I visited deno.com yesterday. I wanted to know if the hundreds of hours I’d spent mastering Deno was a sunk cost. Do I continue building for the runtime, or go back to Node?
deno.com 404 not found error page stating: Sorry, there was an issue loading this page
Well I guess that pretty much sums up why a good chunk of Deno employees left the company over the last week.
Layoffs are what American corpo culture calls firing half the staff. Totally normal practice for a sustainable business. Mass layoffs are deemed better for the moral of those who remain than a weekly culling before Friday beers.
The Romans loved a good decimation.† If I were a purveyor of slop and tortured metaphors, I’d have adorned this post with a deepfake of Ryan Dahl fiddling as Deno burned. But I’m not, so the solemn screenshot will suffice.
† I read Rome, Inc. recently. Not a great book, I’m just explaining the reference.
Deno’s decline
A year ago I wrote about Deno’s decline. The facts, undeterred by my subjective scorn, painted a harsh picture; Deno Land Inc. was failing.
Deno incorporated with $4.9M of seed capital five years ago. They raised a further $21M series A a year later. Napkin math suggests a five year runway for an unprofitable company (I have no idea, I just made that up.)
Coincidentally, after my blog post topped Hacker News — always a pleasure for my inbox — Ryan Dahl (Deno CEO) clapped back on the offical Deno blog:
There’s been some criticism lately about Deno - about Deploy, KV, Fresh, and our momentum in general. You may have seen some of the criticism online; it’s made the rounds in the usual places, and attracted a fair amount of attention.
Some of that criticism is valid. In fact, I think it’s fair to say we’ve had a hand in causing some amount of fear and uncertainty by being too quiet about what we’re working on, and the future direction of our company and products. That’s on us.
Reports of Deno’s Demise Have Been Greatly Exaggerated - Ryan Dahl
Dahl mentioned that adoption had doubled following Deno 2.0.
Since the release of Deno 2 last October - barely over six months ago! - Deno adoption has more than doubled according to our monthly active user metrics.
User base doubling sounds like a flex for a lemonade stand unless you give numbers. I imagine Sequoia Capital expected faster growth regardless. The harsh truth is that Deno’s offerings have failed to capture developers’ attention. I can’t pretend to know why — I was a fanboy myself — but far too few devs care about Deno. On the rare occasions Deno gets attention on the orange site, the comments page reads like in memoriam.
I don’t even think the problem was that Deno Deploy, the main source of revenue, sucked. Deploy was plagued by highly inconsistent isolate start times. Solicited feedback was ignored. Few cared. It took an issue from Wes Bos, one of the most followed devs in the game, for anyone at Deno to wake up. Was Deploy simply a ghost town?
Deno rushed the Deploy relaunched for the end of 2025 and it became “generally available” last month. Anyone using it? Anyone care? The Deno layoffs this week suggest only a miracle would have saved jobs. The writing was on the wall.
Speaking of ghost towns, the JSR YouTube channel is so lonely I feel bad for linking it. I only do because it shows just how little interest some Deno-led projects mustered.
GitHub star history chart comparing NPMX to JSR
JSR floundered partly because Deno was unwilling couldn’t afford to invest in better infrastructure. But like everything else in the Deno ecosystem, users just weren’t interested. What makes a comparable project like NPMX flourish so quickly? Evidently, developers don’t want to replace Node and NPM. They just want what they already have but better; a drop-in improvement without friction.
To Deno and Dahl’s credit, they recognised this with the U-turn on HTTP imports. But the resulting packaging mess made things worse. JSR should have been NPMX. Deno should have gone all-in on package.json but instead we got mixed messaging and confused docs.
I could continue but it would just be cruel to dissect further. I’ve been heavily critical of Deno in the past but I really wanted it to succeed. There were genuinely good people working at Deno who lost their job and that sucks. I hope the Deno runtime survives. It’s a breath of fresh air. B*n has far more bugs and compatibility issues than anyone will admit. Node still has too much friction around TypeScript and ECMAScript modules.
So where does Deno go from here? Over to you, Ryan.
Ryan…
Where is Deno CEO, Ryan Dahl?
Tradition dictates an official PR statement following layoffs. Seems weird not to have one prepared in advance. That said, today is Friday, the day to bury bad news. I may be publishing this mere hours before we hear what happens next…
Given Dahl’s recent tweets and blog post, a pivot to AI might be Deno’s gamble. By the way, it’s rather telling that all the ex-employees posted their departures on Bluesky. What that tells you depends on whether you enjoy your social media alongside Grok undressing women upon request. I digress. Idle speculation has led to baseless rumours of an OpenAI acquisition. I’m not convinced that makes sense but neither does the entire AI industry.
I’m not trying to hate on Dahl but c’mon bro you’re the CEO. What’s next for Deno? Give me users anyone a reason to care. Although if you’re planning a 10× resurgence with automated Mac Minis, I regret asking.