Are you an NGO, human rights defender, or nonprofit fighting authoritarianism? HRF will fund the hardware so you can run sovereign AI — privately, locally, free from corporate and government surveillance. Forever.
Every query sent to a cloud AI passes through corporate servers — potentially monitored, subpoenaed, or cut off by governments. For human rights defenders, that's not a privacy concern. It's a matter of survival. For the first time, HRF is funding hardware so frontline defenders can run AI that is entirely their own — no corporations, no governments, no surveillance.
Run your own server. Connect your team. Every query stays on hardware you own and control.
The status quo forces human rights defenders to choose between powerful AI tools and operational security. Local Intelligence eliminates that choice.
Grants scale to your organization's size and needs. Whether you're a solo journalist or a coalition serving hundreds, there's a configuration that works.
Ideal for individual defenders or 2–3 person teams. Run Llama 3, Mistral, or Gemma locally — no internet required after setup.
The sweet spot for NGOs serving 10–50 users. Two Spark units deliver powerful, reliable local inference for a full team with room to scale.
For larger coalitions supporting dozens of staff and partners. A dedicated inference server running multiple large models simultaneously.
Don't see your configuration? Applications are evaluated individually. If you have specific hardware needs, describe them in your proposal — the program is flexible.
This program is for organizations and individuals on the frontlines of the fight for freedom — those who need AI most and can trust corporations least.
Organizations documenting abuses, supporting political prisoners, or advocating for rights in closed societies and authoritarian states.
Newsrooms and journalists reporting on authoritarian regimes, at risk of having sources exposed through corporate surveillance systems.
Nonprofits and activists promoting free elections, rule of law, or democratic institutions in countries under authoritarian rule.
Individual activists and researchers facing surveillance or persecution who need private, uncensorable AI tools for their work and safety.
Teams building open-source privacy tools, censorship-resistant infrastructure, or AI capabilities specifically for at-risk and dissident communities.
Networks serving dozens or hundreds of frontline defenders who need shared, sovereign AI infrastructure across multiple partner organizations.
The path from application to running your own local inference is straightforward. HRF guides you through every step.
Submit your organization profile, your work fighting authoritarianism, and your hardware or budget needs.
HRF's team reviews for mission alignment, geographic context, and feasibility of your local AI deployment.
Approved applicants receive hardware directly (e.g., NVIDIA Spark units) or a cash grant to purchase equipment locally.
Set up your local inference environment. HRF provides onboarding resources and connects you with our technical network.
Your organization now runs its own intelligence — privately, securely, free from any corporate or government surveillance.
For the first time in history, a human rights defender with a $10,000 budget can run a state-of-the-art AI model — privately, locally, and permanently.
Before this, you needed a data center. Now you can do it with a Spark configuration or a MacBook Pro. HRF wants to put that power in the hands of every frontline defender who needs it.
Applications are reviewed on a rolling basis. There is no deadline — we evaluate each case on its mission alignment and potential impact.
Human Rights Foundation · AI for Individual Rights · 2026