Demand for AI-capable infrastructure is causing server buyers to significantly up orders for enterprise solid state drives (SSDs), and this is having an inflationary impact on the price of flash-based storage units.
These details come courtesy of market watcher TrendForce, which estimates that the growth rate of SSD demand for AI applications is likely to exceed 60 percent, with suppliers accelerating development of higher capacity products.
The inevitable downside of ballooning demand for enterprise SSDs is a spike in prices. Contract prices for this category rose by more than 80 percent over the period from 4Q23 to 3Q24, TrendForce claims.
SSDs play a key role in AI model training thanks to their low latency and high throughput. They are primarily used for storing model parameters including weights and deviations, the analyst says, but are also useful for creating checkpoints - periodically saving a snapshot of AI model progress to aid recovery in case of interruptions.
The flash-based devices are useful in inferencing too, where SSDs allow for data updates to models in real time to fine-tune outcomes. As more information is generated, the storage capacity required also increases, making high-capacity SSDs such as TLC/QLC 16 TB or larger the preferred choice for these applications.
As a consequence, the market for AI server SSDs has seen an upswing in demand for products larger than 16 TB from the second quarter onwards, TrendForce reports. It estimates that overall capacity of AI-related SSDs procured during 2024 will exceed 45 EB (exabytes), where 1 EB is equal to 1,000 petabytes or a million terabytes.
Meanwhile, demand for AI servers is touted to push the average annual growth rate for SSDs to more 60 percent over the next few years, with SSDs that target AI potentially rising from 5 percent of total NAND Flash consumption in 2024 to 9 percent in 2025.
Earlier this year, system builder Dell warned of a hike in both DRAM and SSDs prices of about 20 percent, saying it expected to see "a step function in cost in the second half of the year driven primarily by DRAM and SSD," and it would be forced to "adjust its prices accordingly."
An earlier report from TrendForce also revealed the growth in NAND flash revenue among chipmakers; it was up 28.1 percent to $14.71 billion in the first calendar quarter of 2024, fueled by demand for enterprise SSDs to fit out servers for AI processing.
One SSD maker, Western Digital, recently reported revenue up 41 percent to $3.8 billion for the quarter ending June 30, compared with the same period last year, and said it expects enterprise SSDs to represent a "double-digit percent share" of its portfolio mix for its fiscal 2025.
IDC's Senior Research Director for EMEA Andrew Buss agreed AI demand is propelling the price of SSDs, as well as other components.
"GenAI demand is certainly driving up demand for high performance storage, and this is resulting in some limitations in NAND flash availability which has had the tendency of driving up storage prices," he told us.
From personal experience Buss has noticed consumer SSDs have seen "a significant and sustained uplift in overall price for the capacity."
Trendforce says that in response, suppliers are accelerating process upgrades and aiming for products that user higher density NAND flash components to drive up capacity, and this will eventually produce 120 TB enterprise SSDs. ®
Aug 14): Norway’s Government Pension Fund Global (GPFG) raised its equity investment in Malaysia in June from the end of 2023 as the world’s largest sovereign wealth fund grew in asset size and returns.
The fund also known as Norway’s Oil Fund has 214 Malaysian stocks worth US$2.32 billion (RM10.26 billion) in its portfolio by the end of June, according to its half-yearly disclosure. That compares with 180 companies worth US$2.05 billion on its books at the end of December.
GPFG also has US$178 million worth of Malaysian government bonds, but does not have any investments in real estate or renewable energy infrastructure in the country.
Globally, GPFG generated a 12% return from equity investments but suffered losses of 1% from fixed income, 1% from unlisted real estate and 18% from renewable energy infrastructure.
"The equity investments gave a very strong return in the first half of the year,” said Nicolai Tangen, the chief executive of Norges Bank Investment Management that manages GPFG. “The result was mainly driven by the technology stocks, due to increased demand for new solutions in artificial intelligence.”
Malaysia nevertheless accounted for only 0.1% of all its global investment valued at US$1.55 trillion.
In Malaysia, new investment by GPFG included oil-and-gas services firm Keyfield International Bhd (KL:KEYFIELD), JCY International Bhd (KL:JCY) that makes components for computer storage and telecom services firm OCK Group Bhd (KL:OCK).
GPFG however dumped electronic services firm Kobay Technology Bhd (KL:KOBAY), speaker manufacturer Formosa Prosonic Industries Bhd (KL:FPI) and hypermarket operator Aeon Co M Bhd (KL:AEON), among others.
In terms of sectors, GPFG’s exposure is the highest in financials followed by industrial and technology. The fund raised its investment sum the most in real estate by 93%, energy by 79% and consumer staples by 43%.
The fund’s top three holdings in Malaysia by market value were CIMB Group Holdings Bhd (KL:CIMB), Malayan Banking Bhd (KL:MAYBANK) and Public Bank Bhd (KL:PBBANK).
(recycled post about volatility, just for some general reinvigoration, inspiration and motivation)
So why bother to bodoh2 play volatile counter huh. Why not play strong solid secure counters like Maybank, Gamuda, Ranhill, YTL Power, SimeProp etc. leh.
Because, especially for those who really know, it is indeed the volatile stocks which give the best most lucrative returns. Honestly. How?
Because it's only with volatile stocks that that you can buy so low, then sell so high, sometimes within the space of only two weeks, sometimes a bit longer, depending on market conditions and investors' mood. You in fact RIDE on the volatility, i.e. you actually positively exploit it. Instead of whining crying weeping about it.
You can't do that with Maybank, Gamuda, Ranhill, YTL Power, SimeProp etc. Nope. Oh no. It just does not happen.
Simply because they are just what they are. Too rock solid, too stable, too steady. So, how are you going to make anything from something that you bought at RM5, let's just say, when it just stays at around that same unmoving price for 7 or 8 months. You just can't.
Main precaution being, you would need to find and choose a volatile stock that also has relatively sturdy robust fundamentals, bright prospects and a strong upside potential. Definitely not something in danger of sliding into PN17 in the near term, obviously. For your own safety.
One more thing. It's all well and good to be thinking of making good gains within a month or less. But then OTOH, you also need to be ready to go longer, just in case your one month or less time frame for your exit plan does not work out according to schedule.
Even long time very very long time hard core very very hard core JCY sceptics critics cynics including Mz Vicky are now staunch converted faithful JCY believers and supporters. Which just says it all about the irresistible appeal and power of JCY.
Again sentiment rules. Not news. Not profits. Investors cash out and put inside pocket due to global and local uncertainties. Prices crash like no man's land !
Tech Stocks to Buy on the Rebound - Human and “machine” agree: any of these have the potential to be breakout stocks 3d ago · By Thomas Yeung, CFA, InvestorPlace Markets Analyst
Vicky: "Again sentiment rules. Not news. Not profits. Investors cash out and put inside pocket due to global and local uncertainties. Prices crash like no man's land !"
Regarding sentiment effect, no need for you to keep singing that stale old song like sick old parrot la. Everybody who been in Bursa long enough already would know that stock price is at least 75% about perception.
While you waited until so late because you were always so cynical so critical so sceptical of JCY, then you jumped in only at 68 on Tuesday.
Then yesterday just because you cannot tahan small drop, you jumped ship at 65. Today price surged back up to 68. Banyak padan sama you punya sour grape sore arse hurt butt bitter heart punya muka lor, ki ki ki ha ha lol.😹😂🤣
Key Points: 1. Former Google CEO Eric Schmidt said in a recent talk at Stanford that tech giants are planning for increasingly large investments into Nvidia-based AI data centers.
2. “I’m talking to the big companies, and the big companies are telling me they need $20 billion, $50 billion, $100 billion” in AI infrastructure, Schmidt said.
3. He made the remarks in a video this week, that was posted by Stanford but subsequently removed, on the topic of AI.
I come jcy forum because Calvin spam to my banking stock to ask me to buy hdds...
So, I come here to spam hdds forum to ask Calvin to cut loses in hdds because I don't want to gamble in low margin and high risk hdds which American don't want to Mfg.. And up 400% from 20c!
So you were already gambling in Notion that time lor. You padam erased all your posts in Notion geh, ha ha ha.
You x pandai main JCY lor, if like that one. So, where were you around 10 weeks ago, when guys were buying at 62.5 then selling at 92.5 within 2 weeks leh.
pang72: "I have pocket my profit before hdds crashing"
You made chicken feed profit only lor. If at all. Not like other smarter guys who made bucketloads.
Because your game always based on chasing high then selling not so high lor.
Ini mcm punya game kalo, mmg confirm sudah bnyk kali cut loss panic selling lor, because if chasing high then got high risk of correction after buying also lor.
Guys, dont feel threaten or envy if someone like/support this stock. If they buy, surely they know abt TAYOR. Its their money n their own decision afyer wise analysis.
AI/DC is the trend now, if people want to ride the theme, let them, maybe its an opportunity .
US stocks ended higher on Friday, extending their BIGGEST WEEKLY PERCENTAGE GAINS OF THE YEAR as worries of an economic downturn eased and investors focused on the Jackson Hole Economic Symposium next week.
So, take note guys, especially them chronically continuously naysaying doubting detracting downdragging guys. No more babbling on about IMAGINED "US recession", all right.
This book is the result of the author's many years of experience and observation throughout his 26 years in the stockbroking industry. It was written for general public to learn to invest based on facts and not on fantasies or hearsay....
bryan2003
70 posts
Posted by bryan2003 > 1 month ago | Report Abuse
Demand for AI-capable infrastructure is causing server buyers to significantly up orders for enterprise solid state drives (SSDs), and this is having an inflationary impact on the price of flash-based storage units.
These details come courtesy of market watcher TrendForce, which estimates that the growth rate of SSD demand for AI applications is likely to exceed 60 percent, with suppliers accelerating development of higher capacity products.
The inevitable downside of ballooning demand for enterprise SSDs is a spike in prices. Contract prices for this category rose by more than 80 percent over the period from 4Q23 to 3Q24, TrendForce claims.
SSDs play a key role in AI model training thanks to their low latency and high throughput. They are primarily used for storing model parameters including weights and deviations, the analyst says, but are also useful for creating checkpoints - periodically saving a snapshot of AI model progress to aid recovery in case of interruptions.
The flash-based devices are useful in inferencing too, where SSDs allow for data updates to models in real time to fine-tune outcomes. As more information is generated, the storage capacity required also increases, making high-capacity SSDs such as TLC/QLC 16 TB or larger the preferred choice for these applications.
As a consequence, the market for AI server SSDs has seen an upswing in demand for products larger than 16 TB from the second quarter onwards, TrendForce reports. It estimates that overall capacity of AI-related SSDs procured during 2024 will exceed 45 EB (exabytes), where 1 EB is equal to 1,000 petabytes or a million terabytes.
Meanwhile, demand for AI servers is touted to push the average annual growth rate for SSDs to more 60 percent over the next few years, with SSDs that target AI potentially rising from 5 percent of total NAND Flash consumption in 2024 to 9 percent in 2025.
Earlier this year, system builder Dell warned of a hike in both DRAM and SSDs prices of about 20 percent, saying it expected to see "a step function in cost in the second half of the year driven primarily by DRAM and SSD," and it would be forced to "adjust its prices accordingly."
An earlier report from TrendForce also revealed the growth in NAND flash revenue among chipmakers; it was up 28.1 percent to $14.71 billion in the first calendar quarter of 2024, fueled by demand for enterprise SSDs to fit out servers for AI processing.
One SSD maker, Western Digital, recently reported revenue up 41 percent to $3.8 billion for the quarter ending June 30, compared with the same period last year, and said it expects enterprise SSDs to represent a "double-digit percent share" of its portfolio mix for its fiscal 2025.
IDC's Senior Research Director for EMEA Andrew Buss agreed AI demand is propelling the price of SSDs, as well as other components.
"GenAI demand is certainly driving up demand for high performance storage, and this is resulting in some limitations in NAND flash availability which has had the tendency of driving up storage prices," he told us.
From personal experience Buss has noticed consumer SSDs have seen "a significant and sustained uplift in overall price for the capacity."
Trendforce says that in response, suppliers are accelerating process upgrades and aiming for products that user higher density NAND flash components to drive up capacity, and this will eventually produce 120 TB enterprise SSDs. ®