Firstly, prices at the wholesale level are absolutely going down and that is all the more remarkable because it is happening while LNG-linked gas prices are rising everywhere because of the misadventures of yet another US government in the Middle East.

The chart below shows baseload contracts for NSW and Victoria and the price of caps for NSW.

Caps pay out at the sum of all prices above $300 in a quarter hourly divided by hours in a quarter. Full contract details here. Simplistically caps price how spiky prices are above $300 per MWh, baseload takes the time weighted average. Both are going down - a lot, and that seems to be driven by the reducing in spikes that is caused by batteries competing with gas.

This is now filtering down to “all-in” prices for retail - outright nominal deflation while everything else is off to the moon thanks to Trump’s Iranian adventure. Original documents here.

For small businesses the cuts are into the double digits in most areas

So:

• Wholesale prices are going down

• Retail prices are going down

• This seems to be driven by storage dynamics pricing out gas

There is a great graph of this showing what types of generation make up the peak in Queensland now:

As they say at the poker table, read ‘em and weep.

For those of you who have not watched Burn After Reading the scene is here. A list that I may update over time:

With Enough Rockets and Drones Sovereignty is Possible without Nukes

Iran has beaten the United States to a humiliating defeat, and it has done it without nuclear weapons. There is an interesting contrast in that the one nuclear power in the Middle East - Israel - has been pummeled for a month with conventional weapons. Gradations of escalation matter and when it comes to rockets and drones the Stalin quote “quantity has a quality all of its own” applies.

Blue Ocean Navies Are Not What They Once Were

It is far past time to call BS on the “laundry fire”. If Iran can hit a carrier with a drone or the like, then many other states definitely can. Stand-off distances are now longer, and US gunboat diplomacy is vastly weaker.

This Was an Economic Supply Chain War

Iran won by showing it could inflict vastly more pain on its opponents than they could reasonably take. This is because:

1. US client states had very high share of production in key upstream inputs (ammonia, LNG, oil, urea)

2. That production could be destroyed or damaged

3. That production could be blocked by the enemy

The first problem is hard to fix without interventions like a strategic reserve - and I would remind you all that while Australia passed that legislation quickly the Secure Act remains unfinished business in the US House and Senate.

The second problem looks completely intractable - without directed energy weapons improving or being located outside of drone saturation range there is no clear answer here.

The third also looks intractable - it is hard to stop people making drones and medium or even long-range rockets.

US Defense Suppliers are Completely Unreliable

If the US will not deliver hardware to Switzerland that it has paid for then why should Australia buy submarines from the US? These subs will likely get embroiled in a Taiwan conflict and make Australia a target as with Qatar and UAE - what exactly is in it for Australia? TSMC Townsville would solve the supply chain issue and generate tax revenue from an Australian point of view. In the interim other equipment providers which care about their credibility from Europe, South Korea or Australia should be strictly preferred. The best choice is in country production by local companies, next is localization with an international partner, last-worst is US supply.

US Client States’ Incentives Were Not Aligned With The US

All Gulf states that hosted US bases sought to drive downstream from oil and gas production to add value as well they might because of their structural cost advantage. This just compounded dependencies on them and if they cannot be reliably defended then they cannot be relied upon. The US and Europe likely have to live with this in the short term but longer term this cannot fly. The states themselves will fight tooth and nail to keep the US and Europe hooked because that is what will keep them safe(-ish). Look out for Tucker Qatarlson attacking green ammonia, methanol and non-Middle East sulfur supply soon. It is the job he is paid to do.

Similarly, Israel would prefer the US remain dependent upon the region to keep the US locked into its security which Israel can routinely destabilize with no apparent consequence. On this point Tucker and Netanyahu are on a unity ticket.

US Defense Guarantees are Contingent Upon Inflation

The US political system cannot tolerate inflation - it has played a significant role in the decline of multiple presidencies including Carter, Bush, Biden and now likely Trump. For US’ enemies the message is clear: spike inflation and you can negotiate with the next guy. This makes the US fundamentally weak in the face of any opponent who can generate an inflation shock which is Iran, China and Russia - the big three.

US Inflation is Not Stable Without Secure Supply Chains and Decarbonization

If the US political system cannot stand up to inflation, then the US must fix its vulnerabilities in supply chains that cause it. There is an aesthetic preference on the right for spending more on defense as a macho exercise but to me this looks like people who spend a lot of time on “beach muscles” and skip leg day every time. Leg day is the periodic table and energy - hard, grinding stuff that actually makes you genuinely strong. While US is doing the latest Instagram bicep routine from the guy clearly on steroids, China is doing Bulgarian split squats with a 100kg bar in the weird barefoot toe shoes.

1. No vessels are exiting the Strait (yet): Clearly despite all the talk nobody with $50mm+ of metal to move and another $100mm+ of oil wants to risk it for now. Until that happens then oil markets remain tight and physical premia are still rising versus futures though front months have sold off heavily.

2. Trump remains untrustworthy: To say the current US government is a low trust counterparty is a major understatement. After assassinating numerous Iranian leaders VP Vance’s decision to go to Pakistan is somewhere on the Gavrilo Prinzip to Kenny from Southpark level of bravery.

3. No sign of troop drawdown in the Middle East: There is no sign of reduced deployments indicating a firm deal.

4. The current cope online from Trump adjacent people is that this oil is not their problem because the US produces a lot of oil. There are two rebuttals I have here:

   1. Firstly, that is flat wrong and there is a great paper on this here.

   2. Secondly, if you think anyone in Asia or Europe has a deep ideological problem with bribing a stationary bandit and this represents a shocking change to international norms and so forth see this map of where Roman coins are found:

      

5. US efforts to reduce decarbonization are going to utterly fail and demand for US LNG will get hit hard in the longer term. I discussed this at length on the latest episode of the Polycrisis podcast.

6. Everyone in Israel is very mad and their friends are having borderline meltdowns on Fox. The Gulf states are similarly displeased but it is not clear what they can do about it, if anything.

It is entirely possible the conflict resumes this weekend given attacks have not stopped either from Israel or Iran and that Trump has a track record for changing his mind and screwing people. Similarly, it is entirely possible Trump just cuts and runs leaving a mess albeit one that is oddly climate friendly as people witness the horror of how new weapons make fossil fuel supply chains deeply insecure.

Markets wise crude curves remain steep - it seems that nobody entirely believes this will stick and the loss of capacity remains as well as the lack of flow through the Strait of Hormuz. No doubt another sleepless week in markets and materials.

I have built a model and open-source application that takes a first pass at this problem using an agent based modelling approach and Q learning. It is imperfect in ways I will be upfront about and which you can find outlined here. But even in its current form, it learns things that I think are genuinely interesting — things that match the intuitions of good policymakers but that have never, to my knowledge, been derived from a formal optimization framework. The model and app are available on GitHub at commodity-portfolio-abm. The app is here.

The literature gap

Modern portfolio theory is nearly seventy years old. Markowitz gave us mean-variance optimization in 1952. Since then, the financial economics literature has developed extraordinarily sophisticated tools for portfolio construction under uncertainty: factor models, robust optimization, Black-Litterman, CVaR constraints, hierarchical risk parity. If you want to optimize a portfolio of equities, or bonds, or even financial commodity derivatives, you are spoiled for choice.

But if you are a government or a large industrial buyer trying to decide how to physically procure a commodity — how much to contract from which suppliers, how much domestic capacity to maintain, how much to store, how much to leave to the spot market — the literature offers almost nothing.

This is not because the problem is unimportant. Europe’s catastrophic dependence on Russian gas, the scramble for rare earth alternatives after Chinese export controls, the fertiliser shock following the invasion of Ukraine — these are among the most consequential economic disruptions of recent years. They are all, at their core, problems of portfolio construction in physical commodity procurement under geopolitical risk. And yet the formal tools available to the people making these decisions are startlingly thin.

The credit risk literature comes closest. CVA pricing, wrong-way risk models, Duffie-Singleton, Gordy’s asymptotic single-risk-factor framework — these give us machinery for thinking about counterparty default in financial contexts. However, physical commodity offtakes are different from financial contracts in ways that matter enormously. When your largest supplier defaults, they do not just fail to deliver — their default is the supply shock that moves the replacement price against you. The loss given default is endogenous to the default event itself in a way that creates convexity: the relationship between supplier concentration and conditional expected loss is nonlinear, which means standard credit portfolio models systematically underestimate the diversification benefit. To compare these securities if a bond defaults you lose some of your principal. If you short a stock your losses are uncapped and the downside of commodity offtake defaults are more like stock short sales - you have to buy back in at a higher price.

On the operations research side, there is good work on supply chain optimization and logistics — the kind of work done by scholars like Jun Ukita Shepard, Lincoln Pratson, and Gosens, Turnbull, and Jotzo on coal trade flows and energy logistics. But this work tends to optimize for efficiency under normal conditions rather than resilience under disruption though network flow models can be stressed for scenario analysis. It does not typically model the geopolitical risk dimension or the strategic feedback loops between procurement decisions and supplier behavior.

And the geopolitical risk literature — Caldara and Iacoviello’s GPR index, the resource curse and resource war traditions — gives us measures and narratives but not decision frameworks. Knowing that geopolitical risk is elevated does not tell you how to restructure your offtake portfolio in response.

What is missing is a framework that jointly considers: the portfolio of supplier contracts and their covariance structure under disruption; the option value of domestic production capacity (sometimes expensive but reliable); the insurance value of strategic storage; the feedback effects of procurement decisions on supplier capacity and market structure; and the spot market dynamics that emerge when multiple buyers compete for scarce supply during a crisis.

An agent-based approach

The model I have built attacks this problem from a different direction than the analytical frameworks above. Rather than trying to derive a closed-form solution — which is likely intractable given the feedback loops and non-stationarities involved — it uses reinforcement learning within an agent-based simulation to discover good procurement strategies through trial and error.

The setup is straightforward. A simulated world contains suppliers, each with a capacity, a cost structure (fixed plus variable), and disruption and recovery probabilities modelled as a two-state Markov chain. Countries have demand, domestic production capacity (more expensive but reliable), strategic storage, and access to a spot market. The spot market clears via a price mechanism: when supply is disrupted and multiple countries are competing for scarce supply, the price rises until demand meets available supply, with more price-elastic buyers reducing their purchases more.

The critical feedback loop is on the supply side. Suppliers adjust their capacity based on utilization. If a supplier’s output is being underutilized — because buyers are stockpiling, sourcing elsewhere, or leaning on domestic production — it gradually cuts capacity. This means the “safe” strategy of heavy domestic production and storage can, paradoxically, make the international market structurally tighter and crises worse when they do hit. This is important to consider today when looking at “Bidenist” policies - pulling all capacity onshore can embrittle global supply chains but similarly, Chinese overcapacity that reduces refining capacity abroad then creates problems as we see in Asian refined products today with China stopping exports and the capacity having been lost.

Each country is controlled by a Q-learning agent that observes a compressed state — storage levels, spot prices, disruption severity, and supply tightness — and chooses from a discrete set of actions covering allocation profiles (how to split purchases across contract suppliers, spot, and domestic) and storage decisions (build, hold, or draw). Over thousands of training episodes, the agent converges on a policy that minimizes average procurement cost across the full distribution of disruption scenarios.

What the model learns

Three findings emerge consistently, and I think they are worth highlighting because they match real-world policy intuitions while being derived purely from the optimization:

First, diversification is robustly valuable, and more valuable than naive models suggest. The trained agent consistently spreads procurement across multiple contracted suppliers rather than concentrating on the cheapest one, even when cost differentials are significant. This is the endogeneity-of-LGD result showing up in a different guise: when your largest supplier goes offline, the spot price spikes precisely because of that disruption, and your replacement cost is worst exactly when you need to replace the most volume. The agent discovers this convexity through experience without anyone programming it in.

Second, storage is used strategically, not passively. The agent does not simply fill storage and hold it. It learns to build reserves when spot prices are low and disruption severity is low, and to draw them down during crises or when the market is tight. Storage acts as a dynamic buffer — absorbing surplus in calm markets and releasing it during stress — rather than as a static strategic reserve. The cost of holding inventory is weighed against the tail-risk insurance it provides, and the agent finds a balance that would be very difficult to derive analytically given the path-dependence involved. If this sounds exactly like my work with Employ America and Arnab Datta and others SPR work that is because it is exactly this - reserves need to be dynamically managed.

Third, the agent pays up for domestic capacity. Even though domestic production is modelled as more expensive than international supply on a per-unit basis, the learned policy maintains and uses domestic capacity rather than letting it atrophy. The key insight is that domestic capacity’s value is not primarily in the units it produces during normal times — it is in the units it can produce during a crisis, when the alternative is buying on a spot market that has spiked to multiples of the base price. The agent learns to maintain domestic capacity as a real option against extreme spot price outcomes, even when this raises average-case costs. This has obvious resonance with debates around energy independence, onshoring, and the value of maintaining “uncompetitive” domestic industries as strategic buffers. One deep question from an extension of this model is how you manage this when private sector incentives are short term (“make money or get raided by Elliott”) and governments have electoral cycles. Japan does a good job here with JOGMEC and JERA but few others do.

What the model does not do is equally important. It does not model strategic behavior by suppliers — disruptions are purely stochastic, not adversarial. It does not model transport costs, demand growth, multiple commodities, or the production lags that characterize mining and heavy industry. Countries learn independently without game-theoretic interaction. These are all directions for extension, and I have laid them out in detail in the project README.

Why this matters for policy

The policy relevance should be clear. Governments are currently making enormous bets on supply chain restructuring — friendshoring, critical mineral stockpiles, domestic production subsidies — largely on the basis of narrative reasoning and ad hoc analysis. The analytical frameworks they can draw on are either too abstract (geopolitical risk indices), too narrow (logistics optimization without disruption risk), or borrowed from domains with fundamentally different structure (financial portfolio theory applied without adaptation to physical commodities).

What I think this model demonstrates is that even a relatively simple simulation framework can generate non-trivial insights about optimal procurement strategy. The finding that diversification benefits are larger than naive models suggest has direct implications for how we should think about concentration risk in critical mineral supply chains. The finding that storage should be managed dynamically, not passively, has implications for how strategic petroleum reserves and other stockpiles are operated. And the finding that it can be cost-effective to maintain higher-cost domestic capacity as a strategic buffer provides analytical grounding for industrial policies that are currently justified mainly on national security grounds. How one does this is via a strategic resilience reserve or a command economy. I prefer the former.

The bigger question of how one establishes bank regulation like capital charges or incentives to diversify domestic supply is a trickier one. If no individual country should hold a 20% weighting on Qatar…. then how did the LNG market get there in aggregate? Most governments have not developed the long view of thousands of lifetimes afforded by reinforcement learning, a kind of remembrance of past lives of a Tulku. If we cannot be government by enlightened beings, perhaps we should model more?

The model is open-source, comes with a React dashboard for exploring scenarios interactively, and runs via a FastAPI backend. I would welcome contributions, extensions, and — most importantly — critique. If anyone wants to publish this as a paper with me (I, a lowly non-PhD holder) let me know. The gap in the literature is real and filling it will require work from people with deeper expertise than mine in operations research, mechanism design, and the specific commodity markets where these dynamics play out.

This is a first pass. But I think it is a first pass at the right problem and well posed.

The model is available at github.com/nemoincognito/commodity-portfolio-abm.

Most of these projects have fallen over because it is hard to compete with fossil gas and especially fossil gas supplied projects in the Gulf. We now face another Middle Eastern infinity war and economic collapse because of economic dependence on this part of the world.

This is great foreign policy for the Qataris: take enough market share to be indispensable and hope that keeps you safe. It worked for a time. This is now an absolute catastrophe for its customers, and it is worth completely rethinking blithe assumptions that markets can solve these problems of dangerous physical market concentration and under regulation because we now have two devastating failures in European reliance on Russian gas and now Asian reliance on the Middle East.

Major fossil exporters should be seen as global economic pimps who want to keep customers dependent both for economic rents and their own power. For some it is a mad grab to become too big to fail like Qatar, for the US it is an economic weapon they can wield against Europe and others and which no small amount of diplomatic coercion revolves around. Qatar is subtler than the current US administration but not that subtle. Countries which run an energy deficit should see this for what it is and respond appropriately - decarbonizing and taking a good hard look at risks in global energy supply.

If we are going to have to live through 1970s inflation, we should at least take some inspiration from 1970s grindhouse / blaxploitation and in particular the character of Pam Grier in the film Coffy. While I do not advocate violence the discourse around calling out fossil supply risks, coercion, endless reputation laundering through the art world and so on could be a lot more robust given the stakes. We could at least take something from the scripts of these films.

It was easy for him because he really didn’t believe it was comin’, but it ain’t gonna be easy for you, because you better believe it’s comin’!

The solution of course is modelling the problem, having buffer stocks and resilience reserves, and pushing for domestic capacity or at a bare minimum diversified capacity so that there are no critical nodes in the network.

In other news….

• David Leitch did a superb job outlining how to cut Australia’s dangerous diesel dependency and reduce emissions.

• Ultimatums from Trump, but Iran acting like Rorschach in The Watchmen.

  

A problem with the political economy of the United States today is that it is heavily influenced by the tech right. This group is immensely clever and talented at an extremely narrow range of things more specifically writing software. They are familiar with dependencies and recursions as problems you can fix in an afternoon, not a decade. They are enormously talented in a narrow way like the Zunfte or guilds of Zurich in the Middle Ages. Capable, rich, and wholly unfit to run an empire where you have to think about the agency of others and hard assets and physical limits - not least of all new agency enabled by technology you yourselves developed like drones. I will write more on these parallels another time, but Switzerland is a great place that never had a massive supply chain empire for numerous reasons. If the tech sector wants to be the Swiss they probably should not have hopped in bed with the GOP fossil sector and should have practiced some kind of cleantech / supply chain security isolationism. It would work a lot better than whatever this is - “frat boy at the dive bar starts fight with meth head and gets glassed” kind of foreign policy.

All these folks who are major GOP donors in Nocal need memory and logic. Producing memory and logic predominantly happens in South Korea and Taiwan. It is an assumption online that Korea will keep the fabs open among the West Coast tech twitterati. This would be bad and dumb policy by Korea because it imposes no pressure on Trump or his donors notably the likes of Greg Brockman of OpenAI. If there is no compute for large tech model providers then there is no growth, no subsequent equity round and perhaps a collapse in US growth and capital expenditures. Korea’s government is elected by Korean voters who like to have food and drive their cars, not the oligarchs of Nocal. A good way to explain this to the US is that “no LNG, not enough power, we choose aircon over chip fabs, sorry”. Not a whole lot the US can do about that. The howls from Palo Alto would be extraordinary.

Damage we are already seeing in Qatar has probably ended this datacenter boom anyway - turbine markets were tight along with heat exchangers and having the IRGC blow up half of Qatar’s liquefaction plant is not helping here. Either chips need to use a lot less energy or its very much going to be constrained brutally from here.

South Korea: Restrict chip exports over summer due to power availability. Go wholly nativist on any product and actually seek ways to cause production interruptions in GOP aligned sectors in the US. Claim you need the power for defense and materials as well as aircon. Easy because it is true.

Australia: Australia exports a lot of LNG which is in demand and is horrifically short oil products. Australia should restrict exports to any country not providing it products, starting today most especially China. With regards to the US I think it is worth talking very frankly about basing rights if this is not wrapped up soon. How exactly has the US supported security and stability in the region in the last two years? Could we spend vastly less on submarines and develop our own green hydrogen to ammonia and fertilizer supply chain for less? The answer is yes.

Taiwan: Not dissimilar to Korea but actually vastly worse on grid stability and LNG import dependency because there is a lack of nuclear restarts. Tell Jensen no wafer starts in the summer because you need aircon and watch the US fold. Weaponize TSMC and do not be shy about it. Sell US equities for FX stabilization. That goes for everyone really. Worst case they hand your over to the Chinese navy but honestly that’s the default setting given how they are running down munition stockpiles.

Japan: Startup all the nukes to attenuate the LNG issue. Go after any and everything to hurt US tech. Drive the yen up through changes to capital regs to force sales of US equities and restrict investments in US private equity and VC.

Singapore: Something of a linear combination of the above with some US basing capacity and a lot of tech and a large financial surplus.

None of this is pretty, people will be upset but there is a sense in quotes from Vance that the US can insulate itself from its own idiocy:

This is peak Austro-Thielian isolationism - the kind of belief set you can only hold if you’ve spent your entire existence drinking mélange at Cafe Landtmann Pre-World War I, or doing deals in the valley today. It is born of a kind of systemic autism, unaware of the broader system that the surplus your enjoyment is built on - the extracted grain surplus of the Austro-Hungarian empire then, the energy supply chains and willingness to trade freely in tech services with the US today.

Policy in Asia should be to get the United States to wake up and realize that “we’re in this together”.

Edit: someone pointed out that the deeply secure supply chain “paranoid style” stuff was the whole body of work I have worked on with Employ America. That is at least partially true and it is a physically coherent position unlike “you can just do things” which is sometimes true, sometimes not and most importantly there is no git restore on the high seas.

Before you laugh or cry - let’s explore this a little and “steel man” this behavior. If you believed that you were indeed going to the war to end all wars, then being offered a 25x return on the second coming would be a compelling bet. Then there is the question that if you indeed believe that and that you are entering the kingdom of heaven or being thrown into a lake of sulfur, what is the point of gambling? You cannot spend your winnings because either way you will not need them. If you do not believe this then bet against the second coming and get it wrong your losses are hardly the worst of your troubles. There is a nasty in-between state of the US becoming a klepto-theocratic state that threatens private businesses and which might interfere with market integrity here - Trump declared Jesus or some such. That is why I will not bet on this Pascals wager like proposition.

This strange combination of everything-is-gambling Mammonistic tendencies and Semitic-Levantine apocalyptic thought is everywhere once you start looking. AI GDP impact charts from the Federal Reserve:

indicate wide range of “everyone has every desire met” to “everyone dies” are one example, but so are various forms of doom porn from Citrini to say nothing for the ability to “Bet on Anything, Everywhere, All at Once” and engage in relentless self vs other lifestyle comparisons on Instagram with just about everyone. The underlying social competition and dopamine reward circuitry is not new, nor are these observations about the impacts of new forms of media or gambling but it seems parlays into the apocalypse seems the logical end point here. The apocalypse is important because it gives people a sense of certainty, closure and an inflated sense of importance. In an attention poor world where everyone can shout online “being there and being special” for the end has a deep appeal.

The issue is how do you model and predict people’s actions - especially those with the levers of power if they literally cannot stop talking about the apocalypse in extremely literal terms? “Take him seriously not literally” did not pan out any better than Solana as an asset class, apologies to David Sacks. This is not a silly or irrelevant question: there is a deep emotional logic to all this. These people are in power and they are motivated by these ideas. Taking this all seriously and literally in the anthropological sense is long overdue and it is a serious failing that the press has only managed to engage with this stuff using ridicule or selling cortisol spikes via pearl clutching and panicking. These are serious ideas held by people with serious power: a surgical and clinical view is required because while you might not care much for millenarian cults their cultists, they have an interest in everything and the capacity to change the world.

A first order problem might be that what looks like a disaster to anyone who is not pining for the apocalypse would look like the road to victory to someone thinking apocalyptically. Massive war? Expanding conflict? Famine, disease, pestilence? To the apocalyptic cultist this is running up the scoreboard. People who are shocked or surprised at how happy elements of the US government seem to be about how things are going need to understand the completely reversed perspective here. Ending the world is the point. The people flying the plane do not have the instrumentation turned off, they see it completely upside down. That is vastly worse of course.

So — the US government and its supporters is still comprised of some people who assume that Middle East Wars are bad and lose them elections, but a non-trivial share who think they are going to enter the Kingdom of Heaven the worse this goes. Markets initially priced in a quick conclusion and reversal here and that seems to be wearing thin already. This is to say nothing of the millenarian tendencies on the other side of this conflict. Suffice to say if you are taking out some kind of rational agent public choice framework to call this you may not have a good model of this reality. For that to work you need to either believe those who assume elections will occur and that they have to win them can coerce the apocalyptic side of the administration or hope for the better angels of the IRGC (best of luck) or that perhaps this is the apocalypse after all.

I am not a believer in all this and my sensibilities around time are more cyclical and South Asian. All things change but there are cycles and but never an end as such. If you are not part of team apocalypse or singularity, then the historical record provides plenty of good examples of how to get through hard times. Eric Cline’s work (recently on William Dalrymple and Anita Anand’s Empire Podcast), the history of Buddhism’s spread and survival in Tibet and China all show there is agency in these times where some people really do seem out to destroy themselves and others deliberately.

To get through this states that do not pine for the apocalypse are going to have to consider hardening themselves more going forward - in energy, in food, technology and through a broader range of relationships than the hub and spoke order under Pax Americana. Stockpiling and energy security are a top priority but no doubt only the start and the practical tasks of Malcolm Turnbull-ish Mark Carney at WEF-ish lines of thinking.

The problem is that after the apocalypse, some of us have to go on living.

The entire discipline of computer architecture is, in a meaningful sense, about managing these tradeoffs: getting the right data to the right level of the hierarchy at the right time. A system that tried to do everything in SRAM would be ludicrously expensive - but might be justified in high value applications where memory is the rate limiting step like decode in transformer inference for AI. Hence, a company that did exactly that was acquired for $20bn. A system that relied on optical disks or tapes for everything would be ludicrously slow but may be important for backups that are radiation resistant. In cooking you balance ingredients and process them to make something tasty, in computer architecture you balance component performance and cost for specific tasks and there are tradeoffs between flexibility, performance and cost.

Optimizing the storage for a renewables heavy grid has a very similar challenge. Storage technologies have a number of properties and the jug of water analogy is apt here:

• Unit cost for power or “water flow”: how narrow the neck of the jug is

• Unit cost for energy storage: cost per liter or gallon of water

• Charge efficiency: how much water gets spilled filling it

• Discharge efficiency: how much water gets spilled emptying it

A grid with a lot of renewables has storage needs that vary across timescales in the same way that a processor has memory needs that vary across access patterns. Supercapacitors and flywheels can respond in milliseconds — they are your registers, handling frequency regulation and very short transients. Lithium-ion batteries sit in the sweet spot of one to four hours — they are your L1/L2 cache, shifting solar from midday to the evening peak, handling most of the daily cycling that gets the headlines and the investment dollars. And then you have the deep, slow, cheap storage: pumped hydro, iron-air batteries, flow batteries with cheap chemistries, compressed air. These are your disk or tape — they are there for the multi-day Dunkelflaute events, the week-long wind droughts, the seasonal mismatches that will kill you if you have not planned for them.

The analogy extends further. In chip design, you do not design an ASIC the same way for every application. A bitcoin mining chip optimises for one thing: SHA-256 hashing throughput. A Google TPU’s systolic arrays are built for matrix multiplications. An RF front end for a 5G base station has completely different design constraints from either. The job to be done, in the Clayton Christensen sense, dictates the architecture. So it is with grid storage: the optimal mix of technologies depends entirely on the specific renewable generation profile. It depends on the weather, and the weather depends on where you are.

A grid in southern Spain with abundant solar and modest wind has a very different storage problem from a grid in the North Sea wind belt. Spain’s problem is largely diurnal: too much solar in the day, none at night, with relatively predictable seasonal variation. Lithium-ion handles a lot of that. Northern Germany or Denmark, by contrast, can go through extended periods of low wind and low solar in winter — the infamous Dunkelflaute — where you need days of stored energy, not hours. These are different jobs and they require different mixes of the memory hierarchy. A heavily solar grid might need a ratio of, say, 80% short-duration to 20% long-duration storage. A wind-dominated grid with challenging winters might flip closer to 50/50 or even heavier on the long-duration side, depending on how much interconnection and dispatchable backup you have, or imported e-fuels.

This is where much of the current discourse on energy storage goes badly wrong.

The way most storage technology companies present their total addressable market is essentially: “the grid needs X GWh of storage, and our technology can provide that.” This is roughly equivalent to a memory chip company saying “computers need X GB of memory and we make memory.” It ignores the hierarchy entirely. The TAM for a four-hour lithium-ion system and the TAM for a 100-hour iron-air system are not the same market any more than DRAM and tape are the same market. They complement each other. But because lithium-ion got to market first and has a cost curve that looks fantastic on a slide deck, there is a tendency to either assume lithium-ion will do everything (it will not, or not affordably at multi-day durations) or to claim that long-duration storage will displace lithium-ion (it will not, because it is too slow to cycle daily and too inefficient at 50-70% round-trip versus lithium-ion’s 90%+).

The result is that long-duration energy storage companies frequently overstate their near-term TAM by conflating it with the total storage market, while lithium-ion incumbents understate the need for complementary technologies because acknowledging it means acknowledging the limits of their core product. Investors get sold the same TAM by everyone in ways that would immediately flag you as a huckster in computer hardware. The technology companies themselves can end up targeting the wrong markets due to this loose TAM math — trying to compete on metrics that do not matter for their niche in the hierarchy. A 100-hour iron-air battery does not need to match lithium-ion on round-trip efficiency any more than a hard drive needs to match DRAM on latency. It does need to be cheap enough per kWh of capacity that you can afford to have enough of it sitting there for the bad week in February that comes around every few years in the North Sea.

There is an institution well positioned to clarify this: Open Energy Transition and the broader PyPSA ecosystem. Their open-source energy system models can, in principle, run optimization over full weather years with hourly resolution, sector coupling, and multiple storage technologies with different cost and performance characteristics. That is exactly what you need to determine the optimal storage mix for a given region with a given renewable generation profile. Form Energy has done something like this already — they commissioned Open Energy Transition to build out a PyPSA-Eur model focused on Germany that integrates short and long-duration storage representations and runs optimization scenarios to 2035. This is exactly the right approach. The model takes real weather data, real grid topology, real cost projections, and solves for the least-cost system that keeps the lights on across the full range of weather outcomes. It is far more informative than any analyst’s spreadsheet extrapolation of storage deployment by chemistry.

But this kind of work is far too rare, for two reasons.

First, storage technology companies are generally unwilling to share the performance and cost data needed to represent their technologies accurately in an optimizer. This is understandable from a competitive standpoint: you do not want intellectual property leakage by sharing too much but it means that the models either use generic assumptions or rely on published literature that may lag reality by years and lead to your technology being ignored. If grid planners and funding providers do not know your company exists and an industry runs on long term planning, then your company will not be in the plan. The fast-moving cost reductions in both lithium-ion and emerging chemistries make this lag particularly damaging. A model using 2022 cost assumptions for lithium-ion in a 2026 run is going to get a very different optimal mix from one using current pricing, and the same applies to iron-air, vanadium redox, zinc-bromine, and every other chemistry competing for the “deep storage” layer.

Second, with the notable exception of Form Energy’s work in Germany, very few storage companies are actively seeking to work with energy modelers. This is a strategic error. If you are building a long-duration storage technology, the single most valuable thing you can do for your sales pipeline is demonstrate that an optimized grid in your target markets needs your product alongside lithium-ion, not instead of it. The PyPSA framework already has the technical capacity to do this. Open Energy Transition has the expertise. What is missing is the commercial engagement from the companies building these technologies. They would rather commission consultancies to produce a market sizing report with closed modelling and sources than work with modelers who might produce a more transparent and nuanced one.

The irony is that a properly specified storage optimization model would almost certainly validate the thesis that we need a hierarchy and multiple technologies. No serious modelling of a high-renewables grid with realistic weather variability concludes that a single storage technology meets all needs at minimum cost. Work by Jenkins and Sepulveda is particularly clear on this point. The question is always about the mix: how much of each technology, at what durations, in what locations, and how does that change with the generation profile? These are answerable questions. The tools exist. The weather data exists. What does not yet exist is the institutional will to bring the technology companies to the modelling table and produce the kind of transparent, location-specific, weather-aware storage mix assessments that would actually help investors, utilities, and policymakers make good decisions.

Until that happens we will continue to see storage companies raising capital against inflated TAMs, investors betting on chemistries without understanding the niche those chemistries will occupy in the hierarchy, and grids being planned with either too much expensive short-duration storage or not enough cheap long-duration storage or, most often of all, obscene transmission overbuilds that lock in higher prices for consumers. The companies themselves are poorly served by this because they often target the nearest market at hand (California, occasionally PJM) instead of looking where need, willingness and capacity to pay is highest.

The memory hierarchy was not solved by having every application use the same type of memory. It was solved by a process of analysis, openness of specifications and elegant engineering and continual improvement in computer architectures as components and needs changed. The grid deserves the same treatment but for that it needs an analytic culture of engagement and transparency from technology providers. Among other things it would make investing in long duration storage technologies a more tractable analytic problem.

Many of my readers are familiar with the chain rule from calculus and even more may be familiar with the Kaya identity. To put some structure around discussions of AI bubbles and investment I like to use the following:

The first term is the cost of energy: how much it costs to generate electricity to run the chips and cooling in the datacenter.

The second term is how efficiently power is turned into floating point calculations and - more specifically for LLMs - the tensor operations and matrix multiplication.

The third term is a measure of how well software turns those floating-point operations into tokens.

This is a simplification but not an egregious one of the economics of inference or running the models. Whether there is a bubble comes down to how much people are willing to pay for tokens and whether that is more or less than that cost. Willingness to pay I will not cover here but there are some applications that are clearly valuable and for which there is a high willingness to pay: coding, investment research and the like come mind. There are other applications that are very token intensive that I am not sure there is much of a market for people to pay up for, not least of all those gorilla videos.

And a *checks notes* national socialist cartoon harassment campaign against a poaster.

What ultimately will make or break this business is getting those costs down.

On the power side: I have my doubts. In the US picking favorites in energy sources by blocking offshore wind and solar in an emerging power shortage seems unwise. I do not see orders of magnitude cost declines for any power sources from here. The outlook for help from the power and energy side is decidedly poor from a technologically point of view and made worse by the current policy mix.

On the efficiency side: I am much more optimistic. Photonics are likely to provide at least a few orders of magnitude in power efficiency gains here. Making this happen should be a core focus of the industry more broadly as the physical and political realities of required datacenter build become challenging.

On the software side: Improvements happen quickly and are continuing. The usual rule of thumb for models has been that inference FLOPs requires double the number of non-embedding parameters in the model per token. As model parameter size has exploded this has gone up a great deal. However, various improvements like mixture-of-experts, model distillation and smart routing have reduced this materially by not activating all parameters for every query. This is the equivalent of knowing that if you have a question about medieval literature that you want answered by a university faculty you should not send it to the physics department. This area remains one of active and fruitful development.

State-space models like Mamba will do even more as these approaches are integrated into services like Chatgpt. I expect gains to continue here and intensify.

So, is it a bubble? How will we know? How could it end? A few scenarios are below:

1. Token growth slows a lot: This is absolutely the worst-case scenario of demand collapse or materially falling short of projections. Everything gets hurt here, datacenters are empty or idled and AI is a maximal bust. I see this as wildly unlikely due to current token growth and increasing use. Not likely in my opinion but if token growth as measured starts to flatten or fall this case becomes likely. There is absolutely no sign of this happening at this time.

2. Productivity shock in model architecture or software: State space models have very compelling scaling properties that may significantly reduce compute demand and lead to reduced demand for datacenter and compute capital stock especially via reduced memory requirements. Brutal competition among models as the cost to compete goes down leading to an outcome like solar in the 2010s: the consumer wins, the producers do not. Good for downstream users of tokens but bad for whoever built all the capital. Ugly, but not out of the question especially if the willingness to pay for performance is somehow capped: Cursor seems to be dealing with the challenges of very expensive power users and very lucrative casual users. Debottlenecking this or making it economic to serve power users would be great especially if it led to internal model development being affordable and compelling. Impacts here are acutely demand sensitive as they may allow some applications to be served at a cost that is profitable.

3. Productivity shock in hardware: Very likely, and I have some knowledge but also some personal interests here. It should be noted there are a couple of plausible shocks here:

   1. Photonic compute: Vastly reduced energy use for compute, memory requirement impacts depend on subtle model choices. Great news for incumbent datacenters because you can produce more tokens in the same facility and with the same power fit out. It could be disastrous for the more speculative datacenter pipeline and create acute deflation in the space similar to the impacts of multiplexing on the value of fiber investments towards the end of the telecoms boom. This excellent paper Dynamic Analysis of the Long-Distance Telecom Bubble by Kurebayashi and Osgood gives a particularly thorough analysis of this boom and how the long lead capital expenditures on fiber that were expected to generate a certain throughput had vastly more capacity by the time of completion leading to oversupply.

   2. Interconnect and Memory Bandwidth: A gating factor for most LLMs today is how quickly data can be moved on and off memory or between clusters. This also drives lower GPU utilization as the GPU “wait” for data input / output. Have you ever found that you are consistently ready before your children, that you can mysteriously locate your shoes and socks in a timely manner? That is the life of a memory bandwidth constrained GPU. The problem with this debottlenecking is that as the GPU uses 90%+ of the energy in a rack then energy requirements explode without photonic compute. This is a gradual process and the development of coherent datacenter clusters will allow more compute per unit of datacenter albeit at an energy penalty. Nonetheless, this could also have elements of a supply shock in datacenters but would not necessarily imply that utilities are overbuilding right now.

There are plenty of bubbles that could be happening right now the further you get from the customer because the more links a business is away from the end token demand the more a productivity shock could upset the market. I am not particularly concerned about “hyperscalers” especially as they seem to be offloading some if not much of the credit risk onto alternative credit providers and specialist merchants like neo clouds. What I am particularly concerned about is a bubble in the infrastructure - something that is a pattern that has been repeated time and again in classics like Engines That Move Markets not just because of clear technological paradigm risk similar to the telecoms bubble but because the hostility and regulatory pressure is picking up and making their growth plans extremely challenging.

A recent report by the Australian Energy Market Operator picked up significant news coverage covering the usual gamut of industry wonkery and anti-renewables outrage. The issue as outlined in RenewEconomy is thus:

AEMO says the soaring cost of new grid infrastructure means the 2026 ISP will need to re-visit transmission network projects previously identified as critical, with the exception of projects that are already committed, in a bid to limit the impact on consumer energy bills.

The question is how and why. To that end reading the Transmission Cost Forecasting report is the place to go. They provide a breakdown of components and how they are estimated as a weighted sum of more observable inflation and producer price indices.

The chart below shows the price rise by components most of which are not materially more than CPI and seem to have tracked the general surge in price levels following the end of COVID and the Russian invasion of Ukraine induced energy shock. Notable exceptions are land and property easement costs, underground and overhead lines which are running far ahead of CPI as shown below. Both entail significant civil works and labour, but property and easement costs are purely driven by land values.

Going forward the land and labour components will be the acute challenges here:

Are agricultural land values blowing up in Australia? Yes. Data from the Australian Bureau of Agricultural and Resource Economics farm portal indicates as such.

The problem is that it is very hard to tie this to farm level returns which remain meagre in cash terms according to the same data source. Something is rapidly increasing demand for land that is not tied to margins on beef cattle, sheep, diary or broadacre crops.

So, what is it?

Renewables, probably. This is remarkable for several reasons:

1. Baumol’s Cost Disease Strikes Again: Baumol argued that labour intensive low productivity sectors rise in tandem with high productivity sectors, but this could readily apply to land now. The productivity is mostly on the coast in services and manufacturing, but the rental capture is just as strong inland.

2. Paradoxical Rural Politics: Opposition to renewables can be fierce in regional Australia despite regional Australia being the largest beneficiary not only in terms of jobs and income but also capital values. The stereotypical “Pitt Street Farmer” has little to complain about.

3. There is a better way: Developments in agrivoltaics, vertical solar and allowing animals to graze to reduce grass levels on solar farms are all happening already. Why is there a presumption of lost land at all if there are technical solutions that provide little or no tradeoffs between power generation and agricultural output?

This land appreciation and infrastructure shakedown spiral could end as AEMO and infrastructure investors rethink how much transmission they need in light of falling storage costs. It is not hard to see how a landowner with important rights of way on transmission can extract value, but it seems that such extraction is now so large it is now endangering previous grid plans.

There is a lot more to think about here in particular how much transmission and distribution can be offset by distributed storage and solar, no doubt a great deal more coming on that in the future from AEMO. It is also perhaps a lesson in the limits of being a squattocratic holdout: sometimes the deal does not close, and you end up holding the bag of wool instead of the bag of money.

More on Gallium from the Critical Materials Innovation Hub at the Colorado School of Mines

Link here, playback disabled outside of YT. Some similar conclusions to mine: these problems are cheap to fix if you use waste streams, but the missing markets problem is as large as ever: if China exports Gallium again prices will fall hard, and ex China capacity could be lost all over again.

Solar Wars, Turkey Edition

Looks like even Turkey is losing interest in Chinese dumping in this sector.

Solar Radiation Management

Recent polling by YouGov in the UK was not unfavourable but of course there are some even better options in the tech pipeline. The trick is to bounce as much light back to space using the “atmospheric window”. A good survey paper is here. I think there is a lot more to come hopefully in optimizing the photonic properties of building materials.

Gallium is used in three main applications: Gallium Arsenide for solar panels especially space based installations, Gallium Nitride (GAN) for electronics including power electronics and Gallium Oxide for light emitting diodes (LEDs). As noted in this excellent work by CSIS, the primary military and security concern today is GAN as it pertains to power electronics. For any applications that require high current, high temperature, high voltage or all three Silicon Carbide and GAN have become the preferred technologies from EV chargers to solar inverters to radar. A good background video is here, and a very deep dive into power device comparable topology can be found here. These III-IV group1 wide bandgap materials have largely displaced pure silicon in these applications.

While Gallium may have supply issues today using these materials can allow much more latitude in design and counterintuitively creates less pressure in other supply chains including rare earths. For example, in EVs switched reluctance motors can replace a permanent magnet motor which uses more GAN or Silicon Carbide power electronics replacing rare earth magnets containing Neodymium and Dysprosium. Companies like Conifer are already offering these motors for 1-25 horsepower applications. An insight here is that critical minerals very seldom have good substitutes in a narrow sense2 but can offset one another for “jobs to be done” in the Clayton Christensen sense: you might not have a near substitute for a particular atom or molecule but another topology or design of the product can substitute out those materials for modest trade-offs. This is not always an option, and it is very much not an option in power electronics.

The most pressing and acute issue with Gallium supply right now related to radars and electronic warfare that use power electronics. Here high current and voltage capacity in GAN combined with high switching speed allows the generation of high frequencies at the high power necessary for a radar array. Silicon Carbide has its merits, but it cannot switch fast enough to generate these high frequencies for antenna arrays. Reports from credible sources like DefenseOne indicate high GAN electronics are behind recent success against UAVs by Pakistan using Chinese anti drone equipment and somewhat less credible reports note the importance of active GAN radar in China’s PL15 missile which the Pakistan military now uses. With intact missiles being recovered in India some interesting reverse engineering and analysis is no doubt going on right now. Taiwan in particular has paid attention to these developments according to an article in The Guardian.

Having China control 98% of the supply of Gallium is clearly a problem. It is acutely so now that China has restricted exports of this material putting at risk defence, semiconductor and decarbonization supply chains all at once.

How did we get here? How did the United States and its allies in particular sleepwalk into giving China a supply chain blocking stake on their most advanced weapons?

How Markets Have Failed

There are two elements to this failure: how China achieved a controlling position and why that position was not contested earlier. The most lucid explanation I found of how China came to dominate this market came from a 2014 paper by A.V. Naumov in the Russian Journal of Non-Ferrous Metals where he succinctly described China’s upstream “land and expand” strategy:

The technical and economical strategy of China, which considerably increases the production capacities for gallium despite the occurring excessive demand, deserves consideration. It seems likely that this is the partial case of manifestation of the total trade and economical politics of the government of China, which invests up to $1.5 billion into the development of branches of industry associated with high technologies, notably, alternative power engineering, the production of high-tech equipment, and power-saving and environmentally safe technologies in the period of the 12th five-year plan (2011-2015).

We can distinguish three stages in the strategic commercial politics of China. The first is support for domestic producers of raw materials through a favorable political and economic situation. The second stage is that, as the country occupies the main part of the global world production of one or another raw material, the Ministry of Trade of China starts to limit its export, decreasing the return of the value-added tax (VAT) for export, increasing the export duty, and introducing the export quotes. Finally, at the third stage, the number of export quotes starts to decrease and the tax pressure on the export of raw materials starts to increase. This will support the Chinese producers of the final product and will force other producers throughout the world to carry the production into China in order to guarantee the stable supply and sale. It seems likely that the measures taken in the gallium industry in China in recent years should be considered a preparation for the third stage.

The process can be characterized as dominating upstream supply, crushing pricing to remove competition and then using that locational nexus of supply and assorted tax and trade restrictions to wedge downstream higher value-added activities into China. China had a walking start in this as Gallium is normally recovered from Bayer Liquor from alumina refining and was an easy adjacency for China to move into having rapidly expanded its aluminium sector. Driving pricing down to ~$140 per kg from 2015 to 2019 halted or shuttered any output outside China and left China with the market and dominance in its first large scale application - blue LEDs which combined with Phosphor LEDs allowed LEDs to produce white or warmer tones of light that you can now find in hardware stores.

There was little noise around strategic problems created by LEDs because there were none. Nobody gave much thought to dominance in energy efficient lighting. In 2014 with an approximately 200 tonne global market priced at $120,000 per tonne a market of size $24mm was beneath every market analyst and policy maker as too small, too weird and not of any particular concern. Plants to recover Gallium from waste streams in Australia were shut and nobody cared.

A more persistent issue is the inherent problem of separation of function between governments and companies: companies do not care about a market of this size even if it has profound concentration and inelasticity because they are generally not in the business of geopolitics or handicapping trade shock risks. Sovereign nations that are in the business of geopolitics and that have intelligence agencies can form views on this but do not have mechanisms to intervene in most countries. Notable exceptions are Leninist systems like China or corporatist Japan which has unique resource security infrastructure between JOGMEC and explicit ministries and processes around mineral and material security. Japan’s institutional memory and history of trading houses working hand in glove with the state goes back to Japan’s efforts to secure itself economically prior to World War II about which Michael Barnhart wrote an authoritative account “Japan Prepares for Total War”. Korea has some similar features institutionally and politically.

In the interim, users of Gallium got cheap material and did not ask many questions about concentration risk because of the perceived small monetary amounts involved. That is dangerous and illusory because these materials have highly specific and almost completely inelastic demand. What is the marginal price of a material necessary to avoid $100mm fighter jets getting shot down in a conflict with the main supplier of that material? For Yttrium in thermal barriers $400 of material can block the production of a 200-300MW CCGT that costs $200-300mm. For the company making a 15% margin on that turbine you could argue the value of that material goes from $400 to several million at least when it is a binding constraint on production. The simplistic marginal analysis of the like made by Javier Blas in a recent Oddlots that these markets are small and that we should not care is facile and dangerous. These markets are small right up until they blow up and take out substantial value add downstream. If they can be kept small through supply side policy actions that are similarly low cost and manageable they should be.

What Can and Should Be Done

To solve these issues there are two problems: one physical, one financial. Firstly, economic deposits or sources of material need to be found in secure places or in a portfolio of secure-enough places so that a degree of supply security can be restored. In periods of oversupply material can be stockpiled as outlined in my previous work with Employ America and the Roosevelt Institute but we absolutely do not have that luxury today. The second problem is financial: how can governments incentivize production through cycle even when mining companies with activist shareholders would long shuttered plants? Even Japanese corporates are not immune to the likes of Elliott or other shareholder activists that look to squeeze every dollar of cash flow out of businesses, long term strategic impacts be damned.

The solution to the financial problem I have covered elsewhere as has former Deputy National Security Advisor Daleep Singh and Arnab Datta at Employ America. A strategic reserve that can write caps and floors can ensure passable worst-case economics for production of materials and stockpile excesses in “good” times of Chinese or other dumping. In bad times those reserves can be released, and the baseline level of capacity can be kept running to ensure longer term resilience. How to weigh up stockpiling versus subsidization is an interesting problem and one which I am working on but at this juncture much like a sedentary person seeking to improve their health getting moving would be a good start. The US currently has no legislative authorities or capacity to do this, but the Australian Government does plan to introduce such authorities following the Australian Labour Party’s generational election victory with plans for a strategic reserve which has been received well by domestic security thinktanks.

The physical problem of finding supply that can be brought online quickly is an easy one for Gallium. Australia is the second largest alumina producer in the world and the waste Bayer Liquor from alumina refining is a rich source of Gallium. Rio Tinto has already demonstrated a flowsheet that can do this in Canada that can be copied across to Australia in short order. Australia historically produced Gallium in the 1990s in a small facility setup by Rhone Poulenc, now part of Solvay. Securing 100 tonnes of Gallium at $250 per kg would cost $25mm per annum if it was stockpiled but more likely in the current market context it could be sold at a profit - the problem is that companies’ shareholders do not have the wherewithal to do small projects that might be dumped on by China again. A floor price would have to be offered and the government should also take a cap price to ensure the taxpayers get a return in a supply shock. This would at worst cost single digit to low double-digit millions per year - a pittance in the defence industry and should “through cycle” lead to a neutral fiscal cost or the end of Chinese adversarial dumping. Either represent a great outcome and worth the expense.

This represents an urgent need but would also likely be a source of goodwill with the United States for Australia. As trade negotiations continue and US efforts to stand up such legislation get caught behind various other initiatives on trade and a tax bill Australia can step into the breach. This would give Australia negotiating leverage on trade and particularly steel and aluminium exports. Further extracting other materials and in particular rare earths from these waste streams has an extensive literature but is yet to be seen outside China but may represent good “bolt on” efforts to initial moves in Gallium. There is no time to waste and a significant opportunity for Australia to be a good citizen in trade anti-coercion efforts along with demonstrating value to a United States whose central tenet of its foreign policy today is “what have you done for me lately?”

“Renewables are Making Wholesale Energy Prices Go Up”

This is standard fare at the Australian and the like and is readily rebutted by the Australian Energy Market Operator’s Q4 system dynamics report. As you can see below, prices are only high at certain times of the day - namely around peaks when solar is not available:

Much of these seems attributable to withdrawn black coal supply:

Which in turn has pushed up realized prices for black coal when it is the marginal MWh in the market. This is a great outcome for black coal plant owners, not so great for consumers and most definitely not the fault of renewables.

A more oblique argument might be that renewables destroy midday profitability through excess supply forcing capacity cutbacks of inflexible coal. Maybe, though that should change quickly with the rapid deployment of batteries and come at the expense of very expensive gas that sets the evening peak pricing. As can be seen below batteries are already pushing the marginal pricing power of gas out and hydro remains dominant. Given the practice of coordinated portfolio bidding by the likes of Snowy Hydro and others that own both hydro and gas assets this is as much a market competition and concentration problem as much as anything else especially since such practices have been found to be legal.

More storage owned by parties that do not own extensive fossil portfolios is the answer here and that is what is coming according to Modo Energy and BNEF. The entry of capacity outside of gentailers should be a good thing for competition in peaking hours price formation.

So why aren’t prices going down yet? A few reasons:

1. These batteries are not all installed yet.

2. Global fossil prices are still high following the Ukraine War. While LNG supply is expanding towards year end Australians eat market rate coal and gas now - legacy cost plus coal contracts are a thing of the past now in the NEM.

3. So long as 1 and 2 are true, fossil is high and sets the marginal MWh especially at peaking times.

Tediously for the ALP this should all change over the next 12 months but not in time for the election. One indicator of where this is going is in South Australia where wholesale price components of bills have already begun to fall despite a war and a lack of batteries. That is likely the outlook for most of Australia through 2028.

For grid investment the issue is really one of competition between distributed energy resources owned by consumers versus a regulatory framework which assumes grid services can only be provided by grid owners. Grid owners want more capital expenditures to grow their regulatory asset base to make money. Smart grid technologies and distributed technologies like renewables with storage and vehicle to grid bidirectional charging can obviate a lot of that capital expenditure. How this plays out is less a matter of what is technically possible and more who lobbies more successfully and is well covered by IEFFA here. There are some compelling signs of less regulatory capture going forward from the AEMC but this fight is far from over.

There are some clear cleavages and divergences of interest within the GOP’s broad church. To name a few that I am watching:

Directorate S types versus the Hawks and American Petroleum Institute

If Russian gas returns to the market both via LNG and pipelines - and it is hard to see why Putin would support a deal that does not allow this - then that is a supply shock close to half the size of the entire global LNG market today of around 200 billion cubic meters. It is very hard to see how that squares with the interest of LNG exporters in the US gulf coast. It is very hard to see how US LNG competes with Russian gas in Europe given the Russian gas does not have liquefaction and regasification costs. It is very hard to see how the US walks from Ukraine and gets to tell Europe which gas it can buy. Something has to give here as it pits America First and Russian Friendly foreign policy against the American Petrostate. European gas demand is falling and will continue to do so - especially so if Germany were to restart any nuclear units.

I suspect this is going to be cause of serious rifts within the GOP coalition. To date Elon et al have moved faster and broke things faster than the Senate GOP can respond but that is unlikely to stay the case indefinitely unless the US has a multiyear cultural revolution.

Tech Power Demand vs Owning the Libs by Blocking Renewables

As you have read elsewhere AI uses a lot of energy and AI growth in power demand at the same time as electrification of transport fleets, industry and a generally healthy US economy means power demand is going up. Read more here on datacenters and their infrastructure requirements, and here for US power demand. Most of that growth has been provided by renewables:

Short term forecasts are largely “in the bag” due to projects being constructed but how the US intends to grow power output without renewables is far from clear. GE Vernova and other providers of gas turbines are sold out through 2028 or later at this point and nuclear restarts are going to require significant help from the Loan Programs Office which the current administration appears to be shutting down as quickly as it can. Advanced geothermal is great but nascent and not particularly quick to build especially since organic Rankine plants to use the heat from below have many commonalities with conventional thermal plants. At some point “owning the libs” is going to have to come at the cost of AI dominance or load shedding - a distinctly South African experience which really rhymes with the times.

The only other option is to achieve some radical leaps in chip performance that reduce power demand. I am an investor in this space and a director of Opticore and fundamentally very bullish on these prospects - not least of all because I am fairly sceptical of endless partisan bickering delivering coherent long term-oriented policy. There are others but whether this will all arrive in time before parts of Virginia experience Johannesburg levels of power reliability remains to be seen.

Military Industrial Complex vs Military Industrial Complex

One of the more lucrative trades post-election has been shorting US defence primes and buying European ones. The story is simple enough: US wants to cut spending and pursue isolationism and Europe is going to have to step in, and that is before you consider the possibility (certainty?) that newer defence companies owned by GOP silicon valley donors are probably going to take share. This may be great for the venture capitalists but it is likely to be quite ugly for US defence industrials who are generally very good at spreading their production to marginal districts. There is a rift brewing here between how much more an isolationist US can sell into North Asia or India versus what it loses in Europe and how much it gets chipped away by Anduril and the like. This money flow into key districts has not been messed with since defence consolidation in the 1990s and well…. it may be very contentious between the GOP old school and new school.

There no doubt is more, but these rifts leave plenty of scope for division and destabilization.

The Manosphere / Zynternet / What Happened With Men

I am slightly loathe to discuss this as a:

1. Man

2. A white man

3. A white, hetrosexual man

Because there are still some folks running around screaming in the comments at people like me as if that worked the first time.

I will just stay that if some broad demographic grouping is large enough to swing an election, then it has to be dealt with somehow. I did a lot of martial arts growing up and remember Rogan when he was a UFC announcer and comedian and have at times enjoyed his podcast. Doing a lot of martial arts means you understand some of the practicalities of being a police officer and say, dealing with someone on meth and armed. For anyone who wants to defund the police: turn up to your local dojo and seeing how you go with someone holding a plastic knife. I think a little appreciation for the lived realities of people who do security work can go a long way.

Similarly, what is being packaged by Jordan Peterson and the like is really some combination of stoicism and self-care without any broader reflection. Why is that appealing? Because there is a lack of competition. A lot of the discourse around men and masculinity is “do not be x” or worse “you cannot change and are terrible”. They mostly just want tasks and stuff to do! Bell Hooks has figured this out but strangely a very noisy segment of the left has not. On a more philosophical level a lot of men seek some outlet for what Greek philosophers called thumos and which Fukuyama has written a great deal about. Repackaging old wine (ancient wisdom) in new bottles (apps, podcasts, etc) is an old trick but it still works. People on the left would be well advised to work out how to sell good, constructivist things in better ways.

To that end the below might be part of the way forward. I have heard dumber ideas than ozempic to kick drugs and then straight to some kind of nice family friendly liberal dojo, dharma center or church. If a notable feminist theorist and what some would describe as a “finbro” can agree the strategy needs to change, well, maybe it does.

People Really Dislike Inflation

As noted here, here, here and here. Identifying the problem is now largely done but what to do about it and how to message around it is another. There are few issues here:

1. Can Causes of Inflation be Communicated? Can you communicate the endogenous and exogenous? Most empirical work on the inflationary episode during 2021-2023 has landed where Bernanke and Blanchard did: it was largely attributable to global shocks caused by energy, supply chain snarls （particularly in semiconductors) and food (also energy via nitrogen fertilizers). Is there any way for governments to communicate this or are they invariably going to get the blame? If you buy the inevitability story, then the decline of incumbents in the last year has echoes of Jimmy Carter. Australia has an election soon and this is far from an academic consideration for the Australian Labour Party.

2. What Can You Do About It? A lot, in my opinion. As written about extensively here and at Employ America measures to ensure supply resilience and stability via stockpiling, offtakes and more broadly defined industrial policy can be justified on security and monetary stability grounds alone without appealing to broader left projects. Democrats at time seem to get “the ick” in describing a more uncertain and volatile world and measures to mitigate and hedge against it. It is long past time to wade through the icky feelings and appreciate that these messages chime with messages that work for the right and to seize this ground.

3. Politically Can You Take Credit For What You Do? I believe Democrats can and should take credit for this. The CHIPS act and efforts to ensure secure supply of basic analog semiconductors alone mean that if half of Malaysia simultaneously gets COVID it is unlikely to take out US auto production again. Having the inputs to clean technology onshored means that supply and progress in that area should be more stable and that insanity in the Middle East has no bearing on how much sun falls in Texas or wind blows in Idaho. These really simple facts and messages somehow did not make it into male centric podcast world despite the fact you absolutely can do that as I did in this podcast with the Money of Mine guys.

Border Issues

A firm border policy combined with fairly liberal and nationally self-interested immigration is now bipartisan consensus in Australia and has taken numerous right-wing talking points and scare campaigns off the table. These battles were fought over twenty years ago in Australia and given the clear drift of opinion in the US accepting this reality sooner rather than later is not optional if Democrats want to win. If the latino part of the electorate does not want open borders, then trying to hold this position on behalf of a slim but well-funded group of activists is too high a price to pay. I suspect Democrats will be in a good position to push back on chaotic deportations in two years, but it is going to be important to understand where to draw the line here.

Geopolitics and America’s Place in The World

There are going to be pieces running for thousands of words and Fukuyama is as eloquent as ever. To get down to the nitty-gritty though: much of how the world is configured in terms of energy flows (fossil heavy, travelling long distances) and trade (long and complicated supply chains) presumes a degree of order and a lack of certain kinds of risks. Shipping rates may vary with fuel costs and the demand and supply balance of shipping capacity, but pirates have not impacted freight flows in the Red Sea a great deal in recent history. In comparison the Houthis have done and this represents something of a breakdown in trade risks due to changes in military technology availability and a lack of a hegemonic power that can and will enforce the rules. If the US continues to step back putting the supply closer to the demand and having more energy supply be local and renewable is going to be a persistent trend and that is before we get into any discussion of tariffs. A cold analysis of supply chain risk is likely to intensify many of the impulses of the Biden administration in CHIPS and the IRA but do so elsewhere.

This is going to lead to the rest of the world having to ensure their own security to a greater extent because it is quite clear from recent comments that US commitments are no longer there or capriciously conditional. A quick solution is kicking nuclear non-proliferation to the kerb and that is a much safer bet than anyone wants to admit right now. Similarly, buying US large ticket weapons programs to keep the US happy is much less appealing if the US is a fickle ally. Investing in drones and the platforms that countries actually need along with diversifying supply away from the US is sensible. I doubt this ends well for US defence contractors who get a lot of operating leverage out of exports. Ukraine’s ability to build a lot of capacity in-house during a war is likely to be copied by many others.

There will be many more unsuspected impacts of this but as Gary Gorton notes in his broad work on bank runs the migration from riskless to risky is seldom a smooth or gracious one and has a tendency to accelerate itself as in the collapse of Silicon Valley Bank. That transition now applies to every single US commitment but especially so in the security sphere. Every edge on a supply chain graph now needs to be risk weighted appropriately and the configuration of production, storage and transport reoptimized. This is going to take years.

The official said the investments could be deployed to shore up the resilience of supply chains and finance “illiquid but solvent companies that need greater scale to compete against [People’s Republic of China] rivals”.  

Another use could involve funding the creation of “synthetic reserves of critical minerals”, the official said.

This has been discussed at length in the lithium and critical minerals series I wrote with Arnab Datta at Employ America, and in Arnab’s piece with Daleep Singh in the Financial Times. 

The Economic Problem

China has a tendency to overproduce and dump in strategic sectors. Whether in metals, solar panels or autos and batteries there is a consistent pattern in China prioritizing a sector leading to excessive capacity expansion and dumping abroad. The dumping is as much a result of soft credit constraints in the banking sector whereby banks will lend into businesses with terrible and deteriorating margins because that sector is either a politically powerful SOE or the sector is favored by national level policy. Victor Shih, Michael Pettis and Brad Setser have been on this beat for well over a decade and nothing appears to be changing on the China side. More booms and busts in policy sectors, more non-performing loans and bigger current account surpluses are likely to continue if there is no response. 

The Security Problem

A more concerning development is that once China develops a dominant position in production it frequently uses it to political ends via trade policy. A common strategy here is to ban exports of upstream commodities which they produced at a loss for years in turn killing off ex China competition then using export bans to consolidate the downstream within China. The CSIS series on Gallium is an excellent history of how this happened in Gallium:  

This leading position [in aluminum] has allowed China to establish a dominant share of global gallium production. Deliberate government policies also played a role, with Beijing requiring its booming aluminum producers to install the capacity to extract gallium. From 2005 to 2015 alone, China’s production of low-purity gallium exploded from 22 metric tons to 444 metric tons.

China’s rapid rise in the industry created oversupply in the global market, triggering severe fluctuations in gallium prices throughout much of the 2010s. Leading suppliers in the United Kingdom, Germany, Hungary, and Kazakhstan suffered as a result and were forced to shutter their production, leaving China as virtually the only supplier in the world.

Gallium goes into wide bandgap semiconductors and particularly high-power radio frequency applications (F-35 radars) and anything in power electronics from high wattage phone chargers to EVs. China’s export ban is now driving prices up in a manner similar to what happened when China also flexed its capacity in rare earth metals in the early 2010s. This strategy of developing dominance then using it for broader nationalistic objectives should be familiar to any European who has paid a gas heating bill in the last few years.

The Rare Earth Minerals Wars

After ramping prices up with export controls China then dumped prices to discourage new entrants leading to some spectacular impairments for private sector investors in resources. China almost succeeded in killing all new entrants and were it not for the efforts of JOGMEC, an arm of the Japanese government and a few creditors, an Australian producer and the only operational producer ex China called Lynas would likely have fallen into Chinese hands. Lynas’ equity returns over this period have been poor (LYC AU in white in the chart below) because the commodity has been alternately dumped and restricted (Neodymium-Praseodymium oxide, a rare earth input for high power magnets) which makes for extremely dilutive and inefficient growth in resources. Developing secure supply chains has been a largely thankless task which makes private capital somewhat ill fit to solve these problems because in a head-to-head with the Chinese state and its capacity to crash prices the private sector loses as the biggest most performance insensitive balance sheet wins.

Are We in a Battery Metal War Now?

There are many people who now think China is doing this in battery metals and solar and given the recent financial results for upstream listed companies in those sectors in China （-78% gross margins on polysilicon for one company) and their eye watering losses I am inclined to agree with them (edit: so are the companies themselves at least in solar according to this piece from Caixin out today). China is not the lowest cost player in lithium in particular and running capacity with cash costs in the $17,000-$20,000 per tonne of LCE equivalent when spot is $12,000 is highly suspect. I regret to say I have seen this show before and barring some government action it is unlikely to end well for equity holders. Unless governments outside China are willing to lend into losses like Chinese banks then a different set of tools will be required. 

Trade Protection Alone is Not the Answer

The obvious one is trade protection including tariffs. If a country restricts something being dumped to support the local supply chain that can be very effective in allowing domestic supply to develop. However there are acute problems with this that are emerging today. Tariffs need to be durable and credible and for that there needs to be some political consensus on them and a belief or confidence that the costs of the tariff will be manageable or negligible over time. The longer term cost of tariffs largely depends on how quickly domestic output can be ramped up and what that domestic output costs.These costs may be small and transient for some goods but in many of the capital intensive sectors that China engages in dumping that capacity takes time. Similarly, political consensus or a perceived lack thereof around certain sectors can lead private capital to hold back depending on the political cycle. US solar modules currently cost ~$0.30 per Watt due to tariffs and cost ~$0.085c per Watt elsewhere due to Chinese dumping and yet numerous US capacity expansions are being deferred until policy clarity after the election. This is imposing a cost on US consumers in the interim as a fair “cost plus a passable return” price in the US would be $0.12 to $0.15 per Watt as I argued here.

Smoothed Price and Production Paths 

Offtakes and stockpiling can be used together to smooth this path to building domestic capacity and pushing back against dumping while having minimized inflationary impacts. By providing some degree of security for investors and lenders as domestic output is ramped via fixed prices or collars the state can reduce the risk and cost of capital of building that out in a manner that provides security and predictability outside of a political cycle. This in turn can minimize transitional costs of protection provided the domestic industry is long term viable and competitive. In the case of the solar supply chain the US could provide a guarantee on a net basis including all subsidies such that any IRA roll back would be baked into the fixed or protected price period so that capacity gets built. In the case of metals like Gallium the US could provide similar price protection to get projects up and running and then use a strategic reserve to moderate price fluctuations and inflation going forward. These sectors would become less volatile, cheaper to finance and that would result in better value for consumers just as auctions for renewable capacity drove down the cost of wind and solar power. 

Making Bad Trade Conduct Fail

These measures would make China’s playbook of dumping and dominating upstream to consolidate and provide political leverage wholly ineffective at least for the United States. As seemingly uneconomic as China’s state and banking system nexus can be if they cannot ultimately dominate a sector in the long run by running spectacular losses in the short run then they are unlikely to try. This has important implications for global rebalancing: China responds to and respects little else but hard power both economically and militarily and by securing its supply chains the US will maintain the latter while also showing its steel in the former. It might even inspire similar measures from Europe which to date seems to be slow to grasp the magnitude of the dangers and challenges they face. If there is anything that could lead China to rethink its broader economic practices it is far more likely to be this than pleading requests and shuttle diplomacy.