The supply chain of the next epoch

Fusion, artificial intelligence, and quantum computing are the load-bearing trends of the coming decades. They get the magazine covers, the senate hearings, and the model-portfolio allocations. They are also useless without a tier of adjacent technologies quietly compounding alongside them.

Four are worth dwelling on.

Synthetic biology

The cost of writing DNA has fallen by roughly four orders of magnitude in the last fifteen years and continues to fall. This has quietly turned biology from an observational science into an engineering one. The applications cluster in three places:

• Industrial biology. Engineered organisms producing speciality chemicals, materials, and food ingredients at lower cost and lower environmental impact than petrochemical or extractive alternatives.
• Therapeutics. Cell and gene therapies, mRNA platforms, precision oncology — the parts of pharma that look least like pharma and most like software.
• Agriculture. Climate-resilient crops, nitrogen-fixing cereals, alternative-protein supply chains.

The interesting business question is when synthetic biology stops being a science-tools market and becomes an end-product market. Several firms have crossed that threshold in the last two years.

Robotics and autonomy

Industrial robotics has been a real industry for forty years and a boring one for thirty-nine of them. The current generation is different in three respects: arms are cheap enough that a small-and-medium-business labour-replacement use case finally closes; visual and tactile perception has crossed a threshold; and foundation models for control are doing for robotics roughly what they did for natural language. The combination of these means that generally-deployed robotics — not industrial-cell robotics, but robots that show up in arbitrary environments and do useful work — is moving from research demonstration to fielded product on a visible timeline.

Whether the durable economics live with the platform companies, the robot-as-a-service operators, or the underlying foundation-model labs is an open question we expect the next five years to answer.

Advanced materials

Almost every other frontier on this page is gated by materials. Fusion needs better superconductors. Quantum hardware needs purer substrates and better cryogenics. Batteries need higher energy-density chemistries that don’t catch fire. Solar needs perovskite or tandem cells with stable lifetimes. Aerospace needs composites that handle hypersonic re-entry without ablating.

Materials development has historically been slow because the search space is enormous and the experimental loop is expensive. This is the most direct application of computational chemistry, AI, and (eventually) quantum simulation — closing the loop between proposed compositions and tested ones from years to days.

The companies that own the experimental data and the inference loop on top of it will compound rapidly.

Space

The cost-per-kilogram to low Earth orbit has fallen by an order of magnitude in the last decade and is expected to fall by another in the next. This collapses the unit economics of every business that might want to do work in orbit. Earth observation, communications, in-space manufacturing, and the early stages of an off-Earth materials economy all become tractable in ways they were not.

Space is the rare frontier where the headline cost curve is finally working in the operator’s favour rather than against it.

What we look at

We treat all four as serious frontiers in their own right, not as sideshows to the headline three. Each interacts with the others — biology depends on materials, robotics depends on AI, space benefits from all three — and the most interesting opportunities tend to live in the seams between two or more of them.

The next epoch will be built from a small number of headline technologies and a much larger number of unglamorous ones. We pay attention to both.