A companion essay. This piece continues an argument begun in “Build, Innovate, Grow: What America’s New Fusion Roadmap Means for India.”
In my last essay I argued that the deepest lesson of America’s new fusion roadmap is a change of posture — the state stepping back from building the first power plant, and instead building the shared infrastructure that lets private companies build it. That argument invites an obvious question for India: if our public institutions are to enable a private fusion industry, what exactly should they build?
After some months of looking closely, I think the more urgent question is what they should open. A great deal of what India needs already exists, scattered across the units of the Department of Atomic Energy. The fastest and cheapest way to accelerate fusion here is not a greenfield campaign of new institutes; it is to upgrade, repurpose, and — above all — open a defined slice of the capability the DAE has spent six decades building.
I do not write this from the outside. I carried out my own doctoral fusion research at Berkeley; spent nine years as a scientist at the Saha Institute of Nuclear Physics, where I worked on India’s first tokamak, commissioned in 1987; and later led the ADITYA tokamak at the Institute for Plasma Research, along with the operation and control group of its SST-1 superconducting tokamak. Several of the units I describe below I have worked inside, and I have watched this establishment’s plasma capability grow from a single small tokamak into a national programme. What follows is, in part, an argument that India underestimates what it already owns.
A platform, not a building
The American roadmap calls its shared infrastructure the Tritium-Blanket Development Platform — a distributed network of test stands and loops, public and private, that any developer can draw on because no single company can justify building them alone. India can assemble an equivalent almost entirely from assets it already holds. The right unit of thinking is not a new national laboratory but a platform — a coordinated set of shared user facilities, each anchored in an existing DAE unit and opened, under clear rules, to vetted private developers. Let me walk through fusion’s hardest gaps and where, in the DAE, each could be addressed.