Big ideas often hinge on better tools, and these tools are changing fast. As Zuckerberg and Chan refocus their philanthropy, the aim shifts toward AI that speeds discovery in biology. The Biohub labs they back take center stage, with computing power treated like modern lab benches. They spotlight immune-system breakthroughs and early detection while staying lean on real estate. Partnerships and GPUs, not bigger buildings, now drive the plan, which promises faster progress for medicine without diluting support for communities and schools.
Why Zuckerberg and Chan put Biohub and AI at the core
The shift arrived in a clear note: Biohub becomes the priority, and AI becomes the engine. They say the goal remains bold. Help scientists cure or prevent all diseases this century, and do it sooner as algorithms improve and data flows. The shift applies now in practice.
That intent gained a partner the same day, nearly two years after Chan signaled a science focus. Biohub aligned with EvolutionaryScale to speed understanding of disease through models and data. Better tools can shorten the path from early idea to tested insight, then turn insight into therapies. Work proceeds with urgency.
Since 2015, Zuckerberg and Chan framed giving as long-term. They pledged ninety-nine percent of Meta shares over their lifetimes, worth well over two hundred billion dollars. While the foundation once backed education, policy, and health together, the compass now points squarely at science. This course is deliberate by design.
How the new model works day to day
In 2024, Priscilla Chan told staff the identity is science-first, and the structure will match. CZI and its Biohub Network, which often operated apart, will pull closer to reduce drag. Integration, they argue, trims delays, aligns bets, and lets engineering support the labs that need it most. Teams want fewer handoffs each week.
Inside the labs, compute outranks floor plans. As Mark put it, researchers usually want GPUs more than space or headcount. Because code scales, a stronger cluster can replace square footage, while small teams move faster and avoid heavy overhead. That choice keeps budgets tight.
Even so, Zuckerberg and Chan are not closing doors on growth. They expect a mix of added sites and a central AI team, joined by shared compute. With a lighter footprint, the network can expand reach and still keep coordination tight across programs. It favors fast sprints overall.
Compute power as the catalyst, say Zuckerberg and Chan
Numbers show the push. Biohub plans to lift capacity to ten thousand GPUs by 2028, up from today’s thousand. Chan calls compute the new lab space, because it trains models, simulates biology, and frees time for wet-lab validation. More compute lets models train longer each night.
The research roadmap names four challenges, including using AI to reprogram the immune system. Early detection, better prevention, and smarter treatments could follow from that effort. If models compress cycles, then projects that took years might finish in months, with fewer dead ends. For Zuckerberg and Chan, that challenge anchors the plan.
Ambition runs high, and the language reflects it. Biohub says progress at this pace could unlock frontier medicine, because platforms compound. When code, data, and assays reinforce one another, the loop tightens and ideas turn into tools that scientists can share. Shared code reduces duplication and speeds review cycles.
Leadership changes and a clearer science-first identity
People and roles are shifting to fit the plan. COO Josué Estrada will step down in March, and operations will fold under science lead Marc Malandro. Putting a scientist in the number-two seat signals priorities and should cut steps between labs and leadership. It should simplify cross-team decisions.
Internal notes echoed the theme of speed, though leaders admitted details will take time. Because the map is still forming, teams expect iteration, not grand redesign. Staff heard a once-in-a-century chance to push cell science forward as AI and biology converge. Expect milestones, not big reveals.
Even with a pivot, Zuckerberg and Chan stress continuity. They will keep investing in local communities and public schools, while the biggest bets move to research. A unified mission, they argue, can support both without spreading resources thin across unrelated tracks. They want programs that last.
Shifts from education, and what remains in place
As one of the best-funded philanthropies, CZI’s tilt toward science did not surprise staff who saw earlier changes. Last year, CZI reduced education roles by dozens and paused its policy grants for that area. A broader platform gave way to narrower point solutions, with fewer features and sharper goals to improve traction. That reset shaped expectations.
Hiring reflects the reset. As of November 10, CZI listed multiple AI roles, and Biohub posted lab openings. The message to candidates stays consistent: come for the compute and the mission, rather than a salary built to match Big Tech. Open roles clearly signal priorities.
At the same time, Zuckerberg and Chan say they are not chasing size. They are building an AI core, while Biohub adds sites where it helps the work. Because capacity comes from clusters, not walls, the budget favors GPUs over concrete and furniture. That math guides tradeoffs quarterly.
What this bet on AI-powered biology could unlock next
A tighter focus and larger clusters add up to a simple wager. If compute lets small teams test ideas faster, Zuckerberg and Chan could see breakthroughs arrive earlier. They will still support classrooms and neighbors, yet science now leads the way, because momentum grows where tools align.
