A Sharper PGF: How Filecoin Public Goods Funding Is Evolving

Filecoin Public Goods Funding is moving from one-size-fits-all batches to three purpose-built programs — Kernel, Revenue Development, and R&D — each funded, measured, and paced in the way its work actually behaves. The result is a program with clearer accountability, faster operations, and a much tighter line between every dollar deployed and the network outcome it is meant to produce.

🌱 Where we started

FIL PGF began as a way to reward work after the fact. FIL RetroPGF distributed ~1M FIL across 284 grants over three rounds, recognizing the impact that had already landed. In mid-2025, FIL ProPGF added a forward-looking, committee-driven mechanism, ~$9M USD across 52 grants spanning core infrastructure, tooling, and ecosystem work, each with a public application, a named recipient, and milestones.

Together those mechanisms proved that the ecosystem can fund public goods transparently and at scale. What they also revealed is that very different kinds of work — maintaining consensus resilience, closing paying storage customers, and exploring an unproven idea — were being pushed through one shared cycle, on one shared clock, against one shared bar. That structure was costing us in speed, clarity, and capital efficiency.

🧭 What changed

The program is now more selective and more capital-disciplined, with new structure across strategy, stewardship, data, and operations.

One framework, three categories. All PGF funding is now organized into Kernel, Revenue Development, and R&D. Each answers a different question: what must remain resilient, what turns capacity into paid usage, and what makes sure there is a next generation of Filecoin technology. Every applicant can now see exactly where their work fits before they write a word.

Each category is funded in the shape that fits it. Kernel is treated as a near-fixed cost — a retainer rhythm with annual budgeting and regular audits, so the teams holding up the network get runway instead of paperwork. Revenue Development is high-ROI and milestone-gated, with capital released against demonstrated traction on six-month roadmaps run through Pods. R&D is a portfolio of staged bets funded through rolling RFPs, where a few outsized wins justify the spread. Matching funding mechanics to the nature of the work is where most of the efficiency gain comes from.

A clear evaluation metric per category. Kernel is assessed against agreed success metrics tied to core network resilience, uptime and operational continuity. Revenue Development is measured against network-level KPIs and its contribution to ecosystem growth. R&D is judged on what matures into — or gets absorbed by — Kernel and Revenue Development work. Three categories, three honest definitions of success, no more comparing a randomness beacon to a go-to-market motion on the same scorecard.

Dedicated stewardship. The Blueshift Foundation, a Cayman foundation, now stewards the program and is fully stood up and running. PGF has a durable institutional home rather than a rotating set of caretakers.

One shared record of past funding. Every Filecoin PGF dollar deployed — ProPGF, RetroPGF, and private direct grants — has been mapped and categorized into Infrastructure, Revenue, R&D, and Other, with upstream protocol infrastructure such as IPFS, libp2p, and IPC broken out separately. For the first time the ecosystem can see its full public-goods portfolio in one place, spot overlaps, and make allocation decisions against evidence instead of memory.

Accountability built in. KPI attribution and milestone mapping are in place, with biweekly and monthly reporting already running. Progress is visible continuously rather than at the end of a round.

AI where it removes toil. A chatbot is live on filpgf.io alongside a Telegram bot and AI dossiers for reviewers — cutting review load, shortening turnaround, and letting committee time go to judgment rather than document assembly. Committee simulations, AI primer docs, and queryable per-grant documentation are next.

📅 Retiring “batch”

Batch 3 is the last round to run under a single “batch” umbrella. Going forward, ProPGF funding will be organized by category-specific programs, each with its own cadence, review model, and funding rhythm — referred to independently, as in Kernel 2027, Revenue Development Round 2, and R&D RFPs.

Decoupling the categories means an urgent infrastructure need no longer waits for an unrelated round to open, an exploratory idea no longer has to be dressed up as a deliverable to fit a cycle, and a pod chasing revenue can move on its own clock. Fewer artificial deadlines, less scope creep, better decisions. We will publish the specific processes and calendars for Kernel, Revenue Development, and R&D in a dedicated follow-up post.

On the retroactive side: once the framework is fully solidified, Retro funding could support both Kernel and Revenue Development. We expect to align on an approach shortly and re-introduce RetroPGF in late Q3 / early Q4.

⚙️ Operational excellence as an ROI lever

Great allocation decisions lose value if execution is slow. filpgf.io is being built out as the coordination layer for the whole program, with four objectives: be the single source of truth for program and grant information, be transparent about criteria and decisions, showcase verifiable impact from funded work, and hold the program to published service levels.

Those service levels are now explicit — review assignments notified within a business day, standard milestone reviews closed within a week, complex reviews within two, and approved invoices targeted for payment in roughly two weeks. Grantee and reviewer follow-ups run on a set cadence, tracked in admin views and surfaced in biweekly turnaround reporting. Predictable, fast operations are not administrative garnish; they are how builders decide whether to build here.

🔭 Looking ahead

Our focus now is on operationalising these three pillars and turning them into a living system for Filecoin. Over the coming months, we’ll share detailed posts on how each pillar is being put into practice, the systems we’re building, and what we’re learning along the way.

A major bet we’re making is on AI-driven review and assessment. We’re building systems that can continuously monitor impact, gather context, surface relevant signals, and support reviewers throughout the funding lifecycle. The goal is not to replace human judgment, but to give reviewers better intelligence — making the process faster, more consistent, and increasingly data-driven.

Ultimately, our goal is to build the intelligence layer for Filecoin — a system that continuously learns from funding, impact, and ecosystem activity, and uses that intelligence to make better decisions about where capital should flow. We’re only at the beginning, but these experiments are laying the foundations for a more adaptive, data-driven, and AI-native funding system for the ecosystem.

Stay engaged: follow program updates on filpgf.io, browse funded projects and the latest reports, and join the conversation in the public ProPGF Telegram channel or #propgf on the Filecoin Slack.