PhantaField has closed its angel financing round, eighteen months after receiving its first check from Prof. Chenming Hu. The round brings together semiconductor-industry veterans and deep-tech investors around a single, deliberately contrarian thesis: that the next order of magnitude in computing performance comes from stacking atoms vertically, not from shrinking them further in a plane.
It is a thesis that runs against the grain of where most semiconductor capital flows. The industry's reflex is to chase the next node — smaller features, costlier lithography, diminishing returns. PhantaField's bet is that the cheaper and more durable lever is the third dimension: grow a transistor layer directly on top of the last, and a mature, inexpensive 28 nm process can outperform a bleeding-edge planar one.
What it funds
Proceeds underwrite the two programs that define the company's next two years. The first is industrial: scaling the plasma-photon MOCVD platform from the 8-inch demonstrator to a production 300 mm system capable of qualifying a process, not merely proving one. The second is architectural: building out the team behind what would become Sophon — the compute-in-memory design that turns a stack of grown monolayers into a die that trains and serves frontier models.
Together those two efforts trace the company's whole strategy in miniature: own the tool that grows the atoms, and own the architecture those atoms make possible.