I
Foundational intelligence
Model laboratories and the silicon and data infrastructure that train them. Capital where the half-life of advantage is short and the cost of being right is long.
We accept that most will not, and hold the rest for as long as the work requires. The list below is illustrative; specific portfolio companies, partners, and commitment sizes are not disclosed.
Approximately 10% of the balance sheet.
Model laboratories and the silicon and data infrastructure that train them. Capital where the half-life of advantage is short and the cost of being right is long.
Generative protein design, cell therapy platforms, longevity science. The slow conversion of biology into engineering, on a timescale measured in scientific careers.
Fusion, advanced fission, and the storage chemistry that an electric grid will require. The prerequisites for an energy economy that does not warm the planet.
Reusable launch, Earth observation, and the first generation of in-space services. A new layer of infrastructure built once and used for decades.
Machines that perceive and act in the physical world: industrial, defence, care. Where intelligence finally meets the mechanical work that runs an economy.
Industrial-scale carbon removal, novel materials, the agricultural transition. The companies that will be measured in gigatons, not quarters.
“Frontier capital is not adventure capital. The horizon is long; the discipline is the same.”