Will the Teensy 4.0 and 4.1 be discontinued ?

nounours18200

Well-known member
Hello,

I have read here: https://www.adafruit.com/product/4622
that the Teensy 4.0 & 4.1 will be discontinued, seem

The Teensy boards will be discontinued due to supply changes. Get the ones we have on sale now! For the latest on Teensy, please visit PJRC forums for more information. We’re working on an open-source, community-friendly alternative — stay tuned!

Does this mean that Adafruit will stop selling Teensy 4.0/4.1 , or does this mean that the Teensy 4.0/4.1 will be no longer manufactured ??

Thanks
 
They're not discontinued products. Long story short, Adafruit had a dispute with Sparkfun (who manufacture the Teensys) and ended up having their supply cut off.
 
Is the software (“Tennsy Core”) still being actively developed?

We've had 3 releases so far in 2026, in February, May and June.

Not as visible, we've also had 2 severe supply chain disruptions in 2026, first with flash memory and recently with Schottky diodes. Like in the pandemic, those take a huge amount of time away from software development. :(
 
Over the years, the Teensy platform set the gold standard for microcontrollers, driven by both exceptional hardware and a vibrant community that actively shaped the core.

Looking at recent releases, however, updates seem heavily dominated by toolchain bumps and IDE compatibility fixes, while substantial core features and independent bugfixes have become rare. This trend is mirrored on GitHub, where classic community PRs sit untouched for years, and there sometimes seems to be limited interest in integrating external bugfixes

Maintaining such a complex ecosystem requires quality control, but this heavily centralized "Cathedral" approach is stalling community participation. When well-tested PRs remain ignored indefinitely, contributors eventually give up.

I’m sharing this as constructive feedback because I want to see the platform thrive. It would be amazing to see a bit more of that collaborative open-source spirit return—whether by empowering trusted community members to help triage PRs or establishing a clearer path for integrating community fixes into the core.
 
I'd say the hardware is extreme overkill for most of my uses, but maybe that's just me. I mean a 600Mhz CPU is almost as fast as a PC from 25 years ago. I simply wouldn't know what to do with such power. Also, it facilitates sloppy programming or usage of slow interpreters.

Will Teensy LC remain in production? I haven't purchased one but since most of my stuff is for simple control (i.e. volume and source switching in a pre-amplifier, audio player with limited GUI, wireless rocket ignition, room thermostat) it's sufficient for my use cases. I still have a few Teensy 2.0's lying around here so for now I'm supplied.

I also have a few Teensy 3.5 and 3.6 lying about but haven't yet found a project to use them in.
 
Last edited:
Ah well, I'll have to look elsewhere then.

I'm wondering what projects people are cooking up with a 600Mhz MCU.
 
I'd say the hardware is extreme overkill for most of my uses, but maybe that's just me. I mean a 600Mhz CPU is almost as fast as a PC from 25 years ago. I simply wouldn't know what to do with such power. Also, it facilitates sloppy programming or usage of slow interpreters.

Will Teensy LC remain in production? I haven't purchased one but since most of my stuff is for simple control (i.e. volume and source switching in a pre-amplifier, audio player with limited GUI, wireless rocket ignition, room thermostat) it's sufficient for my use cases. I still have a few Teensy 2.0's lying around here so for now I'm supplied.

I also have a few Teensy 3.5 and 3.6 lying about but haven't yet found a project to use them in.
I am using every bit of that overkill. To the point of wishing it had more RAM1 and RAM2 but i do work around. It is amazing how much this little guy can do!
 
Ah well, I'll have to look elsewhere then.

I'm wondering what projects people are cooking up with a 600Mhz MCU.
My TeensyMIDIPolySynth (TMPS), implements 12-poly 3-voice digital synth capability (where each of the 36 note VFOs allows selective octave control per voice, full mixing of square, sine, triangle, & sawtooth waveforms plus strings per note, each individual note chain includes modulation, filtering, envelope control, & where each modulator LFO includes similar fully mixable waveforms plus pulse, sample/hold, & pink/white noise, filtering, envelope control, plus volume level per voice & overall level, ladder filter & reverb on the final mix of all MIDI generated notes).

Control of all of these capabilities is done from a 7" 800x480 RA8875 TFT touchscreen, with code managing the homebrew UI running on a T4.1 at 600MHz (58000+ SLOC). All of the audio components are processed on a T4.0 running at 816MHz (22000+ SLOC). Hardware wisde, the box supports traditional 5-pin DIN MIDI input/output, USBmidi input/output, & USBsound output, as well as headphone & line-level outputs.

The Teensy 4.x is truly a BEAST & I'm using every bit of the processing speed that it has to offer. I am extremely appreciative of what @PaulStoffregen provides (both hardware & software). Throughtout the development of these capabilities, I have done lots of optimization to get everything to fit in the available RAM (TMPS started out as a 16-poly device, but as capabilities were added, the poly had to be reduced to continue being able to fit in the RAM available).

That's my example & I'm sticking to it . . . LOL

Mark J Culross
KD5RXT
 
I worked in the Hypercube parallel processing 'supercomputer' industry in the mid 80's. We were the first to demonstrate a working 32-node system based on the new 80286/80186 chips at the Oakridge National Labs Hypercube conference. The 32-node system cost about $150,000.

A $30 Teensy 4.x is faster.
 
I use set_arm_clock() to dynamically change the clock rate on the Teensy 4.0 and 4.1. I typically use 72 MHz when I don't need compute for awhile, then bump up to 600 MHz for a short time to do all the straight-line math, then back it down to 72 MHz. I do have a clock with six separate LCD displays. Once a second I bump up to 600 MHz to calculate the graphics and update the displays, then back down to 72 MHz to handle asynchronous events like button presses and GNSS input. If I ever reprogram that clock I'll probably set it to something like 120 MHz and leave it there.
 
Ah well, I'll have to look elsewhere then.

I'm wondering what projects people are cooking up with a 600Mhz MCU.
I built an electronic leadscrew for my lathe with various modes, fully variable thread pitches on-the-fly.

I tried to port it to esp32.....failed horribly. esp32 has to many pin conflict issues
 
If one only could switch off easily the microSD card between writes, as one can do with SPI
CMD7 (and CMD48/49 to access the power management register - if existant)
 
Last edited:
That's amazing. There are only a handful of pins on the ESP that you need to be careful about. All the others can be freely remapped, even I2S.
Teensy is way more straight forward, and when you need to use every pin you can, those handfull of pins mean a great deal
 
Teensy is way more straight forward, and when you need to use every pin you can, those handfull of pins mean a great deal
Which ESP32? I have the opposite experience. I max'd out my T4.1 pin usage and it was a struggle because despite having a couple of alt sets of pins for interfaces, they're still fixed and it was quite a dance to try to get all the interfaces I needed wired up. When I ported this to an ESP32-P4 board, I could mux any pin to any interface, and that particular issue went away, it was quite liberating.
 
Which ESP32?
Exactly.

I think the main issue people run into with the ESP32 in general is that there are hundreds of variations since there is no barrier to entry.

Anyone can easily design a new one, randomly assign GPIO pins on the chip to be whatever physical pin they want it to be on the module and what functionality to assign to that pin out of the box and sell it as a new product. That is why there is such a ridiculously long board list for them which is continually changing.

In that respect, muxing pins differently between different ESP32 modules as defined by the module designer adds an undesirable variability when switching between modules, even the same basic ones like the ESP32-C3 for instance. Then layer on boot functionality to some pins, the different number of pins made available by different modules and then the many different chip variations with different capabilities, along with little to no documentation at the final module level for many of them and it becomes a bit of a tangled mess since there is no single control point like with Teensy.

I have spent time over the last week trying to document some relatively well documented small ESP32 modules (C5, C6, S3) and it isn't easy to get solid info since nothing is 'set in stone' so to speak. Even something as seemingly straightforward as how many analog inputs a module can actually support can be surprisingly difficult to figure out compared to something like a Teensy which I think is probably @MorryStu main point and I am in full agreement on that point.

On the other hand, the ability to mux pins arbitrarily by the end-user as you are talking about is a great feature to have for integration into the end project and would be nice to have on future Teensy.

I would imagine though that there is probably a tradeoff to be made between achieving maximum performance vs maximum GPIO multiplexing flexibility so doubt NXP would/could go to the extend that Espressif does. That may be why they have the relatively limited multiplexing setup in the current chip which is near useless.
 
Exactly.

I think the main issue people run into with the ESP32 in general is that there are hundreds of variations since there is no barrier to entry.

Anyone can easily design a new one, randomly assign GPIO pins on the chip to be whatever physical pin they want it to be on the module and what functionality to assign to that pin out of the box and sell it as a new product. That is why there is such a ridiculously long board list for them which is continually changing.

In that respect, muxing pins differently between different ESP32 modules as defined by the module designer adds an undesirable variability when switching between modules, even the same basic ones like the ESP32-C3 for instance. Then layer on boot functionality to some pins, the different number of pins made available by different modules and then the many different chip variations with different capabilities, along with little to no documentation at the final module level for many of them and it becomes a bit of a tangled mess since there is no single control point like with Teensy.

I have spent time over the last week trying to document some relatively well documented small ESP32 modules (C5, C6, S3) and it isn't easy to get solid info since nothing is 'set in stone' so to speak. Even something as seemingly straightforward as how many analog inputs a module can actually support can be surprisingly difficult to figure out compared to something like a Teensy which I think is probably @MorryStu main point and I am in full agreement on that point.

On the other hand, the ability to mux pins arbitrarily by the end-user as you are talking about is a great feature to have for integration into the end project and would be nice to have on future Teensy.

I would imagine though that there is probably a tradeoff to be made between achieving maximum performance vs maximum GPIO multiplexing flexibility so doubt NXP would/could go to the extend that Espressif does. That may be why they have the relatively limited multiplexing setup in the current chip which is near useless.

Yes, the Teensy just made it so easy.
 
Anyone can easily design a new one ... such a ridiculously long board list for them
Just 398 folders in packages\esp32\hardware\esp32\3.3.11\variants
25 by adafruit​
26 by m5stack​
28 by waveshare​
even if another 100 or 200 are dupes by other makers - that leaves some 50-100 that might be one offs

And the ESP32 variants alone make it tough to find the perfect one when the next one is better ...
 
Back
Top