Last week, on January 16th, Cloudflare announced that The Astro Technology Company is joining Cloudflare. I read the news on a Friday evening with the slightly-confused-but-mostly-pleased. It’s a BIG news. It’s good news. Probably. I think. But most of all it a financial success.

Why this is good news

Astro is a great framework. I moved this blog onto it last September and I’ve enjoyed every part of working on it since. The simplicity and extensibility without any custom APIs and work arounds is something that I actually enjoy.

The Astro team is small. Open source frameworks at that scale always have the same precarious shape - one or two well-known maintainers, a Discord full of contributors, and a hat-around revenue model that hovers between Patreon and freelance work. Cloudflare buying them means full-time engineers, salaries, runway. That’s the kind of thing that turns “great framework with momentum” into “great framework with longevity”.

It also means Astro stops being a side bet for Cloudflare’s product strategy. As the Cloudflare announcement frames it, content-driven websites are the use case where Astro flatly out-performs Next.js, Remix, and the rest. Cloudflare wants to be the home for those sites. Owning the framework that builds them is, frankly, a smart move.

Cloudflare itself

I have a soft spot for Cloudflare that’s only partly rational. Years ago I read about their wall of lava lamps in the San Francisco office - a hundred-odd lamps blubbing away, filmed by a camera, the noise from the footage fed into the entropy pool that seeds their CSPRNG. The world’s most ridiculous source of cryptographic randomness, and also one that genuinely works. It’s the kind of thing a company does when its engineers are having a good time.

That spirit has shown up in their products too. Workers, Pages, R2, D1 - most of which I’ve used at one point or another, all reasonably priced, most of which work the first time. Cloudflare has been one of the few infrastructure companies of the last decade that has consistently shipped things developers want, at a price that doesn’t read like a hostage note.

So if any big company is going to acquire an open-source framework, “the lava-lamp people” is roughly the best-case version of that sentence.

The bit I’m worried about

The history of open-source acquisitions is, charitably, mixed. The pattern goes something like this: small team gets bought, productivity briefly explodes (they finally have full-time hands), the framework gets tighter integration with the buyer’s commercial product, and somewhere over the next two-to-three years the open core starts looking suspiciously like a funnel into the paid tier.

Sometimes that funnel is fine. (Vercel and Next.js, open source is excellent, the hosted product is convenient, you can leave whenever.)

Sometimes it isn’t. <Insert your favourite acquired-then-strangled OSS project here.>

Astro is small enough today that any change in priorities will be felt immediately. The integrations I rely on might not be Cloudflare’s first-class concern any more. The press release is careful to say “Astro will remain open source”, and Matthew Prince’s quote about “protecting and investing in open source tools” sounds the right note. But the actual test is what gets shipped over the next year.

A few things I’ll be watching for:

  • Adapter parity. Does the Cloudflare adapter race ahead while the Netlify and Vercel adapters quietly stagnate? I’m hosting on Netlify and would prefer that not to be a slow-burning problem.
  • The plugin ecosystem. Astro’s strength is how easy it is to write your own remark/rehype plugins, custom loaders, integrations. Does that surface area get more care, or does it shrink in favour of Cloudflare-flavoured affordances?
  • The pace of features that aren’t tied to Workers. SSG and image optimisation tie into Workers?
  • Governance. Will there be an external steering committee, or is this entirely Cloudflare’s house now?

I’m not pessimistic about any of these. I just want to see the answers play out, not just be assured of them.

Doubling down anyway

Despite all of the above, I’m doubling down on Astro. A few reasons -

  1. The framework is already here. Whatever Cloudflare does or doesn’t do with their acquisition, the version of Astro we are using today already does what we need. The worst case is that we keep using this version forever.
  2. The team is the team. As far as the announcement reads, the same people are still doing the work. The only thing that’s changed is whose payroll they’re on.
  3. Cloudflare benefits from Astro succeeding. Their incentive is to make Astro the best content framework on the web. That’s exactly what I want from it.
  4. Optionality is preserved. Because Astro stays open source, even worst-case I have the source, the community, and the option to fork. Open source acquisitions can go badly, but they almost always leave the existing code intact.

I’m cautiously, slightly, mostly excited. Ask me again in twelve months.

Sources: