Amazon recently announced that its Leo satellite internet service is expected to enter South Africa in 2027 through Herotel, the country’s largest fixed internet provider. The partnership, Leo’s first agreement of its kind in Africa, will connect homes and businesses beyond the reach of fibre and fixed wireless networks. Herotel serves more than 350,000 customers in over 550 towns, giving Amazon a route into farming areas and rural communities where distance, terrain and sparse populations make conventional networks costly.
Amazon is entering a market already being transformed by satellite internet. Starlink, the satellite internet service operated by Elon Musk’s SpaceX, served 10.3 million subscribers across 164 countries and markets through about 9,000 satellites as of March, reflecting its rapid expansion throughout Africa, Latin America and Asia.
Satellite services could bring coverage to underserved regions, but whether poorer households can afford them remains uncertain. Yet the promise of universal connectivity carries a deeper risk: The communities brought online by these services may become dependent on foreign corporations that control the infrastructure on which their economies and public life increasingly rely.
Internet access shapes who can study, work, trade, bank, communicate, obtain public services and participate in political life. Network owners decide where they operate, how much people pay and the conditions under which they provide access. In September 2024, Starlink initially defied a Brazilian order to block X, another company controlled by Musk, before reversing course. The episode showed how essential networks give private firms extraordinary leverage over governments.
Amazon and SpaceX are embedded in the national security infrastructure of the United States. SpaceX received a national security launch contract in 2025 covering an expected 28 missions with an anticipated value of $5.9bn to be awarded between the 2025 and 2029 fiscal years while Amazon Web Services (AWS) provides classified cloud services to US defence and intelligence agencies. Their infrastructure, therefore, is already intertwined with US state power. Entrusting them with essential networks across poorer countries would extend American strategic power into their economies, institutions and daily lives.
Of the 2.2 billion people offline, 96 percent live in low- and middle-income countries. Across Africa, only 36 percent use the internet, falling to 21 percent in rural communities. Fixed broadband, where available, costs more than one-quarter of the average monthly income in a low-income country.
Satellite networks reach places neglected by fibre and mobile operators but cannot by themselves overcome the cost of data, devices and electricity. In Nigeria, Starlink raised its monthly residential fee by 50 percent in April 2025 from 38,000 to 57,000 naira ($28 to $42) after an earlier attempt to almost double it met regulatory resistance.
A corporation setting those conditions exercises political power without democratic responsibility. That combination of foreign infrastructure and unaccountable authority has a colonial precedent in Africa.
From 1890 to 1898, a Belgian company built the Matadi-Leopoldville railway to carry rubber and ivory from King Leopold II’s Congo Free State towards Europe. Congolese communities were forced to collect the resources it transported while the colonial regime controlled the route and captured the profits.
Facebook later reproduced that logic digitally.
On February 10, 2015, Facebook and Reliance Communications launched Internet.org in six Indian states before renaming the service Free Basics later that year. Selected news, health, education and employment websites carried no data charges while the wider internet cost money. For millions of poorer Indians coming online for the first time, Facebook risked becoming the internet itself while local businesses and media outside its portal competed against free services.
The prospect provoked a nationwide campaign for net neutrality. Indian regulators ordered Reliance to suspend Free Basics in December 2015, then prohibited content-based data pricing in February 2016. The ruling prevented Facebook from turning poverty into the power to decide what people could read, learn, watch and buy online.
India’s ruling did little to slow Free Basics elsewhere. By July 2019, it was reportedly available in 30 African countries, including Ghana, Kenya and Nigeria. In Ghana, about 70 percent of its services were based outside Africa while major local news sites were absent. In Kenya, users could not reach eCitizen or Huduma Kenya, portals for obtaining identity documents, birth certificates and business registrations. Facebook’s free internet remained distinctly foreign and incomplete.
Free Basics gave Facebook control over a single gateway to the internet. But corporate influence over connectivity increasingly extends beyond services to the infrastructure that carries them. Meta led the consortium that built 2Africa, whose core system reaches 33 countries across Africa, Europe and Asia. Its planned Pearls extension would bring the full system to about 45,000km (28,000 miles). Google’s privately funded Equiano cable links Europe to Togo, Nigeria, Namibia and South Africa.
Amazon plans to connect Leo directly to AWS, allowing governments and businesses to move information into its storage, analytics and artificial intelligence services. In April, it agreed to acquire Globalstar, a US satellite communications company, a deal that would give it satellites, spectrum rights and infrastructure for direct-to-device services.
The danger does not depend on Amazon owning everything carried across Leo. It arises when one foreign corporation supplies several layers of the digital infrastructure on which governments and businesses rely: the connection, cloud storage, analytics and artificial intelligence. Data can travel abroad, be processed into commercial intelligence and return as paid services. When the Global South supplies users and markets while ownership, profits and control remain in the United States, connectivity becomes digital colonialism.
Starlink demonstrates how quickly such power can accumulate. By April 2025, three years after entering Brazil, Starlink had captured nearly four-fifths of the country’s residential satellite broadband market with more than 362,000 individual subscriptions.
Yet Brazil retains public control over part of its satellite infrastructure. Its SGDC satellite has connected more than 12,000 public sites, including 9,380 rural schools and 356 health facilities. This shows how foreign providers can complement public infrastructure without becoming its substitute.
In Southern Africa, Namibia acted before Starlink became indispensable. Its regulator rejected the company’s licence and spectrum applications in March because its wholly foreign-owned subsidiary failed the 51 percent local ownership requirement and other compliance criteria, then dismissed the company’s reconsideration application in June.
Across the border, South Africa illustrates both the role and limits of local participation. While Starlink remains unlicensed, Amazon plans to enter the market through its partnership with Herotel. Under the agreement, Amazon will supply the satellite capacity while the South African company will handle customer service, installation and field operations.
South Africa’s Electronic Communications Act requires individual communications licence holders to have at least 30 percent ownership by historically disadvantaged groups. In December, the government directed the Independent Communications Authority of South Africa (ICASA) to recognise equity-equivalent investment programmes as an alternative to direct ownership. But ICASA said in May that the Electronic Communications Act would have to be amended before such programmes could be fully recognised as an alternative to the statutory ownership requirement.
Herotel will keep distribution and servicing local, but Amazon will retain control over Leo’s satellites, software and corporate strategy. Local participation, therefore, gives South Africans a role in the business, not control over the underlying network.
Greater public control must carry democratic safeguards. In 2025, at least 313 internet shutdowns were recorded in 52 countries, including 30 across 15 African states, cutting access during elections, protests and conflict. Neither a corporation nor a government should be able to suspend lawful access without transparent legal authority, independent review and an effective remedy. Digital sovereignty must belong to the people, not ruling parties.
Licences should require social tariffs, universal service contributions, net neutrality, binding technology transfer and rules for cross-border data transfers. Essential service providers must guarantee continuity while regulators should prohibit exclusive agreements that trap local operators inside foreign networks. Beneficial ownership disclosure must ensure local equity rules benefit communities, not politically connected intermediaries.
Licensing can restrain corporate power but cannot create ownership. Because satellite networks and technology companies operate across borders, national regulation alone is not enough. The African Union should use its African Digital Compact and Data Policy Framework to develop common satellite licensing rules with the African Telecommunications Union on spectrum and competition.
Latin America’s eLAC2026 and the Association of Southeast Asian Nations Digital Masterplan 2030 can support comparable standards. Governments should pool demand and direct regional development banks to finance publicly controlled fibre, internet exchanges, data centres and shared satellite capacity.
Amazon Leo and Starlink can help connect communities abandoned by governments and conventional operators, but connectivity should expand freedom rather than transfer power to foreign corporations.
The Global South’s digital future must not belong to Jeff Bezos and Elon Musk.
The views expressed in this article are the author’s own and do not necessarily reflect Al Jazeera’s editorial stance.
