Connect first, argue about ownership later
South Africa delayed Starlink for years to protect an ownership principle. Ontario built a rural internet plan on Starlink, then had to cancel it. Both got the sequence backwards.
Two governments spent the last two years arguing with the same satellite company, and they made opposite mistakes.
South Africa spent years keeping Starlink out to protect a principle. Ontario let Starlink in without one, then had to pay to undo it. Between them is a genuinely hard question that a lot of countries, and a lot of businesses, are about to face: when you can’t build critical infrastructure yourself, what order do you solve access and control in?
The country that said no first
South Africa requires telecom operators to hand 30% ownership to Black South African partners, a rule from its Black Economic Empowerment program, built to redress ownership patterns left over from apartheid. Starlink refused to comply, and for years that refusal kept the service out of a country where rural clinics, schools, and farms have some of the worst broadband access on the continent.
Elon Musk turned this into a grievance, posting that he “cannot get a license to operate Starlink” in the country he was born in “because I’m not Black”, which is a convenient way to skip past the fact that the rule applies to every foreign telecom operator, not to him personally. But the self-serving framing didn’t make the underlying delay less real. People without a fiber connection stayed without one while the principle was defended.
In December, the government finally created a workaround: an “equity equivalent” path letting companies invest in the country instead of ceding ownership. Starlink’s opening move was to commit roughly $30 million to connect 5,000 rural schools. The rule didn’t get scrapped, and it shouldn’t have; ownership of local telecom infrastructure by local people is a legitimate goal. What changed was the order of operations. The country found a way to let people connect now and keep negotiating the equity question in parallel, instead of making the second thing a gate in front of the first.
The province that said yes first
Ontario ran the experiment in the other direction. It signed a $100 million deal with Starlink to bring internet to 15,000 underserved rural and First Nations locations, no ownership fight, no equity requirement, just a contract with the fastest available option. Then Washington imposed tariffs on Canadian goods, Musk’s role in the Trump administration made him a visible target, and Premier Doug Ford cancelled the deal as retaliation and made clear it wasn’t coming back even after the tariffs were paused.
Ontario had no sovereignty question to answer going in, because it never asked one. That’s exactly why cancelling cost something. The households that were meant to get connected are still waiting, and Ontario ate a settlement fee to walk away from a contract it signed with its eyes open. Access without any thought to control isn’t neutral. It’s a bet that the relationship stays boring, and Musk’s politics made sure it didn’t.
Two sequences, one lesson
Line the two up and the lesson isn’t “protect sovereignty” or “just connect people.” It’s that sovereignty and access solve different problems on different timelines, and treating one as a precondition for the other gets the order wrong in both directions. South Africa made ownership a gate and paid for it in years of people staying offline for a fight that had nothing to do with them. Ontario made access the only question and paid for it the moment the relationship turned political.
The better sequence, and the one South Africa eventually backed into, is to connect first on whatever terms are actually available, and negotiate for control once the thing you’re depending on has proven valuable enough to be worth negotiating over. You can’t credibly bargain for ownership of something you’ve decided to live without. And you’re most exposed the moment you’ve bet on something you don’t own without ever asking what it costs to walk away.
I’ve run into a smaller version of this building financial technology in India. Every fintech company rents infrastructure it will never own: cloud compute, SMS gateways, banking rails, UPI itself. The instinct to insist on owning the stack before shipping anything kills companies that never needed to own it in the first place. The instinct that costs you later is the opposite one: renting something critical without ever asking what happens if the vendor’s incentives stop lining up with yours, or what it costs to leave once you’re dependent. Ontario’s contract is that question answered the expensive way.
Ask a village in the Free State without a fiber line whether it would rather wait for the ownership fight to resolve first, or get connected now and let the argument catch up. I think the answer is obvious, and it’s worth designing policy, and contracts, around that being obvious.