Hey Siri, Are you messing around with the balance of power?

On May 14th, the Biden administration announced massive increases in tariffs on hi-tech imports from China, ranging from electric vehicles, batteries, and solar cells to semiconductors and critical minerals. The scale of increase is as follows: 100% on EVs, 50% on semiconductors and solar cells, and 25% on lithium-ion EV batteries. Regardless of the actual impact of these tariffs on the global supply chain, which economists continue to evaluate, one thing is clear: we are witnessing a further deepening of the ongoing trade-war between the two nations, the political content of which occupies center stage in contemporary US electoral politics.

 

As these tensions continue to escalate, two-steps forward one-step back, the technological industry – more specifically, that of semiconductor design and manufacturing – has become a major frontier within which a shift in balance of power between the United States and China is beginning to play out. With the popular rise in commercial AI programs which require massive computing capacities, and the full incorporation of ‘Artificial Intelligence’ into the war economy, the semiconductor trade-war – or the chip war – is a focal point where inter-imperial competition and rivalry simmers to a brim, threatening occasionally to boil over into a full-fledged conflict. Therefore, it is extremely important today that socialists evaluate the forces of change, disruption, and continuity in the dynamic global supply chain of semiconductors.



 

Semiconductors in the Global Supply Chain


 

Semiconductors, also known as integrated circuits or microchips, are the backbone of the ‘digital economy’, and the basis of almost all technological commodities and devices we use in our daily lives. The global sales value of the semiconductors increased from $335.2 billion in the year 2015 to $526.8 billion in 2023, with a decrease between 2022 and 2023 that is largely accounted for by overproduction and a slowdown in consumer spending on technology after the relaxation of lockdown mandates pursuant to the COVID-19 pandemic.

 

Despite being relatively newer commodities on the global supply chain, semiconductors already occupy the fourth position in trade value, only behind crude oil, refined oil, and cars. 

The supply chain of semiconductors has historically been diffused across the world: from chip design in US, to wafer fabrication in East Asia, and assembly in China. The economy of semiconductor production has today brought together competing countries in a deeply interdependent value chain, entangled in a mesh of economic relations that nobody necessarily wants to be in but is nonetheless needed due to the global nature of the semiconductor ecosystem. Until recently, the supply chain had the R&D and design aspects concentrated around the US sphere of influence, with manufacturing and assembly aspects concentrated around China. In this network, all economies are import dependent at some or the other node in production. If you own an iPhone for instance, you will probably see the words “designed in California, assembled in China” inscribed somewhere.

 

But with significant advancements in the R&D sector in China in the recent years, and a significant upscaling of their electronics industry, the ‘know-how’ is now likely to become less and less of an American monopoly, although presently it continues to be the case. Some recent developments indicate that, we are now seeing the early signs of a movement towards total production of semiconductor-based technologies, from design to assembly, in China and a relative movement of functions away from the US. Such a trend would result in a partial disentanglement of Chinese dependency on imports from the US, with a possibility of increased intensification of the converse. The possibility of this change has been a driving force within US foreign policy in the recent years. For instance, in 2020, the United States imposed sanctions on the Chinese company Huawei accusing the company of “design-theft” as well as of conducting business in Iran in violation of American sanctions. In 2021, the CFO of Huawei, Meng Wanzhou was arrested by Canada upon a US extradition request, at Vancouver International Airport, leading to a new wave of geopolitical tensions. In 2023, Huawei launched its first independent flagship smartphone that was fully designed and manufactured in China, with micro-chip design named Charlotte developed by SMIC, a state-funded technology company. This development reportedly took Washington by surprise. Recent reports also indicate that the SMIC-Huawei collaboration has made further progress and have realistic aspirations to manufacture chips that would be quite close to the current American standards.

 

With around three-fourths of the total semiconductor manufacturing capacity concentrated in East Asia, and with largest majority of rare earth minerals required for semiconductor manufacturing being currently in China, the United States continues to be heavily dependent on China for the sustenance of its tech industry, and any additional loss of control over the ‘know-how’ would significantly impact its hegemony on the global landscape. This shift in total production of any high-demand commodity necessarily comes with a shift in balance of economic and geopolitical power, which has rendered the United States significantly anxious in the semi-conductor trade-wars. 



 

This propensity for a reorganization of the global supply chain, and the possibility of a disentanglement between the US and China, has resulted in the emergence of two kinds of vulnerabilities. One, the economic entanglement between China and the US is one of the mitigating factors that currently prevents an open conflict. A break in these entanglements would be akin to lifting the lid from a simmering pot, with the risk of an open conflict boiling over. An extremely volatile site at which this breakdown could happen is Taiwan, which is currently a leading US-allied semiconductor manufacturer whose value to the global tech supply chain is often called its ‘Silicon Shield’ in relation to China. The second vulnerability, which results from the first, lies at heart of the largest economic sector today under Global Capitalism – namely, the Big Tech Industry. The Big Tech industry today is both driving, and being driven by, these inter-imperial tensions. The semiconductor is to the Big Tech Industry what fossil fuel is to automobile industry. It a fundamental necessity that powers almost all technological products and services that are extensively essential in our social and economic lives. If inter-imperial tensions lead to a disruption in the supply chain, or result in potential microchip shortages, that can disrupt many other forms of production on a global scale.

 

Joe Biden’s blanket of tariffs must be understood also in this light, as part of a longer-term attempt to recover the hegemonic role of the US in the global balance of power, by trying to move towards technological self-sufficiency as whole, but also ‘semiconductor self-sufficiency’ in particular. But such a ‘self-sufficiency’ is an extremely tedious and expensive affair, and not one that is feasible because of the depth of economic inter-dependence today. The US is far behind in production capacity, and it has been predicted that it would need to invest more than one trillion dollars within the next 10 years if it is to achieve manufacturing self-sufficiency with regards to semiconductors.  The call to ‘self-sufficiency’ is also made against the interests of the Big Tech industry. A total national ‘self-sufficiency’ could result in up to a 65% increase in semiconductor prices, factoring in land, labour, electricity, and other capital costs, which would massively increase the overall cost of production in the Big Tech sector. Elon Musk, for example, appears to be aware of this and he has vehemently opposed the new tariffs imposed by Biden on Chinese imports.

 

It must be in light of these ongoing economic trends, that we understand the emergence of ‘Artificial Intelligence’ as a central topic of our socioeconomic and cultural worlds. 



 

The rise of ‘Artificial Intelligence’ and the Semiconductor Industry


 

The massive technical advancements in ‘Artificial Intelligence’ that have occurred in the last few years, were precisely made possible and driven by hardware advancements in semiconductor design and manufacturing. Conversely, the growing demand for semiconductors is also directly connected to the massive increase in computing needs that have been triggered by the commercialisation of ‘Artificial Intelligence’.

 

Among other components, semiconductors are core necessities to fulfill the extensive computing needs of data processing, pattern recognition, and machine learning which are the core functions of commercial AI programs, such as the popular ChatGPT. Even if there is a stagnation in consumption of electronic devices like cellphones and personal computers, as was the case in 2023, the mass commercialisation of ‘Artificial Intelligence’ have already opened a new avenue within which semiconductors are now crucial necessities, adding further fuel to the fire of inter-imperial tensions.

 

It is believed that the co-development of semiconductor hardware and AI software are going to have profound impacts on all aspects of economic life, from health and banking to transportation and manufacturing. We are told by tech industry CEOs and their podcast hosting minions that we must prepare to witness human progress at a previously unimagined scale. But when we compare these predictions with our own life experiences, we quickly realize that these are only delusions peddled by snake oil salesmen, that serve only to distract us from the concrete and already manifest morbid transformations in other areas: namely the war economy, the degradation of our planet, and the moral deprivation of our social world.



 

Artificial Intelligence and the Israeli War Economy


 

An excellent analysisby Anne Alexander in the International Socialism Journal, shines much-needed light into the complex web of Artificial Intelligence, semiconductor fabrication, global hi-tech hardware trade, imperial tensions, and the war economy. She has argued that “AI systems represent vast concentrations of capital” and that Israel’s highly fortified armoured semiconductor foundries tell the story of “decades-long processes whereby the US has helped to incubate Israel as a new centre of capital accumulation”.

 

In Kiryat Gat, only a few miles from Gaza, the tech-giant Intel has a semiconductor fabrication plant or a chip-fab. Build on top of two Palestinian villages that were destroyed during the Nakba, this plant produces semiconductor devices for the Israeli occupation. Last year, Intel and Israel signed a deal to extend this cooperation and open more facilities in 2024 and 2027, investing an additional $25 billion dollars. This deal is only among the many that are rapidly developing in the region. Apple, Amazon and Microsoft also have chip design facilities on occupied territories. Google and Amazon have also signed a joined agreement worth $1.2 Billion with Israel called Project Nimbus, to provide cloud computing services with AI tools for the Israeli Occupation Forces. The Western hi-tech industry in Israel is a direct progenitor of the Israeli war machine, and the ongoing mass investments in semiconductor hardware and AI software serve to create genocidal technologies.



 

Today, we have concrete evidence on how the Israeli Occupation Forces use AI programs, like Lavender, to conduct indiscriminate targeting and bombing with barely any human oversight in the unconscionable genocide of Palestinians. As Palestinian activists remind the world: what is done to Palestinians today will be done to other colonized and dispossessed peoples in the future. This is the present and the future of the military-industrial complex, and we must do everything in our power to resist.



 

Computing the Climate Crisis

 


In addition to these genocidal technologies, the energy cost is also an increasingly concerning avenue of failure. The accuracy of an AI program is directly dependent on the volume of data that is made available to train it. With the cut-throat competition that is ongoing in this sector between tech-giants, the commercial viability of a program is also dependent on reducing the amount of time within which that training can be carried out. High-volume data processing and time-efficient training are both fundamentally predicated upon extent of hardware capacity as well as the volume of energy that is made available. The energy costs associated with training a Large Language Model (LLM) such as ChatGPT is in millions of dollars.

 

While the ruling classes talk about the possibilities of using AI tools to combat climate change or predict wildfires, the reality is that under the current system, those avenues can only help in mitigating the symptoms of the climate crisis, not resolve the climate crisis. Given that we have already likely crossed the 1.5-degree Celsius threshold of rise in global temperatures, the extent of boom in computational needs of the capitalist and the completely unbridled investments that are yet to be made, are certainly scary. Currently, year-over-year, we see an average of 6% rise in carbon emissions directly resulting from digital technologies – one of the sectors with the highest annual rise in emissions.

 

In addition to energy consumption and carbon emissions, the projected growth in e-waste is also a huge concern. In the year 2019 alone, we globally produced 53.6 million metric tons of e-waste, and it is predicted that this figure will rise by 30% before 2030. 

The inter-imperial tensions between US and China also has significant ramifications for the mining industry. If the tensions come to a tipping point where China stops exporting gallium and germanium to the West, that would essentially disrupt the global semiconductor supply chain, resulting in significant losses for the Western tech industry. For this reason, Canada and US are already exploring opportunities to identify and seize rare earth minerals within their borders and in parts of the Global South, and these developments will further expand into and enforce a dispossession of Indigenous territories.

 

Do Socialists dream of Electric Sheep?


 

It is certainly true that the computational power unleashed by advancements in the semiconductor/micro-chip industry as well as developments in software programming, is an astounding transformation of the productive forces in our times. But capitalism and imperial global power, is unfortunately systems that are fundamentally unable to harness those forces for any social good whatsoever. In fact, the necessities of inter-imperial conflict, war and expansion that sustain capitalism mean that those forces can only be used for pure evil in the current system – from dropping bombs to moving us further along in our journey into mass extinction. 

Even in the absence of war and destruction, the deprivation caused by the hi-tech industries in our times extends to our social, personal lives as well. As you may already have noticed, the only contribution of ChatGPT to your life is in the form of extremely confusing customer service at best, and pornographic bots on social media at worst. I

 

The very idea that a CEO thought that plagiarising Scarlett Johansson’s voice and recreating her AI-girlfriend character from the movie Her, would be the ground-breaking ‘innovation’ we need is a sobering reminder of how we have no future under Capitalism. 



 

Artificial Intelligence tools are also diseased with racial biases. When Timnit Gebru, a brilliant software developer for Google, published a paper warning the Big Tech industry of the racial biases inherent to the generative AI tools they were developing, they fired her. An example of racist AI is a beauty contest software called Beauty.AI, that was designed to rate users based on how attractive they look, developed by white businessman and trained on white people as data, which then proceeded to rate dark-skinned people extremely low on its beauty metrics. It is of course the same businessman who ‘pioneered’ this embarrassing project who launched another company that is currently ‘pioneering’ AI-driven identification of medical drugs for diseases. Rest assured, we are in safe hands.

 

Noam Chomsky was one of the first to deny the label of ‘Artificial Intelligence’ to these programs that are so heavily pushed down our throats as authentic intelligence. These programs only scour large swathes of audiovisual and literary data to generate a ‘new’ combination of results. These are, Chomsky said, nothing more than professional ‘plagiarism softwares’. The total reduction of significant technoscientific capacities of our times to one dimensional tools that merely regurgitate tautologies is an unspoken tragedy of our times. Even more, these plagiarism soft wares have created a social pestilence of AI-created art on social media, and the proliferation of predatory gambling apps that are marketed through deepfake endorsement videos from celebrities like Mr. Beast or Drake, indicate the creative limits of ‘Artificial Intelligence’ under capitalism. Visual AI programs like midjourney create the most uninspiring, apolitical art whose emotional impact on the viewer is rarely more than ‘mild annoyance’.

 

In terms of personal/household technologies, the technological contributions are equally dismal. Today, everything that can be plugged into a power outlet has become a ‘smart’ device, from refrigerators, vacuum cleaners to dishwashers. Soon, every ‘smart’ device is potentially also going to be embedded with utter-unnecessary AI programs, whether you like it or not. The marketing departments of all tech companies are currently working overtime to package redundant AI programs as the ‘revolution’ you need in your life.

 

In light of these countless examples, the ongoing attempts to convince the average consumer that ‘the rise of the AI’ has rendered all electronic devices and social experiences obsolete are farcical, and indicative of the fundamental inability of capitalism to meaningfully harness these technological advancements for any social good. 



 

We deserve better!

 

It was comedian Bill Hicks who asked why the ruling classes couldn’t use the same technology that they use to drop bombs to instead geolocate a hungry person and shoot them a banana. The answer, whether he knew it or not, lies in the fact that under capitalism productive forces are only ever harnessed to extract value, exploit human beings, degrade human attention, hollow out meaning from our social lives, and fortify military power. The insatiable hunger of capitalism means that Artificial Intelligence, like all technologies before it, is reduced to being the cutlery at the hands of imperial powers to slice up and devour all life on the planet.

 

We deserve better. We deserve a system that can instead harness those productive forces under free relations of production, centering human joy, freedom, the beauties inherent to our creative powers, and quality of natural life on the planet. The revolution we need is not a talking refrigerator or smiling bomb.