Update on Iran's attack tempo:
Its attack tempo noticeably slowed down earlier this wk as it shifted to longer range & more powerful BM while reducing attacks on GCC countries, especially UAE.

But that look to have changed last 2 days as Saudis & Jordan are getting attention

A simple look at the US tanker fleet action would point to the importance of Saudi Arabia, Jordan & Iraq toward continually supporting Coalition strikes on Iran.

As such, Iran has dialed up attacks on US bases & radar/command sites in this line.

Impact of Sulfur supply interruption will hit Sulfuric Acid production, which is needed for phosphate (fertilizer & LFP battery), chemicals, metal processing (leaching for Cu, Ni, Co, U, Zn, Fe & Ag), refining & more.

Thread on Sulfur industry, China's role & Hormuz risks👇

https://twitter.com/tphuang/status/2028910743404593464

Most common use is fertilizers.
Sulfur is needed to produce Sulfuric Acid.
Sulfuric Acid + Phosphate rock is need to produce phosphoric Acid.
Which is the main feedstock for several types of phosphate fertilizers.
Phosphate demand recently have jumped due to LFP batteries.

In 2025, China continued to import large qty of Rare Earth in order to produce RE alloys & magnets.

RE Oxide in 2025 is up 15% YoY, which has made up for decline of 56% on RE concentrate & 28% in RE compounds. Overall tonnage import remains flat while value increased by 12%, since Oxide are more processed & are higher valued.

China also imports small qty of magnets. Big change in import this yr is caused by no longer getting Rare earth ore from US (for obvious reasons), but more RE Oxide from Myanmar & ore from Laos.

56% of import from Myanmar, 25% from Laos & 19% from US. I'd expect the US import to drop to ~0 in 2026.

A thread on China's auto export for 2025:
7.08m passenger vehicles: 3.26m ICE, 2.21m BEV, 1.06 PHEV & 506k hybrids
1.065m trucks: 681k diesel, 220k gasoline, 85k BEV & 53k PHEV
106k coaches/buses: 64k gasoline, 24k diesel & 17k BEV
Overall, this was 30% growth vs 6.41m in 2024. Overall, the goal of China's auto industry is to surpass Japan. Composition of China's oversea sales need to shift from export to local production.

See below where Japan Inc sold 24.62m overseas in 2018, but 19.8m (80.4%) was produced in oversea factories. Big shift from 1985 when just 15% was produced overseas.

In China, provinces make their own yearly dev plans. I was really impressed in Jan, when I read Inner Mongolia's 2026 plan & here is a look:

It wants to be China's main strategic energy base & produce 1.2B ton of coal, 3.1mt of crude & 31 BCM of NG every yr.

It wants to build 500kV grid along major centers w/ 15 projects. For renewables, it plans to add 40GW to the grid for 170GW in cumulative installation & generate 300 TWh of electricity.

Plans to expand the major renewable bases along desert regions & deliver to other provinces.

Also build 0-carbon industrial parks, lithium industry & modernize coal-to-chemical plants to integrate green energy (like methanol).

Greatly build up the usage of RE magnet in NEVs & wind turbines.

China's wind industry had a banner yr in 2025 w/ 119GW of installation, which was up big time from 37GW in 2022 & 79GW in 2024. Cumulative installation thru 2025 is now up to 640GW.

1130 TWh electricity was generated from wind in 2025. This likely goes up quite a bit in 2026 as most installed capacity was toward end of the yr. See below for wind installations in 2024 (yellow) & 2025 (green).

There was a bump in May due to reforms starting in June, but installation exploded in Dec w/ 36.8GW of installation just in that month.

Bid result would indicate 2026 installation should surpass this total.

Having received electricity generation data from both CEC & NEA for 2025, I used Kimi 2.5 to analyze trend since 2015.

Coal portion has decreased to 51% whereas non-thermal increased to 43% as of 2024. The change has accelerated in past 2 yrs due to solar/wind expansion. Structure of thermal power generation looks like the following. 82.5% of generators are coal, followed by NG, biomass & others (like waste-to-energy plant)

Other thermal generates ~6% of power.
NG capacity has grown the fastest here, but biomass also growing faster than coal.

China's economy will change bigly over the next 20 yrs due to shift in its labor force. Here is Kimi generated labor force projection by 2030 showing work age population drop by 41m over next 5 yrs.

Not only that, 70% of ppl entering work force want to do white collar jobs. So economy needs to create more white collar jobs while dramatically reduce blue collar jobs.

https://x.com/tphuang/status/2020840913992364059

In 2025, China's service sector export had 14.2% growth while import had just 2.5% growth.

Trade deficit shrunk to just 828B RMB. Knowledge Intensive service (like s/w stuff) export grew by 37.8% YoY & surplus here reached 576B RMB. It is 38% of China's overall service export.

Too often, online AI “experts” make generalizations about Chinese AI labs when they each have their own niche. Kimi is not like GLM or Minimax. It is in fact going for large models that excel in deep research & agentic use.

But don’t expect Kimi to excel in multi modal stuff.

https://twitter.com/yifan_zhang_/status/2023085177086316681

Kimi 2.5 popularity has already made it the most popular model on OpenRouter & for OpenClaw users.

Now it is stepping up customer capture & user stickiness w/ Kimi Claw. It has to keep close to SOTA in these critical areas for customer stickiness

https://x.com/kimi_moonshot/status/2023029674549596301?s=46

Long term, ByteDance & Doubao team is always going to have the most cutting edge model from China since it has the most resources & data. Its also not constrained in size of model, so Doubao 2.0 Pro is going to be better than the open src labs, who have to release smaller models. ByteDance/Seed team looked at the MaaS demand & targeted those areas in its model.

Note the corporate demand in processing knowledge content with mixed charts, documents, and other unstructured information.

BD is battling Alibaba & Tencent for business users.

So, we finally have 09V, 110m long, 12-13m beam, likely single/hybrid, w/ X tail & Pumpjet.

Essentially, Chinese seawolf
1st SSN designed after China had reached competence in nuclear supply chain, high end steel, precision mfg & sensors -> finally allowing more the large sub that PLAN always wanted.

USS Jimmy Carter & Virginia class Block 4 are all ~140m long, so they could presumably extend this one that long to accommodate a larger mission bay.

That doesn't need to happen for a while. The lead boat will need to go through sea trials for a while & enter service by around 2030 (if we use the initial 093 timeline). Likely pause for a couple of yrs to fix problems found in testing b4 mass production really ramps up. They will likely build a lot more 093Bs b4 then.

Given that SSN quietness is essentially determined by noise produced by the machines inside the submarine & the passive noise absorbers all around (which itself is determined by interior space). There is no reason to think that China's mechE & material science tech is inferior to what USN had access to back in 1995, I see no reason why this sub would not be at least as quiet as Seawolf class.

ofc, at certain pt, SSNs essentially are as quiet as surrounding waters & then the subs effectiveness can then be determined by factors like its advanced sensors, compute server farm, mission payload and how well it can conceal communication w/ aerial assets. This sub + follow-on 09VI will significant shift global balance of power, but especially in Pacific.

RAN + AUKUS supporters should really think about the implication here.

Seawolf vs Virginia sized sub

https://x.com/RickJoe_PLA/status/2022119092564762910

Every country should strive for energy independence & not depend on hydrocarbon powers.

Reminder that China is world's 4th largest NG producer & has world's largest shale reserves. Production cost of shale is 20% cheaper than pipeline gas.

China doesn't need Russian oil/gas.

https://twitter.com/Tendar/status/2020432259724058940

China also does not depend on Russian pipeline since it gets more gas that way thru Central Asian republics + 12 bcm from Myanmar.

China's increased NG production means it has reached 61% self sufficiency while LNG demand has collapsed. Pipeline import increased due to lower px.

MIIT filings had some interesting new EVs:
1st, we have Tang 9 EV here
Comes in AWD 215+370kW or 300/370 RWD
240 to 250 km/h top speed!
4.9s from 0 to 100
Dimension 5263/1999/1790mm wheelbase 3130, seats 7
BYD going full electric on full sized SUV is big deal
-> battery energy density/cost & charging speed are now good enough to allow for this
BYD used to only do PHEVs for large SUVs. Next is BEV version of the popular Tai-7 Box SUV.
Dimension 5050/1995/1865mm wheelbase 2920
Comes in AWD 215+300 kW or RWD 300 kW

Another medium-large SUV getting BEV version.
Sign of maturing super ePlatform & battery tech.

China's import, export, consumption & production data of Gold from 2025.

932.5t imported in 2025, down 448t YoY, but import bill of 726B RMB was just 0.6% behind 2024.

In fact, Gold was the 4th largest import item in 2025 behind Semi, Oil & Iron Ore.

Gold import masked the size of China's trade surplus in 2025, imagine that. Even w/ China settling some surpluses thru gold, it still ran a $1.2T surplus. This happened while Gold was the largest export item out of America in Oct & Nov, even more than aircraft & oil.

Picture of America sending gold bars to China while getting $ back has never been so accurate.

Much underappreciated in China's energy transition/security drive is its NG plan.

Deep concerted effort in both driving up domestic production & pipeline import to lower import of expensive & geopolitically risky LNG.

Self sufficiency rate now above 60% while pipeline may have surpassed LNG in tonnage this yr.

https://x.com/tphuang/status/2017631911989154195

Important pts from Nikkei article below:
Domestic shale is 50% than imported LNG & 20% cheaper than even Russians pipeline gas.

LNG import fell 11% to 68.43mt, while pipeline has increased -> led to import bill dropping by 12.5% in $ amt, even tho qty just dropped by 2.8%

CCTV reports on exploding order backlog for UHV large capacity transformers from domestic & foreign customers, including North American ones, especially from AI customers.

There are ~3000 transformer companies in China, mostly in Guangdong & Jiangsu. 60% of global capacity.

A Jiangsu factory has order book for until until of 2027. China's domestic transformer mkt grew by > 20% YoY w/ AI computing & UHV transformer orders growing 35% YoY.

Delivery time for US mkt has grown from 50 to 127 wk.
EDWC project is generating huge domestic demand.

My thread on China's silver trade & what's happened in the past 2 months has silver px skyrocketed in China.

1st, Silver has huge & growing industrial demand, especially in the age of PVs, EVs & AI. Consumer demand is sky-rocketing in China -> Why China put in export controls.
75% of silver is by product of mining other metals, which can be separated & processed into Silver concentrate b4 being smelt & refined to 99.99% grade

China produced 3300t of silver concentrate in 2024, just 13% of global production, so it has to import massive amount of Ag to feed smelters. Not so different from other metals.

It’s unclear to me how stringent this will be enforced since Japan is a major intermediate player in supply chain. Does this mean Ga won’t be exported to Japan?

This could affect far more than just Japan. Also, is China ready for possible Japanese retaliation? Dual-use list can be found here
Includes Ga, Ge, Sb, Scandium, yttrium, Lutetium, dysprosium, terbium, gadolinium, samarium & whole bunch of chemicals.
+ upstream material like UHMWPE, graphite, high end Al/Ti alloy, complex fibersxkzj.mofcom.gov.cn/tzgg/art/2026/…

CXMT has filed updated 2025 guidance as part of its IPO filings. It had generated revenues of 8.3B, 9.1B, 15.4B RMB from 2022-2024 & 24.2B in 2025H1.

Now, it expects 55-58B for 2025 & turning profit for the 1st time.
Likely up to $3B in rev in Q4, making it bigger than SMIC. CXMT has 3 12-inch fabs (~20k wpm each) in Hefei & Beijing. It plans to raise 29.5B RMB off IPO out of 34.5B it needs for DRAM tech upgrade, production line upgrade & R&D.

CXMT invested 15.2B in R&D from 2022-2024, 36.6% of its revenue. It seeks to keep that at ~30% going fwd.

ByteDance's Doubao/Seed team has unveiled Seed1.8.
While it is a huge improvement over 1.5, it does noticeably trail Gemini-3 & GPT-5 on most of the LLM benchmarks.

They do seem to focus on Search & "Economically Valuable Fields", where they are SOTA based on benchmarks.


Among other benchmarks, they do quite well on the Visual & Video generation as well as VLM (not a surprise if you've seen their rankings on artificial analysis for this), but not so competitive on coding.

-> Their unique data set from apps are a huge advantage in video & visual

As Gemini caught up to GPT, Google AI Ecosystem is getting a huge boost. What if as Marc says, AI is commodified? As in, OpenAI cannot maintain indefinite lead & command high premium vs Google & Anthropic?

Google can undercut OpenAI/Anthropic until those 2 go out of business.
Using IBM chief Krishna's comment, each GW of DC cost $80B, which means 100GW of DC requires $8T of capex + opex (like electricity cost, taxes & staff) + interest. Interest alone is $800B
5 yr cycle -> $1.6T depreciation/yr.
+ opex & engineers -> over $3T in cost per yr for AI.