Open source is very popular these days, not least due to the open-source strategy that the European Commission announced last week as part of the tech sovereignty package. While there is a lot to criticise about that strategy (exaggerated goals, misleading numbers etc), the one thing I would like to point out is that open source has somehow become a signifier of ‘good’. The Commission exploits that by linking open source to otherwise controversial projects like the EUDI Wallet and age verification. How can they be bad if they are developed as open-source projects?!
This is linked to a widely shared belief that we need to develop open-source ‘alternatives’ to the heavily criticised “evil” platforms by big tech firms. Many progressives mean ‘alternative’ in a double sense: open source both as a basis for better services and software in competition with big tech products, and as a starting point for breaking free from commercial logic. In the following, I argue that open source currently fulfils neither promise and that instead, it trades a supposed “slightly less bad” (even though also that is questionable, as I will show) for a more meaningful debate about how software should be developed and whose needs it should address. I argue that the role of open-source software (OSS) in digital policy is very much like that of the electric car in mobility policy: Slightly less bad, but also in the way of meaningful transformation. At best, alternative products based on OSS are niche projects that successfully evade profit-orientation; at worst, they palliate an expansion and intensification of the profit logic. In most cases, they represent a resource for big tech, follow the proprietary design paradigms and are produced under similar conditions, as I will show in the following.
Overcoming proprietary software is good, as is electrification
Of course, it is useful to overcome the paradigm of proprietary software: Just as electrification is sensible and necessary for a comprehensive network of buses and trains, certain characteristics of OSS are useful to develop digital infrastructures oriented towards the public interest. Compared to proprietary software, OSS facilitates the maintenance of systems and their reuse in new contexts, and is more conducive to a certain degree of cooperation, for example through voluntary interoperability. But just as the electric car does not break away from the paradigm of ‘individual transport’ or bring about needs-based and environmentally sustainable mobility centred on buses and trains, OSS does not challenge the dominance of commercial and exploitative digital corporations or promote the widespread adoption of public-interest-oriented infrastructure. Just as major car manufacturers want to be at the forefront of electrification, thereby hindering broader change, OSS is also heavily co-opted by large digital corporations.
Big tech co-opts open source
The co-optation strategies have evolved over time. For example, Google originally developed more open products on which it selectively imposed restrictions over time. Microsoft started from proprietary products and only embraced OSS as an opportunity for expansion when it purchased the OSS code platform GitHub for USD 7.5 billion in 2018. All major digital corporations rely on OSS as a component or foundation of their products. They benefit from this in several ways, including unpaid labour, standardisation and strategic undermining of competitors.
Firstly, unpaid labour: OSS projects rely on exploitative dynamics by facilitating access to unpaid labour and communal resources. Motivated, predominantly male software developers often work on OSS in their spare time, frequently to brush up their CVs. Also openly available resource like digital ‘commons’ predominantly feed into big tech projects, i.e. public goods such as Wikipedia or OpenStreetMap. For example, Google unilaterally monetises Wikipedia’s content via its search engine and ‘AI’ applications. Wikimedia, the organisation behind Wikipedia, relies on visibility through links and donations and is in a struggle for both as ‘AI’ crawlers and overviews overwhelm its technical infrastructure while reducing click-throughs. A similar dynamic is emerging for so-called ‘digital public goods’, ‘digital public infrastructures’ and also ‘open data’, but I will skip them here for now (more detail in Malte’s and my book in July).
Big tech companies benefit from open source through unpaid labour, standardisation and strategic undermining of competitors.
Secondly, standardisation: OSS projects allow companies to influence technological development through standardisation bodies and to promote the dissemination of their own products. One example is the OOXML format used in Microsoft’s word processing software, which, despite formal standardisation as ISO/IEC 29500, attracts justified criticism because it exhibits proprietary dependencies and inconsistencies in its implementation. Another example is ActivityPub, the Fediverse protocol, which has received renewed attention within the W3C standardisation body since January 2026: Meta, for example, formally holds a seat on this committee, but not the developers of Mastodon, the de-facto most relevant network based on ActivityPub.
Thirdly, strategic undermining of competitors: big tech firms decide to release certain applications as OSS because this allows them to restrict the capabilities of competitors. For instance, by making Android available as OSS (albeit in an incomplete form), Google has effectively made the commercial viability of competing mobile operating systems impossible, while cross-subsidising the development of Android through its dominance in adjacent markets such as pre-installed applications. A similar dynamic is emerging with the release of Chinese ‘OSAI’ models, which are spreading more rapidly and widely than US ‘AI’ models. By allowing unrestricted access to their models, they are, in the long term, reducing the profit prospects of proprietary competitors. Here, too, it is clear that open source is not primarily an expression of technology design geared towards the common good, but rather a means to a (geopolitical) end.
Open source follows proprietary design paradigms
Just as electric cars look practically identical to combustion-engine vehicles, OSS projects (even the more independent ones) usually remain within the same paradigms as the products for which they are supposed to be an alternative. This limited scope also results from the competitive pressure to be noticed at all alongside established applications and to build on established usage habits.
The Fediverse illustrates this neatly: It is, in principle, a positive development as a collection of non-commercial applications connected via a protocol. Yet Pixelfed, a Fediverse alternative to photo platforms such as Instagram, follows the paradigms of its big tech model in many respects. It may collect less data or embed fewer adverts, but it operates according to similar logic within the attention economy: glossy images and regular content increase follower numbers and reach. The fact that this reach is not artificially amplified by algorithms is a step forward, but it is not sufficient to ensure that an Instagram-like photo culture does not emerge, with similarly socially harmful implications. The competitive pressure from big tech pushes even open and non-commercial networks such as Mastodon to align their design with that of the tech corporations. This can be observed between Mastodon and Threads. Threads is the microblogging service of Meta, with several hundred million users worldwide. In 2025, it announced on multiple occasions that it was working on interoperability with the ActivityPub protocol and would thus be connected to Mastodon at a technical level. Whilst this new openness meets the demands of prominent digital policy-makers, it primarily puts Mastodon, as a significantly smaller player, under pressure. For example regarding content moderation, Meta moderates content uniformly with professional content moderators, optimised for efficiency. Mastodon’s moderation is carried out on a voluntary basis and can be understood more as an expression of political practice. However, given the vast difference in user numbers, this means that Mastodon content moderators must either adapt their standards to those of Threads or unilaterally disable interoperability – the latter of which is happening across many large instances, thereby foregoing the supposed benefits of interoperability. Hence, it is fair to say that more direct competition restricts Mastodon’s scope to uphold other principles.
Open source produces under the same conditions – white, male, non-democratic
As the mobility sector shifts from combustion engines to electricity, so does the object of extractivism: from fossil fuels to lithium mines for battery production. But the production remains deeply extractivist. Similarly, the development of OSS resembles that of proprietary software in that it is usually no more democratic or inclusive: The well-paid, permanently employed and largely white, cis and male developers are complemented by the OSS community, which is similarly homogeneous and works within opaque and undemocratic structures. In some OSS projects, unknown developers may propose changes, but decision-making power usually lies with one or a very small number of highly committed individuals. Even successful ‘alternative’ projects such as the non-commercial Signal messenger are OSS, but organised in a similarly hierarchical manner as firms developing commercial, proprietary software. Even in the case of Mastodon, despite its comprehensively federated design, there are barely any institutional democratic structures that allow users to participate in platform design. Switching between instances is their main way to express views.
Moving beyond the open-source horizon
If we criticise the commercial logics of commercial and proprietary software, open source should not limit or distort the horizon of what is desirable, just as the electric car should not be the focal point for the transformation of mobility. Progressive open-source advocates often share the criticism outlined above, but justify their position by arguing that OSS is, after all, less ‘bad’ than proprietary software and can already be implemented as a project today. This may be true in some cases where switching to OSS does not stand in the way of more far-reaching transformation. But focussing on OSS prevents genuine alternative projects from being developed and becoming viable in the first place. We need to acknowledge the lack of transformative power in open source, its co-optation by powerful corporations, and the pressure exerted by capitalist paradigms of technology design. Only then can we think coherently about how technology can be not just less harmful, but also emancipatory and based on solidarity. Malte and I do that in more detail in our book that will be published at the end of July – in German, and it is best to get it from your local leftist bookshop, but otherwise you can already pre-order it here.
Focussing on OSS prevents genuinely alternative projects from being developed and becoming viable in the first place.
Photo by Tesla Fans Schweiz on Unsplash.