googles multiple businesses and gemini isn't the largest one.
anthropic is the anchor external customer of tpu's and nvidia is worth more than all of google. If tpu's actually breakout as a viable alternative over the next few years for multiple clients the business could easily be worth as much as search, maybe more.
You essentially have to run in google to use them and that probably limits their ability to breakout. Anthropic might be doing this deal as a way to shore up their supply chain and cost of both inference and training by leveraging Google's hardware and chip manufacturing expertise.
there are literally not enough tpu's on earth for them to break out, every tpu thats been made is in use, the spike in demand is recent and google has heavy competition for foundry space.
Possibly because they just haven't been able to manufacture enough of them yet to be a viable business to others? They're fighting everyone else for foundry space and time.
TPUs are not that portable and easy for both inferencing and training. It has since improved a lot with their effort on the torch backend (XLA/TorchTPU) and JAX though.
But as far as i know it currently supports just that + tensorflow (which nobody uses it anymore, least here). And last we tried, so much of our kernels needs rework that it’s not worth the effort.
This may change since ironwood but we haven’t tried that generation.
Of course this is well known. Everything Microsoft does is for selfish capitalist reasons and everything Apple does is for altruistic philanthropic reasons.
> Microsoft in 1997 investing $150 million in Apple, saving it from near bankruptcy.
If only Apple could pass the favor forward. But no, they can't be bothered to invest even a single million in Asashi Linux to benefit their own hardware.
Well, the way Apple sees it, their hardware ships with a perfectly good OS. And they know their hardware has so far been so good that even if they ruin said OS, they'd still make money. What we (and Apple, as competition) need is a serious macbook-killer with full Linux support.
It just keeps the lights on for the whole industry.
The tech is great but valuations are out of control. It's cheaper to keep valuations high through these circular financing deals, rather than to allow for any deflation.
It was more complicated than that. Microsoft was only kind of a competitor to Apple. Apple was also a platform for them. They shipped software that was popular on the Mac (mainly Word).
Especially in those days Microsoft was both a platform for software to run on, and a maker of software, and being flexible to emphasize one or the other aspect depending on the way the market is... has been good for them.
What's the explanation behind this? I am sure they use AI in their ad network (matching web sites with ad offerings, maybe generating ads automatically), but is there more to it?
I know AI companies are selling ad training into the models so the models know about your product. I'm not sure if that is what they were referring to, but it could be related.
That was precisely my thought on seeing the news. I did not know about Google's existing entanglements with anthropic, but it seemed like a clear message - Do not panic on the money, do the work.
If you look at their recent actions, they all seem financial as if they have become the monopoly already and can do anything. Maybe it is driven by fear of going bankrupt
Example. Them doing a AB test where they remove Claude CLI from the 20$ pro plan ... they rolled it back now. Other rate limits where they publicly double your quota at NON peak times but lower it during peak quietly. These are tacky and signs of panic.
One such issue is experimentation. But when you see back to back issues, it looks odd.
I love this! I learned my first logic with Crocodile Clips: would simulate and solder a lot of D flip flop based logic before learning about microcontrollers (my first was Atmel AT90S2313)
I assume their chips don't really exist until they're actually supported by ESP-IDF. The ESP32-C5 was announced in June 2022, received initial support in -IDF in August 2025, and more complete support in December. It seems to have only recently started getting third party dev boards.
I have great news for you: after several years of growing frustration with Espressif's inability to launch this chip properly in NA, I found a company called Wireless Tag that presumably felt the same way and just did it themselves:
I've now used this module in several projects. I love it. And I love (x3) the P4. It is amazingly powerful.
A lot of folks talk about the P4 not having radios as a problem; I personally think that it's an advantage. The assumption that every device is a wifi/BT device is baffling to me.
You'd have a very hard time convincing me to use anything but the WT0132P4-A1 at this point. They are cheaper than ESP32-S3-MINI-1U, too.
> A lot of folks talk about the P4 not having radios as a problem; I personally think that it's an advantage.
I believe, the problem is that there isn't a comparable chip with WiFi/BT (ignoring S3, because, you know, Xtensa), not that P4 itself doesn't have them.
I don't use these directly myself in any projects, so I'm citing hearsay... but from what I understand, Espressif officially recommends using an ESP32-C6 as a radio module with the P4. Wouldn't that cover it?
What I have seen directly is a lot of folks reacting negatively to the P4 because it doesn't have radios. They seem to be coming from a "what could it possibly be useful for if it can't [wifi/BT]". While it's easy to see this as a failure of imagination, it does seem true that a lot of folks equate the ESP32 line as what you use when you want to create an IoT device. While that's not wrong or necessarily bad, I've always felt like it's a weird way to pigeonhole an entire SoC family that might be self-limiting.
I guess it hit mass production and then suddenly recalled and is now in sample status again with “X” variant. I designed boards using esp32-p4 and deployed and then suddenly local supply of esp32-p4 went dry. Now waiting for the x variant to go mass production but you can by samples directly from espressif for 3.33$.
I once worked on a problem: GPS tracking shipping containers, since one company had almost 1% lost/stolen each year. I had an idea of using AIS with Si4362 to get positioning data from the container ship itself, but it was nearly impossible to get access to reefer monitoring systems. We ended up just using 4G NB-IoT for coastal tracking and it did solve the problem
reply