I still remember the first time I heard a parent hand a thin PDF to a young doctor and say, “This is everything we know about my child.” That moment stayed with me. It captures exactly why Priscilla Chan has spent more than a decade betting big on rare disease research and the basic science that most funders overlook. What she and Mark Zuckerberg are building through Biohub and the Chan Zuckerberg Initiative feels less like traditional philanthropy and more like a quiet revolution in how medicine gets done.
Why Rare Disease Became the Unexpected Launchpad
Most people think of rare diseases as tiny, isolated problems. Chan sees them as the sharpest possible window into how the human body actually works. When she trained as a pediatrician at UCSF, she kept meeting children whose conditions did not appear in textbooks. Families arrived with homemade research packets. Doctors ordered treatments largely by trial and error. That experience never left her.
She has said more than once that the knowledge base for many of these conditions is shockingly thin. Forty letters different in a genome can change an entire childhood, yet science often has almost nothing useful to offer. That gap is what pushed her back toward basic research after years of clinical training. In my view, that decision still ranks as one of the most clear-eyed moves in modern medical philanthropy.
The Personal Roots That Still Drive the Work
Chan did not arrive at this mission through abstract theory. Her grandfather died suddenly when she was twelve. One morning he dropped her at school; by afternoon the family was saying goodbye. The explanations were vague. Cultural silence around illness made the unanswered questions even louder. Years later, for an independent high-school project, she ordered a cancer textbook and tried to understand what had happened. The details mattered less than the feeling that stayed with her: not knowing why is unbearable when someone you love is involved.
That same drive showed up again during her residency. She cared for a teenager whose cancer returned after an earlier remission. She still keeps the card that patient wrote. Those memories sit beside the more hopeful ones—kids with cystic fibrosis who somehow thrived when others struggled. Positive outliers and heartbreaking losses both pointed to the same conclusion: we simply do not understand the system well enough yet.
From First Child to First Big Commitment
Almost eleven years ago, while preparing for the birth of their first child, Chan and Zuckerberg sat down and wrote out what they wanted to contribute. Survival of the new baby felt like the immediate priority, of course. Beyond that, they asked a larger question: what could they uniquely bring that would improve the lives of all children? Mark brought an engineer’s instinct for breaking systems into parts and rebuilding them. Priscilla brought frontline experience with both classrooms and hospital wards. Together they decided the gap in understanding human biology was the place where their strengths could matter most.
They set a goal that still makes many scientists pause: cure, prevent, or manage all disease by the end of the century. Critics called it impossible. Chan has explained that the word “all” was deliberate. Focusing on a single disease would have encouraged clever one-off connections. Aiming at every disease forced a strategy of deep, systematic understanding. That choice still shapes every major investment they make.
Building the Foundations Most People Skip
Early on, the pair put significant resources into single-cell sequencing. Every cell in the body carries the same genome, yet a heart cell and a hair cell behave completely differently. The difference lives largely in which parts of the genome get read—the RNA transcriptome. Funding that work felt, to some observers, like stamp collecting. Chan and Zuckerberg disagreed. You cannot debug a system you do not understand.
That early bet helped accelerate the Human Cell Atlas. Today the project holds one of the largest open collections of cellular data in existence. Most of the data now comes from other groups. The open-science approach created a flywheel: more data attracted more contributors, which made the resource more valuable, which brought still more participants. During the pandemic the same datasets revealed previously unknown lung cell types especially vulnerable to the virus. Those discoveries arrived only because the foundational maps already existed.
You cannot find the error when you do not understand the underlying system the human body is built on.
I’ve found that this particular mindset—map first, then intervene—separates the most ambitious science efforts from the rest. It requires patience and a willingness to fund work that does not produce an immediate drug. Few organizations stay committed long enough. Biohub was designed precisely for that long horizon.
When Large Language Models Met Cellular Data
The arrival of powerful language models changed the timeline. Suddenly the enormous datasets Biohub and its partners had collected could be interrogated at a scale no human team could match. Chan has described looking up “LLM” years ago and then taking a linear algebra course simply so she could understand the underlying mathematics. That curiosity is typical of how she operates.
More recently the organization began collaborating with teams building evolutionary-scale protein models. Those models computed structures for over a billion proteins in roughly a week—work that once required years of laboratory effort for a fraction of the targets. The practical implication is enormous. If a rare genetic change causes a protein to misfold, researchers can now design corrective proteins computationally and test them far faster than traditional methods allow.
Chan believes this approach will eventually transform even common conditions. Take depression: current practice often involves trying one medication after another while waiting weeks for results. Understanding the cellular and genetic context of each patient could shorten that painful search. Rare disease families, she argues, simply reached this insight earlier because they had no other choice.
Rare As One and the Power of Patient Networks
About eight years ago a colleague named Tanya approached Chan with a proposal that initially felt outside the usual philanthropic lane. Rare disease patient groups were already gathering natural-history data, building registries, collecting samples, and driving research agendas faster than many academic labs. Chan was cautious at first. After seeing the applications, she was convinced.
The Rare As One Network was born. Early cohorts of patient organizations received support to professionalize their scientific efforts. The results have been striking. Some groups moved from unknown mutation to gene-therapy trials in humans within a decade. That speed once seemed impossible. Now it is becoming a model. Chan has grown increasingly confident that the methods these families pioneered will spread to the rest of medicine.
- Patient registries that capture real-world progression data
- Shared sample collections that lower barriers for researchers
- Direct collaboration between families and scientific teams
- Willingness to fund early, high-risk experiments
These practices do not stay confined to ultra-rare conditions. They point toward a future in which every patient cohort, common or rare, helps generate the knowledge needed for better treatments.
Biohub as the Engine for Frontier Science Plus Frontier AI
A year after launching the broader initiative, Chan and Zuckerberg created Biohub as a distinct organization focused on basic science and the integration of advanced computing. The separation mattered. Biohub could concentrate on the tools and datasets that traditional grant systems often undervalue. In April the organization announced a $500 million commitment to an AI platform designed to help scientists understand how human cells function inside the body. That figure alone signals seriousness.
The underlying philosophy remains consistent. Build the measurement tools. Generate open data. Apply the most powerful computational methods available. Then watch the discovery cycle accelerate. Chan expects discoveries powered by current models to reach animal and human studies by the end of next year. That timeline would have sounded optimistic only a few years ago. Today it feels plausible.
What Success Actually Looks Like
Success, in Chan’s framing, is not a single dramatic cure announced on a particular date. It is a steadily rising speed of learning. Rare disease groups have already compressed the path from genetic discovery to clinical trial. Foundational models and richer cellular maps should compress it further. Over five to ten years she anticipates a noticeable shift in how new treatments move from idea to patient. Over longer horizons the same infrastructure should begin reshaping care for common diseases as well.
Perhaps the most interesting aspect is cultural. Ten years ago the goal of addressing all disease drew polite skepticism. Today people more often ask how rather than whether. That change in expectation may prove as important as any single technological advance.
Balancing Urgency with Systems Thinking
Chan has spoken openly about the tension between wanting faster results for patients and recognizing that durable progress requires solid foundations. Mark’s engineering background reinforces the long-view approach: build the learning engine first. Her clinical background keeps the focus on whether the work will eventually matter for real people. The combination appears productive. They adjust strategy as new tools appear—large language models being the clearest recent example—without abandoning the original commitment.
They have also accepted that no single organization can fund everything. Their contribution, while substantial, remains a small fraction of the total scientific enterprise. The strategy therefore emphasizes leverage: create open resources that others can use and improve, support patient communities that already move quickly, and invest in the measurement technologies that unlock entire fields.
Lessons That Extend Beyond the Lab
Watching this effort unfold offers a few practical observations. First, ambition works better when paired with humility about what is not yet known. Second, patient communities often possess both the motivation and the organizational capacity that traditional institutions lack. Third, open data and open tools multiply the impact of any single investment. Fourth, the most powerful computational methods become useful only when high-quality biological data already exist.
I keep returning to the image of that parent handing over a thin PDF. The goal of the work Chan describes is to make those moments rarer. Not by promising miracles, but by steadily expanding the territory where science actually has answers. Progress will remain uneven. Some conditions will yield faster than others. Yet the direction of travel feels clearer now than it did a decade ago.
The next few years will test whether the combination of rich cellular maps, protein-design models, and highly motivated patient networks can deliver the acceleration many hope for. Early signs are encouraging. Groups supported through Rare As One have already reached human trials that once seemed distant. Computational tools continue to improve. The culture of collaboration around open biological data keeps expanding.
For anyone who has watched a family navigate a rare diagnosis, the stakes are personal. For the rest of medicine, the methods being refined today may become standard practice tomorrow. That dual impact—immediate help for the most urgent cases and longer-term transformation of the entire field—explains why Chan continues to treat rare disease research as a strategic priority rather than a niche interest.
The work remains unfinished, of course. Understanding every cell type, every regulatory circuit, and every possible failure mode is a multi-decade project. Yet the foundation is stronger than it was, the tools are sharper, and the community of people pushing the work forward is larger. That combination is what makes the current moment feel different. The bet is still bold. Increasingly, it also looks smart.
Chan has noted that the rare disease community saw the future of medicine first. They refused to wait for someone else to get lucky. They organized, collected data, and demanded systematic approaches. The rest of the field is now beginning to catch up. If the trajectory continues, the families who once had almost nothing may end up having changed medicine for everyone. That outcome would be a fitting legacy for the patients and parents who refused to accept thin PDFs as the final word.
In the end, the story is less about any single technology and more about a sustained commitment to understanding. Map the system. Share the maps. Apply the best analytical tools available. Listen to the people living with the conditions. Repeat. It sounds simple when written down. Executing it at scale, year after year, is the hard and necessary part. Priscilla Chan and the teams around her appear determined to keep doing exactly that.