6 ms·
Memristors' one-year delay will hit IT in the wallet
- mikecane 14y agoWhen I first heard of HP's memristors, I was looking forward to them using it in the TouchPad, to really compete against Apple. Now this. FAIL all around, HP.
- ck2 14y agoThis might also explain why SSD prices are dropping fast. 50 cents per GB now and falling.
- mtgx 14y agoWhat short-sighted thinking. This is why you don't see much innovation from the big companies usually. This is like Apple thinking they shouldn't release the iPad because it would cannibalize their Mac sales. Instead of being an also-ran in flash, Hynix/HP could be the leaders in memristors. But it seems you won't see them thinking like that. Oh well, I guess Samsung will take the lead in that market, too, then, since I think they were the other main company working on memristors.
- unreal37 14y agoAn LOL at "this is why you don't see much innovation from the big companies" followed by a what-if-Apple-did-this example. Apple is a big company. The biggest. The article is B.S. If the release of this technology will replace all RAM, Flash memory, and other memory storage technologies as the author expects, they wouldn't need to delay it. And lastly, if you believe the article, it says it's not in HP's control anyways. It's the manufacture partner Hynix fault.
- batista 14y ago>An LOL at "this is why you don't see much innovation from the big companies" followed by a what-if-Apple-did-this example. Apple is a big company. The biggest. And a LOL for misunderstanding his point. He said that "this is why you don't see MUCH innovation from the big companies", not "why you don't see ANY" innovation. And he used Apple precisely as a _counter_example_ to that lack of innovation, saying that what HP did would be like a hypothetical Apple saying they won't release the iPad because that would cannibalise iPhone sales.
- ryanmolden 14y agoDoes HP have patents here that would prevent others from introducing this tech? Or are they/Hynix simply the only ones with experience with the tech to pull it off? This is such a clear example of one of the fears spelled out in The Innovators Dilemma it isn't even funny. Edit: whoops, mtgx says Samsung is also in this space, so it sounds like patents can't be an obstacle, or at least until the suing starts.
- raverbashing 14y agoOh really, Samsung is going to compete with them? So let's speculate 1 - They delayed this because of manufacturing problems and want to save face 2- They really are that stupid and gave a competitor a 1 year head start (or they made a Gentleman's Agreement - but it wouldn't prevent even another competitor getting in) In short, this is extremely risky.
- ryanmolden 14y ago3) They feel they are far enough ahead of competitors that they can afford to "drain the pipeline of flash based products" and still come out ahead in memristors ultimately. I agree a purposeful delay in order to try and squeeze some drops out of existing stock / sunk costs would be risky, and likely stupid. Clearly no corporation has ever made short-sighted, stupid decisions like that before so this would totally be a first :)
- zrail 14y agoI don't see why this is a big deal. It's a year, so what? We've waited decades, a year isn't going to hurt anything. Hynix is a big business that is heavily invested in flash and it takes a lot to completely change the core of your business over, especially when you have to possibly build new fabs and fab technology.
- ck2 14y agoDouble the storage space and possibly double the reliability for the same price? Also potential power savings. A year is a huge deal if someone was planning on spending a million on a datacenter to make it use SSD for database, etc.
- zrail 14y agoSure, but making plans for a million dollar data center based on nebulous timelines and an unproven technology shouldn't be happening anyway. Your point about it being better is true, but the opportunity cost of going with the known right now is essentially zero.
- evgen 14y agoWhere do you get "for the same price"? I can assure you that even if memresistors cost 1/10 to produce the equiv in other tech they would hit the market at 10x the price. Better tech can get a higher price and as long as the ability to provide market demand is limited by patents or the cost to convert/build fabs then the supplier gets to set the price. Seems like a lot of people in this discussion failed Econ 101.
- Dylan16807 14y agoCompetition pushes the price of a product toward the manufacturing cost of a product over time.
- Tagbert 14y agoCould they not charge a premium price for memristors early on to prevent flash cannibalization and then bring the price down as they transition their output?
- Raphael 14y agoPerhaps they are worried about reverse-engineering, so don't want to release any until they have scaled up manufacturing.
- justincormack 14y agoThats what I thought too. I did think they were going to have a sufficient difference to flash that they would be able to compete in a different way. The HPC market is the classic start point. However, I think there may be other reasons. Maybe they haven't worked out the connection architecture properly?
- ricardobeat 14y agoIt only makes sense to continue selling the older, higher margin products until a competitor comes up. They are asking for it. I would even root for Samsung on this (also because it just finished my laundry).
- tsotha 14y ago>It only makes sense to continue selling the older, higher margin products until a competitor comes up. If you're the only one on the market with memristors the margin is determined by whatever the market will bear, and if they're really that much better than existing technologies it will be much larger. None of this conspiracy stuff makes any sense. They're delaying a year because they can't manufacture the product.
- batgaijin 14y agohttp://www.newscientist.com/article/mg21328535.200-online-spat-over-who-joins-memristor-club.html http://www.newscientist.com/article/mg21328535.200-online-sp... It seems like HP still hasn't been issued patents for it (at least by my googling) so I guess they are happy to just sit on their tech. Also, Memristors are already being actively used for military technology, most notable the SyNAPSE project.
- ck2 14y agoWhy on earth was my title replaced? Sigh. Revolutionary memristors purposely delayed to prevent cannibalization of flash It's 100% accurate and more relevant.
- johnpmayer 14y agoMaybe something HN does automatically to prevent editorializing. Not saying that your title in particular was doing that.
- jvm 14y agoThe ridiculous thing is it's not like the original headline wasn't editorializing.
- mikeash 14y agoIt makes no sense to me. What's the advantage of memristor-based storage, exactly? If it's cheaper to make, then sell storage at a price that undercuts the competition but still makes a better profi. If it's more reliable, then sell storage at a higher price with a better warranty. There's no sensible scenario where memristors are ready to go but they don't want to sell any memristor products yet because it will hurt existing sales. Either they can make a better product, or they can't, and if they can't make a better product then the stuff simply isn't ready.
- jtmcmc 14y agowhat if they have a bunch of flash products already made? They need to get rid of that product first before they make it worth less.
- mikeash 14y agoThey're talking about delaying commercialization from a year out to sometime farther. There's no way they have a year's worth of flash products sitting in inventory.
- pbharrin 14y agoThis is a stupid article. The author makes the assumption that HP and Hynix COULD ramp to high volume production, however it is clear from HP's statement that this is not true: "As with many other ground-breaking technologies being developed at HP Labs, HP has not yet committed to a specific product roadmap for memristor-based products," Let me also say that I worked in the non-volatile memory business for seven years. Any organization that can scale better than its competitors will do so, unless they don't like making money. End of story. If HP and Hynix are not ramping to high volume, it's because they are not ready, not some stupid conspiracy theory that they want to stagnate the development of cloud-buzzword "technology".
- petermonsson 14y agoAh, a breath of fresh sanity from someone who knows the industry. That is very welcome, thank you. The semiconductor industry is brutal in general. Selling to consumers is worse, but the worst by far is the part that sells any type of memory. Forget the conspiracy theories. This space is full of dead companies. Someone will eat you.
- batista 14y ago>The semiconductor industry is brutal in general. Selling to consumers is worse, but the worst by far is the part that sells any type of memory. Really? Because there have been all kinds of price-fixing scandals and accusations for the last 2 decades... Price fixing does not sound very cut-throat to me...
- petermonsson 14y agoThe only DRAM price fixing scandal that I know of is the one from 1998 to 2002 which gave the involved parties litigation on their backs up until 2010. If you know of any others, please let me know. Google did not turn up anything else. Anyway, I want to explain why the DRAM world is brutal: Here is the DRAM top 10 from 2006: http://i.cmpnet.com/eetimes/seminews/2007/chart1_020107.gif http://i.cmpnet.com/eetimes/seminews/2007/chart1_020107.gif And today the state is as follows: http://www.electronicsweekly.com/blogs/david-manners-semiconductor-blog/2012/03/top-ten-less-2-dram-companies.html http://www.electronicsweekly.com/blogs/david-manners-semicon... Samsung: Eating everyone Hynix: Almost breaking even Qimonda: Dead Micron: Making money Elpida: Bankrupt Nanya: Losing more money than it makes Powerchip: Barely alive with 0.1% market share ProMos: Has sold its fabs, soon to be dead Etron: Alive, but with less that 0.1% market share and 15 cent loss per dollar sold. Winbond: Now only in specialty DRAM So, Samsung and Micron are doing well, Hynix is barely OK and the rest are sunk. Yes, I believe that the DRAM competition is fierce and brutal.
- ahuibers 14y agoI haven't seen any lab demonstrations of multi Mbit devices. Until we see that this is years away from production in my experience as a device physics researcher. All storage technologies so far have taken 10 years plus to have large market share and I don't see this as an exception.
- Klinky 14y agoI'll believe memristor tech when I see commercial products of it available. I don't think it's far fetched that a company is slow to shift it's mainstay, profitable business product to an unproven new technology that has no mainstream commercial products out yet. Cutting edge tech is always rife with delays and shortcomings that weren't anticipated until the rubber actually hit the road. While many in these comments are touting "double storage & reliability for the same price", that would be for raw manufacturing costs, not including all the R&D money and the cost of fabrication plants that must be built or retooled. It is highly unlikely consumers will see those benefits anytime soon.
- adaml_623 14y agoThis catchy headline is essentially saying that company X is going to watch their growth and expenditure and maximise their profits in a manner which won't benefit consumers. The headline is interesting because it's stating the obvious, standard business practice but possibly they have hit on the only way to glam up the incredibly boring story of 'product will be delayed by 12 months'.
- ameasure 14y agoHynix (2nd largest memory chip maker after Samsung) and friends were fined just a couple years ago for price fixing memory, http://www.bbc.co.uk/news/10126755 http://www.bbc.co.uk/news/10126755. If competition isn't a problem, why waste money on costly innovation? Reminds me of Eastman Kodak circa 1975 http://en.wikipedia.org/wiki/Eastman_Kodak#Shift_to_digital http://en.wikipedia.org/wiki/Eastman_Kodak#Shift_to_digital. That said, this is pretty weak evidence of collusion.
- ngvrnd 14y agoThis is worrisome: http://arxiv.org/abs/1207.7319 http://arxiv.org/abs/1207.7319