The shortest path to technological sovereignty rarely passes through manufacturing. South Africa has just illustrated this by opening a sixth national quantum research hub, anchored at the University of Pretoria, without seeking to produce a single qubit locally. The wager is different: to train researchers capable of programming quantum computers hosted thousands of kilometers away, of writing the algorithms these machines will execute, and thereby to build expertise that, it is hoped, will remain on the continent.

This choice stands in sharp contrast to the usual temptation toward technological autarky. It also raises a question that theorists of industrial policy have debated for decades: can one build a national capacity by deliberately relying on foreign infrastructure, or is this simply dependence rebranded as strategy?

The Essentials

In 2025, Africa attracted $10.7 million in investments in quantum technologies, compared to global amounts estimated between $12 and $57 billion depending on the methodology employed, with some very high estimates mentioning even larger figures but including strategic orientation funds not yet spent. Within this context of structural underfinancing, South Africa has chosen an original path: concentrating its limited resources on the software and algorithmic layer of quantum, accessing foreign machines via the cloud. The opening of a sixth national hub financed over five years, in Pretoria, embodies this strategy. The principal identified obstacle is human: the retention of locally trained researchers remains weakened by persistent brain drain.

The Financing Gap Tells Only Part of the Story

The numbers are brutal. In 2025, the entire African continent captured $10.7 million in quantum investments. The United States, Europe, and China absorbed the bulk of global investments, estimated between $12 and $57 billion depending on reference sources—McKinsey placing the low end of the range around $12.6 billion, the QED-C recording $56.7 billion in cumulative public commitments, while broader estimates including strategic funds not yet spent advance significantly higher figures. The gap with the African continent remains, in any case, abyssal. Presented in these terms, it seems to condemn in advance any African ambition in this sector.

But this framing is misleading if one stops there. Investment in material capital constitutes one path toward quantum competence, and not the only one. Quantum computers themselves, in their current state, remain unstable, noisy, and accessible via the cloud to anyone with an account at IBM Quantum, Amazon Braket, or Microsoft Azure Quantum. This distributed architecture of access changes the geography of research. A researcher in Pretoria can execute an algorithm on a 127-qubit quantum processor hosted in New York. What he cannot do alone is build this processor or maintain it.

South Africa has decided to concentrate its resources on what its means permit. This is the sense of the choice documented by SA QuTI, the South African Quantum Technology Initiative, and relayed by The Quantum Insider in July 2026.

Six Hubs, a Logic of Networks

The new Pretoria hub is the sixth in a network that is gradually spreading across the country. Each is anchored to a university and specialized on a dimension of quantum: algorithms, post-quantum cryptography, molecular simulation, quantum sensors. Financing is assured over five years, which in the context of African research represents a stability rarely seen.

SA QuTI coordinates the whole. The structure recalls what France put in place with its National Quantum Plan launched in 2021, except that Paris chose to finance simultaneously the material and software layers, with budgets of incomparable magnitude. South Africa, meanwhile, has made a clear-cut choice: to concentrate the effort on what can be done locally with available means.

This logic of networks has precedents in other African technology sectors. The mobile banking industry, of which Kenya was the first global laboratory with M-Pesa starting in 2007, relied on local design of services adapted to local constraints, using telephone infrastructure built by foreign equipment manufacturers. The result was an application-layer innovation that was subsequently exported to other emerging markets.

Quantum is different on one crucial point: the level of expertise required is incomparably higher. Training an M-Pesa developer takes months. Training a researcher capable of writing competitive quantum algorithms takes years of advanced mathematics.

The Strategy of the Software Link Facing Its Critics

The intellectual tension between two readings of industrial policy plays out here most clearly.

Anaïs Voy-Gillis, specialist in reindustrialization and industrial strategy, defends the idea that countries that renounce mastery of the material layer often, in the end, renounce mastery altogether. In her analysis of industrial value chains, the software layer without material anchorage remains vulnerable: it depends on design choices, proprietary formats, and commercial policies of those who manufacture the machines. If IBM decides tomorrow to restrict cloud access to certain countries, or to modify its architecture in a manner incompatible with algorithms developed in Pretoria, dependence becomes visible.

This reading is serious. It applies fully to quantum: quantum computers are not generic servers. Their physical architecture determines the types of algorithms that execute efficiently on them. A researcher who optimizes his algorithms for the topology of an IBM Heron processor has not necessarily developed expertise directly transferable to a competing system.

But the critique has an answer, and it is pragmatic. Daron Acemoglu, in his work on institutions and technology, emphasizes that the pertinent question concerns position in the value-creation chain and the capacity to negotiate, more than sovereignty in absolute terms. A country that trains researchers capable of writing competitive quantum algorithms possesses real negotiating leverage against a country that trains none, even if both use foreign machines. Competence creates leverage, even without the machine.

The European Union, whose preparation for threats related to quantum rests partly on partnerships with non-European actors, moreover illustrates that even large technology powers practice partial and strategic sovereignty rather than integral sovereignty.

Brain Drain, Principal Obstacle

The true vulnerability of the South African model is human.

Training a quantum researcher in Pretoria is expensive and time-consuming. Once graduated, this researcher can join a laboratory in Zurich, a startup in San Francisco, or a public program in Singapore for a significantly higher salary and equipment without equal. The pressure is structural. It is explained by the objective conditions offered by the global market for quantum competencies, which works in favor of already-established centers.

This question of researcher retention is identified as a central obstacle of the program. The envisaged answers combine several levers: doctoral scholarships paired with return commitments, structured partnerships with international institutions that allow mobility without permanent rupture, and the gradual creation of a local ecosystem that makes return or staying more attractive.

None of these answers is new. They are the same ones that India, South Korea, or Brazil employed with variable success depending on the periods. India lost a generation of its best engineers in the 1980s-1990s, before seeing part of this diaspora contribute to building technological bridges between Silicon Valley and Bangalore. The trajectory is not linear.

The question posed by the 2026-2035 horizon is this: can the institutional partnerships that the SA QuTI hubs are forming create sufficient critical mass for the South African ecosystem itself to become attractive to its own graduates? The five-year financing provides a window. It does not guarantee the answer.

Partnerships as Partial Substitute for Machines

The Pretoria model rests in large part on formal agreements with institutions of the Global North. These partnerships offer South African researchers access to machines, training data, and peer communities. They also create, in the best cases, coauthorships and publications that feed the international reputation of local laboratories.

This strategy of international connection recalls what research on innovation in contexts of constrained resources has documented under the name of “frugal innovation”: doing more with less, by concentrating on the links of the chain where comparative advantage is real. For South Africa, this comparative advantage lies in the capacity to train researchers bilingual between local problems and global tools, to test use cases specific to the continent (logistics, health, agriculture) on world platforms.

The continent does not start from nothing on this front. The dynamics of financing technological development in Africa show that the most dynamic African ecosystems have often begun with application before ascending toward infrastructure. The South African quantum trajectory fits this pattern, at a substantially higher altitude of scientific abstraction.

The Africa Quantum Consortium, which brings together several countries of the continent around programs of shared training and research, broadens this network beyond national borders. UNECA, in its recent reports on the African economy, points to quantum as one of the sectors where a coordinated continental strategy could create scale effects inaccessible to each country in isolation.

The Pretoria Wager and Sovereignty in the Quantum Age

Technological sovereignty has always been graduated, even for great powers. The United States uses Chinese rare earths. Europe buys Taiwanese processors. Japan develops AI algorithms on American GPUs. The issue is at which link of the chain a country builds its position and what margin of maneuver this position affords it.

South Africa has chosen the algorithmic link. This choice is defensible in the short term, provided that financing holds, that international partnerships remain balanced, and that trained researchers remain accessible to the local ecosystem, physically or remotely. It remains conditional in the medium term: if material quantum architectures diverge significantly and if cloud access becomes contingent or priced differently depending on geographies, the value of algorithmic expertise depends on the persistence of access to machines.

The sixth Pretoria hub constitutes a first pragmatic response to a real resource constraint. And the manner in which South Africa manages the retention of its researchers and the diversification of its technological partnerships in the coming years will say much about the viability of the model for other resource-limited countries that are already looking toward Pretoria.

Sources

  1. The Quantum Insider, University of Pretoria Quantum Research Hub (July 2026)
  2. SA QuTI, South African Quantum Technology Initiative (no stable verified link)
  3. Nature Africa, Africa’s quantum leap (no stable verified link)
  4. Africa Quantum Consortium, institutional partnership reports (no stable verified link)
  5. UNECA, Economic Report on Africa 2026 (no stable verified link)
  6. Journal d’un Progressiste, La France finance sa filière quantique
  7. Journal d’un Progressiste, L’Europe se prépare à une attaque quantique
  8. Journal d’un Progressiste, Grandir à crédit, selon l’endroit où l’on entreprend
  9. University of Pretoria, official press release on UPQuST
  10. SA QuTI, official website
  11. TechCentral, ‘South Africa’s quantum bet starts to leave the lab’
  12. Ecofin Agency / UNECA, quantum investments Africa vs. world
  13. Élysée.fr, launch of France’s quantum strategy January 2021
  14. IBM Quantum, official hardware
  15. Africa Quantum Consortium, official website
  16. McKinsey — Quantum Technology Investment 2025