I have written before about cloud waste.
About how infrastructure sits idle. About how the meters keep running even when nothing meaningful is happening. About how utilization rates across the industry are lower than most people think.
That piece was mostly about cost and visibility. But there is another angle I have been sitting with for a while.
What if the capacity that is already there — already paid for, already running — could do something useful?
That question is what became WAYSCloud Impact.
The problem I kept coming back to
Cloud infrastructure is provisioned for peak. By definition, that means it runs below peak most of the time. The headroom is real. It is constant. And it belongs to nobody in particular — it is just margin, sitting there to absorb load spikes that may or may not arrive.
At the same time, I kept reading about research that needed compute it could not get. Climate modeling. Epidemiological simulation. Genomics pipelines. Open machine learning projects building tools for medical imaging or language preservation. Serious, meaningful work — the kind that produces public output, not private value — competing for infrastructure that is out of reach.
There is something structurally wrong with that picture.
What we actually built
WAYSCloud Impact is simple in principle.
WAYSCloud customers who have idle headroom in their infrastructure can contribute it to a shared pool. Most of them are already paying for that capacity. By contributing it, they make it available for approved research instead.
We then match every contribution one-to-one. Whatever customers put in, we put in the same amount. The pool doubles before any allocation is made.
Approved organizations — research institutions, non-profits, public research bodies in the EU and EEA — apply on a rolling basis. We review each application individually. There is no queue. There is no first-come-first-served. It is about whether the work fits and whether the capacity is actually there.
The tradeoff is honest: workloads run as preemptible batch jobs. When production needs the resources back, Impact jobs pause. The research we are targeting — long-running simulations, parallel computations, reproducible analyses — tends to handle that well. It is not a surprise constraint; it is the design.
Why I wanted to do this this way
I could have framed this as a sustainability initiative. Or as CSR. Or as a grant program.
I chose not to, because I think those frames carry a lot of soft edges. They tend to become marketing before they become practice.
What I wanted to build was something structural. A real commitment, matched, governed, and sized to what we can actually stand behind — not what sounds good in a press release.
That is why the matching is binding. Why every organization goes through a review. Why the governance model is published. Why we will report aggregate usage openly over time. And why I chose not to make the eligible pool larger than the available capacity can support.
The program is intentionally small. That is a feature, not a limitation.
A note on the utilization argument
There is a version of this conversation that treats idle capacity as essentially free. I want to be careful about that framing.
Idle capacity is not free. It was provisioned, powered and cooled. Customers paid for it. WAYSCloud operates it.
What Impact does is make that existing cost useful to someone beyond the original allocation. The incremental cost of running an approved research workload on capacity that would otherwise go unused is low — but the value to the recipient can be real.
That asymmetry is worth taking seriously. It is also why the matching matters: we are not asking customers to fund something on our behalf. We are building a shared commitment, and putting in our own capacity to match what they contribute.
Where it lives
The program is at wayscloud.org.
If you work at or with a research institution, non-profit, or open research initiative registered in the EU or EEA and you have compute-intensive, batch-style work that would benefit from this kind of access — I would encourage you to read through the program and apply.
And if you are a WAYSCloud customer with idle headroom and you want to put it to use — get in touch.
The capacity is already there. We might as well do something with it.