The r&d isn't expensive because these are legacy/old type chips. There isn't much r&d happening. They're made using old processes.
Any r&d is likely using different materials. Military stuff often uses different materials for specific applications.
The profits come from raising yields, which is much easier on older nodes.
Given the low transistor count, you could probably achieve speeds in the multi-10s of GHz, it would be so small.
A multi-core 99/4A would be hilarious to witness.
Similarly hyping up total R&D spending without a tight timeframe is meaningless, they will eventually spend X$ assuming the company survives indefinitely.
More charitably there’s meaningful differences between nodes, but it’s likely a significant portion of the workforce will be rehires.
https://www.texastribune.org/2025/03/12/texas-congress-corny...
Fabs have pretty decent electricity demand, I actually sited a battery against that one after we did a bit of modeling with the new load vs the solar flows in the area. Siting just means make a land deal, figure out easements, and start working its way through the queue. Might not ever get built though, lot of money and hurdles in the way ha.
Every time I see an industrial announcement like this power is the first thing I think of - where it’s gonna come from and what impact it will have on the existing grid.
If you have a beating pulse and a strong renewable tech IP, you will raise a seed successfully from the Saudi PIF or Cheveron Ventures even in this market. Sadly, this is HN, not Bookface, so most ideas are bad ideas. Sucks too because it was a very fruitful demo days - one of the more impactful over the past several years.
CHIPS invested heavily in OSAT and Packaging capacity - especially in TX. Samsung, Micron, OmSemi, and TI took full advantage of that under the Biden admin while Intel and TSMC were fighting on the airwaves to undermine each other.
Much of the rest has been reinvested in SK and India (TI is part of the SCL Mohali modernization RFP) as part of the QUAD+ initiative.
A lot of us in the private and public sector working in this space aren't idiots.
"we are honored to work alongside them and the U.S. government to unleash what’s next in American innovation.”
They're expecting to get money from government to make this happen. If it will be done is anyone's guess.
I’d love it if things were streamlined when building. Just as a homeowner it’s amazing how hard it can be to get a permit or an inspection. The lame duct tape fix seems to me some weird CMS and ArcGIS, as if that somehow fixes the human and policy problems.
[0] https://www.reddit.com/r/gimlet/comments/a3lyny/reply_all_13...
Step 2: stroke Trump's ego
Step 3: receive money
Step 4: foxconn their way to the bank
How are they going to afford an investment thats ~1/3rd the value of the company? Seems like one of those announcements that no one follows up on to keep them honest?
Much like that “initiative”, I imagine this won’t materialize either, they’re largely hoping government funds get steered toward them and by which point they’ll pocket everything they can while delivering a minimum of what was promised, if anything at all, and they won’t be properly held accountable for it.
He wants credit for everything while doing nothing.
https://www.reuters.com/technology/artificial-intelligence/t...
That was a project by Oracle to build some data centres, many of which had already broke ground and were under construction, and they pulled OpenAI and Softbank in to give it that AI veneer.
So it's stuff that is really happening, and was already happening. Similarly this TI plant was announced under the prior administration as a win of the chips act, and again it's being recycled to current.
This sort of shell game works for a while, but a couple of years the bill comes due.
Construction of the first data center Abilene, Texas site is ongoing. Emily Chang visited the site over a month ago and got video footage [1]. So it's not nothing. Will they actually invest 500 billion in the end? I dunno; I guess we'll see. The main challenges seem to be the availability of electrical power and skilled workers, not demand, investment or ambition.
Obviously $500B is a lot of money, but it's not actually an unbelievable amount of money.
Go to flyover States to see the impact of that money.
I don’t think Illinois has actually awarded any of that funding to providers to build anything yet. It looks like the original schedule was to start awarding grants this summer after planning process from 2021-2025.
I see some of the things funded by the IIJA every day of my life.
And that's not including the projects which are still being worked on getting approved including projects to replace sewage lines in my area, reworking stormwater drainage systems, and additional flood control projects, all of which are planned to be funded through the IIJA and all of which I think my city could really use. 60% of the IIJA funding hasn't been spent yet.
Then you are part of the problem, pull your fingers out of your ears.
The linked article doesn’t even try to complete the argument… there’s not a single reason offered why TI could plausibly have increased total revenues more… instead of just spinning their wheels from a higher cost base.
Or to put it another way, why wouldn’t that revenue growth have gone to their competitors?, which did in fact happen.
And still.
General Electric was (just a few decades ago) the largest non-gold asset, and isn't even Top 50 anymore.
Saudi Aramco was (just a few years ago) the largest non-gold asset, and isn't even Top 10 anymore.
If you want to read about similarities in Xerox's own hubris/downfall, Dealers of Lightning is one of my favorite non-fiction reads of the past few years (about their SFbay PARC Labs of the 70s/80s... and its failures to adapt).
The "Computer Associates" of today is... Salesforce.
Have they really improved Tableau, Heroku, Mulesoft, Sendgrid, Slack, now Informatica?
• Meta (all of it)
• Google (not necessarily the rest of Alphabet)
• Apple — UI updates are not innovation, and there's a lot of bugs
• X (the social media website, no comment on the new parent AI company)
• The car-shaped bits of Tesla — the humanoid robotics bits are at least "future", even if Tesla is not demonstrating anything new in this spaceUnfortunately, my mid-sized US city still doesn't have an Apple Store... so I haven't demoed the VisionPro, yet; but I want to be able to see if its endorseable for my Parkinsons friend (whose eyes still work well); I suspect if I demoed it I'd also purchase one too.
Absolutely seems innovative.
To me, it seems to have the same problems as the Google Glass, in that it's far too expensive and doesn't have a clear idea what its own USP is.
That said, while I've played with a few different VR headsets, I've not had a chance to play with the VP, so perhaps there's something in the quality that would become visible if not for the prohibitive price.
But you don't have to start expensive: cheaper headsets already existed for several years before it came out, and those are perfectly adequate games consoles; and of the reviews I've seen for home cinema and virtual monitor uses, nobody seems to prefer AVP over other headsets half the price.
This madman's extraterrestrial dream™[0] will do more harm than not.
[0] Nightmare - ¡yay let's destroy Earth to build a dependant-upon-Earth colony! /s
Also: if you want to drive a vehicle that actually saves the planet, buy a hybrid (plug-in, or not).
For Mars… well, I like the idea of settling other worlds and expanding and yada yada yada, but Musk has repeatedly demonstrated he's not got the personality to actually do that right and not lock people out of the base for disagreeing with him. Plus, if you Muntz* a life support system or a food supply or the walls on a Mars base, everyone can die before replacements arrive.
At least, so long as they're not of the [buzzword]+[existing product] variety of startup. A decade ago we had "Uber for $[dictionary merge all nouns]", and today you can easily find a lot of people trying to shoehorn "blockchain" into things without any apparent understanding that other ledgers exist. Not those kinds of startup, those are all just copycat pander-to-the-investors schemes at best, and magical thinking at worst.
On a similar note about madness and vision: SpaceX, Rocket Lab, and SpinLaunch. SpaceX clearly has Mars as a big futuristic vision, and unlike say Blue Origin they actually make rapid progress in that direction; Rocket Lab may be small fry compared to SpaceX, but are thinking along similar scale for launch vehicles and have the political advantage of not being Elon Musk, so if Brand Musk implodes as I now expect it to, they may end up doing many similar things in practice; SpinLaunch may not be going anywhere fast, but they're aiming for a very different launch system and are therefore will end up in the history books even if they run out of investors.
A few years ago I would have also said Relativity Space, but these days everyone is 3D printing rockets. I've even seen a small 3D printed rocket engine in person, on display at one of the smaller booths in GITEX Europe last month, so Relativity Space no longer stands out for 3D printing rockets.
I have no idea where any of the AI companies are going. It looks like the sector is in a stage where there's a lot of low-hanging fruit, and therefore rapid improvements are relatively easy to stumble upon. But the same dynamics mean that's a red-ocean environment, whereas (IMO) the best innovation is a blue-ocean environment: https://www.blueoceanstrategy.com/tools/red-ocean-vs-blue-oc...
Hadn't heard the red- vs blue-ocean distinction before. Thanks. The bullet points are similar to Thiel's zero to one thesis. (Right?)
FWIW, I'm currently most excited by efforts addressing climate crisis. Like advanced geothermal, better heat pumps, improving the grid, etc.
If I understand, most would be considered blue-ocean.
Most definitely face the chicken & egg challenge. As in the merit and benefits are apparent. But someone has to underwrite lofting a new market(s) from scratch.
Cheers, Jason
To echo [1] the Xerox example, the C-suite became entirely made-up of former salesmen who'd become addicted to the pavlovian `click`ing sound their massive copiers made in bill-per-page leases... which blinded them from realizing the digitizing world wasn't going to favor their no-risk-taking approach.
`click` - `click` - `click` – late 60's ASMR erotica for Xerox VPs.
[1] amazon.com/Dealers-Lightning-Xerox-PARC-Computer-ebook/dp/B0029PBVCA
Companies can do financial tricks but end users can’t.
Any company large/influential enough to get in front of trump right now is gonna be in the news saying this stuff. The entire admin runs on "the last person to convince trump".
Trump pardoned a bunch of criminals, drug dealers, and other trash people because the right person talked to him at the right time.
This is exactly what people said would happen. It's open season for bribery
Not to mention the fact that we refused to hold the person who tried to overthrow our government accountable for doing so.
If the Germans had taken the same approach in 1945, the Nazis would’ve been back in power by 1950.
His popularity has decreased somewhat, but given a choice between him and the opposition, I suspect that they would continue to do the same thing.
He may indeed pass but I'm not so sure that would change much.
The hyperbole about 2024 being the last election in American history will seem silly in retrospect to most, but those who really believe it now will for the most part never admit they got swept up in hysteria and will instead pat themselves on the backs and credit themselves for America narrowly avoiding disaster. A generation or two later, young people will just roll their eyes when their parents talk about any of it, one way or the other, like millennials did when their parents rambled on about Reagan or Nixon.
Being from 70's generation, with family that had to their encounters with PIDE/DGS, and took part on the liberation forces, I see the supposed hyperbole with different eyes.
Hopefully I am wrong.
When in doubt, bet on the status quo. When the status quo falters, bet on a return to the status quo.
Now, if there are other reasons to do this like for defense or supply chain security, that's a different story. But it will cost you.
Knowing these facts changes your perspective on these semis-in-the-USA news headlines.
Intel makes something like 50% - 75% of theirs in the US.
This seems to be a political term, not one the electronics industry uses.[1] "Foundational chips (also called “legacy,” “lagging edge,” and “mature node” semiconductors) are often defined as chips made with a 22nm manufacturing process or above."
Is there actually a lack of 22nm and larger fab capacity in the US? Or is it just that they're not being used much.
[1] https://selectcommitteeontheccp.house.gov/sites/evo-subsites...
What PA and Fairchild did. We haven’t had a new-entrant opening for fundamental semiconductor design for decades, a fact laid bare in their top engineers’ comp. (No clue if this is what they are actually doing.)
ON Semi is the successor of Motorola's semiconductor business.
What is it that Fairchild and "PA" did that makes them foundational?
Currently it's fabbed in France (STMicro), Germany (GF), and South Korea (Samsung). No plans to onshore that in the US.
Jensen Huang also interviewed at TI but took his first semiconductor job at AMD instead.
I hate doing a first article, seeing nothing wrong on the IC side of the LED board, then apply power and have the IC explode and rocket off the board, taking traces and pads with it (or if I'm lucky, it just craters its packaging.)
And it is BAD. I'm talking 1/20 kind of bad, off multiple reels (I've gone though tens of thousands of them.) Only TI stuff.
The fact sheets (linked on the page) amount to $51bn from what I can see.
https://www.ti.com/about-ti/newsroom/news-releases/2024/2024...
> In December, the [prior] administration finalized a $1.61 billion government subsidy for Texas Instruments to support construction of three new facilities after the company announced plans to invest at least $18 billion under the $52.7 billion CHIPS and Science bill.
[1] "Texas Instruments plans $60 billion US investment under Trump push" https://www.reuters.com/business/texas-instruments-plans-inv...
Edit: Did the HN link change? I thought the link was to a TI press release with few details.
Don't get me wrong, you need that stuff from last millennium—not just 180nm but also the 6 μm I suspect "foundational" actually refers to. You need power regulation, precision measurement, and RF frontends in analog, and analog can't scale down the way digital can. And TI actually retains a world-leading position in those chips, with better and cheaper ICs than anything out of Mainland China today. But it's potentially a game of chasing China's taillights.
SiGe might be an exception; even at 130nm, IHP's SiGe BiCMOS process can hit 450GHz oscillation frequency and 350GHz ft, so even at fairly coarse process node sizes, SiGe could be a strategic enabling technology for submillimeter communication such as Starlink, which is reportedly critical for weapons such as Ukraine's naval drones.
TMWNN•7mo ago
CNBC: <https://www.cnbc.com/2025/06/18/texas-instruments-plans-60-b...>