While I think this might work well for Intel in the long run. So, Intel employees will work in the Dalian fab which will be owned by SK Hynix till 2025? This will be revenue recognized for SK Hynix starting next year? Also, surprising that Micron didn't want to take this up considering they have worked with Intel previously.
Micron wouldn't have much reason to be interested. They already dissolved the IMFT partnership because Micron and Intel disagreed on where they wanted to go with flash R&D. Micron might be able to benefit somewhat from Intel's in-house SSD controller IP, but I'm not sure that's reason enough for them to acquire flash tech they already turned down.
Seems like a bad deal without the inclusion of the Optane business. NAND is a constrained commodity now with market saturation, too many competitors resulting in falling profits. Not a business I would invest heavily into.
I imagine the reason for the stretched out deal, is that it protects SK Hynix against potential US sanctions. A US attack on the fab is a risk they were probably not willing to take.
No disrespect but I really doubt fear of sanctions were a significant part of the purchase. South Korea and Taiwan's huge place in semiconductors and their status as US allies is precisely what allows Trump to sanction Chinese tech firms. Sanctioning Korea would mean sanctioning Samsung and LG in addition to SK which would make the world suddenly very, very, very short on critical necessities like RAM, processors, and displays. And then the US would also be forcing these firms to hold hands with China. Terrible option.
Yeah, I can't help but feel Intel ripped-off SK Hynix with this transfer.
$9 Billion seems stupid when you realise Intel isn't that strong of a competitor in the flash market, especially when Optane is not included in the deal. The only reason I can think of the price being so high is because 1- Intel is greedy and needs the money for R&D to battle Zen3.
The 2- second reason, is probably they can insure sanctions against China doesn't happen for SSDs because of this one deal. Whilst, SK Hynix is a South Korean corporation and not Chinese, this fab is located in Northern China. On this point we can only speculate. A wild conspiracy theory would be that they want the IP from Intel's SSD division to be transferred to Mainland China, even if done so by proxy allied corporation. The recent ImaginationTechnologies (PowerVR) deal does bring up the realities of such scenarios.
Intel has a strong enterprise SSD presence and SK Hynix doesn't. Intel's nand is old tech by now and I think that ESSD market share and know how is where SK want. The fab might be used for other future controller / old product for SK instead of nand.
This might be the worst thing in tech news, people contributing nothing and just complaining about things entirely unconstructively. Saying that won't even do anything, Intel's probably decided approximately when they're launching anyway.
Mobile SoCs with Optane won't happen, or won't matter, because Intel can't do mobile SoCs worth a damn. Somebody else might eventually be able to do a mobile SoC with 3D XPoint support depending on what Micron does with 3DXP long-term. But for now, there's not much indication that 3DXP can do anything interesting when constrained to a mobile-friendly power envelope.
One of the first uses of 3D XPoint was in laptops as swap/cache, so I think you might be jumping the gun a bit. I could see 3D XPoint helping with faster startup, NV data storage and as swap/cache for all sorts of devices because of it's NV and performance properties.
I'm not jumping the gun, I'm just going off my own power measurements. Intel's Optane cache modules aren't super low power devices. They are still about twice as efficient as the best flash-based NVMe SSDs (SK hynix P31) for random reads, but for other workloads and for idle power those Optane modules aren't great. So far, no Optane product has demonstrated power characteristics that indicate 3DXP would be suitable for mobile/smartphone usage, and Intel's definitely not going to prioritize energy efficiency over performance for future generations.
Sorry to link to your article, but the most relevant article here would likely be: https://www.anandtech.com/show/12512/the-intel-opt... and you'll have to excuse my wishful thinking about Optane in low power situation. Thank you for the replies & articles.
Optane seems ideal for a compiler cache for my desktop use case, so looking forward to any 800p and/or H10 successors in the consumer market. I'm not sure if the H10 will fully work with any AMD board :(
NAND flash is still facing a likely dead-end in the long run, as NAND has been scaled down almost as small as it can go before leakage becomes a serious problem
2001: "process nodes are reachign their limits with 130nm, 90nm si so small that quantum tunneling will thwart any further attempts to shrink nodes without changing to materials too expensive to use for proceesors.
1980: "8MB hard drives are pushing the limitations of what is possible with spinning magnetic media. 10MB is too unreliable and its likely hard drives will never reach 20 MB
If this inndustry has taught us ANYTHING, its that when someone says it cant be made smaller it gets made smaller within like 2 years.
Yeah, they'll continue to stack and MLC and PAM and add LP-DRAM to grow density and speed. In the consumer space, the "storage" requirements aren't nearly as a concern as the cloud/enterprise space; so NAND is still viable for the foreseeable future to the consumer.
Wow! Used to be the Intel crowd were claiming TSMC 7nm had parity with Intel 10nm. Now it's parity with Intel 14nm. Next year I expect it will be parity with Intel 22nm, and TSMC 3nm is what's needed to match Intel 10nm...
Define "parity". The definition Intel was keen on for many years (right up until things went horribly wrong) was transistor density. By that metric 7nm matches Intel 10nm in specs (both are about 100 MTr/mm^2) with the important difference that TSMC has actually shipped a variety of products at that density (eg Apple and Huawei SoCs) whereas the best Intel has shipped on 10nm has less than half the claimed density, and the newest 10nm has a density even lower.
Oh, in case you want to point out that AMD's CPUs only go to 4.9GHz (last gen was 4.7GHz IIRC), compared to Intel's fastest at 5.1GHz, Intel's 14nm has been improved over time. TSMC's 7nm will probably also be improved.
You're not really comparing equivalent things. It's not that they haven't figured out *how* to scale NAND smaller and are theorising about what would prevent that (as with your examples) - they already scaled it smaller and it doesn't work well.
Scaling smaller would require fundamentally redesigning how it works, at which point it's not NAND anymore (e.g. X-point).
So... the Total Available Market (TAM) strategy from Bob Swan is no longer viable...? What is Intel going to do if they cannot compete in any or their key markets...?
-Fabs are in terrible shape; -CPUs are infested of hardware flaws; -Servers cannot compete; -Mobile are in disarray; -5G modems are gone; -Memory take the hit now;
-Optane in DDR5 DIMMs. (It isn't going away.) -Networking chips (Intel has the best that there is.) -People don't seem to care about HW flaws (Sad, but true.) -CPU HW video encoding (Some people really love it.) -10nm seems to be viable (Finally) and the 1C boost freq are about what AMDs CPUs can do.
The latter most is misleading, as the 5000 series has better IPC, and most Intel chips can't clock that high, that's why the supply shortage of their "best gaming" CPU.
SSD's are commodity. Their key market for future there is Optane. Where they go with it we will know once 2nd gen has launched and see how much they have improved with PCI-E4 interface. I am expecting Alder Stream to blow away all Nand SSD. If not then it will be disappointing. They still have products in different segments. Networking(Smart NIC, Barefoot Networks, Silicon Photonics), FPGA/ASICs, Mobile Eye. Dont forget most AMD laptops launched recently have Intel Wi-fi chips. So they have products. They have to get their manufacturing in control.
"-CPUs are infested of hardware flaws" Pretty sure they'll be out of the woods on this one (for a while) when they finally get Ice Lake out in enterprise. It won't be the *most secure* platform available, but it won't be "infested" either.
Similarly, they *can* compete in servers, because they are. Their offerings are mediocre but they can currently still make up for that with sheer volume of supply and broad availability from OEMs.
The whole thing looks like financial engineering (Intel's new core competence, as opposed to real engineering).
Robert Cringely, back when he was relevant, mid to late 90s (and Microsoft was king) used to refer to this quote from Frank Gaudette, MS CFO: "Watch for any changes in our accounting,” said Gaudette. “If I need to I can start, depreciating the software and maintain earnings growth for years on flat revenue. Watch for the accounting changes, wait for the next uptick in the stock price, and then sell." Well, the accounting changes came, and the changes in MS' status came with them.
Real engineering benefits all of humanity. Financial engineering benefits whoever gets the bonuses, and the shareholders who are best at navigating the highs and lows....
Sad but true, at the end of the day public corporations are all vulnerable to this type of rot. All it takes is a weak succession and the vultures swoop in.
You might have skipped over a few significant developments in that 25 year period. They switched back to actually trying to innovate and pursuing new markets again (e.g. cloud services, games consoles).
Agreed. They'll magically show an increase in margins from doing this, which will help to offset the massive hits they're about to take in desktop and servers (they still seem to have a lock on notebooks and convertibles thanks to OEMs and Nvidia, we'll see how long that lasts).
Eh, even if NAND stops scaling, this doesn't mean cost won't decrease. Take a look at the solar cells for example, great price per Wp wasn't due to efficiency improvement but due to incredible cost reduction. The same can happen for nand and keep it relevant for ages. Yes, it WILL hit physical wall eventually like every other tech, but I don't think it is going the way of dodo any time soon. Especially as no competing technology has better perf/$ yet.
Yea imo there's assuredly room for yield improvements if nothing else. Also I'm sure controller optimizations could be huge for the space -- once NAND generations stabilize, controllers can be more directly designed for NAND performance profiles (re: SK Hynix's own P31).
Very excited for this, IF it means more affordable higher capacity consumer SSDs in the somewhat near-term. I would even be thrilled with some "miraculous" SK Hynix capacity breakthrough, after acquiring all of Intel's research and patents, that allowed them to double or triple current maximum consumer SSD capacities. *a guy can dream*
To be fair, Intel has already been doing a good job of marketing their QLC to consumers at competitive prices. (That's now their entire consumer SSD strategy, since they stated earlier this year that they were not going to do any more consumer TLC drives.) So even if Intel held on to their NAND business, we could expect the upcoming 144L QLC and future generations to hit the consumer market in follow-ups to the 660p and 665p.
Getting more options for consumer SSDs beyond 2TB is less a matter of technology right now than a matter of needing drive vendors willing to bring niche models to market. Sabrent seems to have helped break the ice a bit there, and now we're seeing other Phison partners launch 4TB or 8TB options.
a fair few moves happening here. So intel and Micron have a pretty sweet relationship. and now Intel has sold its NAND business... That opens up a lot of questions for me . as to why get rid of a business that is really profitable. Intel cant be hurting that bad for cash?? whats the real reason here????
Intel NAND / SSD were falling behind in capacity, I think the money is in the 4-16TB SSD in enterprise right now and Intel is hardly catching up above 4TB, basically selling old stuff at the moment, and will get worse soon.
To be honest, deal of that sort was long overdue. Intel fab buissnes is not competitive nor it has substantial revenue, while other player could use it to leverage it own business in the long run.
$7 billion will be paid soon, in 2021 if I remember. The remaining $2 billion is to be paid in 2025. That makes it very likely the deal with go through, I believe, because SK Hynix is unlikely to walk away when they've alresdy paid over 3/4 the cost and Intel likely doesn't have an option to change their mind. I guess there could be something in the agreement about sending the $7 billion back, but I doubt companies pass $7 billion back and forth very often.
I read the details of the deal here in this article. Seems Intel keeps the NAND IP and operates the fab until 2025. The deal is strange because Intel has little incentive to develop technology they are exiting and which they've already sold for a fixed price in the future. Plus Intel's NAND technology is different from the rest of the industry's includubg SK Hynix. I would guess that SK Hynix will change the fab over to charge trap when it takes it over.
The 4 year period might be a way to get the maximum value of Intel's IP roadmap before it's sunsetted. A good deal of R&D that Intel has already done will probably be applicable 4 years out, after which it will peter out unless new foundational research is pursued starting now, which no one will now do since Intel is exiting the business. Since SK Hynix does not have the expertise in Intel's technology and isn't interested in it, Intel will manage the development and manufacture of the NAND that goes into the product lines SK Hynix is buying from Intel.
Having worked for Intel for several years and have co-workers who have worked for Hynix, I think the reason Hynix extended the deal over 5 years the employees.
The corporate cultures of Hynix and Intel are very different. If the transition happened right away Hynix would risk losing a lot of the employees. The corporate culture at Hynix is not that great. They work their employees very hard and their flash products are known in the industry for their low quality. Hynix doesn't understand the enterprise market and doesn't have the DNA to compete and innovate in that market.
With the employees under the Intel mothership for 5 year, Hynix stands a better chance of retaining them in the long term and potentially having some of the Intel corporate culture transferred to Hynix.
Agree, SKH has quite a turnover and lots of politics between HQ and US. Buying Intel is likely part of the solution if they can keep that business running as it is with the same group of employees. If in 5 years they can get the market they'll pay the rest, if not they can probably walk off and get their money back.
Waste of fucking money, they could have purchased 3 brand new (3 billion a pop) FAB's to massively increase their production capability instead of buying the dying portion of Intel's NAND unit.. not even the Optane division so this is not killing a competitor, it's giving them 9 billion more ways to FUCK you harder in the future lol...
Try 20B for the latest and greatest. Increasing capacity doesn't make you more money (it will flood the market and drag everyone down).
Enterprise SSD have a profit margin of 10-70% depends on the demand and supply. SKH does not have this market and Intel is #2 (pretty much the only thing Intel is still competitive in storage), that's where the price is for.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
64 Comments
Back to Article
Teckk - Tuesday, October 20, 2020 - link
While I think this might work well for Intel in the long run. So, Intel employees will work in the Dalian fab which will be owned by SK Hynix till 2025? This will be revenue recognized for SK Hynix starting next year? Also, surprising that Micron didn't want to take this up considering they have worked with Intel previously.Billy Tallis - Tuesday, October 20, 2020 - link
Micron wouldn't have much reason to be interested. They already dissolved the IMFT partnership because Micron and Intel disagreed on where they wanted to go with flash R&D. Micron might be able to benefit somewhat from Intel's in-house SSD controller IP, but I'm not sure that's reason enough for them to acquire flash tech they already turned down.5080 - Tuesday, October 20, 2020 - link
Seems like a bad deal without the inclusion of the Optane business. NAND is a constrained commodity now with market saturation, too many competitors resulting in falling profits. Not a business I would invest heavily into.Arsenica - Tuesday, October 20, 2020 - link
If SK sees the deal as getting rid of a competitor it may be worth to them. Intel's position in the datacenter SSD market is also valuable.I'm guessing that Intel will keep its plan to keep developing Optane products in their New Mexico fab.
trivik12 - Tuesday, October 20, 2020 - link
There is no way Intel will sell Optane. That is a diffrentiating product for their servers. Its more than just a stand alone product.SK bought Intel NAND for basically Enterprise customer base. Their ruler SSD's are used in most DC. SK can continue to thrive in that market.
BedfordTim - Tuesday, October 20, 2020 - link
I imagine the reason for the stretched out deal, is that it protects SK Hynix against potential US sanctions. A US attack on the fab is a risk they were probably not willing to take.mitsuhashi - Thursday, October 22, 2020 - link
No disrespect but I really doubt fear of sanctions were a significant part of the purchase. South Korea and Taiwan's huge place in semiconductors and their status as US allies is precisely what allows Trump to sanction Chinese tech firms. Sanctioning Korea would mean sanctioning Samsung and LG in addition to SK which would make the world suddenly very, very, very short on critical necessities like RAM, processors, and displays. And then the US would also be forcing these firms to hold hands with China. Terrible option.Billy Tallis - Thursday, October 22, 2020 - link
I think he was referring to the possibility of US sanctions against China affecting the fab in Dalian, China that SK hynix is buying from Intel.Kangal - Saturday, October 24, 2020 - link
Yeah, I can't help but feel Intel ripped-off SK Hynix with this transfer.$9 Billion seems stupid when you realise Intel isn't that strong of a competitor in the flash market, especially when Optane is not included in the deal. The only reason I can think of the price being so high is because 1- Intel is greedy and needs the money for R&D to battle Zen3.
The 2- second reason, is probably they can insure sanctions against China doesn't happen for SSDs because of this one deal. Whilst, SK Hynix is a South Korean corporation and not Chinese, this fab is located in Northern China. On this point we can only speculate. A wild conspiracy theory would be that they want the IP from Intel's SSD division to be transferred to Mainland China, even if done so by proxy allied corporation. The recent ImaginationTechnologies (PowerVR) deal does bring up the realities of such scenarios.
PandaBear - Friday, October 30, 2020 - link
Intel has a strong enterprise SSD presence and SK Hynix doesn't. Intel's nand is old tech by now and I think that ESSD market share and know how is where SK want. The fab might be used for other future controller / old product for SK instead of nand.James5mith - Tuesday, October 20, 2020 - link
Don't care. Just want Alder Stream Optane 2.0 NVMe PCIe4.0 drives this year.romrunning - Tuesday, October 20, 2020 - link
Seconded!0mnibuss - Tuesday, October 20, 2020 - link
Would love to see this and Micron's version, have to worry about power and heat. But I don't see speed as a roadblock.Unashamed_unoriginal_username_x86 - Tuesday, October 20, 2020 - link
This might be the worst thing in tech news, people contributing nothing and just complaining about things entirely unconstructively. Saying that won't even do anything, Intel's probably decided approximately when they're launching anyway.ballsystemlord - Tuesday, October 20, 2020 - link
Do you want to see the prices?ichaya - Wednesday, October 21, 2020 - link
TBH, Mobile SoCs with Optane when?Billy Tallis - Wednesday, October 21, 2020 - link
Mobile SoCs with Optane won't happen, or won't matter, because Intel can't do mobile SoCs worth a damn. Somebody else might eventually be able to do a mobile SoC with 3D XPoint support depending on what Micron does with 3DXP long-term. But for now, there's not much indication that 3DXP can do anything interesting when constrained to a mobile-friendly power envelope.ichaya - Wednesday, October 21, 2020 - link
One of the first uses of 3D XPoint was in laptops as swap/cache, so I think you might be jumping the gun a bit. I could see 3D XPoint helping with faster startup, NV data storage and as swap/cache for all sorts of devices because of it's NV and performance properties.Billy Tallis - Thursday, October 22, 2020 - link
I'm not jumping the gun, I'm just going off my own power measurements. Intel's Optane cache modules aren't super low power devices. They are still about twice as efficient as the best flash-based NVMe SSDs (SK hynix P31) for random reads, but for other workloads and for idle power those Optane modules aren't great. So far, no Optane product has demonstrated power characteristics that indicate 3DXP would be suitable for mobile/smartphone usage, and Intel's definitely not going to prioritize energy efficiency over performance for future generations.ichaya - Friday, October 23, 2020 - link
Sorry to link to your article, but the most relevant article here would likely be: https://www.anandtech.com/show/12512/the-intel-opt... and you'll have to excuse my wishful thinking about Optane in low power situation. Thank you for the replies & articles.Optane seems ideal for a compiler cache for my desktop use case, so looking forward to any 800p and/or H10 successors in the consumer market. I'm not sure if the H10 will fully work with any AMD board :(
TheinsanegamerN - Tuesday, October 20, 2020 - link
NAND flash is still facing a likely dead-end in the long run, as NAND has been scaled down almost as small as it can go before leakage becomes a serious problem2001: "process nodes are reachign their limits with 130nm, 90nm si so small that quantum tunneling will thwart any further attempts to shrink nodes without changing to materials too expensive to use for proceesors.
1980: "8MB hard drives are pushing the limitations of what is possible with spinning magnetic media. 10MB is too unreliable and its likely hard drives will never reach 20 MB
If this inndustry has taught us ANYTHING, its that when someone says it cant be made smaller it gets made smaller within like 2 years.
0mnibuss - Tuesday, October 20, 2020 - link
Yeah, they'll continue to stack and MLC and PAM and add LP-DRAM to grow density and speed. In the consumer space, the "storage" requirements aren't nearly as a concern as the cloud/enterprise space; so NAND is still viable for the foreseeable future to the consumer.ballsystemlord - Tuesday, October 20, 2020 - link
Within 2 years? Like 14nm got replaced by 10nm in only 2 years?Unashamed_unoriginal_username_x86 - Tuesday, October 20, 2020 - link
Thank you for pointing out a singular case that goes against what they saiderotomania - Tuesday, October 20, 2020 - link
Thank you for making sense. (Bonus points - more than once!) Please keep patrolling!Zizy - Tuesday, October 20, 2020 - link
Nobody said 10nm is impossible. Also if you take a look at TSMC they indeed did their jump in about 2 years.ballsystemlord - Wednesday, October 21, 2020 - link
If you recall, TSMC's 7nm brings them to parity with Intel's 14nm.name99 - Wednesday, October 21, 2020 - link
Wow! Used to be the Intel crowd were claiming TSMC 7nm had parity with Intel 10nm. Now it's parity with Intel 14nm.Next year I expect it will be parity with Intel 22nm, and TSMC 3nm is what's needed to match Intel 10nm...
Define "parity".
The definition Intel was keen on for many years (right up until things went horribly wrong) was transistor density. By that metric 7nm matches Intel 10nm in specs (both are about 100 MTr/mm^2)
with the important difference that TSMC has actually shipped a variety of products at that density (eg Apple and Huawei SoCs) whereas the best Intel has shipped on 10nm has less than half the claimed density, and the newest 10nm has a density even lower.
ballsystemlord - Wednesday, October 21, 2020 - link
I was speaking in terms of performance (IE GHz).ballsystemlord - Wednesday, October 21, 2020 - link
Oh, in case you want to point out that AMD's CPUs only go to 4.9GHz (last gen was 4.7GHz IIRC), compared to Intel's fastest at 5.1GHz, Intel's 14nm has been improved over time. TSMC's 7nm will probably also be improved.Spunjji - Wednesday, October 21, 2020 - link
You're not really comparing equivalent things. It's not that they haven't figured out *how* to scale NAND smaller and are theorising about what would prevent that (as with your examples) - they already scaled it smaller and it doesn't work well.Scaling smaller would require fundamentally redesigning how it works, at which point it's not NAND anymore (e.g. X-point).
K_Space - Sunday, October 25, 2020 - link
Moore's law was young and energetic back then. It's at death doors now if it hasn't already.PandaBear - Friday, October 30, 2020 - link
Currently most major players have 96-128 layer class nand in the pipeline, and soon 192 layers.eva02langley - Tuesday, October 20, 2020 - link
So... the Total Available Market (TAM) strategy from Bob Swan is no longer viable...?What is Intel going to do if they cannot compete in any or their key markets...?
-Fabs are in terrible shape;
-CPUs are infested of hardware flaws;
-Servers cannot compete;
-Mobile are in disarray;
-5G modems are gone;
-Memory take the hit now;
ballsystemlord - Tuesday, October 20, 2020 - link
They have strengths.-Optane in DDR5 DIMMs. (It isn't going away.)
-Networking chips (Intel has the best that there is.)
-People don't seem to care about HW flaws (Sad, but true.)
-CPU HW video encoding (Some people really love it.)
-10nm seems to be viable (Finally) and the 1C boost freq are about what AMDs CPUs can do.
RSAUser - Thursday, October 22, 2020 - link
The latter most is misleading, as the 5000 series has better IPC, and most Intel chips can't clock that high, that's why the supply shortage of their "best gaming" CPU.trivik12 - Tuesday, October 20, 2020 - link
SSD's are commodity. Their key market for future there is Optane. Where they go with it we will know once 2nd gen has launched and see how much they have improved with PCI-E4 interface. I am expecting Alder Stream to blow away all Nand SSD. If not then it will be disappointing. They still have products in different segments. Networking(Smart NIC, Barefoot Networks, Silicon Photonics), FPGA/ASICs, Mobile Eye. Dont forget most AMD laptops launched recently have Intel Wi-fi chips. So they have products. They have to get their manufacturing in control.Spunjji - Wednesday, October 21, 2020 - link
"-CPUs are infested of hardware flaws"Pretty sure they'll be out of the woods on this one (for a while) when they finally get Ice Lake out in enterprise. It won't be the *most secure* platform available, but it won't be "infested" either.
Similarly, they *can* compete in servers, because they are. Their offerings are mediocre but they can currently still make up for that with sheer volume of supply and broad availability from OEMs.
Kallan007 - Tuesday, October 20, 2020 - link
Chinese are buying up everything and people complain about doing business with China, well here is a perfect example of why!MrCommunistGen - Tuesday, October 20, 2020 - link
SK Hynix is a South Korean company, so I'm not seeing why this is a perfect example of anything related to China.Kallan007 - Tuesday, October 20, 2020 - link
Ok, I stand corrected! My bad.lmcd - Tuesday, October 20, 2020 - link
Yea lol the SK is South Korea, and the article mentions how they rival Samsung within the region.Spunjji - Wednesday, October 21, 2020 - link
Even if they were Chinese, the post wouldn't make sense. Nobody's obliged to sell their company to China.Roy2002 - Tuesday, October 20, 2020 - link
What? China? And why this is an example? For what? What you are complaining even if Hynix is a Chinese company? What's your point???name99 - Tuesday, October 20, 2020 - link
The whole thing looks like financial engineering (Intel's new core competence, as opposed to real engineering).Robert Cringely, back when he was relevant, mid to late 90s (and Microsoft was king) used to refer to this quote from Frank Gaudette, MS CFO:
"Watch for any changes in our accounting,” said Gaudette. “If I need to I can start, depreciating the software and maintain earnings growth for years on flat revenue. Watch for the accounting changes, wait for the next uptick in the stock price, and then sell."
Well, the accounting changes came, and the changes in MS' status came with them.
Real engineering benefits all of humanity.
Financial engineering benefits whoever gets the bonuses, and the shareholders who are best at navigating the highs and lows....
Cullinaire - Tuesday, October 20, 2020 - link
Sad but true, at the end of the day public corporations are all vulnerable to this type of rot. All it takes is a weak succession and the vultures swoop in.Meteor2 - Wednesday, October 21, 2020 - link
Yeah because Microsoft is doing just *terribly*, 25 years laterSpunjji - Wednesday, October 21, 2020 - link
You might have skipped over a few significant developments in that 25 year period. They switched back to actually trying to innovate and pursuing new markets again (e.g. cloud services, games consoles).Spunjji - Wednesday, October 21, 2020 - link
Agreed. They'll magically show an increase in margins from doing this, which will help to offset the massive hits they're about to take in desktop and servers (they still seem to have a lock on notebooks and convertibles thanks to OEMs and Nvidia, we'll see how long that lasts).Zizy - Tuesday, October 20, 2020 - link
Eh, even if NAND stops scaling, this doesn't mean cost won't decrease. Take a look at the solar cells for example, great price per Wp wasn't due to efficiency improvement but due to incredible cost reduction. The same can happen for nand and keep it relevant for ages. Yes, it WILL hit physical wall eventually like every other tech, but I don't think it is going the way of dodo any time soon. Especially as no competing technology has better perf/$ yet.lmcd - Tuesday, October 20, 2020 - link
Yea imo there's assuredly room for yield improvements if nothing else. Also I'm sure controller optimizations could be huge for the space -- once NAND generations stabilize, controllers can be more directly designed for NAND performance profiles (re: SK Hynix's own P31).Rictorhell - Tuesday, October 20, 2020 - link
Very excited for this, IF it means more affordable higher capacity consumer SSDs in the somewhat near-term. I would even be thrilled with some "miraculous" SK Hynix capacity breakthrough, after acquiring all of Intel's research and patents, that allowed them to double or triple current maximum consumer SSD capacities. *a guy can dream*Billy Tallis - Tuesday, October 20, 2020 - link
To be fair, Intel has already been doing a good job of marketing their QLC to consumers at competitive prices. (That's now their entire consumer SSD strategy, since they stated earlier this year that they were not going to do any more consumer TLC drives.) So even if Intel held on to their NAND business, we could expect the upcoming 144L QLC and future generations to hit the consumer market in follow-ups to the 660p and 665p.Getting more options for consumer SSDs beyond 2TB is less a matter of technology right now than a matter of needing drive vendors willing to bring niche models to market. Sabrent seems to have helped break the ice a bit there, and now we're seeing other Phison partners launch 4TB or 8TB options.
nunya112 - Tuesday, October 20, 2020 - link
a fair few moves happening here. So intel and Micron have a pretty sweet relationship.and now Intel has sold its NAND business... That opens up a lot of questions for me . as to why get rid of a business that is really profitable.
Intel cant be hurting that bad for cash?? whats the real reason here????
Cullinaire - Tuesday, October 20, 2020 - link
did you even read the article? The division was hardly a fountain of profits for the past 5 years.PandaBear - Friday, October 30, 2020 - link
Intel NAND / SSD were falling behind in capacity, I think the money is in the 4-16TB SSD in enterprise right now and Intel is hardly catching up above 4TB, basically selling old stuff at the moment, and will get worse soon.Eliadbu - Wednesday, October 21, 2020 - link
To be honest, deal of that sort was long overdue. Intel fab buissnes is not competitive nor it has substantial revenue, while other player could use it to leverage it own business in the long run.rasheed - Wednesday, October 21, 2020 - link
5 years is a big time, this deal may not go throughYojimbo - Thursday, October 22, 2020 - link
$7 billion will be paid soon, in 2021 if I remember. The remaining $2 billion is to be paid in 2025. That makes it very likely the deal with go through, I believe, because SK Hynix is unlikely to walk away when they've alresdy paid over 3/4 the cost and Intel likely doesn't have an option to change their mind. I guess there could be something in the agreement about sending the $7 billion back, but I doubt companies pass $7 billion back and forth very often.Yojimbo - Thursday, October 22, 2020 - link
I read the details of the deal here in this article. Seems Intel keeps the NAND IP and operates the fab until 2025. The deal is strange because Intel has little incentive to develop technology they are exiting and which they've already sold for a fixed price in the future. Plus Intel's NAND technology is different from the rest of the industry's includubg SK Hynix. I would guess that SK Hynix will change the fab over to charge trap when it takes it over.The 4 year period might be a way to get the maximum value of Intel's IP roadmap before it's sunsetted. A good deal of R&D that Intel has already done will probably be applicable 4 years out, after which it will peter out unless new foundational research is pursued starting now, which no one will now do since Intel is exiting the business. Since SK Hynix does not have the expertise in Intel's technology and isn't interested in it, Intel will manage the development and manufacture of the NAND that goes into the product lines SK Hynix is buying from Intel.
Indervinder - Friday, October 23, 2020 - link
Having worked for Intel for several years and have co-workers who have worked for Hynix, I think the reason Hynix extended the deal over 5 years the employees.The corporate cultures of Hynix and Intel are very different. If the transition happened right away Hynix would risk losing a lot of the employees. The corporate culture at Hynix is not that great. They work their employees very hard and their flash products are known in the industry for their low quality. Hynix doesn't understand the enterprise market and doesn't have the DNA to compete and innovate in that market.
With the employees under the Intel mothership for 5 year, Hynix stands a better chance of retaining them in the long term and potentially having some of the Intel corporate culture transferred to Hynix.
PandaBear - Friday, October 30, 2020 - link
Agree, SKH has quite a turnover and lots of politics between HQ and US. Buying Intel is likely part of the solution if they can keep that business running as it is with the same group of employees. If in 5 years they can get the market they'll pay the rest, if not they can probably walk off and get their money back.nfineon - Sunday, October 25, 2020 - link
Waste of fucking money, they could have purchased 3 brand new (3 billion a pop) FAB's to massively increase their production capability instead of buying the dying portion of Intel's NAND unit.. not even the Optane division so this is not killing a competitor, it's giving them 9 billion more ways to FUCK you harder in the future lol...PandaBear - Friday, October 30, 2020 - link
Try 20B for the latest and greatest. Increasing capacity doesn't make you more money (it will flood the market and drag everyone down).Enterprise SSD have a profit margin of 10-70% depends on the demand and supply. SKH does not have this market and Intel is #2 (pretty much the only thing Intel is still competitive in storage), that's where the price is for.