2025 First Drop
Our first weekly update of the new year; a return of subsea cables, quick look at Chinese AI, and an update on our Missives moving forward.
So, we’re well into the Year of the Snake. And things have of course changed since our New Year's Day post, even as other things continue. As we dive back into Geomastery's regular tempo, for GeoMissives, we’re doing a quick catch up on two areas that have caught our notice and we’ll finish up with some housekeeping details on our Substack and some changes of our own we'll be soon making.
Trawling for cables
Last November, we posted about the quiet hybrid sabotage Russia and China were committing to the world’s subsea cable network; a network that is crucial, relatively unprotected and, with the right equipment, easy to damage.
Since then, cases of cable damages by the same two implicated countries continue to ramp up:
Mid-November, two cables were damaged in the Baltic Sea – hence our post noted above.
During December, a Russian cargo ship loitered off the southern coast of Taiwan, roughly in the area of two cable connections for the island, the EAC-C2C and the SJC2.
On Christmas Day, a power cable between Finland and Estonia was damaged, at the same time four data subsea cables were severed in the Baltic.
Very early in the new year, on 3 January, a cable landing in northern Taiwan was damaged.
On 26 January, a communications cable between Latvia and Sweden was damaged.
Clearly, the two antagonist countries have found an infrastructural breach to prod. And while Scandinavian countries – particularly Sweden and Finland – have made their willingness for direct response clear, most Western-aligned countries are lagging behind. There is a scarcity of ships, equipment and knowledge to continue to repair these damages – there are only 22 designated cable repair ships in the world. NATO member states and US-aligned countries in Asia will likely need to rethink their approach to protecting such critical assets.
This concern – whilst more commonplace in the Baltic Sea against an openly aggressive Russia – is particularly worrying in the South China Sea. There is heightened awareness that there are multiple ways to light the veritable dynamite, but the US, Japan, South Korea, Taiwan and allied countries also need to be resolute in standing their ground and enforcing consequences.
China grasping leads in AI
China’s DeepSeek, a serious competitor to other scaled (mostly US) GenAI models, is causing consternation in the markets and amongst Western governments.
Many are questioning the claims regarding the model’s effectiveness, but that’s not the key technological issue underlying the hugely disruptive effect of DeepSeek's sudden arrival. Rather, it is the cost of the model – it appears much less expensive than its more (comparatively) established market counterparts (That, too, is disputed; at the very least it reflects the benefits of following, not leading, breakthrough innovation.) That DeepSeek achieved such outcomes while heavily constrained regarding the availability of the cutting edge chips undermines the efficacy of the US CHIPS and Science Act, and other efforts to restrain Chinese semiconductor and, by consequence, AI development.
By releasing it as open source, DeepSeek's claims regarding performance can be tested. Already, security researchers have found vulnerabilities. Sam Altman has suggested that DeepSeek ‘distils’ OpenAI’s models, implying IP theft, which, given past Chinese practice over the last 30 years is not improbable. And given the information such models collect – and in this case, stores in the PRC – there are real concerns regarding safety, security and sovereignty.
However, DeepSeek is following the now common Chinese tactic of undercutting Western models – think about how Chinese EV manufacturers and telco/ ICT hardware manufacturers like Huawei have undercut Western industries and businesses, all the while improving their product(s) – and so claiming market share and disrupting Western economic models. And, as in other markets, where one leads, others follow – Alibaba released its own high-performance model on 29 January 2025. Moreover, it belies the West's earlier, long-held assumption that China 'can't do innovation'.
This all reflects the fragility and misalignment of Western policies and economic settings. China has not only leveraged its industrial theft, but ensured there are sufficient incentives and scope for hyper competition in a massive domestic market. In comparison, Western policies are sclerotic, legislation and business models are slow to adapt, and often shaped by market heavyweights to protect their own cemented position from innovation and competition.
The only way to succeed in a world of increasingly intense technology innovation is to diversify, and invest in the seeds, skills and structures that enable participation and the rapid realisation of innovation (iterative and disruptive).
And that is leaving a reducing pool of countries including Australia particularly bereft and vulnerable. Even more than the United States or Europe, Australia's policies have failed – from education to energy to ensuring regulatory environments that encourage entrepreneurs and scaling. The federal government has done little to support an indigenous tech sector, preferring US law and policymaking, models and capability, whether for the military, national security, intelligence, broader government. All of which are dominated by overseas primes and US tech companies, or emerging technology. PsiQuantum, for example, is not Australian, though featuring a couple of Australian-born founders. Nor are we able to rely on large programs like AUKUS to be economic or technological accelerators.
The behaviour of the new US Administration – behaviour that is already uncaring and even coercive of allies – necessitates the Australian Government and Opposition meaningfully tackling a widespread reset, as allies and likemindeds already are exploring. But in an election season domestically, both of Australia's major political parties are looking to double down on their respective past certainties, making such a reset increasingly unlikely. Australians deserve a government prepared to embrace the new realities and challenges of a world shaped by rapid technological change and geostrategic shifts. And as DeepSeek has illustrated, denial is not the way.
As many others on this platform are noting, the words President Trump, China and artificial intelligence will feature large in 2025.
To the house: updates to Missives posting
And lastly, for the moment, we turn to housekeeping. Our first year on Substack has been very well received - thank you readers! We would love to continue publishing as much content as we did in 2024, but in positive tones and like many others in their start to the calendar year, our team is getting quite busy (no surprises why given what else we do with our time). We're also setting forth to deliver the next stage of growth in our Substack readership as well as a broader Geomastery pursuit in generating a trusted ecosystem of likeminded people in challenging, contesting and debating geostrategic issues.
So, we will be slowing down the tempo of topic deep dive posts authored by Geomastery staff while also adding guest authors, and putting these topic deep dives behind a paywall - as noticed by you our readers, we're committed to publishing well researched pieces and as such, they take a comparatively serious amount of time.
We also remain committed to making information and insights freely available in a timely manner. Our weekly updates on geostrategic topics and issues we think are critical but not well covered, in a similar format to this piece, will remain free to all, if slightly less weighty in the middle.
We don’t have all of the perspectives on a lot of these issues, and we don’t pretend to, so we’re hoping to broaden the approach even as we stay focused on what our content can contribute to this space. Hence, the introduction of (paid) guest authors.
To Substack writers looking to join in, you’ll have to forgive us if we already have some writers in mind, so we won’t be slotting people in immediately. BUT reach out, we’d love to start talking – even if casually – to help the Substack community grow. (Ed. Suggestions of what to call our weekly updates could be an entertaining icebreaker.)
In the meantime, we will be dropping a post about questions Australia needs to consider moving forward for our paid subscribers, and free subscribers or casual peruses can look forward to the next weekly update on 10 February. Until then, navigate well.