@GraalVM Founder and Project Lead. VP at @Oracle. Opinions are my own. For DM: contact@thomaswue.dev
Zurich, Switzerland
- Tweets 856
- Following 85
- Followers 4,862
- Likes 2,018
The work from Amazon to bring with Shenandoah a low-pause GC to @graalvm is really great and the early evaluations are quite promising. Thank you @volker_simonis and team!
Just returned from the @graalvm summit, one of my favorite tech events for the last years: cozy, well organized and deeply technical. I presented "Shenandoah for Substrate VM" which will hopefully land in one of the next feature releases together with G1 GC.
Choice is freedom!
1
25
685
Thomas Wuerthinger retweeted
My favourite event of the year came to an end. Just excellent talks. I think we are set up better than ever. Couldn't be more excited about what is coming.
4
11
597
Thomas Wuerthinger retweeted
Just returned from the @graalvm summit, one of my favorite tech events for the last years: cozy, well organized and deeply technical. I presented "Shenandoah for Substrate VM" which will hopefully land in one of the next feature releases together with G1 GC.
Choice is freedom!
4
10
45
2,811
Looks like TypeScript will get their own version of native image. @GraalVM provides this capability for Kotlin and Java to compete with always AOT compiled languages such as Rust or Go already for some time!
Merging Bun AOT next week, then testing on Claude Code
Expecting 2-3x less memory + CPU & a slightly smaller binary
This is the future of server-side JavaScript
Instead of you rewriting in Rust, we rewrote JavaScript's compiler to emit native code ahead of time
1
1
375
Important: Please read the GraalVM Native Image security advisory regarding the eMudhra certificate compromise and take action: graalvm.org/uploads/native-i…
The release notes: docs.oracle.com/en/graalvm/j…
11
36
3,451
WebAssembly execution in your database, powered by @GraalVM 🚀
The Oracle Database Multilingual Engine now can run Wasm libraries right where your data lives. Learn more: blogs.oracle.com/developers/…
8
22
1,992
Thomas Wuerthinger retweeted
Tonight we are kicking off our favorite event of the year, GraalVM Community Summit 🚀
Looking forward to greeting in Zurich nearly 70 of our core committers, partners, and vendors ❤️
5
7
40
1,266
Thomas Wuerthinger retweeted
Missed #IntelliJIDEAConf 2026?
In “Sandboxing AI Agents With GraalVM,” @thomaswue shows how to run AI-generated code in a sandbox with controlled API access.
Watch on YouTube: piped.video/y8HtFI5JNlE
5
7
40
7,395
The new JetBrains IDE Intelligence is using a great architecture 🏆🚀: „Both the daemon and CLI are developed in Java and distributed as a 𝗚𝗿𝗮𝗮𝗹𝗩𝗠 executable, so there is no startup or runtime overhead from the JVM.“
We've been working a lot(!) on our Java LSP offering and Netflix are among the first to publish the benefits that they saw using it in their agentic flows. Check it out!
blog.jetbrains.com/ai/2026/1…
5
8
121
8,966
Thomas Wuerthinger retweeted
The next generation of @GraalVM is shaping up nicely: fast startup, low CPU and memory usage, with no build step or project changes required! 🚀
Here’s a simple benchmark using a single-file program:
3
18
110
4,395
Thomas Wuerthinger retweeted
Spring Boot shops, banks, insurers, payroll systems, all the backends that quietly run the world: millions of JVM apps can now add a 24B-A2B reasoning model as a plain Maven dependency. Every Liquid AI model is supported.
No infra. No ops. No Python.
1009 tok/s prefill (~2.3x llama.cpp)
47.8 tok/s decode (parity)
qxotic.ai
Today we release 𝙻𝙵𝙼𝟸.𝟻-𝟸𝟺𝙱-𝙰𝟸𝙱, the largest 𝙻𝙵𝙼𝟸.𝟻 family member. It is an efficient local model for laptops and PCs to work through multi-step tasks inside popular agent harnesses.
> 24B total parameters, 2B active per token
> Context expanded to 131K tokens
> A 128K-token vocabulary
> 33T pretraining tokens
> It is a reasoning model
> Very easy to fine-tune and customize
> Performs competitively with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks.
> Open base and post-trained weights on @huggingface 🤗
1/
3
7
19
7,581
Thomas Wuerthinger retweeted
Faster Development with @java, Python, and @micronautfw by @graemerocher and @thomaswue piped.video/watch?v=zo6L7bxn… #micronaut #python #pyronaut #java
4
8
466
Introducing @Pyronautfw — a high-performance framework for Python, built with Micronaut, GraalVM, and GraalPy!🚀
By @graemerocher and @thomaswue at #Devoxx: piped.video/watch?v=zo6L7bxn…
1
10
49
3,459
Thomas Wuerthinger retweeted
The video is out!
My #Devoxx talk, "Quixotic: Pure Java Inference at Native Speed" is now available:
piped.video/S7x4DxwRnk4
1
4
22
1,093
Thomas Wuerthinger retweeted
Introducing @Pyronautfw! 🐍🚀
@thomaswue and @graemerocher live at @Devoxx BE, showing Pyronaut in action: Python and Java coming together through @micronautfw, GraalPy, and @GraalVM – with fast startup, a low memory footprint, and a tight developer feedback loop. 🔥
5
17
626
Introducing Pyronaut — a high-performance framework for Python, built with Micronaut, GraalVM, and GraalPy!🚀
— 2.6x FastAPI throughput, 6.5x Flask 🚀
— Build-time validation 🤖
— Ecosystem of libraries 🌍
Get started: pyronaut.io/
5
9
40
2,456