Finding Unlikely Allies

Will Robinson Avatar

Will Robinson

·

Apr 28

Share

The Hygienists Made Pasteur

In 1878, Louis Pasteur realized that microbes were killing us, but no one wanted to believe him. Doctors didn’t think invisibly small creatures could be deadly. Surgeons didn’t want to wash their hands. Germ theory might have been interesting to niche researchers, but it had no momentum to change society.

So how did Pasteur convince the medical establishment? Bruno Latour, one of the founders of Science and Technology Studies, explains in The Pasteurization of France that Pasteur needed unlikely allies. He found a group of people with a completely different problem, figured out that his work could be the lever they’d been looking for, and put it in their hands.

The group was the Hygienists, public health reformers who had been trying for decades to clean up French cities. They weren’t rigorous scientists, and kept accumulating theories for the cause of disease: bad air, bad water, bad drainage, even bad morals. They couldn’t focus their energies because they had nothing concrete to point to. But when Pasteur showed up with a theory for why water and air can go bad, suddenly they had something to champion.

And when Pasteur realized this, he doubled down on them. He staged his anthrax vaccination trial at Pouilly-le-Fort as a public spectacle, with journalists and veterinarians watching. He wrote for the hygienists’ journals, in their language, framing his lab results as answers to their questions about sewers and epidemics. He let them translate his chemistry into their campaigns for vaccination laws and sanitation infrastructure. A few dozen Pasteurians in a Paris laboratory became, through the hygienists’ megaphone, a national movement.

The part that matters: none of this was luck. Pasteur’s science was real, but science alone wasn’t going to win. He won because he identified, early, the one group in French society whose structural problem his work could solve. Then he spent years deliberately feeding them what they needed. The alliance looked unlikely from the outside, which is exactly why it worked. No one was competing for it.

The Same Move, In Crypto

Take Rain (ALL7). Farooq and Charles started a card-issuing platform in 2021 and did something most fintech founders would have thought impossible. They became a Visa Principal Member, the rare status that lets them sponsor card programs without a bank in the middle. On paper, that’s a credential. In practice, it was the start of an alliance.

Visa, a $600B incumbent, had a specific problem: how does a legacy card network stay relevant as money moves to stablecoins? Rain, by existing and working, was the cleanest answer Visa could point to. So Visa pointed.

  • Rain was invited into Visa’s USDC settlement pilot, and announced it jointly, with Visa’s Head of Growth Products speaking on the record.
  • Rain’s CEO went on Visa’s own podcast, hosted by Visa’s Head of Crypto.
  • When Rain expanded into APAC, Visa’s APAC Crypto Lead was quoted in the press release endorsing the move.

This is not a vendor relationship. This is Visa using Rain as a marquee case study in Visa’s own story about its future. Neither company is doing the other a favor. Their structural problems interlock. As of last quarter Rain is valued near $2b, growing at absurd speeds.

In fact, Rain grew large enough to be a platform for a new generation of Alliance startups. For example, last year, Offramp (ALL13) was navigating the friction of bringing dollar-denominated spending to Latin America. They required exactly what Rain had spent four years building. Rain identified Offramp as a marquee launch partner; Offramp gained a Visa-ready card program and a global compliance engine without having to engineer either from scratch. Rain, in turn, received real-world validation of its thesis in the emerging markets that mattered most. Both are Alliance companies. Neither was a coincidence.

How To Actually Do This

When startups at Alliance ask me for advice around distribution, my first question is always: what’s your current stack? My goal is to figure out who might need them to succeed as much as I do. It often takes founders a while to switch their thinking. They’re so focused on solving their customers’ problems, they forget they’re solving their platforms’ problems too.

I don’t ask who they know or who likes them. I ask who needs them to work. Those are the unlikely allies. Most of them don’t know you exist yet. But the day you ship something real, their incentives turn on, and if you’ve picked well, they’ll promote you freely.

And it goes further than marketing. Many of these platforms have their own venture funds. Coinbase, Solana, and Mysten have all invested in Alliance startups, looking to support the companies whose growth also grows their own ecosystems.

It’s almost never enough to just build a great product. Even when preventing entire swathes of life-threatening illness, Pasteur’s research wasn’t enough on its own. Without powerful distribution allies, he might never have socialized his science.

If you’re a founder building something and looking for new allies: Apply to Alliance.

Share


More to Read

Fintech: Request for Startups

Qiao Wang Avatar

Qiao Wang

·

Dec 15, 2025

While it may seem that the low hanging fruit from the digitization of finance has been picked, about half of the world population is still either physical-first or outright unbanked. Furthermore, current technology waves – namely crypto and AI – have the potential to unlock greater value for consumers and businesses.


You're Vibe Coding Wrong

Mohamed Elseidy Avatar

Mohamed Elseidy

·

Jul 10, 2025

Most developers use Tier 1 for UX & rapid prototyping, Tier 2 for daily development, and Tier 3 for complex engineering challenges. The best strategy is to combine them when needed rather than trying to force one tool to do everything. Here's what each tier actually delivers and when you should use them


Your Moat Isn't Your AI Prompt. It's Your Evaluations

Mohamed Elseidy Avatar

Mohamed Elseidy

·

Jun 17, 2025

Anyone can copy a prompt, but they can't replicate the months of iterative learning, the domain knowledge and the edge cases discovered through real user interactions, and the systematic evaluation framework that guided every prompt tweaking decision.