Nigeria’s Proposed 2% Levy on Foreign AI Firms: Digital Sovereignty or a New Cost of Doing Business?

Nigeria’s next battle over artificial intelligence may not be fought over algorithms, safety assessments or the creation of a new regulator. It may be fought over revenue: who earns money from Nigeria’s digital economy, who controls the data that powers it, and how much of that value should remain in the country.

A proposal before the National Assembly seeks to address those questions through a broader digital-sovereignty framework. The Nigerian Digital Sovereignty and Fair Data Compensation Bill, 2025, listed as Senate Bill 722, is reported to target foreign data-service companies through a combination of a levy, local-presence requirements and domestic data-storage obligations.

The reported levy has attracted attention because it has been described as a 2% charge affecting foreign AI companies. But there is an important qualification: the publicly accessible primary material located for this analysis does not establish the precise 2% rate or confirm that the proposal applies specifically to AI companies as a class. The available legislative and policy records describe a wider proposal concerning foreign data-service companies. The final tax base, rate and scope should not be treated as settled without the operative bill text.

That caveat does not diminish the policy question. It sharpens it.

Nigeria has a legitimate interest in ensuring that the companies benefiting from its digital market contribute fairly to its economy and respect its laws. But a levy designed around digital sovereignty can produce very different outcomes depending on what it taxes, how it is administered and whether it strengthens local capacity—or simply increases the cost of accessing global technology.

The central question is whether Nigeria can capture more value from its digital economy without making the infrastructure needed to build that economy more expensive.

The Digital Sovereignty and Fair Data Compensation Bill emerged in a wider debate about the economic power of global technology platforms and the value generated by Nigerian users, businesses and data.

The Senate proposal was introduced on 20 February 2025. Policy trackers describe three central elements: a levy on qualifying foreign data-service companies, a requirement for relevant companies to establish a local Nigerian office, and a proposed requirement to store Nigerian user data within the country.

One reported threshold is annual Nigerian revenue above ₦10 billion. The levy is therefore better understood as part of a proposed framework for larger foreign digital-service businesses than as a charge automatically applicable to every international AI developer.

That distinction matters because AI is not a single commercial category. A foreign company might provide access to a general-purpose AI model, sell cloud computing capacity, license enterprise software, operate an advertising platform or provide data analytics. Some services combine several of these activities.

A rule covering foreign data-service companies could potentially affect businesses that provide AI-related services, even if their principal product is not marketed as artificial intelligence. But whether a particular model provider, cloud platform or software company would qualify depends on the final legal definitions and thresholds.

The bill’s title also points towards a second objective: fair data compensation. That introduces a question beyond ordinary taxation. If a foreign business benefits commercially from data generated in Nigeria, what form should that benefit-sharing take, who should receive it and how should the value be calculated?

These are related but distinct policy problems. A levy collected by government is not automatically compensation paid to individuals whose data has been used. A local office does not guarantee local ownership of technology. And storing data domestically does not, by itself, establish that the data has been collected lawfully or used fairly.

Nigeria’s policy choices need to distinguish these objectives rather than assume that a single charge will achieve all three.

The economic argument behind digital sovereignty is straightforward: a country can generate enormous value through its users, businesses, institutions and information while much of the commercial infrastructure serving that market is owned or controlled elsewhere.

Foreign platforms can monetise advertising, subscriptions, enterprise services, cloud computing and data-driven products. Their local customers may benefit from the technology, but the associated profits, intellectual property and strategic capabilities may remain concentrated outside the country.

For Nigeria, the concern extends beyond tax collection. It includes the development of local technology companies, the availability of digital infrastructure, the ability to enforce domestic law and the bargaining power of Nigerian businesses dealing with global providers.

AI makes the debate more urgent. Modern AI systems depend on computing capacity, models, software platforms, training and operational data. Businesses that rely on foreign infrastructure may gain access to sophisticated tools without having to build those systems themselves. Yet the same dependence can leave them exposed to changing prices, product restrictions, foreign exchange pressures and decisions made by providers over which they have little influence.

A well-designed policy could encourage foreign companies to invest in local operations, build partnerships with Nigerian firms, develop domestic talent and contribute to public revenue. It could also strengthen the government’s ability to identify the businesses operating in its market and hold them accountable.

But digital sovereignty is not achieved simply by charging foreign firms more. It requires domestic capabilities: skilled people, reliable electricity, affordable connectivity, computing infrastructure, research institutions, access to capital and companies capable of building competitive products.

A levy may contribute to that agenda if its design and use support these goals. It cannot substitute for them.

The most important unresolved issue is the reported 2% rate.

The figure has circulated in descriptions of the proposal, but the publicly accessible primary bill material located for this analysis did not allow the rate to be independently verified. The available records support the existence of a proposal to levy qualifying foreign data-service companies; they do not provide a sufficient basis to present the 2% figure as a confirmed provision of the operative bill text.

That distinction is essential for a serious business analysis. A percentage alone does not tell companies what they would owe.

A levy of 2% on gross Nigerian revenue would have a different effect from 2% on taxable profits. A charge on revenue from a defined category of digital services would differ from one applying to all Nigerian revenue earned by a qualifying company. A tax collected from the provider could also be passed through to customers, depending on market conditions and the final rules.

Before companies could assess their exposure, the law would need to clarify at least five points.

First, the tax base: would the levy apply to gross revenue, profits, particular service fees or another measure?

Second, the covered businesses: would liability depend on revenue alone, the nature of the service, a physical-presence test or a combination of factors?

Third, the threshold: how would the reported ₦10 billion revenue threshold be calculated, and would related companies or transactions be aggregated?

Fourth, the treatment of existing taxes: would the levy operate alongside Nigeria’s existing corporate income tax and other applicable obligations, and what rules would prevent inconsistent treatment?

Finally, the administration: which authority would collect the charge, how would foreign companies report revenue, and what dispute-resolution and appeal procedures would apply?

These are not minor drafting details. They determine whether the proposal is a narrow contribution from large foreign service providers or a wider cost affecting the digital economy.

Until the primary text and any subsequent amendments are available, businesses should not calculate liabilities from the reported percentage alone.

The economic consequences of a levy depend on how companies respond.

A large global platform may be able to absorb an additional charge, adjust prices or restructure its operations. A smaller provider may have less room to do so. Some companies could pass the cost to Nigerian customers; others could reduce investment, limit certain products or change the commercial terms under which services are offered.

None of those outcomes is inevitable. They depend on the final design of the levy, competition, customer demand, contractual arrangements and the importance of the Nigerian market to each provider.

But the distribution of costs deserves attention.

Large Nigerian businesses use foreign cloud infrastructure, enterprise software, analytics platforms and AI tools to operate. Startups may depend on external model APIs because building comparable infrastructure locally would be prohibitively expensive. Universities, researchers and small businesses may also rely on internationally provided services to access capabilities that are not readily available domestically.

If the proposal raises the cost of those services, some of the burden could fall on Nigerian businesses and consumers rather than exclusively on foreign shareholders.

The effect could be especially important for local AI developers. Many companies building AI products do not own the underlying models or computing infrastructure. They integrate external tools into products tailored to Nigerian languages, industries and customer needs.

If the cost of using foreign AI infrastructure rises, local developers could face higher operating expenses before they have the scale to negotiate favourable contracts or build alternative systems. A policy intended to support domestic innovation could therefore increase costs for the very businesses it seeks to strengthen.

The counterargument is that foreign providers should not be allowed to benefit from a large market without meeting reasonable fiscal and regulatory obligations. If a levy is predictable, proportionate and transparently administered, businesses may be able to accommodate it without materially changing their investment decisions.

The policy challenge is to determine where that balance lies and to use evidence rather than assumptions to assess it.

The proposed domestic-storage requirement introduces another major question: what would it mean to require Nigerian user data to remain in Nigeria?

Data localisation can support certain public-policy objectives. It may make some data easier for domestic authorities to access under lawful procedures, support the development of local data-centre capacity and encourage investment in domestic infrastructure.

But a blanket localisation rule can also create costs and technical complications.

Many digital services use distributed infrastructure. Data may be backed up across regions, processed by specialised providers and protected through redundancy arrangements. Requiring all relevant data to remain within national borders could force some companies to redesign their systems, contract for additional local capacity or separate Nigerian operations from global services.

Whether those costs are justified depends on which data is covered, the risks the rule is designed to address, the availability of suitable domestic infrastructure and whether carefully defined exceptions are permitted.

Data localisation should also not be confused with data protection. Information can be stored in Nigeria and still be collected excessively, accessed without proper authorisation or exposed through weak security. Conversely, a lawful international transfer can be subject to safeguards designed to protect the information and the people it concerns.

Nigeria already has a binding data-protection framework under the Nigeria Data Protection Act 2023. Any new localisation regime would need to work coherently with that framework and explain what additional protection domestic storage is intended to provide.

The government would also need to consider the capacity of Nigeria’s data-centre sector. Reliable electricity, connectivity, physical security, operational expertise and disaster recovery are all relevant to whether domestic facilities can deliver the resilience expected of modern digital infrastructure.

A localisation mandate without adequate infrastructure could raise costs without delivering a commensurate improvement in security or control.

The proposal’s reported local-office requirement addresses a different problem: the difficulty of regulating a foreign company that earns revenue from Nigerian users but has little or no physical presence in the country.

A local presence can make it easier to identify a responsible entity, serve notices, resolve disputes and engage with regulators. It may also encourage local hiring and investment.

However, a registered office does not automatically mean that a company has substantial operations in Nigeria. Nor does it guarantee that key decisions, technical expertise or control over user data will be located there.

A meaningful requirement would need to specify what constitutes a sufficient local presence, what responsibilities the Nigerian entity would carry and how those responsibilities relate to the foreign parent company. Otherwise, compliance could become a corporate-formality exercise rather than a route to effective accountability.

The government would also need to make the requirements proportionate. A global company generating billions of naira in local revenue has a different capacity to maintain a local operation from a small foreign developer serving a limited number of Nigerian customers.

The reported ₦10 billion threshold suggests an attempt to distinguish larger operators, but the final text would need to make the threshold and its application clear.

The proposal would not arrive in an empty regulatory landscape.

Nigeria already has tax rules addressing certain income earned by non-resident companies with significant economic connections to the country. The Significant Economic Presence Order 2020 established criteria relevant to the taxation of certain non-resident digital and other businesses. The broader corporate-tax framework also applies according to the nature of the business, its activities and the relevant statutory rules.

The new proposal therefore raises a coordination question: would the levy complement existing tax obligations, target a gap in the current system or create an additional layer of liability?

A carefully designed levy could address a clearly identified policy objective that existing rules do not adequately achieve. But a poorly coordinated charge could create uncertainty about which revenue is covered, how obligations overlap and whether similar businesses are being treated consistently.

There is also a distinction between taxing commercial activity and compensating people for the use of their data. Tax revenue generally accrues to government under public-finance rules. Compensation for the use of personal data would require a different explanation of entitlement, calculation and distribution.

If the bill intends to achieve both goals, it should be clear about how the mechanisms relate to each other. Otherwise, the language of “fair data compensation” risks creating expectations that the levy itself may not fulfil.

The proposal is one part of a broader, still-developing approach to AI governance.

Nigeria’s National AI Strategy provides a policy direction for responsible AI development, economic growth, inclusion and technological capacity. Other proposals before the National Assembly address AI governance institutions and the regulation of AI development and deployment. The National Digital Economy and E-Governance Bill, 2025, separately proposes a risk-based framework for AI and other emerging technologies.

These initiatives should not be treated as one consolidated law. They have different legislative histories and objectives, and their provisions may evolve independently.

The distinction is particularly important for foreign AI companies. A levy concerns the financial terms of operating in a market. An AI-governance law may impose duties around risk management, transparency, human oversight or the assessment of potential harm. Data-protection law regulates the processing of personal information. A company could face obligations under more than one framework, but the legal basis for each obligation would need to be established separately.

For businesses, the cumulative effect matters. Multiple rules can be justified when they address different risks. But unclear boundaries can create duplicate reporting, conflicting instructions and higher compliance costs without necessarily improving protection.

Nigeria has an opportunity to coordinate its tax, data-protection and AI-governance frameworks before those overlapping obligations become difficult to navigate.

The international question: can Nigeria attract investment while demanding more?

Nigeria is not alone in debating how digital businesses should contribute to the economies in which they operate. Governments worldwide have considered digital-service taxes, local-presence rules, data governance and measures intended to reduce dependence on foreign technology providers.

But the international comparison must be precise. A levy on digital-service revenue is not automatically an AI tax, and a data-localisation requirement is not the same as a tax measure. Different countries have adopted different combinations of tax rules, privacy obligations, infrastructure policies and industrial strategies.

For Nigeria, the relevant question is not whether another country has imposed a digital levy. It is whether the proposed rules fit Nigeria’s own market, tax system, infrastructure and development priorities.

Predictability is central to that assessment. Investors need to understand which services are covered, how liability is calculated, what exemptions apply and how disputes will be resolved. Local companies need to know whether the rules will increase the cost of the foreign infrastructure on which they depend. Consumers need transparency about whether additional costs will affect prices or access.

A clearly defined, consistently administered framework can provide certainty even when it imposes new obligations. An uncertain one can discourage investment before it generates meaningful revenue.

The proposal raises questions that should be resolved through a published text and an open legislative process.

The rate and tax base must be explicit. The reported 2% figure should be verified against the bill itself, alongside the definition of taxable revenue and the treatment of related entities.

The scope must be technology-neutral where appropriate. If the aim is to regulate foreign data-service businesses, the law should define those services clearly rather than rely on a broad or ambiguous reference to AI.

The threshold must be workable. Companies need to know how the reported ₦10 billion threshold would be calculated, which accounting period applies and how revenue from bundled or cross-border services would be treated.

The localisation rule should be proportionate. Policymakers should identify the data categories and risks that justify domestic storage, consider the capacity of local infrastructure and specify how lawful international transfers would be treated.

The compensation objective needs a transparent mechanism. If the bill is intended to benefit Nigerians whose data contributes to commercial value, the government should explain whether the proposed levy is simply a public-revenue measure or whether a separate compensation mechanism is envisaged.

The framework should be coordinated. The government should clarify how the proposal would interact with existing tax legislation, the Nigeria Data Protection Act and other digital-economy rules.

Finally, the legislative status and current text should be made readily accessible. Businesses and citizens should be able to distinguish the proposal as introduced from any amended version, committee report or final legislation.

The verdict: a sovereignty policy needs more than a levy

Nigeria’s proposed Digital Sovereignty and Fair Data Compensation Bill addresses real questions about the distribution of economic value in the digital economy. Foreign technology companies can play an important role in providing infrastructure and services, but their presence also raises questions about tax contributions, local accountability, data control and domestic technological capacity.

A levy could be one component of a credible response. Yet it cannot, on its own, ensure that Nigerians receive fair compensation for data, that local AI companies become more competitive or that the country develops the infrastructure needed to reduce technological dependence.

Those outcomes require a wider strategy: predictable tax rules, strong data protection, investment in domestic infrastructure, technical education, research, competitive local firms and transparent governance.

The reported 2% rate should remain a provisional figure until it can be confirmed in the operative bill text. The proposal itself is broader than AI, and the publicly available records located for this analysis do not establish that a new levy has become law.

Nigeria’s real test is whether it can translate digital sovereignty from a political objective into a workable economic framework. The best outcome would not simply be more revenue from foreign providers. It would be a digital market in which Nigerian businesses can innovate, consumers can access useful technology, personal data is protected, and the country captures a fairer share of the value its digital economy creates.