Paul Kennedy’s book The Rise and Fall of the Great Powers (1987) examines the interplay between economic strength, military power, and the fate of great powers from 1500 to the late 20th century.

1. Economic Strength as the Foundation of Power: Kennedy argues that a nation’s military and political dominance is rooted in its economic resources. Great powers rise when their economies are strong enough to support military capabilities and global influence, but overextension – spending beyond economic means – leads to decline.

2. Imperial Overstretch: A central concept in the book, imperial overstretch occurs when a great power’s global commitments exceed its economic capacity to sustain them. This imbalance strains resources, weakens the economy, and accelerates decline, as seen in historical examples like the Spanish Empire and Britain.

3. Relative Decline Due to Uneven Growth: Kennedy emphasises that the rise and fall of powers are relative, driven by uneven economic growth rates among nations.  As other states industrialize or innovate faster, a leading power’s relative strength diminishes, even if its absolute capabilities remain intact.

These insights highlight the delicate balance between economic vitality, military ambition, and strategic restraint in maintaining global dominance.

Reserve Currency

China and Russia have been trying to establish alternative reserve currencies to the USD but they have not been able to insofar as their own currencies are concerned.  The increasingly broad ranges of U.S. sanctions of Russian and then Chinese companies due to the Ukraine war and trade and tech war have shifted central bank holdings away from USD into hedging currencies like the AUD, Canadian CAD and Swiss CHF.  China itself has over the last two years moved some of its U.S. Treasuries holdings to Chinese CNY (RMB) assets.  There have been specific instances whereby Beijing issued loans or were trying to pay for commodities in RMB.  But beyond these mutually required deals, general transactions in world trade remain wedded to the USD or Euro.   

The stronger USD injects some dynamic portfolio management factor into central bank forex holdings but in value terms, the decline in USD share of total holdings is marginal, reflecting little movement in nominal USD shares. The U.S. economy relies less on international trade than the PRC economy.  The USD is used most of the time to underwrite global trade.  It is undisputedly the world reserve currency because it has the deepest capital market and U.S. Navy backing it up.  It doesn’t need gold or any precious metal or asset to support its intrinsic value.  The U.S. economy itself is the collateral, and this view is borne out in the historical records of international foreign exchange trading by currency.

The forex figures show steady proportions of total turnover for the listed top traded currencies in April 2022, except for the Chinese CNY, which saw a rise in single currency trades from 4.3% in 2019 to 7.0% in 2022.  In currency pairs, the CNY also saw a rise in USD/CNY trades rising from 4.7% in 2019 to 6.6% in 2022.  EME is Emerging Market Economies currency, which represents the total of 26 most traded currencies excluding the listed currencies in the table plus the Russian rouble. The table confirms the reserve currency status of the USD, rising slightly from 88.3% in 2019 to 88.5% in 2022 as the most traded currency.  Overseas holdings of U.S. Treasuries and other assets remain robust.  Even China is keeping around 55-60% of its total foreign reserves of USD 3.28 trillion in U.S. assets.  This is despite reported selling in the last year due to the trade war.  As of January 2025, China’s holdings of U.S. Treasuries were reported at USD 761 billion, a decrease from USD 859 billion in January 2023.  By Feb 2025, China’s holdings slightly increased to USD 784 billion.  While the exact figure for sales in the year to July 2025 is not fully detailed in data sources, the trend shows China has offloaded a substantial amount, likely in the range of USD50-100 billion over the year, based on the decline from March 2023 to January 2025.  This decline is not out of the ordinary given China’s need to defend the RMB over this period due to the impact of poor economic data, from sharp rises in public sector debt (albeit at local government level) to the collapse of the top tier of the real estate sector, to closure of large numbers of manufacturing facilities – ie all the signs of recession, although official data continued to (unsurprisingly) record strong GDP growth.

China’s U.S. Treasury holdings represent only a portion of its USD-based assets, with additional investments in Agency bonds such as Fannie Mae, Freddie Mac and other U.S. securities.  The exact composition of China’s reserves is not transparent, and some holdings may be obscured through offshore custodians in places like Belgium or Luxembourg.  Further, China’s total U.S. holdings (including non-Treasury assets) could be as high as USD 2 trillion, which would imply about 60% of its reserves are in U.S. assets.  This figure is speculative but shouldn’t be too far off the true extent, especially when compared against the approx. USD 1.1 trillion of Japanese holdings of USD Treasuries (largest in the world), out of a total of USD 1.3 trillion in foreign reserves managed by Bank of Japan.  This shows the depth of the U.S. capital market and the confidence other governments have in the U.S. economy and system.

Technology

The civilisational Warm War is being fought on two fronts in one – kinetic preparation and AI, wrapped up in the battle for nanometre chips.

Kinetic preparations involved all the aspects of our society, economy and military to provide the wherewithal to fight the monolithist bloc and win.  AI, enabled by the nanometre chips, underwrites that potential victory.  The U.S. controls the majority of well-known AI training compute power on this planet and continues to build the biggest, most power-hungry clusters.  China is spending heavily to close the gap.  Recent reporting pegs 2025 AI capital expenditure in China at USD98 billion, up 48% from 2024, with about USD56 billion from government programs and about USD24 billion from major internet firms.  Capacity will grow, but translating capex into competitive training compute takes time, especially under export controls.

Microchips manufacturing process is seriously high tech.  Netherlands based Advanced Semiconductor Material Lithography (ASML) Holding NV is the only company in the world that supplies extreme ultraviolet (EUV) lithography machines that chip foundries use to make wafers to produce custom designed semiconductor chips for their own customers.  To make EUV machines, ASML needs a special laser from American company Cymer, which is the sole supplier of that product.  ASML also needs a special glass from German company Zeiss, which is the sole supplier of the glass.  Since 2019, the U.S. has managed to convince the Netherlands government to stop ASML’s supply of top-end EUV machines to China.  Shortly after, ASML’s less leading-edge deep ultraviolet (DUV) machines were also restricted from being exported to the PRC.  This means China would need to buy microchips from 7-nanometre (NM) to smaller from U.S. companies like NVIDIA.  These chip suppliers design the chips that are made by foundry companies like Taiwan-based Taiwan Semiconductor Manufacturing Company Ltd (TSMC), one of ASML’s top chip foundry customers, the others being Samsung and Intel.

For DUV machines, ASML dominates the market but competes with Nikon and Canon, particularly in immersion and dry lithography segments.  These Japanese companies, which cater for the 7+NM superconductor market, are also cutting back supplies to China.  A key input to the lithography process is photoresist, which to lithography is like cartridges to printers.  Canon and Nikon produce 70% of the photoresist market and supply China on short term contracts.  As photoresist has a short life – 3 months or less – they need to be supplied regularly.  China imports 90% of the photoresist it uses, two-thirds of that coming from 4 companies in Japan.  And despite China achieving some low-yield mass production of 14+NM wafers, silicon wafer supplies are still controlled by Japan, especially the sub-14NM classes.

Both Samsung Electronics (a leading foundry and memory chipmaker) and Intel Corporation (a major CPU and fab operator) are key customers of ASML, purchasing both EUV and DUV systems to support their fabrication processes. These purchases are ongoing, with recent deals focusing on next-generation High Numerical Aperture (High-NA) EUV tools for sub-5NM nodes.  Samsung invested USD1 billion in 2012 for an EUV development stake and in 2018 secured three dozen EUV machines for 2022-24 delivery.  In 2025, it bought two High-NA EUV (EXE:5200B) for USD773 million, for 2NM logic/memory production such as Exynos 2600 and AI chips.  Intel was the first to buy High-NA EUV prototype in 2018, secured all 5-6 units for 2024 for USD2 billion, ordered additional two in 2025 for 14A node and deployed in 2019 the first standard EUV for 7NM.

To produce 5NM and smaller chips, TSMC must buy EUV machines from ASML, make the wafers to sell to NVDIA and other customers to make the chips.  Up until the latest U.S. ban, NVIDIA was holding the lion’s share of the chips market in China.  After the ban, China reacted by cancelling all business dealings with American chip suppliers and using only China made chips.  China’s equivalent to TSMC is SMIC, which is taking delivery from China’s equivalent of ASML which is ASML Shanghai (ASMLS), which has been using EUV and DUV machines bought previously from ASML to develop its own lithography technology.  While steps behind ASML and TSMC, the Chinese companies have been able to produce their own 7+NM chips using previously purchased ASML machines.

The U.S. strategy is to create a vertically integrated supply chain in the U.S. at foundry level.  It convinced the Taiwan government to allow TSMC to establish foundries in Arizona capable of producing 4NM and 2NM chips.  TSMC has completed a semiconductor fabrication plant in Arizona that has started to mass-produce 4NM chips.  Its follow-up fab on the same site is to produce 2NM chips.  The U.S. is pulling ahead while the upheaval in the chip supply market is still being sorted.   While ASML holds worldwide monopoly position for EUV machines and 90% of the DUV market, TSMC holds two-thirds of the world’s foundry market, manufactures probably 90% of 10+NM chips and almost all the most advanced chips that the top American high-tech companies design.  TSMC supplied three-quarters of China’s own semiconductor needs for its electronics industries before the U.S. bans.  The company’s first foray into the U.S. manufacturing landscape was in 2020 when it announced a USD12 billion fab plant in Arizona.  This was in response to U.S. nudging and the pandemic’s impact that raised U.S. concerns over chips supply security.  Then, following the U.S. Chips and Science Act of 2022 offering $53 billion in industry assistance, TSMC upped its investment to USD40 billion with promise of a second fab, which would produce 3NM chips in addition to the 4NM chips from the first fab.  On the heel of mass production being achieved by the first fab in 2024, TSMC raised the stakes again to USD65 billion with planning for a third fab, and then in 2025 unveiled the full grand plan of USD165 billion that transited the individual plant projects to an end-to-end semiconductor eco-system, from wafer fabrication to chip packaging and research, that comprises six fabs and five dozen packaging facilities with a R&D centre by 2028.

Taipei was keeping TSMC’s 2NM chip manufacturing capability in Taiwan initially as an insurance policy for U.S. protection.  But it is shifting towards making such level chips in Arizona and accelerating TSMC’s R&D push to sub-2NM chips.  TSMC’s fab in Taiwan is already moving to producing 1.6NM chips.  It is increasingly clear that the U.S.’s Taiwan strategy is rushing towards meeting Taiwan’s U.S. strategy of tying up the two countries by the chip umbilical cord.  It is not out of bound to suggest that for all intent and purposes Taiwan is now an indispensable link in the U.S. economic supply chain.

The chip ecosystem in the U.S. is not confined to Arizona.  Samsung is completing a USD50 billion semiconductor fab in Texas with the view to supplying top end customers there, one of them being Tesla. The Samsung Texas saga provides insight into the difficulties of setting up chip foundries.  Such scale investment usually requires a cornerstone customer to begin with.  TSMC did a deal with Apple to supply the latter exclusively for the first year of production.  This guarantees Apple’s next generation i-phones a temporary monopoly position.  Samsung approved the Texas fab without a foundation customer.  The construction also became an engineering nightmare due to various characteristics of the location not being perfectly aligned with a semiconductor fab’s requirements.  For that reason, hats off to the way Samsung has spent money and engineering, procurement and construction time and expertise resolving the issues it encountered.  After a few years delay, the company has recently inked a contract with Tesla to supply chips for Tesla’s full-self-driving mode on some exclusive basis.   

As chip transistors reach atomic size, the cost of pushing the frontier mounts exponentially.  Currently, only TSMC has attained 2NM manufacturing capability, with Samsung behind at 5NM and Intel at 7NM.  Other foundries around the world are Global Foundry with 12NM, UMC with 14NM, PSMC 16NM, DB Hi-Tech 90NM, Vanguard/Intel 90NM, Tower Semiconductor at 130NM.  The chip sizes provide buyers with the ability to design and manufacture their own electronics products.

Lacking the deep private capital market that the U.S. commands, Beijing has been pouring massive amounts of money into shoring up China’s chip making capacity by using Huawei as buyer and SMIC as manufacturer.  China’s ASML Shanghai Microelectronics Equipment focuses on building lithography machines, being able to develop 28NM DUV machines with the target being closer to 7NM DUV systems for SMIC to use.  It is reported that a version of the 7NM chip is already in production and used in Huawei’s latest smart phone.  Market estimates are that by 2035 China could be competing directly with TSMC.

NVIDIA has been approved by DC to export its H200 chips to China with the U.S. government taking a 25% clip on the price.  This followed lamentation by Jensen Huang that he couldn’t understand how policy makers could think it’s a good idea losing the China market.  He was pincered between Washington’s export constraint and Beijing’s retaliation on the company’s “potential breach of market competition policy” in China.  In this regard, Huang repeated Cook’s and Musk’s and many other high-end corporate leaders’ views that losing China is not in the interests of America.  Yet, there have been no comments on the fact that Beijing has played a fierce game at keeping Western businesses out of China while taking their IP and passing it on to domestic companies.  Nor do they admit that sales success of their high-end products means China has been getting more efficient more quickly at building its military and logistics around the world as well as oppression systems at home.  If the U.S. passes on top-end techs to China, after helping China master low- and mid-end techs to corner the world’s industrial and manufacturing base, then the obvious question for this one-way trading regime is, what’s in it for America?  The comment that the U.S. needs to keep AI developers, half of them in the world residing in China, display their talents using U.S. technology stacks instead of somebody else’s, is not convincing.  Giving China access to high-end chips makes it more efficient in closing the gap with U.S. technology unless NVIDIA could guarantee that reverse engineering or having access to those chips will not assist China in leapfrogging its way to beating the U.S. in chip making by replicating the U.S. tech stacks.

We all accept trade is good, that’s never been an issue.  But the constant IP theft, cyberwarfare, overt and covert espionage and social unrest fomentation, transnational police intimidation and sometimes outright crime, the use of Western technology for domestic suppression and oppression, pervasive human rights abuses, all those factors bear on the parties’ willingness to trade and cannot be swept under the carpet.  Eliminating them, like eliminating slavery in the West, is part and parcel of the raison d’être of civilised society and today’s international trade.  If we think trade is just commerce, which can be separated from socio-political justification, then why didn’t the West trade with Soviet Russia or the Axis forces in World War 2?  If the West couldn’t trade with adversaries in a Cold War or Hot War, what is it doing trading with them in a Warm War?

What has made a trade and tech war inevitable is also the unique character of the contested technology – AI.  Unlike piecemeal technologies of the past – except perhaps electricity – AI is an all-encompassing tech.  It’s not like smart phone or EV technology that a country can share with another for the sake of obtaining lower cost production for better margins.  AI is a “total tech”, meaning the country with the most advanced AI has the potential to literally rule the world.  It is not technology that can be shared but must be owned.  International trade logic does not apply with this “complete” technology.  But it is also true that a nation cannot ask its businesses, its commercial entities, to fight a government-to-government war.  The job of leading the protection of national security in an adversarial contest is that of government, not business.  Just like with government intervention at home by throwing tax dollars left, right and centre to buy votes from public programs, distasteful as it is to accept government subsidies, a business must do it so that it doesn’t fall behind its competitors who accept the bribes.  If Apple or Tesla or NVIDIA doesn’t build plant in or sell to China, it risks falling behind in the global competitive market for their products.  It is the government’s job to lead, to draw the line for all businesses so that the barricades can be raised and applied to all Western businesses in a competitively neutral manner.

The benefit from trade with China has flowed mostly to the financially largest corporations and their shareholders.  But eventually, the CCP’s game-plan of displacing Western businesses of any type or shape in China by Chinese companies that are nurtured and protected by Beijing will end up pushing Western businesses out of the country anyway.  China is for Chinese business, and the rest of the world is fair game. 

This has happened to high-tech social media companies, international payment system companies, banks, car makers (even the most efficient Japanese ones) and plenty other Western companies that once dreamed about China.  Some among the U.S. business elites still cling to the hope that eventually, post-current trade war, the two superpowers would continue to live Siamese twins’ lives.  Hedge funds still recommended only 1-2 years ago to overweight China markets for their American portfolios.  Microsoft and Apple, the two largest technical and professional trainers in China, took their time to exit China until Trump won the second term.

But things have changed for good.  The U.S. is not alone in distancing from China supply chains; China is doing the same in reverse.  And the E.U. is obliging.  The Dutch government’s move to decisively disconnect Nexperia’s IP from Chinese ownership and Chinese labs, only two years after it had allowed Nexperia to acquire Delft-based start-up Nowi, is part of a no-holds-barred contest.  Not just companies but countries around the world are taking position.  They are accepting that the U.S. and China are seriously decoupling, against all the conventional wisdom of fear of dis-integrating these two economies from one another given their supply chain interconnectedness.

For all the criticisms of Trump as a boorish construction billionaire from New York, he seems to appreciate history and world power contests far more than any of the post-Reagan Presidents.  Trump must play with the cards he’s dealt with, leftovers from previous administrations that kicked the cans down the road.  The U.S. has been riddled with truly naïve financial and corporate elites and intellectual tribes, who have thrown USD1-2 trillion into China to build it to where it is today and showered the CCP with praise of linear leadership compared to Democracies’ serpentine progress.  They still refuse to acknowledge it's a bet that looked too good to be true and that has gone sour for America.  Now that he is finally in charge – his first term being discounted due to his inexperience in government affairs and the ravenous monolithic convergence pouring all its power and resources into making his Presidency unworkable – Trump is pushing a history making global agenda.  Interestingly, world leaders can see his second term as deterministic in which way the world will be headed.  They have pledged their countries’ fortunes and destinies to him, each in the hundreds of billions of investments in the U.S.  It is this allocative efficiency that Trump’s critics miss.  Global capital is flooding to the U.S.  Looking at China’s technical and productive efficiency can be intimidating, but it’s the wrong place to check.  China and its BRICS partners do not have net demand capacity to soak up the excess supply capacity of each other.

One of the most effective leadership aspects that Trump is bringing to the table is in energy supply.  Without lowest-cost power supply, the West is at serious risk of being defunct in AI.  The lead that the U.S. currently holds in AI development is very fluid.  China’s DeepSeek shows that technical and productive efficiency is borderless.  The tech world is talking about a fundamental shift in AI development from transistor based to memristor based, from binary to organic logic.  Even using a different approach to building large language learning models (LLMs) within the transistor paradigm can cut memory and storage capacity required in jumping from node to node as DeepSeek shows.  Maybe this is not something that NVIDIA and other chip designers may have missed, it’s something they could be working on among other logic options.  It’s just that the payback period for the chips business is so short, 1-3 years, that punting tens of billion dollars on a short-lead binary transistor technology is OK and a must to provide cashflow and stay in the game while they invent others.

What the West still holds in spades is allocative efficiency.  The U.S. stands head and shoulders above the rest of Western governments in driving this agenda.  It is letting the market decide where to focus capital and in the current environment it is lowest cost energy supply systems.  In others, from Europe to U.K., Canada to Australia, governments are still stuck in the hypothetical emission fear of yesterday while a perfectly suitable zero-emission technology, nuclear energy, is sitting there waiting.

Australia is throwing more tax dollars to transit an electricity supply system that has doubled and tripled power prices in the market in a decade while the media extols lower renewable energy cost!  Our energy policy guarantees higher retail electricity prices.  Emission-based energy supply contortion is old paradigm, neither the U.S. nor the BRICS are paying attention to it other than lip service to keep the horse-trading at the U.N. going.  The U.S. has cast aside the Net-Zero nonsense to defend Western civilisation.

Building semiconductor fabs and datacentres, with secure exclusive supply of EUV lithography machines, is a necessary but not sufficient step for AI dominance.  The race for AI dominance is staggering in financial and energy investment terms.  A single project like Stargate began construction with a bill of $500 billion.  This is in the same state as SpaceX’s Starbase, another $0.5 trillion complex.  These projects, together with chips demand from Apple, Tesla, Microsoft, Google, Amazon and Meta, have made semiconductor production plants possible to be built in the U.S.  Additional amounts of capital that have been pledged or earmarked in the immediate future for investment in the U.S., from local and foreign funds, rival the entire Australian GDP.  Leading AI players in the U.S. are looking at installing their own nuclear energy plant to power their data centres and the independent, adaptive AI technologies including the computational intensity and software that are driven by algorithms.  This is in a market with a vast shale gas and oil industry and where the U.S. government has already re-liberated fossil fuel supplies as a cornerstone of U.S. energy policy.

While waiting for nuclear power addition, big techs are using all forms of energy supply.  Musk’s AI venture, xAI, is racing to build the world’s largest AI supercomputer cluster, the Colossus project, in Tennessee.  To get power supply within months, the project acquired an entire natural gas-fired power plant from Europe and shipped it to site to fuel what is expected to be 1 million NVIDIA GPUs with a demand of up to 2 gigawatts (GW) of power – enough to supply half of Perth.  The approach bypassed U.S. regulatory and construction delays by importing ready-made infrastructure.

In a statement to Congress, the ex-CEO of Google Eric Schmidt said 99% of all electricity could be used to power superintelligent AI compared to 3% at present.  He also mentioned that superintelligent AI was coming sooner than expected and warned that China could gain an edge if the U.S. could not keep up with computing and power needs.  Hyperbole aside, the message is clear.  To meet its AI industry’s needs alone, the U.S. needs to bring in 29 GW of power generation by 2029 and 67 GW by 2030.  The latter figure is Australia’s entire power system.  In the long term, the marginal cost of AI as a product or service will tend to zero given the unbounded efficiency gain embedded in the technology.  But an energy molecule will remain an energy molecule, so the cost floor for AI is the marginal cost of power supply.  The race is in driving the cost of this molecule to as low as possible.

China is racing to build its own AI momentum with the advantage of low power cost from a supply system dominated by coal and gas.  The country’s rapid build-up of solar and wind farms is for manufacturing job maintenance purposes while other key industries like real estate are suffering recession.  China builds for the sake of building, its reliance on solar-wind power production as a portion of the projected total power consumption creeping up only slightly in the next 10 years.  Accepting decoupling as inevitable, Beijing’s reaction to U.S. chip bans by banning U.S. chips is not a short-term negotiating tactic.  Even as the U.S. allows NVIDIA chip export to China to resume, Beijing considers that move to be a band-aid.  It is pouring more state money into fast-tracking domestic chip making.

China’s plan several years back has been to stock up ASML machines to buy China time to make its own.  Its inventory of EUV and DUV machines ordered in 2023 is doing just that.  Buying China time.  Fast forward to early 2025 and the chip war blew up on ASMLS’s screens in China – more accurately, the screens went dead.  Any connection with ASML was cut, even tech support.  There was finger-pointing at Chinese buyers trying to tinker with ASML machines causing the machines to self-disable, an inbuilt safeguard against reverse-engineering.  But the damage is done all the same.  Below the top-end fight where capital mobilisation is vital and where the U.S. holds a decisive advantage, the vast use of AI in the broad economy is just as critical and this is where Beijing is mounting its battles.  Advanced weaponry doesn’t need high density sub-10NM chips, it is produced using 100+NM chips.  An air-to-air missile like the AIM-120 AMRAAM or ground-to-air missile like a MIM-104 Patriot relies on specialised semiconductors for guidance, seekers, radar processing, and control systems.  These chips prioritise radiation hardening (rad-hard), reliability, and operation in extreme environments (high G-forces, temperatures, EMPs) over the raw performance of consumer-grade tech.  As a result, they typically use larger process nodes compared to smartphones or AI servers, which chase sub-5NM for density and speed.  Predominant range for guidance and seeker electronics is 45-130NM or larger, up to 250NM for highly rad-hard components.  These are mature nodes that balance performance with durability.  Smaller nodes of sub-10NM are more susceptible to single-event upsets (SEU) from cosmic rays or nuclear effects, which can flip bits or cause failures.  Rad-hard designs use techniques like silicon-on-insulator (SOI) or wide-bandgap materials such as GaN for RF, often on older fabs for proven resilience.  The 7NM is popular in experimental or non-critical AI/drone integrations, but not yet standard in operational missiles due to hardening challenges. Niche / advanced subsystems use 14-28NM for high-compute tasks like signal processing or AI-assisted targeting in newer variants.  Most weapons, however, can work with chips up to 250NM.  Up until recently the larger nodes have been preferred as they can be acquired from trusted U.S. foundries like Global Foundries for sabotage-proof operations.  With TSMC and Samsung lodged on U.S. soil, there could be niche purchases although most sub-7NM are for smart, small high performance consumer electronics.

The lithography, semiconductor and chip market is going through radical changes.  The tech war is establishing two parallel supply chains, technologies, and approaches to problem solving.  In March 2025, Sekeri, a Shenzhen company, unveiled a domestic immersion lithography system capable of patterning features below 7NM.  The demonstration showed live wafer processing electron microscope images of successful patterns and yield data that defied known lithography limits.  Another breakthrough by a group of Chinese researchers enabled more lithography enhancement by combining multiple exposures of older equipment with AI driven pattern correction algorithms.  A few months later, a company called SMK said it was producing 5NM chips using entirely domestic equipment.  The yield rates, the rates of success in turning out usable chips after the intricate manufacturing process, are reported to reach 65% compared to TSMC’s 85-90%.

It is hard to confirm the details but generally the gaps between Western and China chips remain, though both sides have seen significant accretive market value for the leading companies.  For now, the market is seeing distinct opportunities for both sides.  Analysts compare NVIDIA’s export-grade H20 with Huawei’s Ascend 910B and find the NVIDIA part still has advantages in memory capacity and bandwidth, which matter for training large models.  Software maturity gaps around Huawei’s stack remain, that reduce effective throughput, even when nominal specs look close to older NVIDIA parts like A100.  These issues make it harder for Chinese labs to match U.S. training runs at the same wall-clock cost.  However, with export approval for the higher grade H200, it is an open question again whether this would turn out to be good or bad for the U.S.  The H20 (often called HGX H20) is a downgraded variant specifically designed by NVIDIA to comply with previous U.S. export restrictions for sale to China.  It has reduced specs, including 96 HBM3 memory, 4 TB/s bandwidth, lower core count (78 Streaming Multiprocessors and TDP of 450W).  On the other hand, the H200 is a superior, full performance Hopper-generation AI GPU released in 2024, with up to 141-144 GB of HBM3e memory, 4.8TB/s bandwidth, high-compute performance (3958 TFLOPS in FP8 Tensor Core) and a TDP of 7700W.  We shall see how Beijing and the market respond to DC’s latest gesture of goodwill.    

Sustainment of a Superpower

In terms of geographical security, unlike China, India, Russia or Europe, the U.S. is blessed with two-side oceans that invite it to access both sides of the world, the Pacific and the Atlantic, to meet up in the Indian Ocean and Mediterranean half a world away, where the hot spots potentially lie.  Inside the U.S., another stark difference with other countries or continents is its waterways, as represented by the Mississippi, the longest water and water tributary system in the world.  Plenty has been written about the power of the mighty Mississippi already, but it is worth us repeating some of its critical aspects as this river system makes the U.S. the lowest cost transport economy and most productive agricultural economy.  The U.S. can feed itself, a very large population, and many other countries with just 1% of its workforce in agriculture, thanks to its fertile hinterland being close to all the bulk water transport connections scattered by the Mississippi that cost one-tenth of overland transport.

Spanning 3,730 kilometres from its traditional source at Lake Itasca in northern Minnesota to its mouth at the Gulf of Mexico, the main stem of the Mississippi forms the backbone of one of the most extensive and navigable inland waterway systems in the world.  Including all its tributaries like the Missouri and Ohio rivers, U.S. Geological Survey estimates that the entire system reaches an incredible 500,000 kilometres, second only to the entire Amazon system that traverses 7 countries in South and Central America.  The Mississippi drains about 40% of the contiguous U.S. across 31 states, making it a natural integrator of the nation’s geography.  Its navigability allows for efficient barge transportation of bulk commodities, which is more cost-effective and fuel-efficient than rail or truck alternatives.  Economically, the river is vital to U.S. dominance as a global agricultural and industrial powerhouse.  It facilitates the movement of 300 million tons of goods annually, including grains, petroleum, coal, and construction materials.  The river basin produces 95% of U.S. agricultural exports, supporting industries that employ over 400,000 people in recreation and related sectors alone.  This internal connectivity has historically enabled the U.S. to develop a unified domestic market, reducing transportation costs and fostering economic expansion westward.  Without this system, the U.S. would face higher logistics expenses, potentially eroding its competitive edge in global trade.  Militarily, the Mississippi enhances U.S. strategic depth and resilience.  As a critical shipping route, it allows for rapid internal mobilisation of resources during conflicts or emergencies, minimising reliance on vulnerable oceanic supply lines.  Inland waterways carry 14-18% of U.S. freight, which is essential for national defence – enabling the transport of military supplies, fuel, and equipment without exposure to foreign blockades.  Historically, control of the river was pivotal during westward expansion and the Civil War, securing territorial integrity.

In modern times, it bolsters U.S. hegemony by providing geographic advantages – the river system’s vast, interconnected nature makes invasion logistically daunting, while supporting a self-sufficient industrial base that underpins military production and power projection abroad.

Naturally, adversaries of the U.S. would see the country’s internal water system as one of its most powerful natural endowments.  At the Louisiana District Export Council (LDEC) general meeting held on 24 Jul 2025, Supervisory Special Agent Ben Dreessen from the New Orleans FBI office gave warnings about the threat of China to American trade, including that China was “targeting the Mississippi River system.”  Dreessen outlined China’s five-year plan for achieving worldwide monopolies in strategic industries through intellectual property acquisition and other means.  Specifically on the Mississippi River system, he cautioned that China was targeting it as part of a broader strategy, including establishing businesses and buying properties in major port cities like St. Louis and Chicago, as well as focusing on ports in southern Louisiana as a gateway to the U.S. interior from the Gulf of Mexico.  He emphasised that this could enable China to gain access to and control supply chains, leverage economic ties for political influence, collect intelligence, and potentially disrupt U.S. commerce.  Dreessen also discussed related cyber threats like Volt Typhoon and Salt Typhoon aimed at U.S. infrastructure. The warnings came as trade between New Orleans and China had surged over the decade to 2023 according to data from Louisiana Department of Transportation and Development.  In that period, the state recorded a 387% increase in loaded total 20-foot equivalent units – nearly a fivefold jump in trade volume between New Orleans and China.  More broadly, Dreessen noted that China was “anticipating major conflict with the U.S.,” and was working to interfere with the American military, “induce panic” and “impede decision making in the White House”.  Developments since that meeting have focused on escalating U.S.-China trade tensions around maritime activities, including the U.S. imposing new port fees on Chinese-built, -owned or -operated ships starting 14 Oct 2025 (potentially costing Chinese carriers billions), with China retaliating by announcing similar fees on U.S.-linked vessels from the same date.  

The U.S. has also rolled out a major maritime strategy in Sep 2025 to counter China’s global port influence, emphasising investments abroad rather than domestic concessions.  Prior to July 2025, there were isolated cases like the Chinese state-owned COFCO taking full ownership of a grain terminal on the Mississippi River in 2024, but nothing comparable has occurred since the LDEC meeting.  Overall, U.S. policy appears geared toward restricting rather than enabling further Chinese inroads into domestic port infrastructure.

Lawmakers and federal agencies have increasingly described China’s strategy as one in which the government tightly controls major trade and investment channels.  Beijing’s policies often require U.S. firms to partner with Chinese SOEs and transfer proprietary technology to gain access to Chinese markets – a practice seen in sectors ranging from aerospace to semiconductors.  American companies tended to comply because they misperceived China’s governmental and legal framework as like the West given China’s membership of the WTO.  Only later that Western businesses and governments accepted the reality of CCP-ruled China, which resembles a fascist or mafia state instead of a “normal” country, with clear intent to rival the U.S. as an ideological enemy instead of a commercial competitor.  Examples cited by Congressional analysts include U.S. firms partnering with Chinese companies to develop the C-919 passenger jet, state funding for semiconductor manufacturing to reduce dependency on foreign suppliers, and government mandates requiring localisation of electric vehicle battery supply chains.  China also expanded control of critical minerals, pharmaceuticals, and biotechnology through overseas acquisitions and state-backed investment.

In the energy sector, Chinese purchases of U.S. LNG have at times included related investments in American export terminals – deepening Beijing’s exposure to, and potential influence over, U.S. energy infrastructure.  Such commercial moves could be seen as benign from a same-value country as Japan, but with China that horse has bolted, with the West convinced that China’s goal is in vanquishing the free market liberal society world, preferably from within.

The U.S. is also self-sufficient in energy supply, relying on low-cost shale gas and oil and natural gas reserves that can see the nation to the end of the century, even considering any prospects of peak-oil.  Its northern neighbour can supply another century of fossil fuels, putting aside the U.S. capacity to mass build nuclear power plant as soon as its new regulatory framework is ready, as promised by the Trump administration.  At the end of World War 2, the U.S. was the last major participant standing with an intact homeland.  With industrial capacity at its peak, the economy making up more than half of the world’s GDP, the U.S. opened its market to allies to establish a new security network to fight the Cold War.  With the USD being the reserve currency by default and by design, the country saw fit to wean the world off the traditional view that such currencies ought to be backed by gold or some store of value assets.   

The 1944 Bretton Woods system was predicated on the U.S. dollar (USD) serving as the world’s reserve currency, with the USD directly linked to gold as a store of value.  Under the agreement, the USD was convertible to gold at a fixed rate of $35 per ounce, and other currencies were pegged to the USD, creating a system of fixed exchange rates.  This gold-backed USD provided stability and confidence in international trade and finance, as participating countries could rely on the dollar’s convertibility to gold.  The system required the U.S. to hold sufficient gold reserves to back the dollars in circulation, a commitment that became strained by the 1960s due to increasing U.S. deficits and global dollar demand, ultimately leading to the system’s collapse in 1971 when President Nixon suspended gold convertibility.  When the U.S. effectively ended the Bretton Woods system by suspending the convertibility of the USD into gold – often referred to as the “Nixon Shock” – the argument for moving away from the gold standard was multifaceted and rested on economic, practical, and geopolitical reasoning, with the U.S.’s overall power, including its military and economic dominance, providing implicit backing for the dollar’s continued role as the world’s reserve currency.     

By the late 1960s, the U.S. faced a “Triffin Dilemma” in that the demand for USD as the global reserve currency led to persistent U.S. trade deficits, as the U.S. supplied dollars to the world.  However, maintaining gold convertibility at $35 per ounce required sufficient gold reserves, which were dwindling.  By 1971, U.S. gold reserves had fallen from 20,000 tons in the 1950s to 8,100 tons, while foreign dollar holdings grew, threatening a run on U.S. gold if countries demanded conversion.  The system was unsustainable because the U.S. could not maintain enough gold to back the expanding global supply of dollars needed for trade and reconstruction including post-WW2 Marshall Plan and aid to Asia.  Pegging the dollar to gold constrained U.S. monetary policy.  The fixed exchange rate limited the Federal Reserve’s ability to adjust money supply or interest rates to address domestic issues like inflation or recession.  By 1971, the U.S. faced stagflation (rising inflation and unemployment), partly due to Vietnam War spending and domestic programs.  A fiat currency – one based on common acceptance rather than on a particular store of value asset like gold – would allow greater flexibility to print money, adjust interest rates, and manage economic challenges without the rigid gold constraint.  The U.S. argued that the dollar’s value no longer needed gold backing because it was underpinned by the strength of the U.S. economy – the world’s largest, most industrialised, and most stable at the time.  The USD was already deeply embedded in global trade, with oil and other commodities priced in dollars (eg, the emerging petrodollar system).  While not explicitly stated in policy announcements, the U.S.’s unmatched military power played a significant implicit role. The U.S. Navy ensured the security of global trade routes, particularly for oil, which was increasingly dollar-denominated after agreements with OPEC nations (notably Saudi Arabia) in the early 1970s.  This petrodollar system reinforced demand for USD, as countries needed dollars to buy oil, replacing gold with a new form of backing: global economic necessity and U.S. geopolitical dominance.

By 1971, the USD was the linchpin of global finance, used in international trade, reserves, and debt issuance.  Most countries were willing to accept the dollar without gold backing because of its liquidity, the depth of U.S. financial markets, and the lack of a viable alternative (the Soviet rouble was not convertible, and European currencies were fragmented).  Thus, the end of Bretton Woods marked the pinnacle of American power, with the smooth transition to a fiat system cementing its position as the underwriter of global economic and military security.  This military power continues unabated today, although U.S. GDP now accounts for one-quarter of global GDP due to successful U.S. economic aid to most of the world for the last eight decades.

When assessing U.S. economic power, we need to include the waterfall effect of the U.S. economy.  The U.S. “empire” is not the same as empires of the past, which mostly attempted to hoard resources and wealth at home, not abroad.  Colonisation transferred resources to home base, rarely engaged in substantial two-way trade for the benefit of the colonies.  It’s a later version of the tributary system China used for centuries.  But the U.S. created an international trading system that benefited both itself and other nations, based on U.S. capital investment, using the USD as fiat currency, and backed by U.S. military might.  It ran a de-facto global U.S. economy, with bits of every other nation’s GDP being parts of U.S. GDP.  Every nation is a U.S. strategic client, otherwise, its policing capabilities cannot be applied to adjudicate disputes between nations.  Hence, the “U.S. economy” is much bigger than its official 25% of world GDP as often quoted in statistics.

Whether you are an oil-rich Islamic state in the Middle East or a developing country in Southeast Asia, or a Baltic member of the E.U., you need to adhere to the U.S. economic order so the world can go round.  The moment you decide to exit this order, you lose the unwritten U.S. protection and take your chances with the wolves in the wild including sub-Saharan pirates.  And if you happen to cross paths with the U.S. in future on very unfriendly terms, you could be bombed back to the stone age without losing a piece of your territory.  The U.S. does not need your land; it only needs to set you back 10-20-40 years until the next encounter.  Warfare in countries like Vietnam, Iraq, Afghanistan, never really represented U.S. “losses”.  After each encounter, the U.S. withdrew without an ounce of hesitation, with each of those countries destroyed for decades, while the U.S. economy hardly skipped a beat.  The emotional trauma that the media inundated the world with, about the devastation wrecked on American society and psyche, is true as far as individual tragedies are concerned, the families that lost their members in the battlefields, but the impact on the nation, the economy, was never lasting.

This is not to say that the U.S. global network is anything but humane and beneficial to its members.  Looking at the world through dark-coloured lenses, the U.S. system is the best – or least detrimental to humanity – that can be offered.  All countries in the world, large or small, are accorded similar security protection.  Smart countries like the Asian Tigers, with tiny land and limited natural resources, have managed to grab on to the U.S. coattail for the prosperity ride.  Countries that had made mistakes to be an enemy of the U.S., like Germany or Japan, could still come back to the fold if they are willing to engage in win-win cooperation with the U.S.  Germany was split in halves between two empires, and once the Evil one fell the U.S. facilitated the return of East Germany to its brethren.  A non-U.S. empire would never have done that; it would have tried to exact its pound of flesh for having spent 45 years funding NATO to break the back of the Soviet Union economy.  Vietnam, on the other hand, tried to re-unify under a communist regime in breach of the 1973 Paris Peace Agreement sponsored by the U.S.  Hanoi paid a 30-year price, the first 20 years until 1994 when President Clinton re-normalised relations, and another 10 at least for it to climb out of the hole it dug itself into.  It took until 2016, over 40 years since the end of the war, for the U.S. to lift the arms embargo on Vietnam as part of the resetting of military strategy towards China.  From rice exporter to starvation threatened, Vietnam in the end had to come back seeking peace with America asking to re-join the world community – meaning the capitalist community.  Modernisation in Vietnam in the 1990s followed China’s Modernisation in the 1980s, both cases with approval from the U.S. for it to take effect.

The Vietnam War is a salutary lesson for countries around the world.  You can call it whatever name you like, revolutionary war, Cold War proxy war, nationalist war, communist war, independence war, but in the end the capitalist world wins as you must ditch all your collectivist hubris to escape famine.  In this sense, non-Vietnam Southeast Asia has the U.S. and its allies to thank for.  Military intervention in Vietnam contained the spread of Chinese communism by bogging it down in Vietnam, the transmission channel for Beijing.  By the time the war was done, Thailand, Malaysia, Singapore, Indonesia had had enough time to consolidate capitalism with U.S. and Western help.  Vietnam’s invasion of Cambodia in 1978 was unavoidable due to the China-backed Khmer Rouge’s provocation and incursions into Vietnam along the border.  The West also was relieved that Hanoi did the dirty job for the world in annihilating the genocidal Pol Pot regime.  But it was very unlikely that Hanoi would have tried to expand to Thailand or Malaysia on its own volition, as its war had always been sold to the Vietnamese people as an independence war against the U.S.  Once Saigon had fallen, the rationale for sustaining national hardship went away.

What Hanoi could have done if it weren’t delayed by 20 years in its conquest of the south was to spread Maoism and Leninism to support the communist movement in Indonesia and Malaysia, with unknown consequences.  The U.S. and allies’ human sacrifices in the Vietnam War were not in vain; they did save the rest of Southeast Asia even though they failed to save South Vietnam.  The regrettable Nixon-Kissinger strategic move on China made poor timing for Saigon, which had been hamstrung by Washington DC’s “Limited War” strategy that prevented Saigon and its allies from taking the war to the north, a ludicrous constraint imposed on South Vietnam by DC politicians and diplomats running the war show.  If the U.S. hadn’t taken the Korean War to the Yalu River, drawing in a massive Chinese response, the war might not have ended up where it all began, at the 38th parallel.  South Korea would not have had the chance to rebuild on the back of destruction that North Korea had to recover from and the human wave losses that China itself suffered in pushing back the Americans.  North Vietnam, on the other hand, felt secure that it would never be invaded, so spent every few years restocking for an all-out assault on the South, in 1968, 1971-72 and finally 1974-75.  The only time when Hanoi wavered was at end 1972-early 1973, when the U.S. took strategic bombing to the North.  Armchair analysts said such campaigns would not work, pointing to World War 2 statistics that appear to show the bang-for-buck was not worth it in terms of physical destruction of enemy assets for the bomb tonnage dropped.  But the psychological effect on the people was immediate.  Carpet bombing had a tremendous impact on German and Japanese civilians and thus their regimes – even though U.S. strategic bombing took pain to avoid civilian concentrations and aimed at military-industrial targets, hence the term “strategic”.  It took only a fortnight but Hanoi almost capitulated according to post-war accounts by Hanoi residents.  Moscow was quick to intervene and proposed the 1973 Paris Accord framework, halting the bombing and saving Marxism in Vietnam, just as the CCP had saved it before.

This is a hard-earned military lesson for the U.S. that the country learned well in its subsequent First Gulf War.  General Norman Schwarzkopf, commander of U.S. Central Command (CENTCOM), was appointed overall supreme commander of the multinational coalition forces.  The author of It Doesn’t Take a Hero fame recalled the pervasive political constraints on military campaigns as a young officer in the jungles of Vietnam.  He sought and was approved a free hand to exact victory from Desert Storm.  He took back Kuwait in 4 days and eliminated three-quarters of the 150,000-strong Iraqi Republican Guard.  In the Second Gulf War, Iraqi Freedom, supreme commander Tommy Franks conducted a similarly all-out campaign to decimate Saddam Hussein’s forces in just weeks.  It took 20 days for the first U.S. combat units to arrive in Baghdad and another 10 days to topple the regime.  It’s the kind of military campaign that the U.S. military is structured to carry out.  Non-war, non-conflict scenarios like the Somalia extraction (Black Hawk Down) mission in 1993 that it failed to do, re-taught the half-baked military lesson of the Vietnam War, something the U.S. has tried to avoid so far.

The Afghan withdrawal was a disastrous tragicomedy executed by an incompetent administration and a group of small and politicised military leadership.  But the purpose of war was achieved, with that country abandoned to its prejudices with no more impact on the world scene.  Once the new Trump administration was installed, the military got back to its purpose-specific operations and how the U.S. uses its military to degrade and devastate, not take, enemy territory.  As a Chinese general has said, China would never allow a situation in which the U.S. could go to war with China the way it did with Iraq.  The U.S. will win an all-out military war hands down.  This is the reason why monolithic convergence is trying to conquer it from within.  And if the U.S. falls, the rest of the West falls.

Right now, U.S. capital dominates the world.  U.S. weapons control the world.  U.S. hegemony just wants your business and cooperation, not your land.  The U.S.’s Constitutional Republic is there for you to use.  Even Ho Chi Minh recited parts of the U.S. Constitution in his declaration of independence in Hanoi’s power vacuum of 1946.  Millions of migrants have flooded into the U.S. illegally because they know the worst-case scenario is free deportation with three months’ pay.  No firing squad or hard labour prison camp or re-education camps a la gulags or Uyghur.  The West is the most accommodating civilisation to adversaries there is, in war and in peace.  It beggars belief that any external regime, let alone its own citizens, would think of distancing from it, trying to antagonise or destroy it, unless they belong to the insane unconstrained vision side of history.