Democratizing Compute: Why Access to HPC Matters for Everyone

‍

Computing for Humanity (CFH) is dedicated to ensuring that no great idea is held back by a lack of computing resources. In a recent conversation with Antonia Maar of Intersect360 Research —  a Europe-based analyst covering global HPC and AI markets and author of the newsletters Intelligence, Explained and EU AI Act, Explained — we explored why broadening access to high-performance computing (HPC) is critical for innovation, equity, and impact. “Access shapes participation, and participation shapes what gets built,” Maar emphasizes. If researchers and citizen scientists cannot run models or process data, even the best ideas remain unrealized. Today’s HPC and AI markets have exploded – the global AI infrastructure market alone grew 121% to $193 billion in 2024 – yet most of that compute power is concentrated in a few hyperscaler companies.

Expanding access is not charity but smart strategy. As Maar notes, “democratization isn’t charity, it’s efficiency,” because great ideas can come from anywhere, not just the wealthiest labs. When more researchers have compute resources, the entire innovation ecosystem gets stronger. Computing for Humanity lives this by providing donated clusters and a free cloud platform (MyResearchCloud) so “no human advancement should wait in line for computing time”.

The Cost of Uneven Access

When access to HPC is uneven, research narrowly follows privilege. Antonia warns that unequal access “makes privilege look like merit” – only those with computing power appear to produce the best ideas. Over time, talent flows toward the handful of institutions or companies that have resources, creating an “innovation monoculture.” For example, in Europe and elsewhere, talent flows to where the compute is — often toward U.S. hyperscalers and well-funded institutions. To avoid leaving brilliant ideas behind, we must level the playing field. This echoes CFH’s mission: as our founder puts it, extending the life of every “core” is crucial – “every core we add to try solving a problem will get us just that much closer to a solution”. By reusing corporate servers and sharing them with academia, CFH ensures that promising research isn’t stalled for lack of hardware.

Lowering the Bar: Mid-Tier Resources and Community

How do we bring more people into HPC? Maar advises to “lower the floor, not just the ceiling.” Instead of only building flagship supercomputers, we need accessible mid-range clusters and cloud resources that let newcomers experiment and learn. For most academic or citizen-science projects, CFH’s approach – providing affordable virtual servers on MyResearchCloud – fits exactly this need. In fact, our platform serves “90–95% of underserved researchers,” offering free or low-cost CPUs and GPUs.

Access also means community. “Access isn’t only about machines,” Maar points out, but also about networks, visibility, and a sense of belonging. That’s why global efforts like the HPC-AI Leadership Organization’s (HALO) and the Women in HPC network exist: they connect early-career scientists and underrepresented groups with mentors and peers. CFH similarly prioritizes inclusive access – by combining technology with outreach and training, we help more voices enter and thrive in research computing.‍ ‍

Global Trends: Infrastructure and Strategy

The push for equitable HPC access is mirrored at the policy level. Europe’s EuroHPC initiative, for example, co-funds supercomputers and even AI “factories” across member states, ensuring wide national coverage. In Canada, the Digital Research Alliance combines resources from universities and governments to share HPC facilities. Maar notes that Europe’s tiered co-funding model has produced well-structured infrastructure, while Canada’s more flexible approach requires oversight so as not to become a bottleneck (long wait times can erode usability).

She highlights an emerging “middle-tier gap”: the computing market is splitting into huge hyperscale clouds and small local clusters, leaving medium-scale needs underserved. This middle ground is actually an opportunity: CFH targets precisely that sweet spot by offering medium-power clusters and cloud VMs for projects that are too large for personal computers but too small for massive supercomputers. ‍

Another under-appreciated constraint is energy and power. Modern AI and HPC centers increasingly require as much electricity as heavy industry. In fact, data centers are projected to consume more than double their electricity by 2030. The challenge is not raw power production, but having reliable electricity at the right time and place. As Maar notes, future growth of HPC – in Canada, Europe, or elsewhere – will depend on sustainable energy planning and grid capacity, not just on buying more hardware.‍ ‍

Successful Partnerships and Clear Governance

Public-private partnerships can accelerate infrastructure access if done thoughtfully. As an example, CFH’s collaboration with ThinkOn (a data center provider) is cited as a “clean value exchange”: CFH brings mission and purpose, while the partner contributes technical capability. Maar points out that success comes when each party is clear about its role.

She also warns against “disguised philanthropy”: well-meaning donations that create dependency. The key is transparent governance and shared accountability over the long term. In practice, this means binding agreements about resource use and exit plans, and recognizing that research infrastructure needs commitments measured in years, not quarters. CFH embodies this by operating as a non-profit custodian: donors and volunteers contribute resources that CFH then manages for the public good, under a clear mission.‍ ‍

Women, Leadership, and Inclusion

Maar spoke candidly about the structural barriers women still face in innovation ecosystems. Women are increasingly present in the conversation, she notes, but “presence isn’t influence.” Real barriers persist around sponsorship, strategic visibility, and access to decision-making. She describes a “technical credibility tax” — women in this space are routinely expected to prove competence in ways their male counterparts are not. “The industry still talks too much about pipeline and not enough about culture and retention,” Maar argues. The real question is whether the field makes it realistic for women to stay, advance, and shape decisions.

‍ ‍For women building tech careers from non-traditional backgrounds, Maar’s advice is empowering: reach out and connect. Her key tips include:

- Network across domains: Tech work often sits at the intersection of technology, policy, and people. Being able to translate between worlds is a real strength.
- Build community early: Join groups (like Women in HPC or policy forums) and start contributing (e.g., write blogs, attend meetups) even before you need help.
- Challenge norms: If you encounter biases or outdated assumptions, raise them. Often people simply don’t realize, and calling it out can lead to improvement.

“Support can’t only show up on symbolic occasions,” Maar concludes. “It has to show up in the in-between moments too. The system may be slow. But we don’t have to be.” At CFH, we share this commitment to structural inclusion — ensuring that diverse voices shape not just the conversation, but the decisions.

Infrastructure for Impact

“What does infrastructure for impact mean?”, we asked Maar. For her, it means systems that expand what people are able to do. Infrastructure isn’t just hardware racks; it includes the governance, training, and community support that make technology truly accessible. A powerful computer without documentation or help is under-used; similarly, rigid infrastructure that only serves one type of project leaves others behind.

CFH’s work exemplifies infrastructure for impact. We repurpose enterprise servers into research clusters and offer a persistent platform (MyResearchCloud) that grows with projects. We think in decades, not quarters, ensuring that scientists can count on stable access. This reflects a philosophy she praises: the notion of supporting infrastructure intentionally to broaden participation.

As Maar notes, infrastructure has a way of becoming invisible—until it’s missing. In the 21st century, compute capacity is like hospitals or clean water: essential for modern life. Yet grant funding for computing can be fragile. That’s where philanthropy plays a crucial role. As she says, “infrastructure is a force multiplier”: funding one piece of infrastructure yields hundreds of research outputs over time. In contrast, funding a single lab yields a single result. ‍

For donors, this means their impact can echo far beyond one project. Every server or core donated to CFH can accelerate discoveries in medicine, environment, and more. In fact, CFH has even named computing resources after legacy donors to link gift-giving with scientific progress. As our founder Roy Chartier puts it, “every core we add… gets us just that much closer to a solution.” ‍

As Maar puts it: “Once access becomes real and usable, it changes who gets to contribute. And when that changes, the whole ecosystem becomes stronger.” That is the principle behind Computing for Humanity’s work: expanding access so that promising ideas are not limited by infrastructure. If you want to be part of that effort, we invite you to get involved.

‍ ‍