Hello,

I caught wind of a new DC-link capacitor technology that enabled high temperature operation, which I know from my own traction inverter design forays has been the weakest link in long term reliability. So, I invited THEM to write about their tech as a contributed article, which I then put a very light editing touch on and got into our publishing queue for our readers to know and use. Perfect if you're designing a high voltage inverter, converter, or VFD, in my opinion. I'd love to see them do a 600V version of the tech, but 900V rating is where they start in their roadmap.

If you have some whiz bang stuff you'd like to write about as a deep technical dive for our readers, please feel free to hit me up with an email with your ideas. Can't be fluffy, not a sales pitch, not looking for a PhD thesis, and needs to be an interesting read for your peers worldwide - I can work with you on some of those aspects.

It's always been a pain to convert third party device models to LTSPICE, especially for those of us that don't do it very often and who have better things they want to do, so we've got a step by step article on how to do it which should be a perfect reference to squirrel away into your design tricks folder.

Rohm's included EROM models for some of its shunt resistors, enabling accurate modeling of device temperatures in Flotherm.

With Tesla's backsies on offering Full Self Driving and with OpenAI on shakey ground, 2026 and 2027 will bring an interesting shakeup of the bottom feeders in AI processing - we provide an idea of where things may go as the churn and burn happens over the next year or so. I'm a big fan of ML and generative design, and was pre-"AI", but the biggest way to turn me off is to tell me about the "AI" in your handheld belt sander.

enjoy,

-andyT

From Around the Web

DuckDuckGo poll says 90% of responders don’t want AI

https://www.pcworld.com/article/3044488/90-of-duckduckgo-users-dont-want-ai.html

Extended use of LLM-based AI leads to atrophy in the pre-frontal cortex, which creates dummified users and higher dependence on....AI. Meanwhile, all  available data has been mined, so AI is now training on AI and increased processing power is no longer scaling.

https://m.youtube.com/watch?v=z2TH5ietCZQ

Unlike the popular media's take, this video by Nate Jones alleges that the recent Amazon layoffs of 30,000 people is merely a trade of human capital for infrastructure capital—buying GPUs at a time when the company had a miserable -$4.8B (a net outflow) in free cash flow:

https://m.youtube.com/watch?v=7sk3qmIQZnI

Mediocre guitar player incrementally designs a magnetically suspended guitar bridge which sort-of works. He doesn't seem to understand that force can be increased by decreasing magnet gap length and relies on weak materials in high tensile strength materials tasks:

https://m.youtube.com/watch?v=ueCO4spGNPs

Watch Top Gear's "Stig" drive McMurtry Automotive's EV to clean up both an F1 car's record and the track itself like a vacuum cleaner:

https://m.youtube.com/watch?v=NDfKhBcGh9w

Nvidia’s plan to invest up to $100 billion in OpenAI to help it train and run its latest artificial-intelligence models has stalled after some inside the chip giant expressed doubts about the deal:

https://archive.ph/BXlAP#selection-558.0-558.1

Court orders restart of all US offshore wind construction. Trump admin’s “it’s classified” ploy put on hold in five different cases allowing wind farm construction to proceed. All five farms are expected to be completed before the appellate courts can act.

https://arstechnica.com/science/2026/02/court-orders-restart-of-all-us-offshore-wind-construction/

Cue the meme of the guy holding a woman's hand and turning and admiring a woman they had just passed...Despite its latest software having almost zero improvement over the prior release, glass-housed OpenAI throws stones and is allegedly unsatisfied with some Nvidia chips and is looking for alternatives, sources say:

https://www.reuters.com/business/openai-is-unsatisfied-with-some-nvidia-chips-looking-alternatives-sources-say-2026-02-02/

As AI struggled with power delivery and cooling, foundry services providers like TSMC found their urgent-to-ramp-up AI chips stored on shelves. They're wisely not falling for the hype or the excuses for sloppy coding as "Silicon Valley Thinks TSMC is Braking the AI Boom":

https://m.youtube.com/watch?v=2lLFBun1qR0