> logic technology can extend for the first time below the 1 nm node, advancing the era of angstrom-level scaling, where dimensions approach the size of individual atoms. While transistor nodes now refer to a generation of manufacturing technology versus an exact physical dimension, IBM’s 0.7 nm technology—also referred to as 7 angstroms—demonstrates how continued scaling remains possible.
Continuing the well established trend of making bold claims about physical dimensions that have nothing to do with any of the structures in the chip, and the name scales better than the tech.
What they actually deliver is a "nanostack architecture" built with ~5nm features that according to them is comparable to a hypothetical real sub-1nm chip.
It's an impressive achievement nonetheless but it looks like the industry has a few too many marketers.
My read on it was that they are trying to imply a transistor density (in a 2D plane sense) that is comparable to a 1nm process? But they achieve that through stacking (3D, not 2D) since the features aren't actually anywhere near 1nm?
Correction: it’s not stacking. They do things like FinFET (turning the gates in the third dimension) and gate-all-around which increases the density of transistors per unit area. But they don’t have layers of transistors. At least in the logic and analog processes.
I really can't see where the 0.7nm is coming from. The white line looks like it's just an edge of a feature that is "15 rows of silicon atoms", which by some quick arithmetic on Wolfram Alpha has to be AT LEAST ~1.6nm, and the way the rows of atoms appear to be packed in that image and by the provided scale, it seems to be significantly more. Using the white line as a meaningful measurement seems to me to be more misleading than any other interpretation here.
A 0.7 nm planar transistor made of silicon has no performance, because a device so small cannot function as a transistor.
The intended meaning of "0.7 nm" is that if you compare the transistor density per area of a "0.7 nm" manufacturing process with that of a "350 nm" process (like used for some Pentium II CPUs, at a time when "350 nm" was a real length), the ratio between the transistor densities is (350 nm / 0.7 nm)^2 = 500^2 = 250,000.
Comparing with the number of transistors of a Pentium II, a 0.7 nm CPU should be able to contain about 5000 billion transistors. This is consistent with the fact that the latest 3 nm NVIDIA Rubin GPU has 336 billion transistors and a 0.7 nm circuit must have a density around 16 times greater than a 3 nm circuit.
However, for many of the modern node names used by some companies even this computation is not really true, because marketing may have chosen an arbitrary name that is smaller than for the last process of the main competitor.
For now, IBM has not provided any kind of information that could prove their claim that their new CMOS process has the transistor density corresponding to "0.7 nm" (i.e. 16 times greater than the TSMC "3 nm" CMOS process).
> Continuing the well established trend of making bold claims about physical dimensions that have nothing to do with any of the structures in the chip, and the name scales better than the tech.
We care about PPA (power, performance, area) and not how large or not-large features actually are. Comparing gate lengths between a 1980s planar transistor and a 2010s 3D FinFET or GAA transistor is obviously nonsense, the relatively aligned node names of the industry actually do make sense as a shortcut here.
It's been decades since published node sizes had any connection to actual feature size. Sadly this is just how it works in the semiconductor industry now.
Smart money would be on the first person who realized that the expertise required to understand the technical details had grown beyond that possessed by 51%+ of stock investing (trading?) population as weighted by transaction volume.
>So essentially, since 1997, the node name has not been a representation of any actual dimension on the chip, and it has erred in both directions by almost a factor of 2.
We've already been through years of "7nm isn't actually 7nm" across different fabs - completely different measurement conventions, none corresponding to real feature sizes. Now sub-1nm? If it is real then at that scale we're probably in the several atom width territory.
As it can be seen from the photos, horizontally the features are much bigger than 5 nm.
For silicon, the gate length of a FET has a lower limit somewhere between 10 nm and 15 nm.
The current CMOS manufacturing processes have not reached the limit yet. For making smaller transistors, a transition to other semiconductor materials will be necessary.
The vertical thicknesses of various layers may be of only a few nanometers or even of a fraction of a nanometer, but that does not matter directly for the circuit density.
The supposed node size refers to horizontal dimensions, not to vertical dimensions.
Vertical dimensions of around 1 nanometer or less could be achieved already many decades ago, because they depend on growth speed and on time, not on lithography, like the horizontal dimensions.
The industry should have stopped decades ago to talk about the "size" but they should have characterized a CMOS process by its density, e.g. in logic gates per square mm.
However, an actual concrete number would be disliked by marketing, because they could no longer claim that their "1 nm" process is better than the "2 nm" process of another vendor, if their density is not really better.
The scale bar on the far right "photo" (micrograph?) doesn't make sense. It is only slightly less than half the scale bar on the middle photo (10 nm), but the image is clearly scaled up by much more than 2x. Individual silicon atoms are circled in the right photo, but the covalent radius of silicon is about 0.11 nm, so they should be much smaller if the scale bar is accurate.
The scale bar also got about 50% longer, which would imply a 3x zoom. That also seems about right based on the relative feature sizes. Same thing happened between the first and second image.
Better metrics are transistors/mm^2, performance/watt, and raw performance, since at this point "nm" is fluff and easily game-able.
Different companies measure it differently too. This was a while ago, but I remember reading that Intel 10nm was more or less close to TSMC 7nm. I'm sure this is still true to varying degrees.
Comparing nodes across foundries is a kind of a coin toss. At least, from the name. You actually need to go into the specifics of the pdk and process to understand what features there are. Can’t rely on the name for anything.
Unlike marketing terms, "nm density" is actually useful measure.
It describes density measure where you can compare it to planar transistors from the 28-nanometer (28 nm) node around 2010 to 2011 and before. A "0.7 nm" node has equivalent transistor density as if we could have shrunk standard flat transistor node down to 0.7 nanometers.
It's a physical quantity per some unit of spatial measurement so the units still don't match up b/c in one case the transistors are stacked per volume & in the other case per area.
> Historically, "node" sizes (like 28nm or 7nm) directly correlated to the physical length of a transistor's gate. Today, names like 3nm or 2nm reflect a marketing generation. The actual transistors are significantly larger than these nanometer labels, meaning density varies between companies
> Research organizations like IEEE have proposed new metrics, such as transistors per cubic millimeter (MTr/mm^3), to accurately map future 3D scaling. However, commercial chip foundries resist this change because it would make it harder to calculate commercial yields and thermal density limits using standard industry formulas.
> It's a physical quantity per some unit of spatial measurement so the units still don't match up b/c in one case the transistors are stacked per volume & in the other case per area.
However the active devices are still just one layer. This isn't like 3D NAND where you actually have transistors on top of each other. So the comparison only considers the area for both kinds of transistors.
The comparison is good. Humans are also stacked in volumes and we still measure population density over a surface because the third dimension is less significant in this context.
I got the answer from the AI I was looking for & it makes sense. You can try to map the volumetric density to areal density but the mapping is not canonical so it doesn't say anything about the physical reality of actual transistor density since the reality is that it is a volumetric measure that gets fudged for marketing purposes. 3D volume for chips is going to keep increasing so they will eventually transition to measuring density over volume instead of area.
3d population density would be an interesting measure, though. It would provide a better feel for how cramped a populated volume actually is.
It would have to be the displacement volume, to account for cities with only a few really tall buildings, and also somehow adjusted for buildings with high ceilings.
All "density" means is that it's a quotient; I use spectral densities daily and that's "count per Hertz" which weirdly enough works out in normal units to be seconds.
Density does directly scale with both of those in the form of more chips per die (=> lower cost) and smaller capacitance (=> less dynamic power dissipation).
If you want to reduce "effectiveness" of some process down to a single number, then density is far from the worst metric to pick.
Why is that all you care about? Stepping down a node gets you dramatically improved timing and design feasibility. The reduced density means you can pack the same design into less area. Your most challenging timing paths now have to traverse a shorter distance, and you can fit more of them relative to certain node-size invariant structures
Just to be clear, this doesn't mean that anything on the die actually measures 0.7nm — it means that it's roughly double the density as the previous node generation. At some point the industry decided to keep talking about "nanometers" even though the actual transistor sizes have been decoupled from the node name for years.
Node sizes have been "generations" for a long time. The number is just the previous number divided by sqrt(2), rounded in an arbitrary direction. Sqrt(2) because reducing each dimension by that number leads to double the density, and it's non-obvious enough to sound like a real measurement
I am just astonished at the quality of the structure. Not only that, but the quality of the "cut" to then be able to take this picture. I was closely involved in EM imaging more than 30 years ago, this is just wonderful to see such pictures now.
IBM has been the company with the most patent registrations in the US for I think 29 of the last 30 years. They're one of the largest industrial research organizations in the world. They're doing more hard science research than almost anyone else.
Which is so weird, right? Like what is IBM now and how does a research lab make sense with the rest of their business?
The money-making parts of IBM are: legacy software and hardware (declining), consulting (low margin, low leverage), enterprise software (mostly redhat, not really growing). It's hard to explain how IBM research is accretive to any of that.
I don't know enough about their business to say, but I'm thrilled at even the idea that someone might actually value long-term success over quarterly earnings.
IBM at one time had nearly everything in modern tech under their roof like Xerox and squandered it. There is no comeback for them.
They were the second American chip company that said no to Steve Jobs when hinted about designing smaller better chip mobile devices the other two were Motorola and Intel Apple had to eventually do it themselves. Apple Silicon
Licensing is a substantial source of revenue, and their servers have very impressive (think Telum’s caching) innovations, even though they rely on third-parties for manufacturing the chips themselves.
They are also betting on quantum computing to become commercially relevant.
The hardware division has 80%+ margin and still makes the systems that process 75% of all financial transactions. Their processors for those systems are on par or better than any other, I don’t think that is a business at all. This cash cow is not going away any time soon and gives them the profits to make bets on the future of computing.
It's just a shame that none of it seems to pan out, and in the areas where I know what they're talking about, it all sounds like cynical nonsense to me.
I really wish I had followed through when I was ask by some of the guys at IBM Research to apply when we had worked together on a partner project. Though I didn’t have a degree which I seem to remember was a sticking point, this is in the mid 2010s
The labs might not be that different from consulting, the NYT reporting on this notes they run R&D labs so they can license the tech they develop to people who actually make chips.
Most of their fabs were divested to GlobalFoundries, but they still have pretty significant fab capability and capacity- I suspect at least partly to have a us-based chip-making for military ("Trusted Foundry").
I mean, you can't get smaller than an atom, there is some amount of plausibility of using individual atoms as at least the occasional computing element.
Beyond that, engineering a quark-gluon plasma as a processor? I'd watch that Star Trek episode. (we might fantasize about stuff like that but we're roughly monkeys smashing rocks together in a cave vs. building an iPhone sort of gap away from that kind of thing unless somebody has a really good idea)
You could, in principle, use photons and/or electrons. We got pretty damn close in the vacuum tube era, and photonic computing has been a popular research topic for a while.
You also have quantum computing, which I think can/does use subatomic particles? Not sure about that one
You can’t make smaller chips features with photonics. Visible light photons have a wavelength between 400 and 800 nm, much larger than current chip features. When you go to higher frequencies they get smaller, but they are really difficult to produce and control.
It depends on the type of quantum computer. In some a physical qubit is a single atom, but then to make it reliable they need to add error correction resulting in logical qubits consisting of at least 100 or so physical qubits.
Another type of quantum computer uses qubits consisting of "quantum circuits" which are actually huge macroscopic constructions (> 1mm).
> You could, in principle, use photons and/or electrons. We got pretty damn close in the vacuum tube era, and photonic computing has been a popular research topic for a while
Wait, what? How does this work in principle for storage? You can store electrons but you're saying you can store photons too?
We do use electrons. That's what flows through transistors to do computations. Or, vaguely, the distribution of the electric field....
>We got pretty damn close in the vacuum tube era
Uh, what?
There's only so many fundamental interactions in the Universe. Computing requires you to be able to distinguish two states and our current methodology is built around some sort of black box three input machine that can output either state, a switch.
That switch is the part that cannot be scaled down infinitely. The reality we are familiar with doesn't exist at atomic scales. "Things" don't even have properly defined boundaries at a certain level, and thermal noise is a huge issue.
IMO a much more direct limiter of our current computing capability is lack of manufacturing ability, and heat. We were lucky that transistors were so amenable to lithography as a concept, that they work so well in 2D and as a surface feature, as that is what drove our advances the past 100 years and enabled computing to be such a normal thing. The combination of a "Solid state" effect, the electric force having very convenient properties, and lithography being so amenable to scaling things in various directions is how we got here.
But lithography doesn't scale into 3D. We've been hacking around that by doing more layers but that scales awfully, has very strict limitations, and makes the heat problem infinitely worse, to the point of making it impossible to work around.
If we could assemble things atom by atom exactly how we want, we could vastly improve our theory and practice, and build really intricate processor chunks with effective cooling channels or something, and computing would scale so much more. Maybe. Maybe some other problem would suddenly start dominating in that world.
Biology literally is nanotechnology, but it takes massive tradeoffs in exchange. It might never be possible to manufacture, at scale, stuff atom by atom. The Universe doesn't promise us infinite progress in technology. Quite the opposite.
It doesn’t, no. The most successful platform actually uses superconducting devices as large as millimeters, you can literally see them with the naked eye.
The issue with “just” photons and electrons is that you need something else to force them to behave like you want. And photons are large and non-interacting, really the opposite of what you want for computing. Great for communications of course.
I always thought the true limit was the Planck length against which an atom is giant. There's a whole zoo of sub-atomic particles but I don't think we know how (or if) we can apply those for practical computing.
You could make maybe ten transistors or so, but no more. That technique is quite literally pushing atoms one by one with a sharp needle. Not scalable, though maybe useful for some quantum computing platforms’ fabrication since we’re at early stages.
And you could write nice sci-fi about subatomic transistors, but forget making them in this reality.
Yes, and we're already there. We've been there for quite a while, in fact.
Once you make the gate of a transistor small/thin enough, quantum effects take over. Electrons will randomly teleport into and through the gate causing the transistor to conduct when it shouldn't. I don't have numbers to hand, but it's on the order of a few atoms wide. There's really nothing that can be done about it either, as far as we know. Electrons just aren't physical objects at this scale, you can't simply exclude them from any given volume of space. The electron wave function will simply just appear wherever it wants (within the electron probability cloud). The only way to stop it is to make your insulating junction thicker than the probability cloud.
The experiment to observe this behavior is pretty simple though (Young's double slit), and it was conducted more than 200 years ago. The explanation came much later but it's not like the phenomenon was hiding somewhere.
It’s both ridiculous and quite amazing really. The hint that there is something less random underneath it that we just haven’t figured out (and lack the resources to explore at this time) is tantalising.
Even if there isn’t, the way it seems all based on the uneven flow of state over spacetime is deeply fascinating for someone who studies computing.
Reality is far, far stranger than we give it credit for. Makes you wonder what other completely bonkers secrets the universe has for us.
And frankly, the sheer insanity of quantum teleportation is why I don't buy any argument that faster than light travel is impossible. Not because "teleportation", but because every time we think we understand the rules of the universe, it laughs in our face. The universe is wacky beyond our wildest dreams.
I find it ridiculoua that we believe it is random probability instead of trying to find (and maybe later mitigate) the real sources of this randomness.
> you can't simply exclude them from any given volume of space...
... inside a silicon crystal.
You can keep the electrons into as small a volume as you want, but you need something there forcing them, and doped silicon will only force them so much.
In fact, those transistors are smaller than what a silicon crystal can do, and the electrons are only held there because they are made of more materials than only silicon.
This is why I suspect we will be neat the upper limits of this around the late 2030's. We are just running too close to the fundamental limits. And so far there isn't anything really radical even on the horizon as a solution for this.
all production ics have a flat layout with 3d transistors. the physical limit is how close can the features on one wafer get before theres too much tunneling.
there is a clear way forward with wafer stacking instead of shrinking layouts but so far we only have amd v-cache style asymmetric designs with memory stacked on top of compute. for a fully connected stack that acts like a single chip you need insane precision to align the wafers and a way to remove heat from the middle so the chips dont fry themselves.
if someone finds a way to stack cpu cores without thermal issues that will be a real revolution. huawei might be close with their logic folding but nobody knows if it really works and what the heat problems are like.
I always feel like I'm not quite getting quantum stuff no matter how much I read and learn: what does this advancement have to do with quantum computers?
Don't worry about not grokking quantum computing stuff, neither do any of the people who invest in it as well as many people who work on it.
1. The OP has nothing to do with quantum computers.
2. Quantum computing deals in coherent quantum states: associated with N qubits there are 2^N complex amplitudes. You can measure by sampling the square-magnitude of the complex amplitude which turns it into a Probability Distribution. Quantum computing "gates" cause interference in the complex amplitude of entangled qubits cancelling out incorrect results, such that if you maintain coherence for long enough and sample the final state and measure the probability distribution, you get a computationally useful result. The key challenge in quantum computing is extending the coherence time of a larger and larger number of qubits, which is why you hear so much about quantum error correction. Recent results from Google showed a scaling law for "surface codes" using multiple qubits to create an error-corrected topological qubit with extended lifetime. There is no telling how far this scaling law will go, but as long as Gil Kalai is in the next room, it is unlikely there will be actual useful quantum computation for a while.
I’m sure they will license it. It’s better for them if everyone in the industry can innovate on everything around it. All the process tech companies will make it more cost effective, for instance, which helps IBM as well.
Broadly speaking yes, this is the business model. IBM has been at this for many years with technology transfers, licensing agreements, support and other arrangements. Rapidus, Samsung, GlobalFoundries, ST, SMIC, AMD, etc. have all used IBM R&D work at various times for various nodes and products.
The cutting edge of semiconductors is a writhing mass of copulating tapeworms, and IBM lives deep inside that ball. For IBM, what this means is that when you buy one of the ASML machines to make products with this process, you'll pay IBM for the knowledge and support to actually get it working, or give them a cut, or something else, TBD, as circumstances warrant.
Ah, thanks for that explanation. I was wondering how IBM could fund cutting edge research in semiconductors when they haven't been a semiconductor manufacturer for many years.
How do these 3D architectures scale regarding yield? (The naïve expectation would be that adding vertical layers should have exponential impact on yield.) Is this commercially viable, near term?
Approximately everyone (at least in the F500) outside of Big Tech uses them. For example, Costco's entire inventory management system runs on IBM i (so, POWER). You can see the classic terminal look around the store. Banks run a TON of z and i. You'll never see them because they're essentially always in data centers, but I guarantee you interact with them even if it's very non-obvious because there's 50 microservices between the UI and the actual system of record.
And some of those i and z systems have some obscene engineering in them. I cannot find it now but i remember seeing a cross section of their 40 layer PCB simply so they could get more memory/IO performance to the CPU. POWER may have dropped out of the consumer space but it still has a vital place in servers.
Two big problems 1) NOBODY knows what IBM's definition of "sub 1nm" means 2) IBM bullshits so much more than anyone including Intel (remember the "teleportation" ads years ago) that nobody is going to waste time researching what they mean in reality
That's not what I meant. IBM published some scientific papers which explain the details. If you don't have the background to understand the papers and you don't trust the marketing that still doesn't justify rhetorically flipping the table.
People love to hate on IBM, ask why it is relevant today, while several of their beloved FOSS projects are in some way, form or fashion, made possible with IBM's money.
> IBM and its partners conduct this work at a leading semiconductor research facility in Albany, New York, which will soon be home to a High Numerical Aperture Extreme Ultraviolet (High NA EUV) lithography tool, essential for the future of logic scaling. Developed by ASML, this technology enables ultra‑precise circuit printing, supporting the creation of smaller, more powerful chips.
I'm guessing that this is the technology that is developed by Cymer (ASML subsidiary) in California, correct? Is there competing technology? I know xLight is trying to make some inroads on their own version of this EUV tech. I have not heard about any progress though.
The industry does use a collection of more practical measurements, like transistor density. Marketing pieces for the news tend to use this kind of jargon precisely because it can be fudged & it sounds like it means something else than it really does to the average person. It's also simple enough to avoid needing to really explain what kinds of numbers are impressive etc, everyone just knows less than 1 nm is tiny and they've heard X nm for decades to compare to at this point.
> IBM sees a path to production in as early as the next 5 years.
5 years is a long time for a product roadmap, so there are probably some significant unsolved problems remaining, and the timeline depends on whether IBM can solve these problems.
For anyone who needs it, a friendly reminder that CPU nm marketing is a complete fabrication and the physical size of transistors has zero relation to the marketing claims. These are not, in fact, physically sub 1 nm, despite the bombastic claims.
They don't describe the exact physical size (that would rather defeat the point of the marketing), but you can see the photographs at the bottom have a scale measured in tens of nm.
At some point in the transistor scaling, the electrons started leaking across the gate, we've switched from 2D design to 3D structures to prevent that, so the actual physical gate pitch for like the TSMC 3nm is around 45 nm in distance.
Currently thrown around numbers mean the "equivalent performance/density" or something like that.
I wonder at what point does it become deceptive in the legally biting way to market process nodes like this. One can wax and wane a bunch about industry terminology status quo, but this is mental.
IBM regularly announces silicon breakthroughs like this but I'm not aware of those ever becoming products. Is IBM mainly in the business of licensing their technology to big silicon manufacturers with stuff like this? Is it just marketing for their consulting business?
IBM Z series mainframe Telum CPUs are designed by IBM but manufactured by Samsung. IBM no longer owns any fabs. I assume they have some kind of technology licensing deal.
It's a lab. It's where ASML brings up the prototype machine and gets it working, with IBM talent working out the problems and getting it ready for commercial operation. They won't make chips at scale there: the facility isn't designed for that part. The thing to understand here is that isn't a simple, clean, comprehensible business arrangement. The Albany facility is highly subsidized by the state. IBM has their hooks deep in the operation and occupation of the site. Such facilities are extraordinary with capabilities that talent that are unique and fabulously expensive. That's why ASML is there, and not just doing it in some village in the Netherlands. It's why when Obama, Biden, Trump or whomever tells ASML to whom they will and won't be selling hardware, ASML listens.
> It's why when Obama, Biden, Trump or whomever tells ASML to whom they will and won't be selling hardware, ASML listens.
My understanding is that ASML's acquisition of Cymer in California (the actual EUV light source technology) in 2014 was only permitted under a strict technology sharing and export agreement with the US government. And that the technology development and production had to remain within the US.
The USA CHIPS Act and NY State have provided $100 billion+ in funding with the expectation that ASML's core R&D and "prototyping" like this will be done in the US in partnership with US companies (like IBM).
Not bad, now you just have to fold it 86 times to reach one Planck length. The only issue you'd run into is it would have to be 77 quadrillion kilometers thick
A little bit of a nitpick, but wouldn't that be a picometer instead of angstrom node? Like, isn't a "pico-" the next magnitude smaller than "nano-", or am i wrong?
Otherwise, that chip tech sounds really awesome - at least for the future!
You had the right idea. Angstroms are not an SI unit. The SI units jump by three orders of magnitude at this scale: picometer, nanometer, micrometer, millimeter.
(In the same way that meter jumps three orders of magnitude to kilometer[1], or millions to billions to trillions, etc.)
[1] Technically there are intermediate SI units between meter and km but nobody uses them. There are not intermediate SI units between the tiny ones.
Decimeter is used occasionally for densities, because 1 g/cm^3 is the same as 1 kg/dm^3 but the latter is a little easier to imagine. The cube decimeter is also used under the name of... liter.
Likewise, there is also deca- and hecto-. Hectograms are used for shopping.
Decameter (dam, 10 m) is never used, but there is a non-SI unit of area based on it, called the are. Nobody uses the are, but its multiple the hectare (1 square hectometer) is common in some countries when talking about land plots. It's a little less than 2.5 acres, for people in the US.
Everyday necessity. The gap between mm and m is too large, there are many things in daily life that are better expressed in cm. SI units must strike a balance between three factors: not having so many denominations nobody can remember them; not having so few denominations that using them adds too much wordiness to daily life (150mm or 0.15m are wordier than 15cm); and a degree of familiarity with the everyday units people used before metric, to smooth the transition and encourage adoption.
Useless fact I just learned from Wikipedia: Ångström/Angstrom (in Sweden of course we still use the original spelling) has its own UNICODE symbol, Angstrom sign: Å (U+212B) not to confuse with the Swedish letter Å (U+00C5). Looks slightly different in my browser.
Looks like that's deprecated. From the next sentence:
However, version 5 of the standard already deprecates that code point and has it normalized into the code for the Swedish letter U+00C5 Å `latin capital letter a with ring above`
> logic technology can extend for the first time below the 1 nm node, advancing the era of angstrom-level scaling, where dimensions approach the size of individual atoms. While transistor nodes now refer to a generation of manufacturing technology versus an exact physical dimension, IBM’s 0.7 nm technology—also referred to as 7 angstroms—demonstrates how continued scaling remains possible.
Continuing the well established trend of making bold claims about physical dimensions that have nothing to do with any of the structures in the chip, and the name scales better than the tech.
What they actually deliver is a "nanostack architecture" built with ~5nm features that according to them is comparable to a hypothetical real sub-1nm chip.
It's an impressive achievement nonetheless but it looks like the industry has a few too many marketers.
On the otherhand, no investor really cares what it's called, they just need to know it's next gen.
My read on it was that they are trying to imply a transistor density (in a 2D plane sense) that is comparable to a 1nm process? But they achieve that through stacking (3D, not 2D) since the features aren't actually anywhere near 1nm?
If they're adding a dimension, the marketing should reflect that.
I know they won't go for an anything that makes as much sense as 5nm3, so I vote for "1nm hyper space"
“TeraThread”
Should population density in cities also be measured in hab/km³ rather than hab/km²?
We are used to stacking people vertically in cities.
Correction: it’s not stacking. They do things like FinFET (turning the gates in the third dimension) and gate-all-around which increases the density of transistors per unit area. But they don’t have layers of transistors. At least in the logic and analog processes.
yeah, where on the pictures is the 0.7nm feature? The linespacing is around 5nm. Is it the white line which is 0.7nm?
I really can't see where the 0.7nm is coming from. The white line looks like it's just an edge of a feature that is "15 rows of silicon atoms", which by some quick arithmetic on Wolfram Alpha has to be AT LEAST ~1.6nm, and the way the rows of atoms appear to be packed in that image and by the provided scale, it seems to be significantly more. Using the white line as a meaningful measurement seems to me to be more misleading than any other interpretation here.
It's the equivalent performance of a 0.7 nm planar transistor. It's not about the feature size.
A 0.7 nm planar transistor made of silicon has no performance, because a device so small cannot function as a transistor.
The intended meaning of "0.7 nm" is that if you compare the transistor density per area of a "0.7 nm" manufacturing process with that of a "350 nm" process (like used for some Pentium II CPUs, at a time when "350 nm" was a real length), the ratio between the transistor densities is (350 nm / 0.7 nm)^2 = 500^2 = 250,000.
Comparing with the number of transistors of a Pentium II, a 0.7 nm CPU should be able to contain about 5000 billion transistors. This is consistent with the fact that the latest 3 nm NVIDIA Rubin GPU has 336 billion transistors and a 0.7 nm circuit must have a density around 16 times greater than a 3 nm circuit.
However, for many of the modern node names used by some companies even this computation is not really true, because marketing may have chosen an arbitrary name that is smaller than for the last process of the main competitor.
For now, IBM has not provided any kind of information that could prove their claim that their new CMOS process has the transistor density corresponding to "0.7 nm" (i.e. 16 times greater than the TSMC "3 nm" CMOS process).
> Continuing the well established trend of making bold claims about physical dimensions that have nothing to do with any of the structures in the chip, and the name scales better than the tech.
We care about PPA (power, performance, area) and not how large or not-large features actually are. Comparing gate lengths between a 1980s planar transistor and a 2010s 3D FinFET or GAA transistor is obviously nonsense, the relatively aligned node names of the industry actually do make sense as a shortcut here.
You have to admit it's getting progressively sillier though.
It's been decades since published node sizes had any connection to actual feature size. Sadly this is just how it works in the semiconductor industry now.
Who started it?
Smart money would be on the first person who realized that the expertise required to understand the technical details had grown beyond that possessed by 51%+ of stock investing (trading?) population as weighted by transaction volume.
https://www.eejournal.com/article/no-more-nanometers/
Sometime around 2011 when Intel named their process node 22nm which the gate length was 26nm
Your article says it started earlier:
>So essentially, since 1997, the node name has not been a representation of any actual dimension on the chip, and it has erred in both directions by almost a factor of 2.
We've already been through years of "7nm isn't actually 7nm" across different fabs - completely different measurement conventions, none corresponding to real feature sizes. Now sub-1nm? If it is real then at that scale we're probably in the several atom width territory.
As it can be seen from the photos, horizontally the features are much bigger than 5 nm.
For silicon, the gate length of a FET has a lower limit somewhere between 10 nm and 15 nm.
The current CMOS manufacturing processes have not reached the limit yet. For making smaller transistors, a transition to other semiconductor materials will be necessary.
The vertical thicknesses of various layers may be of only a few nanometers or even of a fraction of a nanometer, but that does not matter directly for the circuit density.
The supposed node size refers to horizontal dimensions, not to vertical dimensions.
Vertical dimensions of around 1 nanometer or less could be achieved already many decades ago, because they depend on growth speed and on time, not on lithography, like the horizontal dimensions.
The industry should have stopped decades ago to talk about the "size" but they should have characterized a CMOS process by its density, e.g. in logic gates per square mm.
However, an actual concrete number would be disliked by marketing, because they could no longer claim that their "1 nm" process is better than the "2 nm" process of another vendor, if their density is not really better.
Just get better marketers to say your 2nm process has more gates per sqmm than your competition 1nm process.
Exactly. WatsonX AI quantum angstroms for e-business.
Now with Tivoli Monitoring!
At some point, people would not know if they have the deal or not have the deal in this new era of quantum e-business ...
The scale bar on the far right "photo" (micrograph?) doesn't make sense. It is only slightly less than half the scale bar on the middle photo (10 nm), but the image is clearly scaled up by much more than 2x. Individual silicon atoms are circled in the right photo, but the covalent radius of silicon is about 0.11 nm, so they should be much smaller if the scale bar is accurate.
The scale bar also got about 50% longer, which would imply a 3x zoom. That also seems about right based on the relative feature sizes. Same thing happened between the first and second image.
>The vertical thicknesses of various layers may be of only a few nanometers or even of a fraction of a nanometer
We should flip the transistors sideways then!
Marketing scams has been the norm for the silicon industry for a big decade.
How would you fix that? This is a global scam. Big Markets regulating them: that would have to be the USA and EU.
Better metrics are transistors/mm^2, performance/watt, and raw performance, since at this point "nm" is fluff and easily game-able.
Different companies measure it differently too. This was a while ago, but I remember reading that Intel 10nm was more or less close to TSMC 7nm. I'm sure this is still true to varying degrees.
Comparing nodes across foundries is a kind of a coin toss. At least, from the name. You actually need to go into the specifics of the pdk and process to understand what features there are. Can’t rely on the name for anything.
Unlike marketing terms, "nm density" is actually useful measure.
It describes density measure where you can compare it to planar transistors from the 28-nanometer (28 nm) node around 2010 to 2011 and before. A "0.7 nm" node has equivalent transistor density as if we could have shrunk standard flat transistor node down to 0.7 nanometers.
Density is mass per volume so how are you comparing it to a planar transistor? Your units don't even match.
https://en.wikipedia.org/wiki/Transistor_count#Transistor_de...
Not all densities is mass per volume. eg. population density.
It's a physical quantity per some unit of spatial measurement so the units still don't match up b/c in one case the transistors are stacked per volume & in the other case per area.
> Historically, "node" sizes (like 28nm or 7nm) directly correlated to the physical length of a transistor's gate. Today, names like 3nm or 2nm reflect a marketing generation. The actual transistors are significantly larger than these nanometer labels, meaning density varies between companies
> Research organizations like IEEE have proposed new metrics, such as transistors per cubic millimeter (MTr/mm^3), to accurately map future 3D scaling. However, commercial chip foundries resist this change because it would make it harder to calculate commercial yields and thermal density limits using standard industry formulas.
https://share.google/aimode/Z5BqUjlZWFNphm6Z6
> It's a physical quantity per some unit of spatial measurement so the units still don't match up b/c in one case the transistors are stacked per volume & in the other case per area.
"Planar" and "3D" in this context refers to the shape of the transistors themselves. In a planar transistor the functional structure is spread out in the area, like this: https://en.wikipedia.org/wiki/File:MOSFET_functioning_body.s... while 3D transistors spread into the volume: https://en.wikipedia.org/wiki/Multigate_device#/media/File:D...
However the active devices are still just one layer. This isn't like 3D NAND where you actually have transistors on top of each other. So the comparison only considers the area for both kinds of transistors.
The comparison is good. Humans are also stacked in volumes and we still measure population density over a surface because the third dimension is less significant in this context.
I got the answer from the AI I was looking for & it makes sense. You can try to map the volumetric density to areal density but the mapping is not canonical so it doesn't say anything about the physical reality of actual transistor density since the reality is that it is a volumetric measure that gets fudged for marketing purposes. 3D volume for chips is going to keep increasing so they will eventually transition to measuring density over volume instead of area.
Population density isn’t planar, eg, why cities are denser than towns: their centers are more than 2D.
Population density is literally measured over a planar metric: people/km^2
You are suggesting it's measured over a cube volume?
There are some really dense people
Some people are dense... now in three dimensions!
3d population density would be an interesting measure, though. It would provide a better feel for how cramped a populated volume actually is.
It would have to be the displacement volume, to account for cities with only a few really tall buildings, and also somehow adjusted for buildings with high ceilings.
density is quantity per unit measure.
mass per volume is one example.
Density is this reply.
All "density" means is that it's a quotient; I use spectral densities daily and that's "count per Hertz" which weirdly enough works out in normal units to be seconds.
Why not use something absolute, like nand-gates per volume?
I am not a chip designer, doesn't area matter way more than volume? Vertical space is basically free; it's horizontal space that is at a huge premium.
Plus heat dissipation is a limiting factor, which scales with area.
Transistor density in units of MTr/mm. (Million Transistors per square mm) is also used. The formula is
MTr/mm = 0.6×(NAND2 Tr Count)/(NAND2 Cell Area) + 0.4×(Scan Flip Flop Tr Count)/(Scan Flip Flop Cell Area)
Why do you think it's not commonly used?
Yeah, the actual sizes are right there in the pictures, and never < 1 nm.
Why is density any bit important? All I care about is the price per transistor, and the power usage(mostly gate charge and leakage current?).
Density does directly scale with both of those in the form of more chips per die (=> lower cost) and smaller capacitance (=> less dynamic power dissipation).
If you want to reduce "effectiveness" of some process down to a single number, then density is far from the worst metric to pick.
Why is that all you care about? Stepping down a node gets you dramatically improved timing and design feasibility. The reduced density means you can pack the same design into less area. Your most challenging timing paths now have to traverse a shorter distance, and you can fit more of them relative to certain node-size invariant structures
isn't density a way to lower both the price and energy dissipation (so better heat management & energy efficiency)?
only a matter of time before some marketer figures out they can get promoted by branding a generation of chips 0nm
What industry doesn't have a few too many marketers? Take everything with a grain of salt.
I believe it could be a stepping stone to significantly higher density in the future.
Just to be clear, this doesn't mean that anything on the die actually measures 0.7nm — it means that it's roughly double the density as the previous node generation. At some point the industry decided to keep talking about "nanometers" even though the actual transistor sizes have been decoupled from the node name for years.
Generations.
Gen Alpha were born since the naming became detached from actual physical size. And parts of Gen Z (before) and Gen Beta (after).
Naming generations is a futile, arbitrary process, whereas the measurement unit for engineering processes should be precise and well defined?
I was saying not just “years”, but generations (approx two, straddling a full one) have passed since the naming reflected anything real.
In hindsight, I phrased that poorly.
Node sizes have been "generations" for a long time. The number is just the previous number divided by sqrt(2), rounded in an arbitrary direction. Sqrt(2) because reducing each dimension by that number leads to double the density, and it's non-obvious enough to sound like a real measurement
Looking at this picture: https://filecache.mediaroom.com/mr5mr_ibmnewsroom/201436/IBM...
I am just astonished at the quality of the structure. Not only that, but the quality of the "cut" to then be able to take this picture. I was closely involved in EM imaging more than 30 years ago, this is just wonderful to see such pictures now.
For what it's worth, here's my 7000+ word deep dive into the technology.
https://morethanmoore.substack.com/p/ibms-announces-07nm-pro...
Hi Dr. Cutress! I loved reading your deep dives on Anand, they're what got me interested in semiconductors. Hope you're well <3
weird q: the photo shows a wafer with partial chips rendered on the edges?
Remember that IBM paid Global Foundries $1.5 billion to take their fabs and design services group.
GF did not pay IBM. IBM paid GF to take the fabs away.
https://www.reuters.com/article/technology/ibm-to-pay-global...
It was 15 years ago. Whole management got replaced, they are quite ambitious. Let's see if how this works out now.
The plan doesn't involve running fabs again. Too expensive. IBM licenses their manufacturing technology (so far to Rapidus in Japan).
The most surprising part for me is that IBM still somehow owns silicon labs, I was sure it's effectively a consulting company by now
IBM has been the company with the most patent registrations in the US for I think 29 of the last 30 years. They're one of the largest industrial research organizations in the world. They're doing more hard science research than almost anyone else.
Which is so weird, right? Like what is IBM now and how does a research lab make sense with the rest of their business?
The money-making parts of IBM are: legacy software and hardware (declining), consulting (low margin, low leverage), enterprise software (mostly redhat, not really growing). It's hard to explain how IBM research is accretive to any of that.
I don't know enough about their business to say, but I'm thrilled at even the idea that someone might actually value long-term success over quarterly earnings.
IBM at one time had nearly everything in modern tech under their roof like Xerox and squandered it. There is no comeback for them.
They were the second American chip company that said no to Steve Jobs when hinted about designing smaller better chip mobile devices the other two were Motorola and Intel Apple had to eventually do it themselves. Apple Silicon
Licensing is a substantial source of revenue, and their servers have very impressive (think Telum’s caching) innovations, even though they rely on third-parties for manufacturing the chips themselves.
They are also betting on quantum computing to become commercially relevant.
The hardware division has 80%+ margin and still makes the systems that process 75% of all financial transactions. Their processors for those systems are on par or better than any other, I don’t think that is a business at all. This cash cow is not going away any time soon and gives them the profits to make bets on the future of computing.
Isn't consulting one of the most high-margin areas in general?
It's just a shame that none of it seems to pan out, and in the areas where I know what they're talking about, it all sounds like cynical nonsense to me.
I really wish I had followed through when I was ask by some of the guys at IBM Research to apply when we had worked together on a partner project. Though I didn’t have a degree which I seem to remember was a sticking point, this is in the mid 2010s
The labs might not be that different from consulting, the NYT reporting on this notes they run R&D labs so they can license the tech they develop to people who actually make chips.
Most of their fabs were divested to GlobalFoundries, but they still have pretty significant fab capability and capacity- I suspect at least partly to have a us-based chip-making for military ("Trusted Foundry").
One of the images has "15 rows of Si atoms".
Is there a limit to how small things can go? A single atom?
Is there a physical/molecular limit to Moore's Law?
I mean, you can't get smaller than an atom, there is some amount of plausibility of using individual atoms as at least the occasional computing element.
Beyond that, engineering a quark-gluon plasma as a processor? I'd watch that Star Trek episode. (we might fantasize about stuff like that but we're roughly monkeys smashing rocks together in a cave vs. building an iPhone sort of gap away from that kind of thing unless somebody has a really good idea)
You could, in principle, use photons and/or electrons. We got pretty damn close in the vacuum tube era, and photonic computing has been a popular research topic for a while.
You also have quantum computing, which I think can/does use subatomic particles? Not sure about that one
You can’t make smaller chips features with photonics. Visible light photons have a wavelength between 400 and 800 nm, much larger than current chip features. When you go to higher frequencies they get smaller, but they are really difficult to produce and control.
It depends on the type of quantum computer. In some a physical qubit is a single atom, but then to make it reliable they need to add error correction resulting in logical qubits consisting of at least 100 or so physical qubits.
Another type of quantum computer uses qubits consisting of "quantum circuits" which are actually huge macroscopic constructions (> 1mm).
> You could, in principle, use photons and/or electrons. We got pretty damn close in the vacuum tube era, and photonic computing has been a popular research topic for a while
Wait, what? How does this work in principle for storage? You can store electrons but you're saying you can store photons too?
We do use electrons. That's what flows through transistors to do computations. Or, vaguely, the distribution of the electric field....
>We got pretty damn close in the vacuum tube era
Uh, what?
There's only so many fundamental interactions in the Universe. Computing requires you to be able to distinguish two states and our current methodology is built around some sort of black box three input machine that can output either state, a switch.
That switch is the part that cannot be scaled down infinitely. The reality we are familiar with doesn't exist at atomic scales. "Things" don't even have properly defined boundaries at a certain level, and thermal noise is a huge issue.
IMO a much more direct limiter of our current computing capability is lack of manufacturing ability, and heat. We were lucky that transistors were so amenable to lithography as a concept, that they work so well in 2D and as a surface feature, as that is what drove our advances the past 100 years and enabled computing to be such a normal thing. The combination of a "Solid state" effect, the electric force having very convenient properties, and lithography being so amenable to scaling things in various directions is how we got here.
But lithography doesn't scale into 3D. We've been hacking around that by doing more layers but that scales awfully, has very strict limitations, and makes the heat problem infinitely worse, to the point of making it impossible to work around.
If we could assemble things atom by atom exactly how we want, we could vastly improve our theory and practice, and build really intricate processor chunks with effective cooling channels or something, and computing would scale so much more. Maybe. Maybe some other problem would suddenly start dominating in that world.
Biology literally is nanotechnology, but it takes massive tradeoffs in exchange. It might never be possible to manufacture, at scale, stuff atom by atom. The Universe doesn't promise us infinite progress in technology. Quite the opposite.
It doesn’t, no. The most successful platform actually uses superconducting devices as large as millimeters, you can literally see them with the naked eye.
The issue with “just” photons and electrons is that you need something else to force them to behave like you want. And photons are large and non-interacting, really the opposite of what you want for computing. Great for communications of course.
photons and electrons are the water, not the pipes and valves
what matters is the size of the pumbing
A physicist can fill in but there might be some stuff made out of subatomic particles for example.
https://en.wikipedia.org/wiki/Exotic_matter
> you can't get smaller than ...
I always thought the true limit was the Planck length against which an atom is giant. There's a whole zoo of sub-atomic particles but I don't think we know how (or if) we can apply those for practical computing.
https://en.wikipedia.org/wiki/There%27s_Plenty_of_Room_at_th...
https://en.wikipedia.org/wiki/Landauer%27s_principle
> https://en.wikipedia.org/wiki/There%27s_Plenty_of_Room_at_th...
Yes, single-atom manipulation has already been demonstrated:
* https://en.wikipedia.org/wiki/IBM_(atoms)
Can you make transistors using that technique? Can you smaller?
You could make maybe ten transistors or so, but no more. That technique is quite literally pushing atoms one by one with a sharp needle. Not scalable, though maybe useful for some quantum computing platforms’ fabrication since we’re at early stages.
And you could write nice sci-fi about subatomic transistors, but forget making them in this reality.
Yes, and we're already there. We've been there for quite a while, in fact.
Once you make the gate of a transistor small/thin enough, quantum effects take over. Electrons will randomly teleport into and through the gate causing the transistor to conduct when it shouldn't. I don't have numbers to hand, but it's on the order of a few atoms wide. There's really nothing that can be done about it either, as far as we know. Electrons just aren't physical objects at this scale, you can't simply exclude them from any given volume of space. The electron wave function will simply just appear wherever it wants (within the electron probability cloud). The only way to stop it is to make your insulating junction thicker than the probability cloud.
>"The electron wave function will simply just appear wherever it wants (within the electron probability cloud)."
I don't know which is more ridiculous, the fact that reality works like this, or, that a species of apes was able to figure this out.
The experiment to observe this behavior is pretty simple though (Young's double slit), and it was conducted more than 200 years ago. The explanation came much later but it's not like the phenomenon was hiding somewhere.
It’s both ridiculous and quite amazing really. The hint that there is something less random underneath it that we just haven’t figured out (and lack the resources to explore at this time) is tantalising.
Even if there isn’t, the way it seems all based on the uneven flow of state over spacetime is deeply fascinating for someone who studies computing.
Reality is far, far stranger than we give it credit for. Makes you wonder what other completely bonkers secrets the universe has for us.
And frankly, the sheer insanity of quantum teleportation is why I don't buy any argument that faster than light travel is impossible. Not because "teleportation", but because every time we think we understand the rules of the universe, it laughs in our face. The universe is wacky beyond our wildest dreams.
I find it ridiculoua that we believe it is random probability instead of trying to find (and maybe later mitigate) the real sources of this randomness.
There's a whole branch of theoretical physics dedicated to this.
> you can't simply exclude them from any given volume of space...
... inside a silicon crystal.
You can keep the electrons into as small a volume as you want, but you need something there forcing them, and doped silicon will only force them so much.
In fact, those transistors are smaller than what a silicon crystal can do, and the electrons are only held there because they are made of more materials than only silicon.
This is why I suspect we will be neat the upper limits of this around the late 2030's. We are just running too close to the fundamental limits. And so far there isn't anything really radical even on the horizon as a solution for this.
Aren't we reaching the point where there is no "solution" in terms of density and physical dimensions?
It's like the speed of light being a constant, or the Planck length being the smallest that can be subject to standard physics.
Quantum computing, which is a complete change in the actual physical model of computing, appears to be the only alternative.
all production ics have a flat layout with 3d transistors. the physical limit is how close can the features on one wafer get before theres too much tunneling.
there is a clear way forward with wafer stacking instead of shrinking layouts but so far we only have amd v-cache style asymmetric designs with memory stacked on top of compute. for a fully connected stack that acts like a single chip you need insane precision to align the wafers and a way to remove heat from the middle so the chips dont fry themselves.
if someone finds a way to stack cpu cores without thermal issues that will be a real revolution. huawei might be close with their logic folding but nobody knows if it really works and what the heat problems are like.
How does IBM commercialize this? Do they license this out to fabs?
Sit on a patent and try to scrape earnings from others, maybe? That is, license or litigate.
They licensed 2 nm to Rapidus so yes.
boost sales for their systems division, POWER CPUs, mainframes, maybe Quantum stuff
I always feel like I'm not quite getting quantum stuff no matter how much I read and learn: what does this advancement have to do with quantum computers?
Don't worry about not grokking quantum computing stuff, neither do any of the people who invest in it as well as many people who work on it.
1. The OP has nothing to do with quantum computers.
2. Quantum computing deals in coherent quantum states: associated with N qubits there are 2^N complex amplitudes. You can measure by sampling the square-magnitude of the complex amplitude which turns it into a Probability Distribution. Quantum computing "gates" cause interference in the complex amplitude of entangled qubits cancelling out incorrect results, such that if you maintain coherence for long enough and sample the final state and measure the probability distribution, you get a computationally useful result. The key challenge in quantum computing is extending the coherence time of a larger and larger number of qubits, which is why you hear so much about quantum error correction. Recent results from Google showed a scaling law for "surface codes" using multiple qubits to create an error-corrected topological qubit with extended lifetime. There is no telling how far this scaling law will go, but as long as Gil Kalai is in the next room, it is unlikely there will be actual useful quantum computation for a while.
I’m sure they will license it. It’s better for them if everyone in the industry can innovate on everything around it. All the process tech companies will make it more cost effective, for instance, which helps IBM as well.
A sad ending for one of great American companies...
> Do they license this out to fabs?
Broadly speaking yes, this is the business model. IBM has been at this for many years with technology transfers, licensing agreements, support and other arrangements. Rapidus, Samsung, GlobalFoundries, ST, SMIC, AMD, etc. have all used IBM R&D work at various times for various nodes and products.
The cutting edge of semiconductors is a writhing mass of copulating tapeworms, and IBM lives deep inside that ball. For IBM, what this means is that when you buy one of the ASML machines to make products with this process, you'll pay IBM for the knowledge and support to actually get it working, or give them a cut, or something else, TBD, as circumstances warrant.
> a writhing mass of copulating tapeworms
What a metaphor! I will file that one away for some glorious future opportunity.
Credit to Alan Cox. I filed it away after he wrote it in a LKML message many years ago.
Ah, thanks for that explanation. I was wondering how IBM could fund cutting edge research in semiconductors when they haven't been a semiconductor manufacturer for many years.
How do these 3D architectures scale regarding yield? (The naïve expectation would be that adding vertical layers should have exponential impact on yield.) Is this commercially viable, near term?
Keep hearing that IBM makes these incredible chips but don’t see anyone using IBM chips. What do they do with them?
usgov
Approximately everyone (at least in the F500) outside of Big Tech uses them. For example, Costco's entire inventory management system runs on IBM i (so, POWER). You can see the classic terminal look around the store. Banks run a TON of z and i. You'll never see them because they're essentially always in data centers, but I guarantee you interact with them even if it's very non-obvious because there's 50 microservices between the UI and the actual system of record.
And some of those i and z systems have some obscene engineering in them. I cannot find it now but i remember seeing a cross section of their 40 layer PCB simply so they could get more memory/IO performance to the CPU. POWER may have dropped out of the consumer space but it still has a vital place in servers.
The product here is the research and licensing the tech.
Their line of POWER chips are used in their mainframes
No, their mainframes use the zArch which goes back to the System/360 from 1964.
https://en.wikipedia.org/wiki/Z/Architecture
IBM sells huge servers with POWER architecture CPUs but they are not what people are referring to when they talk about IBM mainframes.
Ericsson used a lot of power chips in their telecom hardware 10 years ago at least, haven't worked with their stuff since then so I haven't kept up.
Two big problems 1) NOBODY knows what IBM's definition of "sub 1nm" means 2) IBM bullshits so much more than anyone including Intel (remember the "teleportation" ads years ago) that nobody is going to waste time researching what they mean in reality
I know what it means. Something isn't automatically bullshit because it's outside your field of expertise.
“I’m able to parse a marketing term” isn’t a great claim here. Pointing out it’s an unclear term largely abused is valid.
That's not what I meant. IBM published some scientific papers which explain the details. If you don't have the background to understand the papers and you don't trust the marketing that still doesn't justify rhetorically flipping the table.
People love to hate on IBM, ask why it is relevant today, while several of their beloved FOSS projects are in some way, form or fashion, made possible with IBM's money.
which beloved FOSS were funded by IBM?
They effectively funded the legal defense against SCO that was trying to claim ownership of the IPR to Unix/Linux.
They bought and funded (and still fund) Redhat which funds Fedora and GNOME and Wayland and systemd.
Seems they own Red Hat, which is some no-name company that has nothing to do with Linux.
> remember the "teleportation" ads years ago
Never heard of this, care to elaborate?
I think the OP references Intel ads with Jim Parsons from like a decade ago.
https://www.youtube.com/watch?v=PQWpYQm60P4
There were more of them, including some with other celebrities.
Maybe it is done to pump their stock with low effort? It seems that way for many companies.
> IBM and its partners conduct this work at a leading semiconductor research facility in Albany, New York, which will soon be home to a High Numerical Aperture Extreme Ultraviolet (High NA EUV) lithography tool, essential for the future of logic scaling. Developed by ASML, this technology enables ultra‑precise circuit printing, supporting the creation of smaller, more powerful chips.
I'm guessing that this is the technology that is developed by Cymer (ASML subsidiary) in California, correct? Is there competing technology? I know xLight is trying to make some inroads on their own version of this EUV tech. I have not heard about any progress though.
Correct
Cymer builds the EUV light source, but the biggest enabler for High NA EUV is using anamorphic optics (ie asymmetric horizontal and vertical magnification) from Zeiss: https://www.asml.com/en/news/stories/2024/5-things-high-na-e...
More marketing lies from IBM. Sad to see the decline and corruption of a once OK company.
Selling a 5 nm vertically stacked chip as equivalent to 0.7 nm
Shades of Watson and other IBM lies.
Why doesn't the industry use something like transistor density per cubic cm? This would extend to 3d cases and impossible to fake
The industry does use a collection of more practical measurements, like transistor density. Marketing pieces for the news tend to use this kind of jargon precisely because it can be fudged & it sounds like it means something else than it really does to the average person. It's also simple enough to avoid needing to really explain what kinds of numbers are impressive etc, everyone just knows less than 1 nm is tiny and they've heard X nm for decades to compare to at this point.
Is the 3D transistor proposed by IBM on the same track as Tao's Law proposed by Huawei?
has anyone found a paper with details?
also, I was expecting to see cfets mentioned.
They dont give the exact size hmm
Relevant quote:
> IBM sees a path to production in as early as the next 5 years.
5 years is a long time for a product roadmap, so there are probably some significant unsolved problems remaining, and the timeline depends on whether IBM can solve these problems.
For anyone who needs it, a friendly reminder that CPU nm marketing is a complete fabrication and the physical size of transistors has zero relation to the marketing claims. These are not, in fact, physically sub 1 nm, despite the bombastic claims.
So the title should be corrected. The did not debut sub nm chips at all.
That ship sailed long ago. I think it was around 32nm-22nm node when the marketing term started diverging from the physical feature size.
More reason to frame it correctly, like "IBM debuts another 5 nanometer chip technology, selling it as sub nanometer chip technology"
The marketing nm better represent the density and performance of the transistors than the actual feature size, especially in this case.
>These are not, in fact, physically sub 1 nm, despite the bombastic claims.
Why? What's their real size?
Not doubting you, just trying to understand and also trying to assess how exaggerated the marketing is.
They don't describe the exact physical size (that would rather defeat the point of the marketing), but you can see the photographs at the bottom have a scale measured in tens of nm.
At some point in the transistor scaling, the electrons started leaking across the gate, we've switched from 2D design to 3D structures to prevent that, so the actual physical gate pitch for like the TSMC 3nm is around 45 nm in distance.
Currently thrown around numbers mean the "equivalent performance/density" or something like that.
How come we don't hear anything about Watson?
Well Watson went from a product, to product line branding. Then changed to Watsonx.
Doing a search on that there are loads of news articles.
IBM debuts a PDK and some test structures, as they have no production fabs
They have small scale fabrication capabilities or they wouldn’t be able to validate the technology enough to sell it.
Tell me where I said they didn’t have fabs
Sub-Atomic level chips soon, then?
Isnt that quantum computing?
Yes, shortly followed by negative space chips
I wonder at what point does it become deceptive in the legally biting way to market process nodes like this. One can wax and wane a bunch about industry terminology status quo, but this is mental.
It'll be a brave man who takes on the IBM legal department over terminology in widespread use.
Since a transistor can't be smaller than a single atom, maybe it's time to start optimizing our software again.
IBM regularly announces silicon breakthroughs like this but I'm not aware of those ever becoming products. Is IBM mainly in the business of licensing their technology to big silicon manufacturers with stuff like this? Is it just marketing for their consulting business?
My understanding is they are largely an IP business. That said this release mentioned an ASML machine on prem, so?
I believe that IBM makes the chips for their Z Series mainframes. I mean, that's low volume production, but they need small feature size.
IBM Z series mainframe Telum CPUs are designed by IBM but manufactured by Samsung. IBM no longer owns any fabs. I assume they have some kind of technology licensing deal.
https://www.ibm.com/products/z/telum
> IBM no longer owns any fabs
Per IBM: "IBM Research at Albany [...] includes more than 100,000 square feet of semiconductor fabrication space"
I guess that is technically a R&D fab not a production one, but they definitely have in house fabrication capability
It's a lab. It's where ASML brings up the prototype machine and gets it working, with IBM talent working out the problems and getting it ready for commercial operation. They won't make chips at scale there: the facility isn't designed for that part. The thing to understand here is that isn't a simple, clean, comprehensible business arrangement. The Albany facility is highly subsidized by the state. IBM has their hooks deep in the operation and occupation of the site. Such facilities are extraordinary with capabilities that talent that are unique and fabulously expensive. That's why ASML is there, and not just doing it in some village in the Netherlands. It's why when Obama, Biden, Trump or whomever tells ASML to whom they will and won't be selling hardware, ASML listens.
> It's why when Obama, Biden, Trump or whomever tells ASML to whom they will and won't be selling hardware, ASML listens.
My understanding is that ASML's acquisition of Cymer in California (the actual EUV light source technology) in 2014 was only permitted under a strict technology sharing and export agreement with the US government. And that the technology development and production had to remain within the US.
The USA CHIPS Act and NY State have provided $100 billion+ in funding with the expectation that ASML's core R&D and "prototyping" like this will be done in the US in partnership with US companies (like IBM).
IBM's contributions to computing hardware and software are incalculable.
So many breakthroughs in hard drives, chips, transistor density, and other aspects of computing have come out of their labs.
Great to see them continuing to innovate.
But, yeah, usually they partner and license. Over the years, they've spun off more and more of their hardware businesses.
Don't forget copper interconnects for ICs. https://www.chiphistory.org/ibm-s-development-of-copper-inte...
It's great that they found a working business model for a pure r&d lab, and with such awesome results.
I wonder why isn't this more common.
Not bad, now you just have to fold it 86 times to reach one Planck length. The only issue you'd run into is it would have to be 77 quadrillion kilometers thick
A little bit of a nitpick, but wouldn't that be a picometer instead of angstrom node? Like, isn't a "pico-" the next magnitude smaller than "nano-", or am i wrong?
Otherwise, that chip tech sounds really awesome - at least for the future!
There are 3 orders of magnitude between nano (^-9) and pico (^-12). An Angstrom is ^-10m.
Aaahhh, ok, thanks!
You had the right idea. Angstroms are not an SI unit. The SI units jump by three orders of magnitude at this scale: picometer, nanometer, micrometer, millimeter.
(In the same way that meter jumps three orders of magnitude to kilometer[1], or millions to billions to trillions, etc.)
[1] Technically there are intermediate SI units between meter and km but nobody uses them. There are not intermediate SI units between the tiny ones.
Why above 1mm do we go by tens instead of thousands?
We have centimeter (10 mm) then decimeter (100mm) then meter (1000mm). Then we jump to thousand again (kilometer).
>We have centimeter (10 mm) then decimeter (100mm)
Does anyone actually use those? I think I would throw up a little in my mouth if I saw either of those on a mechanical drawing.
Centimeter is the commonly used metric for small distances in everyday parlance, just like an inch.
Decimeter is used occasionally for densities, because 1 g/cm^3 is the same as 1 kg/dm^3 but the latter is a little easier to imagine. The cube decimeter is also used under the name of... liter.
Likewise, there is also deca- and hecto-. Hectograms are used for shopping.
Decameter (dam, 10 m) is never used, but there is a non-SI unit of area based on it, called the are. Nobody uses the are, but its multiple the hectare (1 square hectometer) is common in some countries when talking about land plots. It's a little less than 2.5 acres, for people in the US.
Everyday necessity. The gap between mm and m is too large, there are many things in daily life that are better expressed in cm. SI units must strike a balance between three factors: not having so many denominations nobody can remember them; not having so few denominations that using them adds too much wordiness to daily life (150mm or 0.15m are wordier than 15cm); and a degree of familiarity with the everyday units people used before metric, to smooth the transition and encourage adoption.
Answer that question and you'll get the whole impetus for logarithmic scales.
Useless fact I just learned from Wikipedia: Ångström/Angstrom (in Sweden of course we still use the original spelling) has its own UNICODE symbol, Angstrom sign: Å (U+212B) not to confuse with the Swedish letter Å (U+00C5). Looks slightly different in my browser.
https://en.wikipedia.org/wiki/Angstrom
Looks like that's deprecated. From the next sentence:
However, version 5 of the standard already deprecates that code point and has it normalized into the code for the Swedish letter U+00C5 Å `latin capital letter a with ring above`
1 picometer = 0.001 nanometers, 0.01 angstrom
1 angstrom = 0.1 nanometers, 100 picometers
1 nanometer = 10 angstroms, 1000 picometers
Because 1 angstrom equals 10⁻¹⁰ meters and 1 picometer equals 10⁻¹² meters, the relationship is:
1 Å = 100 pm. 1 pm = 0.01 Å.