OpenAI partners with Malta to offer ChatGPT Plus to everyone

Malta makes ChatGPT Plus a nationally accessible service: a European precedent

OpenAI has announced an unprecedented partnership with Malta to “bring ChatGPT Plus to all citizens,” according to the company’s own wording in its official communication. The agreement, presented as a collaboration at the scale of a state, is not limited to providing a premium subscription to a mainstream conversational agent. It also includes training programs designed to encourage “practical and responsible” uses of artificial intelligence, with the ambition of spreading the tool across the population, while also structuring its adoption.

The signal is strong. Since the launch of ChatGPT at the end of 2022, OpenAI’s announcements have mainly focused on partnerships with companies, specific public administrations, universities, or developers. Here, the change in scale is clear: it is no longer about equipping an organization, but an entire country. Even if Malta, with just over 540,000 inhabitants according to recent demographic estimates, remains a small member state of the European Union, the political significance of the announcement goes far beyond its domestic market. For OpenAI, it is an opportunity to demonstrate that a conversational AI service can be treated as a layer of civic infrastructure. For European governments, it is a full-scale test of access, training, governance and, in the background, digital sovereignty.

The choice of Malta is not insignificant. The archipelago has regularly positioned itself as a testing ground for regulatory and technological experimentation. The country had already sought to stand out in blockchain at the end of the 2010s, with a proactive discourse on innovation and attractiveness. In the field of AI, its small size, relative administrative centralization, and international orientation make it an effective laboratory for rapid deployments. For OpenAI, a national partnership in Malta makes it possible to observe large-scale adoption in a European environment, English-speaking in practice for a large part of the population, and institutionally compatible with EU frameworks.

The message is all the more important because it comes at a stage of normalization in the generative AI market. In 2023, the issue was emergence. In 2024, the battle shifted toward integration into software suites, search engines, smartphones, and business workflows. In 2025, the question becomes more political: who pays for access, who trains users, who defines legitimate uses, and under what guarantees of security, confidentiality, and continuity of service? The Maltese partnership puts these questions at the center, because it transforms an individual subscription, historically designed for voluntary users, into a collective access mechanism supported at the national level.

It is important to measure what ChatGPT Plus represents in OpenAI’s ecosystem. The paid offering has long been the gateway to the company’s most advanced models, with more generous usage caps, priority access to new features, and performance superior to the free offering. In the public’s perception, Plus is not just an extra layer of convenience: it is often the real premium product, the one that gives access to the “full” OpenAI experience. By potentially offering it to all citizens of a country, OpenAI and Malta are therefore changing the very nature of the object. We are moving from a logic of private consumption to a logic of national capacity.

The original source, published by OpenAI under the title “OpenAI and Malta partner to bring ChatGPT Plus to all citizens”, emphasizes two areas: broadening access and supporting usage. This dual dimension is essential. The recent history of public digitalization has shown that equipping people without training them produces uneven, even counterproductive, results. Generative AI further amplifies this risk: without a culture of evaluation, verification, and understanding model limits, mass access can result in a multiplication of errors, cognitive dependencies, or superficial uses. By integrating training into the partnership, OpenAI and Malta are trying to address a frequent criticism aimed at rapid AI deployments: innovation is not the same as public policy.

What the agreement contains: access, training, and experimentation at the scale of a state

The information communicated by OpenAI describes a national partnership aimed at extending access to ChatGPT Plus to all Maltese citizens. Even if not all operational details have been made public with the same level of precision as a traditional infrastructure contract, the announcement sets out several structuring elements. The first is the universality of the ambition: the tool is not reserved for an administrative elite, a few educational institutions, or subsidized companies. The stated principle is one of broad availability, conceived as a lever for spreading skills.

The second element is training. OpenAI highlights programs intended to encourage concrete and responsible uses of AI. This wording covers several dimensions. On one hand, it is about showing citizens how to use the tool for useful tasks: writing, summarizing, learning, information retrieval, administrative assistance, preparing entrepreneurial projects, language support, or helping small organizations. On the other hand, it is necessary to integrate the now-standard guardrails: verifying responses, protecting personal data, understanding hallucinations, limits on sensitive topics, and distinguishing between decision support and automated decision-making.

The third element, less explicit but central, is the transformation of ChatGPT Plus into a quasi-public digital access service. The term is not officially used by OpenAI, and precision is important: ChatGPT is not becoming a public administration, nor a public service in the traditional legal sense. But the political gesture does indeed consist in treating a commercial service as a capability of general interest. In some respects, this development recalls the way states have historically supported access to the internet, school computer equipment, or educational platforms. The novelty lies in the nature of the service: it is not a network or a portal, but a versatile cognitive intermediary.

The Maltese case also raises very concrete implementation questions. How is eligibility verified? How is the national digital identity linked to an OpenAI account? What categories of data can be used in exchanges? What retention, logging, or protection settings are applied? What governance is in place for uses in education, healthcare, public administration, or the nonprofit sector? At this stage, the public announcement opens more avenues than it provides exhaustive answers. But that is precisely what makes it a laboratory: the interest of the partnership lies as much in its future modalities as in its principle.

Economically, the operation also has symbolic significance. ChatGPT Plus is marketed, in many markets, at around $20 per month before taxes. If this price level is applied theoretically to a population of more than 540,000 inhabitants, the result is a rough order of magnitude above $10 million per month, or more than $120 million per year. Of course, a national partnership is never simply a mechanical multiplication of the public price by the number of inhabitants: it involves specific conditions, real activation rates, usage caps, financing arrangements, and possibly a gradual ramp-up. But this order of magnitude is enough to show that the announcement is not just a communications stunt. It suggests a sufficiently significant commitment to become a case study in the economics of generative AI.

For OpenAI, the interest is multiple. First, the company secures a national testing ground in the European Union, at a time when questions of compliance, liability, and governance are becoming central with the AI Act. Next, it expands its base of premium users within an institutional framework, which can produce valuable aggregated usage data on the real adoption of AI in everyday life. Finally, it strengthens its strategic narrative: OpenAI is not only selling APIs or subscriptions, it is seeking to become a general-purpose utility infrastructure, capable of entering public policy.

For Malta, the calculation is just as readable. The country can position itself as a European pioneer of mainstream AI, attract the attention of investors, regulators, and the media, while offering its population access to an advanced tool without letting individual cost become a barrier. In a country where the economic fabric is made up of many small businesses, services, tourism, international trade, and digital activities, the productivity argument is obvious. Generalized access to AI can support freelancers, students, SMEs, intermediate professions, and public employees, provided that uses are properly framed.

In the background, OpenAI’s announcement raises a simple but decisive question: can a state consider access to a cutting-edge conversational agent as a collective good, in the same way that it supports access to digital education or online administrative services?

Why Malta, and why now: the strategic context of OpenAI and Europe

To understand the significance of the agreement, it must be placed in the context of OpenAI’s recent evolution. The company led by Sam Altman began as a research lab focused on general AI, before becoming one of the most influential players in the global generative AI market. The launch of ChatGPT in November 2022 triggered an unprecedented acceleration in public adoption. In just a few months, the service reached tens of millions of users, then hundreds of millions of uses worldwide, redefining expectations around information retrieval, writing, software assistance, and content creation.

This growth quickly forced OpenAI to change stature. The company was no longer facing only research challenges, but also issues of infrastructure, compliance, monetization, and regulatory diplomacy. The partnership with Microsoft, already structuring, took on decisive importance in the provision of computing capacity and in the integration of models into professional products. But at the same time, OpenAI had to learn how to deal with governments, data protection authorities, sector regulators, and educational institutions. Europe has often been at the forefront of these discussions, particularly after the questions raised in Italy in 2023 about ChatGPT’s compliance with the GDPR.

In this context, Malta offers particularly interesting ground. A member of the European Union since 2004 and of the eurozone since 2008, the country sits at the intersection of several dynamics: small administrative size, strong international openness, institutional bilingualism with Maltese and English, and a recurring desire to position itself as an agile regulatory hub. Its domestic market is too small to threaten OpenAI’s balance in the event of imperfect experimentation, but sufficiently structured to provide a credible political signal at the European level.

Timing also matters. The European AI Act is entering its phase of gradual application, and AI players know that the next battle will be fought less on the raw demonstration of capabilities than on the ability to fit into frameworks of trust. A national partnership like Malta’s allows OpenAI to show that it can work with an EU member state not only as a technology supplier, but as an adoption partner. That is a major difference. Many AI companies have signed agreements with ministries or agencies; few can say they support access at the scale of citizenship.

This announcement must also be read in light of competition. Google is pushing Gemini across its Workspace, Android, and Search ecosystem. Microsoft is deploying Copilot in Windows, Microsoft 365, and development tools. Anthropic is positioning itself on safety and professional uses with Claude. Meta, for its part, is distributing its Llama models in a more open logic, while integrating Meta AI into its mainstream platforms. In the European market, Mistral AI embodies a continental ambition for sovereignty, with in-house models, industrial partnerships, and a discourse favorable to a European alternative to American platforms.

Faced with these strategies, OpenAI is clearly seeking to consolidate another line: making ChatGPT the universal entry point to AI for the greatest number. The Maltese partnership fits perfectly into this logic. Where Microsoft sells an assistance layer integrated into digital work, and where Google couples AI with its existing services, OpenAI is trying to have its conversational interface recognized as a full-fledged usage environment, central enough to justify a public access policy. It is a broader ambition than a simple subscription.

For Europe, the announcement comes at a moment of strategic tension. On one hand, European institutions want to encourage AI innovation, fund computing capacity, support startups, and avoid total dependence on American players. On the other hand, the real uses of generative AI are already massively structured by non-European platforms. The Maltese case makes this tension visible. Offering ChatGPT Plus to all citizens can accelerate familiarization and productivity. But it can also reinforce the centrality of a non-European supplier in the daily digital life of a member state.

A laboratory for public administrations, education, and citizens: promises and limits

The most tangible interest of the partnership is probably on the usage side. For citizens, generalized access to ChatGPT Plus can reduce a form of emerging digital divide: the one separating users who have advanced AI tools from those who remain limited to free versions, traditional search engines, or less capable services. In practice, generative AI is already an efficiency multiplier for some students, consultants, developers, content creators, executives, and freelancers. By lowering the financial barrier, Malta could extend this benefit to broader categories of the population.

In education, the effects can be considerable, but ambivalent. National access to a premium tool can help with personalized learning, course reformulation, language practice, informal tutoring, and the inclusion of students who need differentiated explanations. For a bilingual country like Malta, the ability to navigate between languages, registers, and contexts can represent a practical advantage. But school is also one of the areas where AI raises the most controversy: cheating, outsourcing intellectual effort, difficulty assessing real skills, standardization of output, and dependence on sometimes erroneous answers.

This is why the training component of the partnership is decisive. Training in AI does not just mean learning how to write good prompts. It implies developing a critical culture: knowing when to ask for a summary, when to demand sources, when to verify a figure, when to refuse to use the tool for a sensitive decision, and how to protect one’s data. If Malta succeeds in articulating access and literacy, it could produce a more robust model than the purely opportunistic deployments seen elsewhere.

In public administration, the potential gains are often mentioned but remain difficult to quantify. Public employees can use conversational assistants to draft letters, summarize documents, prepare briefs, classify requests, translate content, or improve the quality of user relations. In a small state, where human resources are not unlimited and versatility is often necessary, a tool like ChatGPT Plus can act as a cross-cutting accelerator. But public administrations also face strong obligations in terms of traceability, confidentiality, neutrality, and accountability. A conversational agent cannot become a decision-making black box.

The Maltese partnership will therefore be watched on one specific point: does the tool remain a productivity assistant, or does it drift toward outsourcing essential public functions? The boundary is crucial. Helping rephrase an administrative text does not have the same implications as suggesting responses to social assistance requests, guiding citizens through complex procedures, or intervening in contexts of health, justice, or taxation. The closer AI gets to the core of public action, the more governance requirements increase.

For businesses, especially SMEs and freelancers, the agreement could produce a rapid effect. In many sectors, generative AI already makes it possible to save time on prospecting, communication, customer service, documentation, monitoring, simple data analysis, or translation. In a country like Malta, strongly oriented toward services, tourism, financial activities, and digital, the spread of ChatGPT Plus can become a lever for competitiveness. The question will then be whether this advantage remains local or whether it quickly dissolves into a global commoditization of tools.

The technical and social limits, meanwhile, do not disappear with a premium subscription. Models remain fallible. They can produce convincing but inaccurate answers, reproduce biases, oversimplify complex problems, or give an impression of excessive mastery. Moreover, equality of access does not guarantee equality of benefit. The most trained, most educated, or most familiar with digital tools often derive more value from AI than audiences more distant from digital technology. Without targeted support, a universal program can paradoxically deepen certain inequalities of use.

The real test will not be how many citizens activate ChatGPT Plus, but how many learn to use it in a useful, critical, and autonomous way.

Digital sovereignty, platform dependence, and the European framework: the underlying questions

The most sensitive angle of the partnership concerns digital sovereignty. When a state facilitates mass access to an AI service operated by an American company, it makes a strategic choice that goes beyond innovation alone. It accepts, at least partially, that an essential layer of its population’s digital cognitive assistance rests on a platform outside its own sphere of industrial control. This does not necessarily mean a loss of sovereignty in the strict sense, but it creates a form of functional dependence.

This dependence can take several forms. The first is economic: if citizens, schools, businesses, and public administrations organize part of their practices around ChatGPT, the cost of exit increases. Changing supplier becomes more difficult, not only for technical reasons, but also for cultural and organizational ones. The second is normative: the rules of use, the model’s limits, product priorities, and functional developments remain largely defined by OpenAI. The third is geopolitical: in an environment marked by technological rivalry between major powers, the concentration of advanced AI capabilities in the hands of a few private American players poses a structural challenge to European autonomy.

The Maltese case is all the more interesting because it is taking place within the European Union, which is precisely trying to build a doctrine of “open technological sovereignty.” The idea is not to close the market to foreign players, but to preserve capacities for choice, negotiation, and substitution. Yet a national partnership with OpenAI can be read in two ways. In an optimistic reading, it allows a member state to experiment with AI at scale while developing local skills, thereby strengthening its future ability to arbitrate between different suppliers. In a more critical reading, it helps normalize the dominance of non-European platforms at the very moment when Europe is seeking to build its own alternatives.

The comparison with other initiatives is instructive. In France, Mistral AI has established itself as the symbol of a European ambition in foundational models. The company has raised considerable amounts, forged partnerships with Microsoft and other industrial players, and is pushing its models into both businesses and the public sector. In Germany, Aleph Alpha long occupied a comparable position, even if its trajectory has been more uneven. At the institutional level, the European Commission is multiplying initiatives around AI factories, high-performance computing, and support for local ecosystems. The partnership between OpenAI and Malta therefore arrives in a landscape where the question is no longer only “should AI be adopted?” but “with whom and under what conditions?”

The issue of data is just as central. OpenAI has gradually evolved its offerings, privacy settings, and commitments for professional users. But a deployment at the scale of a country requires a higher level of clarity. What types of conversations are retained? What guarantees are there for minors? What framework governs uses in public services? How do GDPR obligations interact with personalization, memory, or model improvement features? In a small state, governance incidents can have rapid effects on public trust.

For the French-speaking market, these questions resonate strongly. In France, Belgium, Luxembourg, French-speaking Switzerland, or Quebec, the adoption of generative AI is progressing, but it often remains fragmented between individual initiatives, business experiments, and sectoral directives. The Maltese model could inspire some public decision-makers, particularly in local authorities, education systems, or digital inclusion policies. But it could also provoke reluctance, precisely because it relies on an already dominant American player. In the French context, where debates over sovereign cloud, public procurement, and dependence on hyperscalers are already intense, a “ChatGPT Plus for all” financed or supported by the state would appear as a highly political decision.

Finally, it should be noted that sovereignty is not reduced to the geographic origin of the supplier. A state can gain capacity for action if it uses a foreign platform to train its population, structure its uses, and develop internal skills, provided it retains room for maneuver. Conversely, it can lose autonomy even with a local solution if that solution is opaque, underperforming, or insufficiently governed. The Maltese partnership will therefore be judged less on the symbolism of the flag than on the quality of the guardrails, contractual clauses, transparency, and long-term strategy.

What this precedent changes for French-speaking Europe and the mainstream AI market

The significance of the announcement goes far beyond Malta because it opens a new market category: subsidized or organized national access to premium conversational AIs. Until now, the sector had been structured around three dominant models. The first was the individual subscription, typical of ChatGPT Plus or certain competing premium offerings. The second was B2B or B2G sales, with enterprise or public administration licenses. The third was indirect integration into broader products, such as Microsoft 365, Google Workspace, or smartphones. The Maltese partnership introduces a fourth model: mainstream AI as a collective capability supported at the scale of a country.

For market players, this could alter competition. If other states become interested in this type of arrangement, suppliers will have to offer not only high-performing models, but also governance, training, compliance, and support offerings. The difference in raw quality between models will remain important, but it will no longer be enough. Public administrations will want dashboards, service commitments, security guarantees, language options, support mechanisms, and educational frameworks. In other words, mainstream AI could become more “public” without becoming public, by adopting the attributes expected of a service of general interest.

For the French-speaking world, several scenarios are emerging. A first scenario would see states or regions drawing inspiration from Malta to launch targeted programs: free or heavily subsidized access to AI tools for students, job seekers, teachers, public employees, or SMEs. A second scenario would be more cautious: rather than choosing a single supplier, public authorities would fund interoperable access platforms or digital vouchers allowing users to select their tool. A third, more sovereigntist scenario would consist in first supporting European or open-source solutions, even if that means accepting sometimes less consistent performance in the short term.

France could be a particularly intense ground for debate. On one hand, the French state is pushing AI familiarization, supporting research, funding infrastructure, and highlighting its champions. On the other hand, the real spread of tools in public administration and education remains uneven, with many local experiments but little unified doctrine. The Maltese precedent could fuel discussions on a right of access to advanced tools, particularly in territories where digital inequalities persist. But it would immediately run into questions of public procurement, data protection, language, industrial dependence, and consistency with support for European players.

French and European companies will also watch the potential knock-on effect. If an entire country becomes familiar with ChatGPT Plus, that creates a user base that is more skilled, more demanding, and more inclined to integrate AI into work processes. Local software vendors will have to adapt. Some will seek to connect to the OpenAI ecosystem. Others will bet on sector differentiation, confidentiality, or business specialization. Consulting firms, training organizations, and integrators will also have an increased role, because access to the tool does not solve the question of turning it into measurable productivity gains.

Finally, there is a democratic issue. If conversational AI becomes a common intermediary between citizens and information, between users and services, or between learners and knowledge, then deployment choices also fall within the realm of public debate. Who decides which tools are recommended? What transparency exists about their limits? What place is there for alternatives? What control exists over systemic errors? The partnership between OpenAI and Malta has the merit of making these questions visible. It shows that AI is no longer only a subject for engineers, startups, or large corporations, but a full-fledged object of public policy.

In the long term, the Maltese precedent could herald a deeper reconfiguration of the market. In the same way that access to broadband, educational cloud, or office tools eventually became a matter of structuring choices for states, access to advanced cognitive assistants could become an issue of national competitiveness. Countries that train their populations early in these tools will have an adoption advantage, even if technologies evolve rapidly. But this advantage will only be durable if it is accompanied by an ability to change suppliers, audit uses, develop alternatives, and embed AI in a broader industrial strategy.

That is where the Maltese initiative takes on a European dimension. If it succeeds, it could inspire other governments to consider mainstream AI as an infrastructure for access to knowledge, productivity, and services. If it fails, it will serve as a warning about the risks of becoming too quickly dependent on a single platform. In both cases, OpenAI has already obtained something essential: bringing the question of a premium conversational agent subscription into the field of national policy. For French-speaking countries, the debate now opening is not only about adoption. It is about the kind of digital society we want to build in the era of generative models, between immediate convenience, strategic autonomy, and the collective capacity to govern tools that are gradually becoming the mental infrastructure of everyday life.