To Conquer the Primary Energy Consumption Layer of Our Entire Civilization

3 days ago 1

Three years ago we set out to make cheap synthetic natural gas from sunlight and air. At the time I didn’t fully appreciate that we had kicked off the process of recompiling the foundation layer of our entire industrial stack.

Last year, we made cheap pipeline grade natural gas from sunlight and air and expanded our hydrocarbon fuel road map to include methanol, a versatile liquid fuel and chemical precursor for practically every other kind of oil-derived chemical on the market. Unlimited synthetic methane and methanol underpinning global energy supply is a good start, but it is now abundantly clear that we have a strategic need to re-localize our supply chain for all critical minerals.

We wanted to understand the effect of geography on fuel production cost, so we performed a series of detailed simulations in a range of locations. These simulations allowed us to calculate, for any given location, solar PV, battery, and load cost, the optimal array and battery size, asset utilization, and overall effective power costs. Each simulation took real world solar PV production data at 5 minute intervals over an entire year.

One key question for equipment operators considering an off grid power system using only solar and batteries is: How much will power really cost? It depends on how sunny and seasonal your location is, but it also depends on how much you want to operate your equipment in order to maximize revenue per dollar spent. This graph solves that problem, automatically calculating the optimal asset utilization for any asset cost class, and displaying the corresponding cost of power.

More formally, the graph shows the optimized effective amortized power cost per MWh consumed by a load characterized by its capex measured in $/MW. For example, an inexpensive electric kettle is about $20/kW, while an AI training datacenter is more like $50,000/kW. Utilization is given by the saturation of the data point, and we’ve selected six locations for a representative range of weather and seasonal conditions.

Power system costs are baselined at $100/kW for solar PV and $100/kWh for the batteries, which we think are representative of near future conditions. For reference, the cheapest solar modules are currently $82/kW and cheapest batteries are $94/kWh. The results scale smoothly, so if you strongly believe that the true numbers are $0.20/W and $200/kWh, simply multiply the axes values by two – but it doesn’t change our conclusions.

There are several salient features.

First, in nearly all places and for nearly all load types, solar+batteries are cheaper than any other potential power source, replacing the concept of an expensive, undesirable “green premium” with a “green dividend,” a rare win-win for using advanced new power technologies. We have always believed that the revealed preference of the market is that less polluting technology is enthusiastically adopted, if and only if it also saves the customer money!

The second is that power costs do increase as we drive to a large number of “9s” of reliability towards the right edge of the graph, but even for loads as expensive as AI training centers, the optimal utilization is probably closer to 99.9% in terms of weights updated per dollar spent. In practice, rather than the datacenter going dark for 8 hours a year, 99.9% utilization means some partial throttling in winter over longer time periods.

Third, there is a clear bifurcation between battery supported loads over $2000/kW that run at utilization greater than 75%, and diurnal, battery-free solar loads under $400/kW that run at utilizations below 30%.

We’ve seen recent studies on off grid solar power and the cost of reliability with solar in general but as far as we’re aware, ours is the first study to comprehensively demonstrate the tremendous opportunity in low capex intermittency-friendly matched loads. We intuitively understood this for synthetic fuel, but fully generalizing this for all consumers of energy in all geographies was… non-trivial.

WIth the benefit of this insight, we can redraw the graph above with some directional commentary. We now have conviction that low capex, intermittency-friendly loads can unlock a roughly 5x bigger, greener dividend. Green is the color of money, after all!

But what are these loads? What do they do? What do they look like? They must be relatively small, power intensive, cheap, and intermittency-friendly. Their value add is due to privileged access to cheap power. But colocating some miniaturized chemical plant, factory, or other business right next to a solar array is not without its challenges, so what processes and products are worth the hassle?

Let’s take a look at the energy landscape of our industrial economy.

This graph shows which industries derive their value from consuming energy (bottom right) vs some other high value-add operation per unit energy (top left). They are materials, utilities, and oil and gas. The primary production layer of industry, which is concerned with processing naturally-occurring raw materials into much higher value intermediate commodity products.

For example, of the ~6 processes involved in making aluminum cans from red mud, 60% of the total energy consumption occurs just in the Hall-Héroult process, in which the refined alumina is electrolyzed. Not coincidentally, this is also the step with the single largest value add, so exploiting cheap energy reduces aluminum production cost, improving access for buyers worldwide.

This insight prompts the following recipe for processes that Terraform’s unique capabilities provide exclusive economic access to.

They have relatively generic raw inputs that are readily available: sunlight, air, water, various commonly occurring rocks and minerals.
They adapt high maturity existing industrial processes to consume 2-3x more electricity in order to get to lower capex, cutting down from $2000/kW to $200/kW.
They adapt existing industrial processes to be intermittency-friendly, that is, capable of throttling up and down to match variations in available solar power due to weather and the day/night cycle.
They adapt existing industrial processes to operate at the 1 MW scale immediately adjacent to the array, which is much smaller than typical plants. Miniaturization enables cost control through manufacturing rather than construction. It also helps with intermittency, since it’s easier to turn a small ship than a big ship.
The output products are relatively valuable, storable, stable, and now contain most of their embodied energy. We are not advocating relocating the entire industrial stack to endless fields of solar panels! But Terraform’s products will require relatively little additional energy input in high value-add secondary manufacturing steps, which will occur in more conventional centralized factories with less sensitivity to power costs relative to other factors of production. Most of our value add is occurring in the energy intensive processing step using Terraform’s proprietary technology.

Let’s examine the set of energy-intensive primary products for compatibility with this recipe. I’ll include the chemical formula of the most energy intensive steps in production.

Hydrogen production (2 H₂O → O₂ + 2 H₂)
Synthetic natural gas (2 H₂O → O₂ + 2 H₂, CO2 + 4 H₂ → CH₄ + 2 H₂O)
Synthetic methanol (2 H₂O → O₂ + 2 H₂, CO2 + 3 H₂ → CH₃OH + H₂O)
Synthetic coke/graphite/graphene (pure carbon) (CO₂ -> C + O₂)
Ammonia (6 H₂O + 2 N₂ → 4 NH₃ + 3 O₂)
Cement (CaCO₃ → CaO + CO₂)
Steel (2 Fe₂O₃ → 4 Fe + 3 O₂)
Non-ferrous metals (eg 2 Al₂O₃ → 4 Al + 3 O₂)
Desalination (something like Na⁺.2X H₂O → Na⁺.X H₂O + X H₂O)
Exotics (secret projects)

There are some other highly energy intensive primary industries that do not match this recipe, either because their energy inputs (ie. low grade heat) are cheap to store or their value creation is not sensitive to energy cost.

Paper. Energy consumption is mostly heat, which is easy to store, so there’s no benefit from intermittency-compatibility. That said, it is possible to make a water-proof paper substitute from CaCO3 and polyethylene.
Glass. Similarly to paper, there’s no energy intensive chemical reduction step occurring.
Data center operations. While energy intensive, the load is so expensive there’s no benefit to operating intermittently and the product is relatively insensitive to power cost, provided that there is power supply.

The energy intensive step in common is chemical reduction, or ripping the oxygen atoms off, which is the inverse of combustion. Cement is an exception, but our direct air capture process performs calcination (ripping CO₂ molecules off of limestone) so we already have a mechanism for this highly energetic process. Desalination is another exception, but in this case we’re ripping hydrated salt ions off respectively the oxygen and hydrogen parts of water, and keeping the water, and the business model is a bit different.

These products underpin our entire civilization. They consume a prodigious quantity of energy, usually in the form of coal. Or, in the case of coke, literally are coal. In all cases, these are commodity products, with zero market risk and steadily rising consumption. Their price and scarcity are fundamental constraints to the progress of the human condition.

We see a path to cost parity with all existing production processes in the US by 2031, often sooner. Terraform’s proudly energy intensive philosophy positions these processes to reap ongoing and substantial production cost advantages courtesy of solar’s continuing price declines.

Since the beginning of the industrial revolution, growth has been constrained by energy availability, and primary energy production was the imperative. With the solar PV revolution, we have an inversion of this paradigm, similar to the effect described by Aggregation Theory. The new imperative is to find better ways to consume this unexpected bounty in productive ways.

To occupy the solar cost sweet spot at 30% utilization with $200/kW capex loads obliterating oxides forever.

To conquer the primary energy consumption layer of our entire civilization.

Reflecting this strong conviction, Terraform Industries is ramping up maximally aggressive development efforts for solar-adapted processes to make all these products. Terraform is the place where the world’s most far sighted visionaries come to build for a millennium of material abundance. Our pioneering work in this sector has made us the natural leader and ideal partner to develop this portfolio and allocate capital across each of these verticals.

Here’s a non-exhaustive list of current career opportunities at Terraform.

Technical Chief of Staff
Technical Recruiter

Electrolyzer System Lead
Electrolyzer Production Mechanical Engineer
Electrolyzer Product Technician
Electrolyzer Test Technician

Direct Air Capture System Lead
Direct Air Capture Cyclonic Calciner Engineer
Direct Air Capture Mechanical Engineer
Direct Air Capture Technician

Reactor Injection System Lead
Reactor Injection Compressor Engineer
Reactor Injection Controls Electromechanical Engineer
Reactor Injection Technician

Methane Reactor Lead Engineer
Methane Reactor Design Engineer
Methane Reactor Mechanical Technician

Methanol Production Lead Engineer
Methanol Reactor Chemical Engineer
Methanol Reactor Design Engineer
Methanol Reactor Mechanical Technician

Photovoltaic Systems Lead Engineer

Ops/GSD Lead
Project Development and Construction Lead
Terraform Sales Lead
California Permitting Lead

Coke Project Lead Engineer
Coke Mechanical Technician

Ammonia Project Lead Engineer
Ammonia Reactor Chemical Engineer
Ammonia Reactor Mechanical Engineer
Ammonia Reactor Mechanical Technician

Cement Production Lead Engineer
Cyclonic Systems Mechanical Engineer
Cement Process Technician