Energy windfall tax proposals are all the rage again.

• Germany Weighs Windfall Tax as Iran War Drives Price Surge

• Australia Launches Gas Tax Review as War Fuels LNG Windfall

• War Forces Reeves To Keep North Sea Windfall Tax In Place

• EU ministers weigh oil price cap and windfall tax to rein in soaring energy costs

• Democrats urge windfall tax as big oil set to make billions from Iran war

Capital expenditure for oil & gas majors already sits at ~48%, down from 73% back in 2015, due to a shift towards capital returns. Windfall taxing their profits in boom years only compounds this underinvestment.

Also, the only upstream investment that ramped over the past decade was in the Middle East, where the war is now busy damaging energy infrastructure.

Podcast/Video

I’m sure I’ve included this speech before, but it’s that important, here it is again:

The Future of War with Former Navy SEAL | Erik Prince

Drone + missiles + satellites is the new military paradigm.

On sale now: China is mass-producing hypersonic missiles for $99,000

A Chinese company is in production of a hypersonic missile, with a sticker price comparable to that of a luxury sedan.

The launchers for the YKJ-1000 resemble common shipping containers, are mobile, and so the platform can be deployed to virtually anywhere.

Military analysts admit that this development radically changes the logic of warfare, with massive cost/benefit advantages now on the side of offense.

Russia is sharing satellite imagery and drone technology with Iran, WSJ reports

Russia operates approximately 110 military satellites (1,559 total). Iran fields just 10–13 operational satellites, none with resolution better than 1 meter.

Quote

“Almost half of the US data centers planned for this year are expected to be delayed or canceled. One big reason is the shortage of electrical equipment, such as transformers, switchgear and batteries.”

-Bloomberg: America’s AI Build-Out Hinges on Chinese Electrical Parts

Subscribe now

It’s just pinch point after pinch point with the Chinese holding the cards, be it GOES, NOES, rare earths, gallium, germanium, antimony, etc.

Tweet

Is oil facing an inverted March 2020?

Charts

If there were a “Political Darwin Award”, Germany and Taiwan would be fighting for first place with this chart (coal still makes up 22% and 35% of their respective grids).

The Darwin Awards are a long‑running, tongue‑in‑cheek “honour” for individuals who accidentally remove themselves from the gene pool through spectacular stupidity.

Taiwan hasn’t even made it a year before trying to restart two reactors.

Something I’m Pondering

I’m pondering whether this war, like the 1973 Oil Embargo, will usher in a new paradigm of energy security. Take the example of 75% of Japan’s and 40% of France’s electricity coming from oil in the lead-up to the 1973 oil Embargo.

Hope you are all having a great weekend!

Cheers,

Ferg

P.S. Always a pleasure chatting to Trav and JD as I’m a huge fan of the show and their regular email, which you should all get! (Mia teased I’m cross-eyed in the below AI pic, and now I can’t unsee it haha).

This is the 20x the title is referring to, and yes, I’ll also be hanging onto my gold mining juniors….

Subscribe now

Article(s)

Always a worthwhile read: Goehring & Rozencwajg Q4 When the Strait Closed

A wildly divergent view from Anas Alhajji (no paywall): The Hormuz Crisis: An Alternative View*

*A view I struggle to reconcile as if the US is in control why can’t they defend their military bases? “Many of the 13 military bases in the Mideast region used by American troops are all but uninhabitable” or why are we seeing an increase in missile and drone penetration rates?

“A degraded Iran firing fewer and better aimed missiles and drones at carefully selected, fixed targets is getting more effective at imposing costs,” Grieco said. “In terms of how many are getting through, the trajectory is moving in the wrong direction.”

Podcast/Video

Luke Gromen | Open or Closed: The Strait of Hormuz is the Only Thing That Matters!

For tracking the conflict, BCA’s dashboard is useful: BCA’s Iran Conflict Daily Dashboard.

Subscribe now

Quote

“Supply chains will move from just in time to just in case”

- Paul Singer

Where does LNG now fall on this “just in time to just case spectrum”, given it was the fuel promoted by the industry as a reliable bridge from coal (and nuclear in the EU) to full reliance on renewables + batteries?

Many developing countries will be carefully reconsidering the security of their energy mix, having seen the positions the EU and now Taiwan find themselves, having closed nuclear and coal generation.

Tweet

This was an incredible piece that Louis was kind enough to drop the paywall on: Shattered Assumptions And The Energy Quandary.

Besides missiles and drones, the Iranians have deployed some epic trolling.

Charts

The US earnings forecasts ripping higher caught my eye.

Breaking down the drivers of these EPS upgrades.

Something I’m Pondering

I’m pondering the above EPS upgrades in light of three questions.

1. Can TSMC get the power to maintain production?

   TSMC produces around 90% of the world’s leading-edge semiconductors used for AI and quantum computing applications. TSMC’s top five customers: Nvidia, Apple, AMD, Broadcom, and Qualcomm, account for an estimated 65-70% of TSMC’s total revenue. There is almost no alternative manufacturing capacity for their most advanced chips.
   

   TSMC also consumes roughly 10% of Taiwan’s total electricity consumption, and Asia’s consumption is ramping from here into the peak cooling season in July. As natural gas (all imported LNG) accounts for 48% of Taiwan’s electricity (for which they have 10 days storage), Qatar provides 34% of that LNG, which equates to roughly 16.5% of Taiwan’s electrical generation.

2. Can TSMC get the helium it requires to maintain production?

   The Ras Laffan facility in Qatar supplied 30% of the world’s helium and will take one month plus to come back online if restarted tomorrow, before considering 14% of annual helium export capacity was destroyed in the 19th March missile attacks.

   Approximately 200 specialised cryogenic ISO containers filled with liquid helium sit stranded in the Middle East. Liquid helium boils off within 35 to 48 days, meaning these inventories will be lost entirely if shipping does not resume by mid-April.

3. Are the Gulf States US AI investment commitments still ‘sound’ given their current situation with lost revenues, domestic rebuilding and increased defence requirements?

   Gulf Cooperation Council states have announced approximately $250 billion in AI-specific investment commitments through early 2026, representing roughly ~15% of the $1.4–2.0 trillion projected for US AI infrastructure spending during 2026–2028.

Then again, it looks like the Nasdaq may now be in the process of repricing these risks.

Hope you are all having a great weekend!

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Sometimes it’s easier to make the point with headlines (dates highlighted).

I couldn’t resist a meme.

Podcast/Video

To balance the above, I’m definitely keeping an eye on how rapidly China is overcoming renewable constraints on their grid.

These videos were eye-opening.

• China’s renewable power dependency has a major flaw. This transformer just solved it.

• Revolutionary generator transforms Chinese factories into power plants.

• China builds new water battery to power entire electrical grids.

Subscribe now

Quote

Edge isn’t permanent. It decays. The question isn’t whether something works. It’s how long before the market adapts and makes it stop working.

Trading Psychology Stick (from Kevin Muir’s 2025’s TRADING WISDOM)

This quote struck a nerve with me, as whenever I found a “winning angle” in trading, it slowly disappeared over time. The only angles I found that are “evergreen” are having a longer time horizon and behavioural arbitrage.

Tweet/Notes

Great insight into the issues with opening the Hormuz Strait.

Anas is preaching to the converted.

Charts

Waiting for a chart breakout….

I found this breakdown (Unheard: Is Iran winning the war?) on drones and missile launches interesting.

Something I’m Pondering

I’ve been pondering this quote, but with my trading/investing information diet.

A question to determine if you’re thinking critically: what proportion of the books and articles you read do you disagree with? If it’s near zero, then either you’re agreeing with whatever you read, or you’re only reading whatever you already agree with. Both paths lead to idiocy.

-Gurwinder

It’s very easy to build an echo chamber in Substack, podcasts, YouTube, and Twitter.

One investor I really enjoy reading is Erik from YWR, as we come at the market in a similar way, looking for hated out-of-favour sectors, yet draw some very different conclusions. I pride myself on sorting through roadkill no one will touch, yet YWR: Untouchables has some stuff I refuse to touch (renewable energy companies and European autos.) We also have quite different macro outlooks (YWR: 4 reasons we make new highs), and I love seeing his process and how he draws his conclusions.

Solar is another area I am trying to find more experts (I had a great chat with David Dixon, Rystad’s senior renewables analyst). I still hold the view that solar makes little sense in countries in the Northern Hemisphere. What gives me pause is some of the battery and grid breakthroughs I’m seeing, particularly in China (as I shared in the above videos). If sodium-ion or even water batteries prove economic at scale, I’ll need to rethink my entire view of solar’s potential.

Hope you are all having a great week!

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

AUSMASA: Skills Shortages in Mining

I guess it should come as no surprise that the chart of Australian mining graduates looks similar to the chart patterns of many of the commodities I follow, in its bombed-out, crawling along the bottom of the chart.

It’s not just Australia either:

• The United States has lost nearly half its mining engineering enrollment — from 1,561 students in 2015 to approximately 600 by 2022. Twelve mining engineering programs have closed since 1985, reducing the total from 25 to 14. The remaining schools collectively graduate fewer than 200 mining engineers annually, whereas an estimated 400+ are needed.

• The United Kingdom represents the most extreme case. The Royal School of Mines at Imperial College abandoned mining as a standalone discipline. Camborne School of Mines, the last UK institution offering undergraduate mining engineering, paused new intakes in 2020 after its final cohort dwindled to just 4 students. Of the 1,250 mining engineers on the UK Engineering Council register, 80% are over 50 and 40% are at least 60. The UK currently has no traditional full-time undergraduate mining engineering program. (I stand corrected thanks Josh as the the course has now been re-opened to new graduates!)

Podcast/Video

Jeffrey Gundlach: Our Next Turning Point

I particularly enjoyed this section from the 38-minute mark on, where he dug into “Volatility laundering”. DoubleLine recently put out this piece on it.

Subscribe now

Quote

This ties to Gundlach’s observation that there has been a lot of “marking to myth”.

“Larry, beware of the three L’s and the three M’s.”​

“Liquor, ladies, and leverage.” (the three L’s)​

“Mark to market, mark to model, and mark to myth.” (the three M’s)​

-Lawrence McDonald describing a conversation he had with Charlie Munger.

Tweet/Notes

Marine Traffic: Strait of Hormuz Ship Tracking Map

Yes, a few ships are slipping through with their transponders off, but consider the prior average was roughly 138 vessels per day, of which ~60% were tankers.

Unless vessels start moving, tacos are off the menu.

Charts

Following the earlier article on mining skills shortages, one has to wonder whether mining automation will be the way to overcome the skills shortage.

A little case study I’m monitoring is Fortescue’s full embrace of cheap Chinese loans and mining equipment vs BHP’s ongoing disputes with China (China widens BHP iron ore restrictions as talks drag).

China will dominate the global automation/robotics market, having perfected it in its domestic market (before considering critical mineral access).

Something I’m Pondering

I’m pondering new ways to use AI to improve my research/investment process.

They say the first casualty of the war is the truth, so I’ve been playing around with ways to get a well-rounded picture.

Take the example below, using Claude Opus 4.6 with extended + research mode.

A hack I use constantly is the following when first looking at a company. It’s amazing what it throws out and quite often leads you to a competitor who is better positioned. A good exercise is to run your current portfolio positions through this and see how confident you are being on the other side of the risks it identifies.

Hope you are all having a great week!

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Three different views/scenarios on AI.

• AI Won’t Nuke SaaS: Why the White-Collar Apocalypse Is Just Tech Bros’ Dreaming Nonsense - Craig Tindale

• The 2026 Global Intelligence Crisis - Frank Flight, Citadel

• THE 2028 GLOBAL INTELLIGENCE CRISIS - Citrini and Alap Shah

For me, Luke Gromen frames it best:

Podcast/Video

Great interview with the GOAT: Hard Lessons: Stan Druckenmiller: Invest, then investigate

Niche content for offshore oil bulls: Esgian Rig Webinar: Floating Rig Market: What lies ahead?

Subscribe now

Quote

Somewhere in Beijing, inside the Ministry of Commerce headquarters on Chang’an Avenue, there exists a filing cabinet. It contains export license applications for gallium. In 2024, every single application from American companies went into that cabinet. None came out. Not one kilogram of gallium was approved for direct export to the United States during the entire duration of the ban.

-Shanaka Anslem Perera The 99% Monopoly: The Chip Metal America Cannot Buy

Tweet

Despite the headlines, the US has actually become more import-reliant over the past year, from 100% reliance on 15 commodities in 2024 to 100% reliance on 16 commodities today.

Charts

I find the argument that something historically highly cyclical is no longer cyclical amusing, i.e., semiconductors to the upside and oil to the downside....

New Rules: Semiconductors Move From Cyclical To Structural

Something I’m Pondering

I’m pondering how the complete absence of a bullish narrative markets often marks a bottom.

No one had anything bullish to say about oil services or producers in April/May last year.

A bottom is always obvious after the fact.

Sentiment is just a 3-month trailing average of the price action.

-Gerald Minack

This was a great note on energy by Michael Goldberg at Collins St Asset Management:

Our #1 Commodity Call for 2026: What’s Old is New Again! Authored by Michael Goldberg, January 2026

I hope you’re all having a great weekend.

Cheers,

Ferg

P.S. I had the honour of jumping on the Market Huddle and chatting with Kevin Muir, whose blog Macro Tourist was one of the first blogs I started reading a decade ago!

Subscribe now

This is wild when you consider the global implications for mining costs.

The state-owned giant, operated by China National Coal Group, is thriving even in an era of falling coal prices – and posted a 40 per cent net profit margin in 2024.

For comparison, Wall Street titan Morgan Stanley achieved just over 20 per cent last year.

With artificial intelligence (AI) and 5G oversight almost everywhere, Dahaize is the smartest coal mine ever built, according to industry body China National Coal Association.

Since production began in 2023, its 980 employees – a skeleton crew compared to traditional mines – have overseen fleets of 5G-connected robots, autonomous tunnelling systems and AI algorithms that coordinate almost everything, from coal extraction to railcar loading.

Each worker now produces nearly US$1 million worth of annual output, according to Liang.

“A crew of four is all we need,” Liang said.

Tunnel-inspecting drones scan shafts in eight minutes – a task that would take humans many hours – while robotic arms handle pipe repairs and many other functions.

Autonomous trucks, guided by an underground positioning system, navigate foggy, dust-choked tunnels to ferry coal to AI-powered washing plants, where one worker can process 1,100 tonnes of coal per day.

Even as China’s coal prices slid 18 per cent in 2024, Dahaize’s revenue hit 9.1 billion yuan (US$1.25 billion), with 3.8 billion yuan in net profit – a margin comparable to luxury brands.

Podcast/Video

This had some great insights: Global Liquidity Has Peaked – What Happens Next for Gold & Financial Markets? | Michael Howell

The US and China look to be on opposing trajectories in their liquidity cycles.

Subscribe now

Quote

“Liquidity is the fuel that drives asset markets. Asset markets basically go up and down according to the inflows or outflows of liquidity into markets… Those money flows tend to be predictive. They lead fixed income markets by around about 6 to 9 months. Equities maybe a tad longer, and things like Bitcoin is actually a very sensitive asset and that tends to be led by liquidity within about a 3‑month time frame.”

-Michael Howell

If that theory proves correct, it paints a pretty grim outlook for equities and fixed income over the next 6 months.

Tweet

LLM Stats Leaderboards: (It’s the first I’ve heard of Zhipu AI: China’s Zhipu AI launches new major model GLM-5 in challenge to its rivals, February 12th 2026 ).

While the performance difference between top models is splitting hairs at this point, the cost difference is anything but, with GLM-5 currently 13-20% of the cost of Claude Opus 4.6, and is open source!

Charts

NVIDIA is still waiting patiently for those Chinese approvals of its H200 chips.

Nvidia CEO Says Chinese Government Yet to Approve H200 Imports (29th Jan 2026).

China is a market Nvidia really needs if they are to live up to the average analyst estimate of doubling sales and EBITDA by FY2028…

Something I’m Pondering

I’m pondering the “capacity illusion” in China: its capacity is understated, as it runs coal plants at a low utilisation (~50% capacity factor, which could be dialled up to 70-75% if required). Wind and solar’s capacity factors are already low and overstated when you consider they’re averages, i.e., more sun in summer than in winter for solar.

This ability to increase coal utilisation is why China can pull a +8.2% (2018) or +9.4% (2021) jump in electrical generation when required.

The US has no such luxury with an aging grid that, despite many investment headlines, is still sitting near all-time lows in the number of miles of new transmission lines built.

I hope you’re all having a great weekend.

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

China’s coal-fired power generation declines for the first time since 2015

I landed on the above article while trying to understand how China is building coal plants at a rapid pace, yet coal demand is forecast to decline in a linear fashion as they are being built.

According to Wood Mackenzie, coal-fired power capacity factors were as high as 60% in 2011, then declined to 52% in 2024 and 48.2% in 2025. Wood Mackenzie expects the utilization to further decline to 32% by 2035 as portions of the fleet transition to reserve status.

So the assumption is that China’s coal fleet will keep expanding while its utilisation rates will plummet, and demand will fall with it. It’s certainly dropped substantially from peak utilisation in 2011 of 60%, but has fluctuated around the ~50% capacity factor the last decade, granted its share of electricity has dropped from 79% to 58% as of 2024 (2025 still an estimate).

The idea being that coal’s role is structurally shifting from baseload to flexibility provider, which is true, as roughly 600 GW of China’s coal fleet is undergoing flexibility retrofits to ramp up and down in response to variable renewable output.

The questionable part of all this is what if electricity demand picks up?

For example, the explosive growth of artificial intelligence and data centres in China could drive unexpected spikes in electricity demand. While China is investing heavily in renewable capacity, rapid demand growth may force continued reliance on existing coal-fired power plants to meet immediate needs.

According to Wood Mackenzie, the aggregate capacity of data centres will reach 78 GW by 2030—a 105% increase from 38 GW in 2024. Since a large proportion of this new demand will concentrate in densely populated urban centres, coal-fired power may remain indispensable for maintaining grid reliability.

The chart below is a perfect example, as in 2022, you’ll get headlines of;

“66% of electrical growth was renewables in 2022”.

Yet when demand surprises like 2021 & 2023 you don’t see headlines;

“Coal was 72% of electricity growth in 2021”.

If you look at China’s power generation instead of installed capacity, it paints a clear picture.

Podcast/Video

I really enjoy Kevin Walmsley’s channel: China's top universities are opening to foreign students. That's a big problem for US schools.

• Best Global Universities for Engineering

• Best Global Universities for Computer Science

• Harbin Institute of Technology (HIT Tuition costs) vs US MIT Tuition Costs
  

Subscribe now

Quote

‘The AI race is our Chinese against their Chinese.’

-Naval Ravikant

No American grads on Meta’s top AI team, all 11 hires are immigrants

• Meta created a new Superintelligence Lab with an initial team of 11 senior AI researchers.

• All 11 of these researchers are immigrants, i.e., none completed their undergraduate education in the United States.

• Seven of the 11 were born in China or did their undergraduate studies at top Chinese universities such as Tsinghua, Peking, Zhejiang, and USTC.

Tweet

AI can optimise capital-intensive systems, such as manufacturing, robotics, and commodity processing. While in a capital-light system, AI “optimises” by eating profit margins (along with the valuations assigned to them).

Just watch what Musk is doing, not what he’s saying:

• ‘Optimus chain’: Chinese suppliers form the backbone of Tesla’s humanoid robot initiative

“With about 50 to 70 per cent of manufacturing and core component production expertise residing in China, we expect Chinese players to take on greater roles in the global humanoid robot supply chain,” said Cheng Xin, a partner at US consultancy Bain & Co.

• Tesla Will Derive 80% of Its Value From Optimus Robot, Musk Says

It’s just drop-shipping when 70% of the supply chain is outsourced to China, which is responsible for the majority of the progress in the field (China has issued five times the number of patents related to humanoid robots than the US).

Drop-shipping is a great business model until your drop-shipper decides they will compete with you and eat your margins.

It’s humorous that the answer to Tesla having its lunch eaten by Chinese EVs (with an estimated ~40-50% reliance on the Chinese supply chain) is to pivot to robots with 70% reliance on the Chinese supply chain.

Charts

Oil demand growth forecasts hinge on non-OECD growth assumptions.

Take the below International Energy Outlook 2023 (from EIA, as they didn’t publish 2024 or 2025). As of 2026, we are already going to exceed the 105.5million barrels a day (taking the midpoint of IEA 105mbd and OPEC 106.5mbd).

Something I’m Pondering

I’m pondering the different approaches of the Chinese and Americans to software monetisation.

The US model was built on strong IP protection, high-margin software licensing, and decades of consumer conditioning from Microsoft, Adobe, and SaaS pioneers, who treat software as a standalone product commanding fat margins of 70%+.

The Chinese model was forged by widespread piracy, intense price competition, and a cultural acceptance of near-zero margins.

It’s funny when you consider decades of piracy have positioned the Chinese businesses well to deal with a world of AI piracy.

Chinese software piracy rates exceeded 90% from 1994 to 2004, dropping to 78% by 2010, according to the China National Intellectual Property Administration. Or consider Microsoft’s revenue in China, the world’s second-largest PC market, which was once less than 1% of its global total. MIT research found Chinese students would switch to internet piracy even if street piracy were eradicated. This has led to a generation of consumers who are conditioned to view software as fundamentally free.

I’ve used a Tesla vs BYD example a few times in Tesla charged USD$8000 for its self-driving while BYD threw it in for free on cars that cost as little as USD$9600.

China treats software as infrastructure that enables hardware sales, ecosystem lock-in, and downstream service monetisation.

Because for some insane reason, the Chinese companies don’t focus on “maximising returns to shareholders” but “minimising profits so you can kill your competitors.”

-Angelica Oung

Talking of killing competitors open source Chinese models like Kimi and Qwen are currently 90% cheaper then leading edge US models and have access to all that coal capacity….

Kimi K2.5 and Qwen-Max charge around $0.15–$1.20 per million input tokens, while Claude Sonnet 4.5 costs $3–$15 per million tokens; that's savings of up to ~90–95%.

I’m having a lot of fun sorting through Chinese companies, as I feel many sectors are mispriced for the next decade.

I hope you’re all having a great weekend.

Cheers,

Ferg

P.S. It’s always fascinated me the idea of perceived safety, i.e. Nifty fifty (they are such high quality companies you can’t lose money…) to the Pristine Collectral of Treasury bonds. Compared to a Chinese industrial with close to its entire balance sheet in cash, a decent dividend and a highly competitive business on a global scale, yet is perceived as extremely risky…

Subscribe now

Article(s)

Trump’s trade deal with South Korea explained in headlines:

• South Korea eyes non-binding deal to steer $350 bln U.S. investments.

• Korea to delay $20bn US investment on currency pressure.

• South Korea offers tax breaks on overseas stock sales to stem currency’s slide.

• Trump Vows to Raise Tariffs to 25% on South Korean Goods.

South Korea has deployed an unprecedented fiscal intervention—offering temporary capital gains tax exemptions on overseas stock sales to retail investors who repatriate funds—as the won approaches crisis levels last seen during the 1997 Asian Financial Crisis and 2008 Global Financial Crisis. Announced in December 2025 and formalized in January 2026, the policy represents a novel attempt to stem capital outflows without traditional capital controls, addressing a structural shift in which Korean retail investors have become the largest foreign buyers in the U.S. equity market, accounting for 11% of foreign net purchases in 2025.

Podcast/Video

I love both Erik and Craig’s work: YWR:The Return of Matter

Subscribe now

Quote

The factory in Ohio remains half-built, its machines silent, while the paperwork for its most critical components circulates through an endless loop of administrative “reviews.”

-Craig Tindale (The Chokepoint Mandate)

Similar to the “excess capacity” label given to the Chinese industry when it strategically undercuts competitors to gain market share.

“Regulatory limbo” will be a phrase you’ll hear more frequently, whether it’s critical minerals import licenses or H200 chips, which I touched on a month ago regarding Nvidia.

Nvidia’s CEO says China is still finalising licence for H200 chip (January 29, 2026)

“The H200, the actual license for H200 is being finalised. And I’m hoping that also the Chinese government would allow Nvidia to sell the H200, so they have to decide. And I’m looking forward to a favourable decision,” he told reporters at Taipei’s downtown Songshan airport.

“I think that H200 is very good for American technology leadership. It’s also very good for the Chinese market. And the customers would very much like to have H200,” he said.

“And so I’m looking forward to a good decision. And so we just have to wait patiently,” he added.

Citing sources, Reuters reported on Wednesday that China has given approval to ByteDance, Alibaba and Tencent to purchase more than 400,000 H200 chips in total.

However, the approvals came with conditions which one source said were too restrictive, with customers not yet converting the approvals to purchase orders.

Tweet

This chart is so wild I had to fact check China actually installed more power capacity in 2025 than India’s entire installed capacity….or two Germanys if you prefer (before I even start on their widening gap between installed capacity and generation).

China’s Four-Year Energy Spree Has Eclipsed Entire US Power Grid

Charts

US semiconductor sales rolled over hard at the end of last year. While China and Asia semiconductor sales are grinding higher.

Something I’m Pondering

I’m pondering all I’ve learned since writing this piece just over a month ago: Setting Fire to Paper Markets and following reading Craig’s latest piece: The Revenge of Matter.

Chinese futures markets are explicitly constructed to “serve the real economy” not speculation.

• Chinese futures exchanges mandate RMB-only settlement

Western exchanges like the LME and COMEX see less than 1-2% of contracts result in physical settlement, the Shanghai Futures Exchange (SHFE) operates under a fundamentally different model where physical delivery upon contract expiration is mandatory rather than optional (no cash settlement escape hatch). Traders holding positions at expiration must either deliver physical metal or accept forced liquidation with 20% penalties.

• Regulatory Restrictions on Retail Participation (anti-speculation)

Chinese futures markets employ sophisticated mechanisms that channel speculative activity away from near-delivery contracts while preserving physical settlement capacity for commercial users:

Individual investors—representing 97% of all Chinese futures market participants—are prohibited from trading nearby contracts. This regulatory fence forces retail speculation into distant contracts (typically 3+ months from expiration) while reserving near-term contracts for institutional and commercial hedgers with genuine delivery intentions.

• Margin Escalation

Western exchanges set position limits as a percentage of open interest without distinguishing speculators from hedgers.

SHFE escalates margin rapidly as expiry approaches.

Takeaway?

What if US dollars can no longer guarantee access to required processed commodities?

What if to guarantee access you need Yuan settlement and we are seeing the start of a new Petrodollar style system based around China’s monopoly of mineral processing?

I hope you’re all having a great week.

Cheers,

Ferg

P.S. I’m just finishing up my latest post on exchange operators, and feel countries will soon be faced with a choice.

Subscribe now

Article(s)

These are the big flows I’m trying to front-run.

China’s $1.2 Trillion Windfall Quietly Seeps Into Global Markets

The shift in China’s foreign assets from the central bank to the private sector is ushering in a new era of global capital flows, with risks of a sudden capital reversal and increased sensitivity to yuan appreciation, and potentially leading to a global financial system reliant on liquidity sourced from China.

Chinese Savers Have $23 Trillion and Few Options Beyond Stocks

The idea that China’s small investors will shift a chunk of their $23 trillion savings pile to the stock market is a tantalizing one for global firms, who are showing signs of returning after years on the sidelines.

US and China Flip the Global Script as Capital Flows Reverse

China’s outbound investment is a consequence of a trade surplus that has continued to grow as efforts to increase domestic consumption falter and Chinese companies seek new markets. Where once China’s trade surplus ultimately ended up in US Treasuries via a state-controlled system, the capital is now funding new foreign factories.

It’s wild when you consider not only are China’s surpluses no longer being recycled into Treasuries, but the rest of Asia no longer needs to store their surpluses in USD to ensure they have access to the commodities they need to grow, since China/Russia is more than happy to do deals in their own currencies.

Podcast/Video

This had a bunch of great slides: Jeffrey Gundlach’s Just Markets: Clue

Quote

“While the West financializes deficits, rivals industrialize surpluses.”

-Craig Tindale

Subscribe now

Tweet

I love Louis’s framing of the Chinese setup as reminiscent of the post-GFC United States.

Charts

After watching Jeffrey Gundlach’s presentation, I dug up this chart from one of his previous presentations. Luke Gromen sums up the Trump/Fed drama in:

This fight is not about Fed independence; at its core, it is because the US debt and fiscal situation is going critical and neither the Trump Administration nor the Fed want to take the blame for what comes next.

Which is monetisation of US deficits i.e. print the difference while inflation rips.

Commodities are sniffing out this inevitability one by one.

Something I’m Pondering

I’m pondering how many times I’ve read “excess capacity” in articles relating to China when what was really happening was China strategically undercutting global competitors to gain market share.

Craig Tindale touched on it in this piece, with the Quantified Asymmetry Between Market‑Led and State‑Directed Systems.

Take the chart below of China’s market share of processing across multiple commodities.

An example is 3-5 years transformer delays from where China isn’t cited as the main constraint with Siemens Energy, GE Vernova, Hitachi Energy, that is, until you peel back a layer.

• Manufacturing facility limitations (40-50% of delay)

  China represents approximately 60-70% of the "manufacturing facility limitations" constraint that causes 40-50% of transformer delays.

• Specialised skilled labor shortages (20-25% of delay)

• Copper windings sourcing and cost (10-15% of delay)

  China represents approximately 30% of the “copper windings sourcing and cost” constraint.

• Grain-oriented electrical steel (GOES) and tariff uncertainty (15-20% of delay)

  China has dominant market share in both GOES (56%) and non-oriented electrical steel (NOES) (72%).

With any reshoring plans, it’s going to be a “blend of constraints”, be it skilled labour, manufacturing/processing, critical minerals, or energy.

None of which you can financialise your way out of.

I hope you’re all having a great week.

Cheers,

Ferg

P.S. One of the more striking visuals is how the US defense sector consolidated into a few names while at the same time outsourcing the supply chain to China.

I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Article

Craig has put out three absolute bangers in a row, which I highly recommend you take the 1-2hours to read, starting with The Return of Matter, then Cost of Capital, Industrial Irreversibility and now his most recent piece:

THE HARD BIFURCATION: The Convergence of Financial Liquidity and Physical Insolvency

Please excuse the analogy of a parasite, but it is inescapable. The parasite is within the western system, and until we recognise this, we will forever accept it.

The parasite is structural. It is the class of systems that extract value from the Financial Ledger while imposing costs on the Material Ledger. It consumes throughput without maintaining capacity. It arbitrages price signals while degrading the physical base that makes those signals meaningful.

This parasitic layer rewards asset stripping, offshoring, regulatory arbitrage, and financial engineering. It treats factories, skills, energy systems, and feedstock chains as expendable. Returns are realised immediately. Depletion is deferred. The ledger records profit while the substrate erodes.

Over time, this dynamic hollows the host. Capital appears abundant. Capacity decays. The system survives by liquidating its own foundations. Recognition fails because the Financial Ledger continues to clear. The Material Ledger records loss.

Until this structure is identified and constrained, decline presents as normal operation. The parasite is mistaken for efficiency.

The US’s deeply negative Net International Investment Position is probably the chart which illustrates the above in the clearest way. The US has been running persistent current account deficits, meaning it consumes more than it produces and finances this gap by selling assets to foreigners or borrowing from them.

Podcast/Video

Inside China Business has quickly become one of my favourite YouTube channels (H/t James for putting me on to it!) How Huawei is winning the race for global 5G telecom dominance: US firms didn’t even show up

Quote

“I am often asked ‘how is Europe going to fund the purchases of the ever increasing amount of bunds and joint-EU-bonds they are going to issue?’ The answer is staring us in the face: they will fund it out of the S&P500.”​

-James Aitken (The End of The Beginning)

Or the first of domino, which Kevin Muir made me aware of via this piece on Korea.

Tax Support Measures to Revitalize the Domestic Capital Market and Mitigate Structural Supply-Demand Imbalances in the FX Market

First, temporary tax incentives (for a one-year period) will be granted on capital gains from overseas stocks for individual investors who sell overseas stocks held as of December 23, 2025, convert the proceeds into Korean won, and make long-term investments in domestic equities.

This obviously ties right back to the early US NIIP chart.

Tweet

I pulled up the below chart to check the breakdown, and it’s actually got ridiculous in the past year, as while Education & Healthcare Services and Leisure & Hospitality remained relatively stable at 56-57% during 2023-2024, it surged to 94.5% in 2025!

Charts

What’s a takeaway from Craig Tindale’s work? Well, for me, it’s that Western electric vehicle and wind companies are well and truly screwed. As they will have to compete with the Chinese globally and go from being drowned in subsidies to competing with their own government’s security requirements for the critical minerals they lack, production and processing capacity.

The forecasts of the below OECD countries massively expanding offshore wind additions isn’t happening… Before even considering the effect a surge of inflation will have on the non-existent economics of offshore wind projects.

Something I’m Pondering

The question I’ve got the most recently is, am I buying Canadian oil producers (especially post Venezuela sell-off) and their relative attractiveness to oil services, which are my main expression of the oil trade.

While miners, oil & gas producers, and Canadian oil & gas producers are all dirt cheap and have great outlooks, if we are talking smelly roadkill, then oil services is in a league of its own.

Below is the payoff required to get back to April 2011 (peak commodity boom).

Or take the oil services ETF (XES) vs Canadian Producers ETF (ZEO), with both ETFs being equal-weight.

Canadian producers have crushed oil services performance-wise.

Which is why I find the below divergence so important, as services were tied to oil moves up until August 2025 when this relationship broke down, and services have been on a tear. The majority of investors likely threw in the towel with oil services on liberation day, after round tripping from 2021.

Get out your microscope, as if this were the bottom (which I think it was). We are in for a decade-plus of outperformance. It’s ironic that I can’t even chart the offshore oil services ETF because it got closed down (which is part of my bottoming checklist).

I hope you’re all having a great week!

Cheers,

Ferg

P.S. I have a theory on the debasement of value on social media…

Mia and I discuss this occasionally as she gets daily requests for “collaborations” with the idea being you send influencers free clothes to post to their followers and drive sales.

It’s crazy to me how nonexistent the conversion rate is with “collabs,” even with influencers who have millions of followers, yet a post can’t even sell a dress…

I point this out, as the trend was already clear, likely on the back of low barriers to entry (you just need a smartphone), creating excess supply, i.e. the younger generation all want to be influencers, why work a white collar or god forbid a blue collar job when you can be paid to take selfies..

This “excess supply” is about to get a whole lot worse with AI. Take this Bali entrepreneur who creates AI influencers and grows their followings (Emily Pellegrini, Fiona Pellegrini, and Bailey Moore).

There was a barrier to entry for creating luxury content: you actually had to have the money or access to yachts, private planes, etc., to create it.

Now that what is real is being blurred with AI, so is the ability to monetise it when anyone can magic it up on a computer.

Enough dribble from me, have to get back to waterskiing behind my chopper…

Subscribe now

J.P. Morgan Outlook 2026: Smothering Heights by Michael Cembalest

Cembalest also put out this on his blog, which is worth a read: Ten things about Venezuela: on oil, geopolitics and drugs

Podcast/Video

A quick explainer video on DeepSeek’s new paper, mHC, which was released on the 31st December 2025: DeepSeek Just Crushed Big Tech Again: MHC - Better Way To Do AI

DeepSeek recently published a new AI architecture paper called Manifold-Constrained Hyper-Connections, presented by CEO Wenfeng Liang. The paper focuses on improving how information moves inside large AI models.

For the last decade, all AI models have used a single, narrow “express lane” to pass information between their internal layers. DeepSeek’s new paper, mHC, is a blueprint for turning that single lane into a multi-lane “superhighway”, allowing much more information to flow through the AI at once. In the past, whenever researchers tried to build these wider highways, the AI would become “unstable” and crash because the data would get too chaotic.

DeepSeek solved this by adding mathematical guardrails (the “Manifold” part), limiting how streams are mixed so each layer can only rearrange information, not amplify it.This keeps the model stable across many layers. The result is small but consistent accuracy gains with little added cost, showing that preserving and combining information more cleanly through better internal structure can deliver real advantages without making models bigger or more expensive.

The result is an AI that is significantly more powerful and capable of handling complex details, but costs almost nothing extra to build or run.

In reading about mHCs’ multi-lane superhighway, which allows much more information to flow through the AI at once, I couldn’t help but draw the parallel to “multiplexing”, which allows much more information to flow through fibre-optic cables at once…

For those unfamiliar with multiplexing, this piece touches on it, Boom and Bust in Telecommunications (page 10).

These arguably have been the most impressive advances in telecommunications in recent years. In 1996, a strand of fiber transmitted data at approximately 2.5 gigabits per second (Gbps). By 2000, the capacity of the same fiber could reach 100 Gbps. This increase in capacity resulted from developments in “multiplexing,” the transmission of more than one channel of information over a single medium . Instead of 2.5 Gbps over one wavelength, companies could replicate this flow over 40 wavelengths on the same fiber. Fiber capacity has since increased further, with equipment maker Cisco in July, 2002, claiming a maximum capacity of 320 Gbps over relatively short distances.

Quote

“If you don’t have a significant draw down in your past, it’s in your future.”

-Todd Rapp (CEO Fortress Multi-manager group)

Subscribe now

Tweet

This is wild if the West’s largest technology “moat” has potentially been breached: Exclusive: How China built its ‘Manhattan Project’ to rival the West in AI chips

Charts

A few charts illustrating how stretched things have got.

The average analyst estimate of 47 analysts is that Nvidia doubles its sales from $213B to $407B by FY2028, which is pretty aggressive when you consider the following points;

Jensen Huang says Nvidia went from 95% market share in China to 0%—‘I can’t imagine any policymaker thinking that that’s a good idea’ (October 21, 2025)

As of January 2026 Nvidia is “re-establishing presence” in China….

• The Trump administration has greenlit H200 exports (with their 25% fee).

• Nvidia CFO Colette Kress confirmed that the company has submitted license applications to ship H200 chips to China and is awaiting government approvals​

• Beijing has not approved the license; China’s government has not yet provided formal approval for H200 imports.

• Do analysts really believe Nvidia is going to be allowed to re-capture substantial lost market share?

  • China Tells Companies to Stop Buying Nvidia Chip With AI Uses

  • China bans foreign AI chips from state-funded data centres

  • China formally directs public institutions to only use domestic AI chips

Something I’m Pondering

I’m pondering how developing countries would be crazy not to convert dollar based debt to yuan here.

I mean, simply look at how wide the spread between the Chinese 10-year and the US 10-year has become!

Happy New Year to you all!

Cheers,

Ferg

P.S. I had the pleasure of jumping on Palisades Gold Radio to discuss with Stijn a lot of the points I dug into in the piece below.

Subscribe now

While I referenced this brilliant piece: The Return of Matter: Western Democracies’ Material Impairment in my latest post, I skipped over what I believe is the ultimate scarce asset in skilled labour (in the West that is).

The Human Bottleneck

Real smelters, SX lines, calcining trains, and high‑temperature chemical plants are run by people with tacit knowledge that does not live in textbooks:

• Metallurgists who know how to nurse an unstable furnace back into spec without cracking a refractory lining.

• Process engineers who have personally tuned a 200‑stage SX train rather than just modelling one.

• Maintenance crews who have spent a decade keeping acid plants, off‑gas systems, and high‑vacuum equipment alive in dirty conditions.

Those people are thin on the ground in the US and EU because we paid them to retire or move to other industries. Their apprentices were never hired. Universities still teach metallurgy and chemical engineering, but graduates are pointed at batteries, software‑wrapped “process analytics,” or ESG consulting instead of refineries and smelters. When we talk about “rebuilding the midstream,” we are implicitly assuming a labour force that does not yet exist.

It took a decade for the US to lose six million manufacturing workers, and it will take as long, if not longer, to attract them back, IF the incentives are right (pay a high real wage), before touching on robots, etc.

Yes, historically, the US doubled manufacturing jobs when on a war footing between 1940 and 1945 through the conversion of civilian industries to military production, government incentives, low-interest loans, accelerated depreciation allowances, and access to government-financed facilities.

A crucial factor that enabled this was that the unemployment rate stood at 17.2% in 1939 (the back end of the Great Depression), with approximately 9.5 million Americans still out of work (by 1943, unemployment reached a mere 1.2%). The war economy essentially absorbed the entire Depression-era unemployment pool into manufacturing and related industries. Not exactly the set-up we have today, with 4.6% unemployment, before questioning how many of those are actually fit (obesity, drugs, etc) to work in manufacturing.

Podcast/Video

Luke Gromen: Some thoughts on BTC and silver

Unless the Fed and/or Treasury pre-emptively begin post-COVID-sized liquidity injections, financial conditions in the global economy will likely continue to tighten until a crisis forces said “nuclear printing.”

Quote

Capitalism works not just by allowing upward mobility, but by accelerating its downward equivalent.

-Nassim Taleb

Subscribe now

Tweet

The question for me is straight to nuclear printing, or crisis, then nuclear printing?

Charts

As I was working on my latest piece and creating the chart below, it struck me that there is very little commentary on Chinese control in another form of processing in refining.

While on the surface, China’s 18.5% of global refining doesn’t seem a cause for concern. Once you start to piece together their control via financing, the amount of oil they have “encumbered” via long-term contracts, equity investments, loans-for-oil deals, and active stockpiling paints a different picture (especially when considering the US inability to refine its own light sweet, as 70% of US refining capacity is optimised for heavy, sour crude).

Something I’m Pondering

I’m pondering how tech leadership has quietly passed from US to China.

ASPI’s Critical Technology Tracker: 2025 updates and 10 new technologies

In eight of the 10 newly added technologies, China has a clear lead in its global share of high-impact research output. Four—cloud and edge computing, computer vision, generative AI and grid integration technologies—carry a high technology monopoly risk (TMR) rating, reflecting substantial concentration of expertise within Chinese institutions.

The historical data for these new technologies tells a familiar story: an early and often overwhelming US lead in research output in the opening decade of this millennium, eroded and then outmatched by persistent long-term Chinese investment in fundamental research. In total, China now leads in 66 of the 74 technologies tracked, with the United States leading in the remaining eight—an imbalance that underscores why trusted partners need to act together to leverage comparative advantages, reduce concentration risk and shape the trajectory of critical technologies together.

The ratio of US tech to Chinese tech suggests the market is rapidly repricing this reality, as I’m trying to capture in (Hated Tech, Hated Tech 2.0 and The Coming Asian Boom).

I hope you all had a great Christmas and wish you all a happy New Year!

We managed a little getaway with friends.

Cheers,

Ferg

P.S. While I did my best to explain whats happening this Youtube breakdown on silver is fantastic: PANIC MODE: Shanghai Hits $78.49 — $10 Arbitrage Gap Next (h/t Chris M) and most recently SILVER HITS RECORD $85 IN SHANGHAI — Comex Losing Control

Subscribe now

JP Morgan: The Deep End By Michael Cembalest (December 2nd, 2025).

If you don’t have time to read it (it’s 56 pages), these charts tell the story (there is a lot of “trapped” paper/fantasy gains in PE and VC).

It also referenced this older piece, which I enjoyed:

JP Morgan: It’s Mostly a Paper Moon By Michael Cembalest (December 5, 2023)

Podcast/Video(s)

The above has me thinking about this observation from Mike Green that wealth creation is a mirage if it can’t be liquidated: The ‘everything bubble’ presents a bigger risk than AI: $11 billion strategist | Mike Green Simplify (timestamped).

“Now the question is can they actually get that money out? That’s an entirely different question. So when you again you know the phrase Ponzi is often used right Ponzi’s work great as long as money is being put in... if that’s being pushed higher that creates wealth for you feel richer than you actually are because when the marginal buying shifts to selling there’s really nobody underneath... We’ve gotten rid of all the short sellers. We’ve got rid of the value managers... Passive investing works in both directions.”

As to what to do about it? This was a great conversation on why Steven is getting back into the volatility game (no paywall): Listen: Grant Williams’ Interview with Steve Diggle.

Subscribe now

Quote(s)

“Robots won’t trade well at inflection points. They will be fine at momentum but not at being contrarians. It’s called the pain trade for a reason. Investors need the willingness to be wrong for a while.”

-Daniel Bloomgarden, Head of Research, Multi-Asset Solutions

I also wonder how poor data quality will likely undermine AI investment models when you consider then very first digital financial databases was CRSP in the early 1960s (just stock prices), the first electronic stock market came in 1972 with the NASDAQ, the SEC’s EDGAR started in 1993 while the The Global Emerging Markets Database (GEMS) was only established in 2000.

The majority of the data-rich period has occurred under a single market regime, which exacerbates the issue: developed-market equities and large-cap securities are where data history is richest. In emerging markets, smaller companies, and commodity markets, data constraints could significantly limit AI’s effectiveness.

Tweet/notes

Myrmikan Research: This Time It’s the Same

This was a depressing observation.

Charts

This chart caught my attention, as most are aware of the high decline rates of fracking/tight oil, which represents ~10% of global oil production. Kazakhstan accounted for 41% of global uranium production in 2025 (25,750 tonnes). Now consider their production profile in the chart below.

Something I’m Pondering

Having searched the Fuck you money clip on YouTube, the algorithm suggested a clip of the concept of “fuck you family”. Chris Williamson had outlined the idea in this piece: Freedom, Liberals & Love.

Similarly, “fuck-you freedom” is kind of downstream from fuck-you money, but also can be achieved through cultivating a lack of reliance on other groups.

But I’ve recently learned about another type of fuck-you liberation.

One which is significantly cheaper, more accessible, more common and maybe even more powerful.

The “fuck-you family”.

Many fathers I’ve spoken to have told me about how their priorities were completely changed upon starting a family.

All the previous status games they played seemed petty.

The admiration and gamesmanship they used to play in an attempt to impress people in power or those with status seemed juvenile and shallow.

Much of their anxiety around whether different people liked them or thought they were cool evaporated.

The only people they needed to care about impressing were now asleep in their house.

Maybe it’s just the stage of life I’m in, but this observation hit the nail on the head for me. The only time I feel a pang of envy isn’t seeing the flash car or the big house, it’s when the happy family with four kids sits across from us in our local cafe.

They just make life so fun and fulfilling.

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Article(s)

Two great commentaries well worth your time:

• Horizon Kinetics Q3 2025 Commentary

• Goehring & Rozencwajg: Moving from Carry to Anti Carry

This chart caught my attention as the Permian was the last growth engine of US shale.

Podcast/Video(s)

As a shareholder, I got a lot out of this one: No Spin, Just Tin. Brett Smith on MetalsX and Every Tin Project.

I also enjoyed Michael Cembalest’s insights in this interview: JPMorgan’s Playbook for a 10-15% Correction (or Worse) — ft. Michael Cembalest | Prof G Markets

Subscribe now

Quote(s)

“Historically, the Fed has almost never cut rates when prices paid surveys are going up. Why would they, right?

I mean, you’d be easing into an inflationary environment.

It looks like they’re going to do that, and it’s a huge gamble. The gamble is that the inflationary increase is tariff-related, and once it makes its way through the system, it will not ignite some kind of wage-price spiral, and inflation will be coming down by the spring, and the Fed will be vindicated.

But that’s a big bet, and that’s the bet that you know that we’re looking at as we’re heading into you know the end of the year and early next year, and the markets are pricing in a couple of cuts here, uh, because of that.”

-Michael Cembalest

Tweet/notes

Another piece well worth your time reading, which Louis has kindly put no paywall on: Unraveling China’s Productivity Paradox.

Charts

A few charts from playing around with Perplexities charting ability.

Something I’m Pondering

I’m pondering whether the liquidation of the Global Offshore Oil Services Index ETF marked a bottom for oil services. Even though oil has remained under pressure, the oil services space as a whole has risen over the past few months.

The closing of the KOL ETF was certainly the bottom in the coal market with hindsight.

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Article(s)

Lyall’s knocked out another great piece: Private Equity/Credit: The Bubble and its Implications

Rasmussen’s fabulous words, the whole PE industry is essentially a giant exercise in “volatility laundering”. While low volatility is a desirable trait for risk-averse institutions, if the low volatility is entirely fictional and an artifact of failing to mark-to-market positions, you are receiving no benefit other than the capacity to delude yourself.

Private equity funds in fact own portfolios of small and relatively low-quality businesses that were acquired in a competitive bidding environment, are highly leveraged, and are held in extremely high fee, non-transparent, and illiquid vehicles. The idea that such assets would be lower risk, less volatile, and uncorrelated to listed equities/economic conditions, is patently ridiculous.

Understanding everything Lyalls laid out makes it amusing when I see charts boasting of the super risk-adjusted returns of private equity.

Podcast/Video(s)

Steve Diggle, Founder & CEO, Vulpes Investment Management

This was a great interview, and it’s absurd that fewer than 1,000 people have taken the time to watch it.

Steven touched on a point I’ve only heard Mike Green mention: investors assume they will always have deep liquidity, which is not guaranteed in HFT market-making, and that passive fund flows can reverse, exposing a lack of bid depth.

• Liquidity Cannot Be Assumed: Markets may become discontinuous during stress events.

• Passive Investing as a Potential Amplifier: Passive funds could create “doom loops” during sell-offs when massive redemptions overwhelm natural buyers and amplify market dislocations.

• Fed Put Confidence is Misplaced: After the 2022 inflation scare (9% peak), the Federal Reserve has less room to manoeuvre than markets assume; the Fed is unlikely to rescue speculators unless the economy faces severe problems.

  Subscribe now

Quote(s)

“So you know we we we worry about discontinuous markets and one of the things I would always say to people is you know you can’t guarantee that you’re going to be able to get out of things you don’t want.

So, a good thought experiment to run on your portfolio every now and then is: what if all trading stopped in perpetuity?

Would you like what you have?

If suddenly markets became private and if you hate what you have, then you’re not an investor, you’re a speculator.”

-Steven Diggle

Tweet/notes

Kopernik 3Q 2025 Investor Conference Call

While not a Tweet/note, I’ll put it here because it’s super valuable to me as I check in quarterly on what Kopenik is adding and trimming.

Charts

I find it interesting the conviction investors have in future data center projections and associated trades.

One potential scenario is Josh Wolfes prediction:

I am convinced that 50% of the inference that we do is the punch line will be on the device, and that means that you will not need anywhere near the compute infrastructure that Jensen and Nvidia would like you to

I have no idea if he’s right; all I know is that the majority are acting like the above blue line is a foregone conclusion.

“When everybody thinks alike, everyone is likely to be wrong.”

-Humphrey B. Neill

Talking of which, this is the latest bomb to drop….

Google dropped Gemini 3.0, which was completely trained on its own TPUs, negating the need for Nvidia’s GPUs…

• 4-6× lower hardware acquisition costs through internal manufacturing

• 20-30% lower total cost of ownership from reduced power and cooling

• 30-50% reduction in actual power consumption of the compute hardware (TPUs vs GPUs doing the same work).

• 80% reduction in total electrical infrastructure requirements for the facility (servers + cooling + backup systems combined).

Jensen and Altman are now facing a chronic case of Premature Extrapolation.

Something I’m Pondering

I’m pondering the common theme of a lack of “price discovery” be it in private equity as Lyall detailed, passive flows as Mike Green has documented or even with regards to acknowledging actual inflation/paying real wages as Luke Gromen has noted “US government nor Corporate America has any intention of paying workers a positive real wage v. essentials.”

Yes, inflation will be a little lower than it otherwise would be for a little longer, but it means workers will NOT go to trades in sufficient numbers to reshore US industrials and rebuild the US grid, which means in the not-too-distant future, US domestic inflation rates may SOAR to face-peeling levels as US domestic grid infrastructure and US competitiveness break down.

Suppressing price discovery doesn’t remove the risk, it just hides it out somewhere in the future, aka “volatility laundering”.

It’s the bush fire analogy in attempting to banish fire/volatility; all you do is create a build-up of dry matter/risk, making the eventual fire that much worse.

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Brilliant piece by Louis Gave: The Wrong Question To Ask

In past reports and podcasts, I have argued that access to high-end semiconductors is like driving a Ferrari to work: very nice if you can afford it, but hardly essential when a Toyota gets the job done. That was the lesson of DeepSeek and Qwen (Alibaba’s large language model). China showed it could stay competitive in the artificial intelligence wars without spending fortunes on top-tier chips.

Podcast/Video(s)

All things AI (Brad Gerstner sits down with Satya Nadella and Sam Altman).

While Sam got all the focus for throwing the toys out of the cot when asked how he plans to fund $1.4Trillion of spending commitments with $13b of revenue…

I was more interested in Satya Nadella’s comments here:

The biggest issue we are now having is not a compute glut, but it’s a power and it’s sort of the ability to get the builds done fast enough close to power. So, if you can’t do that, you may actually have a bunch of chips sitting in inventory that I can’t plug in. In fact, that is my problem today, right? It’s not a supply issue of chips. It’s actually the fact that I don’t have warm shells to plug into.

Subscribe now

Quote(s)

“Index funds guarantee exposure to winners — A small number of stocks always drive the majority of returns (currently FANG + Nvidia; historically AOL, Cisco, Microsoft, etc.). Since predicting future winners is nearly impossible, owning an index guarantees you capture them.”

-Morgan Housel

Call me superstitious, but Mr Market is always listening whenever the word “guarantee“ is uttered. Housel’s states his allocation is 80% Vanguard Total Stock Market ETF (100% US stocks) and assuming an even split between his Vanguard Value Index (VTV) and Vanguard Total International Stock (VXUS) for the remaining 20%.

This allocation assumes that the futures’ big winners will be in the US (89.6% US allocation).

What if they aren’t?

Tweet/notes

I get a lot of value out of this Twitter account and its regular updates of Chinese passenger vehicle volumes, both domestic and export. One thing that always jumps out at me is that, besides BYD, the exports are almost entirely ICE and PHEVs.

You see this clearly when you break down light vehicle sales by catalysed and BEV (BEV is 13.6% of global sales). Breaking it down by market, BEVs have had minimal penetration in emerging markets, where all the growth is.

The takeaway is I’m won’t be selling my rhodium, platinum and PGM miners any time soon!

Charts

Mining is hard, and bringing new mines online is taking longer.

It’s the same for conventional oil and gas projects, which have been taking longer and longer.

Lastly how long might it take for the US to add back the manufacturing workers it lost in the 2000-2010 period?

The most rapid industrialisation occurred during the Second World War, with the conversion of civilian industries to military production and a host of government incentives. Post the Great Depression, the unemployment rate stood at 17.2% in 1939, and by 1943, it reached a mere 1.2%. Debt-to-GDP was as low as 44% in 1941 before hitting 119% in 1946 (high unemployment and low debt-to-GDP certainly aren’t where the US finds itself today!).

Something I’m Pondering

I’ve just read the IEA World Energy Outlook 2025 and its good to see they added back in a current policy scenario which has oil demand growing to 113 million barrels by 2050 which works out as 0.47% demand growth per annum.

What’s a reasonable growth rate? OPEC is forecasting demand to hit 123 million barrels per day from 104 million this year, and simply extrapolating the 30-year average lands you at 140 million barrels per day, while excluding China’s growth drops it to 130 million barrels per day.

One thing is clear to me: the IEA will continue revising its oil demand figures after the fact.

It’s great to see journalists point out that while they are calling for a massive surplus, they can’t account for 1.47 million bpd of oil!

Where are the oil barrels? IEA gap deepens confusion over looming glut

MISSING BARRELS

The murky crude picture got even muddier on Tuesday when the IEA report also noted that it was unable to account for 1.47 million bpd of oil in its global balances for August, the equivalent of 1.4% of annual demand. By comparison, the IEA “unaccounted for balance” figure for July was 850,000 bpd, or 370,000 bpd for the second quarter overall.

I wouldn’t worry they’ll eventually find it….

The IEA regularly discovers what analysts call “missing barrels”—oil that was consumed but not initially counted in demand estimates. In May 2025, the IEA made substantial historical revisions:​

-2022 demand was undercounted by 260,000 barrels per day (b/d)

-2023 by 330,000 b/d

-2024 by 350,000 b/d

These cumulative revisions totalled approximately 343 million barrels (nearly three and a half days of global consumption). This pattern transformed what the IEA initially reported as supply surpluses in 2022-2024 into actual deficits, fundamentally altering market perceptions.

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Energy system transformations and energy security by John Kemp

Some fantastic charts that help visualise the fact that Asia has driven ~80% of the increase in global energy use, with Asian energy consumption set to keep growing ~4% a year, while the rest of the world’s energy growth is sub 1% per year.

All energy sources are growing rapidly in Asia, with the big question for me being whether storage can allow solar to keep growing (I plan to dig into this).

Podcast/Video(s)

Morgan Housel - The Art of Spending Money (EP.466)

“All behavior makes sense with enough information.” Your spending habits are rooted in past experiences, emotional wounds, and unconscious programming shaped during childhood. Understanding this provides both self-compassion and empathy for others’ seemingly irrational financial choices.

Subscribe now

Quote(s)

“You can measure everything about a bubble except the most important part: When investors will stop believing in it. The end of the bubble is just the end of enthusiasm. And enthusiasm isn’t a tamable statistic. It’s a hormone that owes nothing to the logic of your data.”

-Morgan Housel

Tweet/notes

The contrarian signals are coming in hard and fast.

Charts

This is why I’ve focused so heavily on Asia recently as it’s in the driver’s seat for global growth and energy demand.

This chart illustrates the point well: the world’s energy consumption, excluding Asia, has been essentially flat over the last decade.

Something I’m Pondering

In my last few “Finds” I’ve pointed out how flawed some economic data is.

• US Manufacturing: How GDP Hides Industrial Decline

• US job openings: The Ghost Job Crisis Exposed

• Oil Demand: Where are the oil barrels? IEA gap deepens confusion over looming glut (hint; missing barrels = Asian consumption growth).

This recent release on the federalreserve.gov website pointed to a whooper of a “discrepancy” in that Cayman Islands as % of foreign holdings of treasuries was 17.6% not the previously thought 4.85% meaning its the largest foreign holder of treasuries and absorbed 37% of net issuance…

The Cross-Border Trail of the Treasury Basis Trade, Accessible Data

Why this matters?

Based on Federal Reserve research and historical hedge fund trading patterns, approximately 65% of Cayman Islands hedge fund Treasury holdings are likely attributable to the basis trade, which is heavily leveraged (typically 10-20x). At the same time, the MOVE index (VIX of Bonds) has been setting new lows after the liberation-day spike.

High leverage with low volatility… what could go wrong?

Cheers,

Ferg

P.S. I wrote this piece to outline a few of the rules I stick to religiously to protect myself from a dose of FOMO…

Subscribe now

Coatue Public Markets Update Deck and Kopernik: Hindsight is 1920 (Sep 2025)

I enjoyed these Coatue slides.

Podcast/Video(s)

I enjoy hearing Erik’s well-thought-out bull case for the markets to temper my own bearishness: The Bull Case: Real Growth Beyond “Project Zimbabwe” | With Erik @YWR

Subscribe now

Quote(s)

“Quick wealth is fragile wealth. I love the idea that the speed in which you made your wealth is the half-life for how fast you can lose it.”

-Morgan Housel (A Few Laws of Getting Rich)

Tweet/notes

“The Flight to Shit” continues….

The brutal reality of a fertility rate of 1.0.

Charts

This trend has my attention, as there is a window to take advantage of it and hedge the portfolio in the process (a topic for my next piece).

Variant Perception has done some great work highlighting this opportunity.

Elevated single-stock volatility is being masked at the index level by unusually low correlations. This creates a fragile set up for potential spikes in index volatility. Correlations are already near historical lows, so there is limited room for further compression. The risk now lies in a rebounds in correlations, which translates into higher volatility at the index level.

-Variant Perception

Something I’m Pondering

I’m pondering how a lot of the US data we have to work with is crap.

Last week touched on this piece: How GDP Hides Industrial Decline

• Arbitrary quality adjustments and input substitutions that make it “a fictitious measure of output with no meaning.”​

• Fail to measure the percentage of components actually made in America versus assembled from foreign parts.

• GDP counts sales receipts, not actual production (BYD produce equivalent cars for half the price of Tesla, $30,000 in U.S. “value-added” might represent one car while the same amount in China represents two cars).

This week it’s Ghost jobs, as I have a friend who has been hunting for a new online job and explained to me how a third of the job postings don’t even exist.

Which isn’t great, given that US job openings have been trending lower before you even take these into account.

1 in 4 Job Listings on LinkedIn Are Likely ‘Ghost Jobs,’ According to a New Report. Here Are the Cities Impacted the Most. or this The Ghost Job Crisis Exposed

While its a global issue the US is the world leader in ghost jobs currently.

The analysis found that 27.4% of all U.S. job listings on LinkedIn are likely ghost jobs with no intentions to hire. At 27.4%, the report found the U.S. had the most ghost jobs of all other countries, compared to Canada (24.9%), the U.K. (14.2%), and Australia (10.9%).

Why is it happening?

The simple answer seems to be that it’s low-cost research, and there is nothing to stop companies from doing it.

Building “talent databases” and salary and competitive research

• Collecting resumes for positions that might exist someday, possibly, if budget approvals come through and hiring freezes lift.

• Ghost jobs function as free market research, allowing companies to understand salary expectations, skill availability, and competitor hiring patterns without conducting expensive studies.

According to survey data, 58% of companies that post ghost jobs do so to collect resumes to keep on file for future hiring needs. This practice allows companies to build talent databases at minimal cost while keeping candidates engaged with false hope.

Takeaway?

I’ve observed in a few pieces that you can get away with a lot of “pretend” with easy money and low inflation (would any of this exist with structurally higher inflation?)

Pretend jobs, pretend job openings, pretending US hasn’t hollowed out its manufacturing base, IEA pretending oil demand isn’t growing, pretending the world can run on wind and solar, pretend science, pretending financial engineering can replace actual engineering….

Own scarce real assets that stand to benefit from an Asian Boom.. (I’ve brought a new company in a new industry this week).

Cheers,

Ferg

P.S Loved jumping back on Money of Mine for a chat with Trav and JD.

Subscribe now

Lyall Taylor’s latest piece just dropped, and it doesn’t disappoint: Artificial Intelligence or Authentic Stupidity?

Podcast/Video(s)

These were taken from Lyall’s above piece, and admittedly, I didn’t actually watch these so much as run them through Perplexity Pro, requesting 10 bullet points with timestamps (then listen/watch if the bullet point is intriguing).

• Richard Sutton – Father of RL thinks LLMs are a dead end

• NVIDIA: OpenAI, Future of Compute, and the American Dream | Brad and Clark Tang sit down with Jensen Huang

Subscribe now

Quote(s)

“One of the main causes of that decline is the onerous domestic-build requirement of the Jones Act, which prohibits U.S. shippers from operating vessels constructed abroad. American-built coastal and feeder ships cost between $190 and $250 million, whereas the cost to build a similar vessel in a foreign shipyard is about $30 million.^”

-The Jones Act: A Burden America Can No Longer Bear

This ties to a piece in the Charts section.

Tweet/notes

I plan to dig into solar in a piece soon.

This Tweet touches on an important point, which is rarely discussed: the spread between solar generation and final energy consumption is blowing out (you can only install so much before additional MW are worthless unless battery storage is economic).

Charts

This piece shouldn’t really come as a surprise in much the same way as US inflation data is manipulated:

How GDP Hides Industrial Decline

Quick Summary

• America Has Lost Manufacturing Dominance Across Critical Industries - Real production data shows devastating declines: U.S. semiconductor market share fell from 37% to 12%, solar panels from 50% to near zero, shipbuilding from 5% to 0.2% of global tonnage, and per-capita manufacturing output is down across wood products (-9%), machinery (-27%), appliances (-28%), furniture (-45%), and textiles (-71%).​

• Nominal Comparisons Between Countries Are Equally Misleading - When comparing U.S. vs. Chinese manufacturing, GDP counts sales receipts, not actual production. Since Chinese companies like BYD produce equivalent cars for half the price of Tesla, $30,000 in U.S. “value-added” might represent one car while the same amount in China represents two cars, meaning China’s actual output could be 4-8 times greater than GDP suggests.​

• Current Manufacturing Statistics Don’t Track What Actually Matters - Existing metrics fail to measure the percentage of components actually made in America versus assembled from foreign parts. A Dell server “assembled” in North Carolina with Korean, Chinese, and Taiwanese components counts as 100% American output, masking the true extent of industrial hollowing-out.​

• GDP’s “Real Value-Added” Manufacturing Data is Fundamentally Broken - The most commonly cited statistic showing U.S. manufacturing up 71% since 1997 produces nonsensical results where steel tonnage is down 18%, but “value-added” is up 125%. The metric creates wild distortions through arbitrary quality adjustments and input substitutions that make it “a fictitious measure of output with no meaning.”​

• Quality Adjustments Turn Subjective Bureaucratic Decisions into Fake Objectivity - GDP calculations embed hundreds of arbitrary decisions by government economists (like claiming a 25% car quality increase equals 166% GDP growth) but present these as objective data. When methodologies change mid-stream because numbers “look wrong,” it reveals these are subjective judgments disguised as hard statistics.

Something I’m Pondering

This is a big deal: BHP Billiton Shifts to RMB Iron Ore Settlement with China.

I touched on how important it is in The Coming Asian Boom.

The removal of the US dollar handbrake

The other side of the US running large deficits and over-consuming has been Asian countries running surpluses to ensure they have the US dollars to buy the commodities/energy they need.

Asia’s oil import dependency is ~81% so increasingly being able to purchase their energy needs in their own currency takes the handbrake off their ability to grow more rapidly (boosting consumption and trade with each other).

Other major miners will have to follow suit or risk losing market share.

Trav from Money of Mine did a great job breaking it down here on Twitter, and I highly recommend taking the time to read it.

Cheers,

Ferg

P.S I was surprised to see rhodium ranked as by far the highest US Supply Disruption Risk.

For anyone wanting to get up to speed on the rhodium trade I’ve removed the paywall for The “New Energy Vehicles” Pivot piece which deep dived rhodium and how I’m positioning in it.

Subscribe now

Something everyone should have on their radar: SPVs, Credit, and AI Datacenters

By structuring it this way, via special purpose vehicles (SPVs) in which they have joint ownership, companies like Meta don’t have to show the debt as their debt. It is the debt of those guys over there, that SPV. Not us. Granted, they retain shared control, and they get to use the AI data center, and nothing there happens without their say-so, but still. It’s not ours.

This is accounting trickery, of course. It is a transparent attempt to raise large amounts of money without balance sheet damage by putting the debt in a vehicle you indirectly control, but that, for accounting reasons, doesn’t have to be disclosed as your debt on your balance sheet. The accounting term of art is “control without consolidation”.

Podcast/Video(s)

Plain English with Derek Thompson: This Is How the AI Bubble Could Burst

Subscribe now

Quote(s)

“The big insight in AI that I’ll share is that I am convinced that 50% of the inference that we do is the punch line will be on device and that that means that you will not need anywhere near the compute infrastructure that Jensen and Nvidia would like you to believe. So even though I was super bullish on Nvidia going back to 2016, today he would like you to believe that you need hundreds of thousands of H100 clusters for training and for inference. You do need it for training, but the models particularly even open source models today are getting so good that the real value is in the repository of data that somebody might have longitudinally. Bloomberg with their financial data, a pharma company with their published and unpublished scientific results, Facebook Meta with all of your Instagram likes and Elon with all of your tweets.”

-Josh Wolfe

Could this create an outcome similar to multiplexing with fibre optics during the Dotcom boom?

Apple’s hesitancy to join the AI CAPEX in size, combined with the paper they released (LLM in a Flash: Efficient Large Language Model Inference with Limited Memory) supports this hypothesis.

Tweet/notes

In trying to find the source of the below chart, I found Michael Cembalest’s recent piece, which is a great read: The Blob: Capital, China, Chips, Chicago and Chilliwack.

Charts

The below report is helpful for anyone trying to get up to speed on robotics trends.

International Federation of Robotics: World Robotics 2024

US industrial robotic installations have been fairly anaemic and heavily concentrated in automotive (where they are being eaten alive by Chinese competition globally).

China, on the other hand, continues to ramp up industrial robot installations rapidly, with domestic suppliers scaling production rapidly. Robotic installations are also far more broad-based in China, with electronics being the lead customer.

This chart illustrates how rapidly China has evolved in the global robot sector.

Something I’m Pondering

I’m pondering if I have enough exposure to Chinese tech.

Yes, I ramped my exposure when I dug into it in Hated Tech and Hated Tech 2.0, but I’m still feeling underweight after reading this piece: The Industrial Explosion, with the below two charts in mind.

In reading the above, I’m reminded of this observation from The Electric Slide.

In economic terms, energy, intelligence, and action are strong complements in the production of anything.

And in the immortal words of Joel Spolsky, “Smart companies try to commoditize their products’ complements.”

America is, implicitly or explicitly, making a bet that whoever wins intelligence, in the form of AI, wins the future.

China is making a different bet: that for intelligence to truly matter, it needs energy and action.

If you control energy and action, making intelligence abundant strengthens your position.

China has all the ingredients to massively benefit from commoditised AI.

The US lacks these ingredients with its hollowed-out manufacturing base and aging grid. If anything, the risks are to the downside due to AI’s impact on the US’s service-heavy economy/employment.

Cheers,

Ferg

P.S. I recently caught up with Lucijan at Triangle investor for a chat.

Lastly, after last week’s tin special, living in Indonesia and knowing how things work, I’m a bit peeved I didn’t see this one coming… Prabowo oversees US$427 million confiscated asset handover to PT Timah.

Subscribe now