting cost, increasing capacity, reducing complexity—and laying the groundwork for the next generation of the platform.
We believe that great commerce software shouldn’t require massive infrastructure budgets or dozens of specialised hosts. It should quietly work—scale-up easily, stay performant, stay efficient.
And that’s exactly what we’ve engineered.
Get in Touch
Want to learn more about how our ARM-first architecture translates into real world benefits for your store? Just drop us a line—we’re always happy to talk tech or show you the numbers.
© Badger Commerce Limited. All rights reserved.
Badger Commerce is a subsidiary of Kedos Consulting Limited, specialists in e-commerce, retail technology, microservice design, architectural design of large enterprise solutions, and mobile apps.
