TEC Review 1: Market Power

Market Power

The politics of specialised silicon ‘chip’ manufacture is dominating technology and market news. As restrictions of specialised US-made chips to China come into force, these altered market forces have prompted interesting methodological and economic debates about the cost-efficiency of large scale machine learning datasets and the energy that powers them.

It has been reported that as access to advanced US-chips becomes more difficult, Chinese tech companies are responding by training their language models on much smaller datasets. The significant reduction in computing power for smaller data sets not only reduces the kinds of energy required by these processes, but also forms part of Chinese companies to significantly reduce the cost of their tools for developers to use to build their applications. Chinese companies have reported that their lighter models are between five to ten times less expensive than those developed by US companies on larger datasets.

However, in the background another big tech/AI shift is occurring that looks likely to have equally far-reaching effects.

Over the past month ‘Big Tech’ has continued its preference for ‘small nuclear’ to power their servers.

Small Modular Reactors are a form of nuclear powered electricity production that have an output of approximate one-third that of traditional reactors. The advantages of SMRs over larger nuclear reactors are largely related to their much smaller size, and therefore the comparatively reduced size of the sites on which they are based, increased ease of transporting the reactors, and the equally smaller (and, it is claimed, cleaner) running costs of the reactors. Because of their smaller size, SMRs power output is also easier to adjust according to local need.

In mid-October, X-energy announced that Amazon would invest approximately $500 million, with the aim to ‘bring more than 5 gigawatts online in the United States by 2039, the largest commercial deployment target of SMRs to date’. Alongside its investment in X-energy, the Financial Times has reported that Amazon are also exploring SMR projects with Dominion Energy in Virginia.

The move towards SMRs is not driven solely by the tech sector, in fact government spending has dwarfed that of the tech sector. The US government has already announced billions of dollars worth of investment in SMRs and looks set to continue underwriting the risks in the sector (see the US government nuclear awards map). Whilst there have been reports of huge boons for the nuclear stock sector - SMR manufacturers Oklo Inc and NuScale Power saw share price rises of 99% over one week in mid-October - the link between the scale of public investment in the sector and the market response has been largely obscured through a more myopic focus on the servers. As an increasing number of analysts have shown, behind the moves of the US tech-sector is the much larger shadow of US military-fiscal power, whose financial might points to market changes as much as the interests of tech companies.

Another important factor in the roll-out of SMRs will be access to the raw materials - eg. uranium - upon which even these smaller reactors are dependent. In a political landscape increasingly marked by both territorial and trade wars, the links between energy security, our digital dependency upon the tech sector and their servers, and military policy become ever more entangled.

What might matter for those of us outside of the US is how our policy makers make sense of such changes and the lessons they learn from it about the role of the state and whose interest the (their?) tech sector is set up to serve.

The European Union has produced a range of tech regulation documents and proposals over the past year seeking to mitigate the impact of potential adverse effects of a US/China dominated sector. At the same, European policymakers continue to decry the lack of competitiveness of the European Union. Martin Sandbu has used the recent European Policy Analysis Group report to highlight the embeddeness of older industries in defining the character European R&D as opposed to important shifts in the US. Since 2000, US private R&D was led by Ford, Pfizer, and General Motors, but in 2024 was Alphabet, Microsoft and Meta; the EU by contrast was led by Mercedes-Benz, VW and Siemens in 2000 and in 2024 are led by…Mercedes-Benz, VW and Bosch today!

While SMRs take-off amid huge state-investment and take-up from the US’ largest private investors, Matteo Draghi’s recent European competitiveness report has called for a range of measures to improve European growth and increasing energy and security dependency.

As part of his suggestion, Draghi has said that public sector investment is important for stimulating the scale of private investment needed for serious innovation. However, the scale of public investment occurring in the US and China relative to private investment seems not to be ignored. Equally, in Draghi’s report and the resulting commentary, much less has been said about the changing structure of industrial development in the USA and China in the 21st century and, equally absent in such commentary has been a discussion of the desired structure of industry relative to governments and markets in the context of an increasingly militarising and polarising Europe.

Space Race

As significant moves have been taken in the powering of big tech and governmental servers, so to have potentially important mergers between the world of big tech and big techs relation with earth and space travel.

Amid reports of the links between big tech surveillance and individualised price-gouging moves by companies, it has been reported that the US taxi firm Uber are considering a $20 billion takeover of the travel technology giants Expedia (FT). Expedia own a huge number of fare aggregators and search engine websites including Hotels.com, Vrbo, Travelocity, Hotwire.com, Orbitz, Ebookers, CheapTickets.

While framed as part of a move to diversify Uber’s activities, it may also be seen as a consolidation of the travel sector and its power to develop even more sophisticated dynamic pricing strategies, in which previously discrete forms of travel and pricing become ever more entangled.

*Eóin Phillips is professor at La Salle and principal co-researcher in the Observatory of Quantum Technology (Smart Society Research Group). He teaches and coordinates across the university’s courses in economics and technology studies. His research specialises in the intersection of science, technology and economics, from the eighteenth century to the present day. He gained his MPhil and PhD in History of Science and Technology at the University of Cambridge and before joining La Salle researched and taught at the University of Cambridge, Ruskin College, Oxford, and at the Autonomous University of Barcelona (UAB).