The Center for Therapeutic Genetics, announced Tuesday and backed by $34.5 million from ARPA-H, brings together a founding group that reads like a roster of the field. It includes David Liu, inventor of base and prime editing; Timothy Yu, a pioneer of individualized genetic therapies; Cat Lutz, who leads rare disease model development at the Jackson Laboratory; and Wendy Chung, chief of pediatrics at Boston Children's.
The most revealing thing in the launch materials is not a scientific claim. It is Lutz's framing: for many rare diseases, the question is no longer whether we can treat them, but whether we will build the systems to do it.
That is a statement that the bottleneck has moved, and it is worth taking seriously, because it reframes what kind of problem this is.
The current model asks families to become biotech companies
To see why an organizational fix matters, look at how a child with an ultra-rare mutation gets treated today.
With relatively few gene therapies on the market, families of children with rare genetic disorders have few options but to partner individually with researchers and raise millions of dollars to develop treatments for their conditions.
Read that as a job description. A parent receiving a devastating diagnosis must identify the right academic lab, secure its interest, raise seven figures through philanthropy or media attention, coordinate manufacturing, and navigate an FDA pathway designed for products with thousands of patients. All while caring for a critically ill child, frequently against a clock, since many of these conditions cause irreversible damage in the first months of life.
Some families succeed, and those cases become news stories. The success rate is not really a measure of scientific feasibility. It is a measure of how many families can execute a biotech startup under the worst circumstances of their lives.
The innovation is a platform, not a molecule
What CTG proposes is that the components be reused. Central to the model is a platform strategy, in which design tools, disease models, manufacturing processes, safety data, and clinical protocols developed for one program carry to the next.
This sounds like administrative detail and is the entire point. In a bespoke therapy, most of the work is not the part unique to the patient. Designing a specific guide RNA for a specific mutation is comparatively fast. What consumes time and money is everything around it: establishing that the delivery vector is safe, validating manufacturing, building animal models, writing clinical protocols, assembling a regulatory submission, standing up monitoring.
Under the current arrangement, each family pays for all of that infrastructure as though no one had ever done it before. Under a platform, those fixed costs are incurred once and amortized. The first therapy is expensive; the fiftieth costs a fraction, because only the patient-specific component is new.
Yu's version of the ambition is that the methods developed for one rare disease will carry to the next, and eventually to the many, turning what were remarkable exceptions into something routine. The center also intends to share methods, data, and training with other clinicians so they can do the same, which matters more than the therapies it develops itself. A platform others can adopt scales; a center that treats its own patients does not.
Why a commercial company could not do this
The nonprofit structure is not idealism. It follows from arithmetic that no business model resolves.
A therapy for a mutation carried by three people worldwide has no market. There is no price at which development costs are recovered, because the denominator is three. Orphan drug incentives help for conditions affecting thousands, and they do nothing for n-of-1 medicine, where the patient population is a single child.
That leaves two options. The first is what exists now: philanthropy organized ad hoc by individual families, which works occasionally and selects poorly. The second is treating the underlying capability as public infrastructure, funded by government and nonprofit sources, with the marginal cost per patient driven low enough that treating one more child is not a fundraising campaign.
CTG is a bet on the second, and ARPA-H funding for a nonprofit consortium is exactly the instrument the arithmetic calls for. It also depends on that funding remaining durable across budget cycles, which is not guaranteed. A separate federal rule proposed the same year would give agencies broad authority to terminate any active grant for convenience, and a platform built to compound over years cannot compound if any individual award can be pulled at any time.
The regulatory half of the problem
A platform alone would not be enough, because the second bottleneck is regulatory, and it happens to have moved in the same year.
You cannot run a randomized controlled trial with one patient. There is no control arm, no statistical power, no way to satisfy the evidentiary standard drug approval was built around. That is not a bureaucratic obstacle so much as a genuine epistemological problem: how do you know a therapy worked when the sample size is one?
In February 2026, the FDA issued draft guidance for sponsors seeking approval for targeted individualized therapies by generating substantial evidence of effectiveness and safety when randomized controlled trials are not feasible due to small patient populations, specifically addressing genome editing and RNA-based therapies.
That guidance and this center are complementary halves. A platform that cuts development cost is not useful if no resulting therapy can be authorized, and a flexible regulatory pathway is not useful if each therapy still costs millions to reach. The science, the manufacturing infrastructure, and the evidentiary framework all had to arrive together, and they appear to be converging.
The equity problem this addresses without saying so
There is an uncomfortable feature of the current model that the platform approach quietly targets.
When access depends on a family's ability to raise millions and recruit an academic collaborator, the selection criteria are not medical. They favor families with money, professional networks, media savvy, English fluency, proximity to a major academic center, and the capacity to spend a year on fundraising. A child with the same mutation whose parents work two jobs in a rural county does not get a therapy, and the difference has nothing to do with the biology.
Driving marginal cost toward something a hospital could absorb, and distributing the methods to clinicians elsewhere, changes who is eligible. That is a larger effect than any individual therapy the center develops.
Reasons for restraint
Several caveats are worth holding, and the founders acknowledge some of them directly.
The center is in its founding phase and is not yet accepting patient referrals or providing clinical care. Nothing has been demonstrated at scale.
The funding is modest relative to the ambition. $34.5 million is a fraction of what a single conventional drug program costs, which is consistent with the platform thesis, since the whole premise is that shared infrastructure is cheaper, but it also means the model must work quickly to justify further investment.
The technical challenges are real and not solved by organization alone. Delivery remains the central unsolved problem in gene editing: reaching the right tissue, particularly the brain, at sufficient efficiency without immune complications. Base and prime editing are precise, and precision does not resolve delivery.
The meaningful test is not the first therapy. It is the second and the tenth. If CTG's model works, each successive program should cost less and take less time than the one before, because the reusable components accumulate. The people who invented the relevant tools are now saying the frontier is logistical, and building an institution premised on that. If they are right, the constraint on treating rare genetic disease was never entirely what we could discover. It was what we bothered to industrialize.