The memory market starts to show a slowdown after months of steady price hikes according to TrendForce's latest memory spot price trend report. DDR5 memory prices, which had been going up consistently, are now stabilizing as they've hit fairly high levels that sometimes go beyond contract prices. While DDR5 still sees small price increases compared to DDR4, buyers and traders are growing more wary and leaning towards DDR4 products. The common DDR4 chips (i.e., DDR4 1Gx8 3200 MT/s) saw a big 12.8% rise this week going from $2.18 to $2.46, but industry experts think these big price jumps will ease off in the third quarter.

NAND flash memory prices are cooling off after big jumps since late February. NAND flash chip costs have shot up, but customers are losing interest as more products flood the market. This change in supply and demand is showing up in the numbers - 512 GB TLC chips dipped by 0.18% this week to $2.73, as buyers take a "wait and see" approach. The big picture hints that the memory market might be settling down after months of steep price hikes. Nevertheless, major North American cloud providers' AI investments drive enterprise SSD demand surge, creating a supply shortage and supporting up to 10% quarterly price increases in Q3 2025.

Martin Lau, President of Tencent, has divulged that his company has accumulated a "pretty strong stockpile" of NVIDIA AI chips. In a mid-week earnings call, the Chinese executive reckoned that this surplus will come in handy—upon the company unleashing its full-on upcoming "AI strategy." Lau was responding to a question regarding ripples caused by a recent introduction of revised licensing requirements for "high-end GPUs." His lengthy reply seems to align with "leaked April time" information; when industry analysts theorized a massive $16 billion spend—reportedly, big Chinese tech firms had splurged out with swift acquisitions of NVIDIA H20 GPUs. Lau commented on present day conditions: "it's actually a very dynamic situation right now. Since the last earnings call, we have seen an H20 ban, and then after that there was the BIS new guidelines that just came in overnight...If you look at the allocation of the usage of these chips, obviously they'll be used for the applications that will generate immediate returns for us. For example, in the advertising business as well as content recommendation product, where we actually would be using a lot of these GPUs to generate results and generate returns for us. Secondly, in terms of the training of our large language models, they will be of the next priority and the training actually requires higher-end chips."

Team Green's engineering team has likely been strong-armed into designing further compromised hardware; as "exclusive" sanction-conforming options for important enterprise customers in China. Tencent seems to have enough pre-ban specimens to tide things over, for a while. The firm's president envisioned a comfortable position, for the foreseeable future: "over the past few months, we (started) to move off the concept or the belief of American tech companies—which they call 'the scaling law'—which required continuous expansion of the training cluster. And now we can see even with a smaller cluster you can actually achieve very good training results. And there's a lot of potential that we can get on the post-training side which do not necessarily meet very large clusters. We should have enough high-end chips to continue our training of models for a few more generations going forward." Huawei's controversial Ascend 910C AI accelerator seems to be the top alternative contender; tech watchdogs believe that this design's fortunes will be closely tied to the rising dominance of DeepSeek. Fairly recent leaks have indicated impressive progress being made within China's domestic AI accelerator infrastructure.

Despite recent whispers of TSMC losing a key 4 nm node process customer, industry analysts reckon that Taiwan's premier foundry business will remain in a comfortable leading position for the foreseeable future. "Optimistic" expert opinion points to "revenue growth for supply chain players," driven by the rapid progress of the firm's 2 nm manufacturing prowess. Naturally, their cutting-edge manufacturing capabilities—supposedly bolstered by GAAFET—are being tracked with keen interest. According to a fairly fresh Ctee Taiwan news piece, the usual big players are reportedly queued up and present within factory order books. The likes of Apple, NVIDIA, AMD, Qualcomm, MediaTek and Broadcom are mentioned. Mid-way through last month, Team Red officially announced a big collaborative milestone: "(our) next-generation AMD EPYC processor—codenamed "Venice"—is the first HPC product in the industry to be taped out and brought up on the TSMC advanced 2 nm (N2) process technology."

Ctee's report cites recent statements by C. C. Wei. Apparently, the TSMC CEO has stressed (on multiple occasions) that there is "unprecedented" demand for his company's 2 nm production pipelines—far exceeding previous levels for 3 nm. In addition, TSMC reps—who are currently touring the States; hosting technology symposiums—have revealed 2 nm (N2) defect density trends. Ctee outlined these intriguing details: "(N2's) defect density (D0) performance is comparable to that of the 5 nm family, and even surpasses the 7 nm and 3 nm processes of the same period, making it one of the most technologically mature advanced nodes." Wei's foundry team seems to be well ahead of main competition; insiders reckon that alleged equipment upgrades signal a push into 1.4 nm territories. Crucially, mass production of 2 nm (N2) wafers is expected to begin later this year—a cross-facility push was uttered by industry moles.

As disclosed in a new press release—issued by the US Commerce Department—TSMC's North American operation has started another expansion. Last month, Taiwan's leading chip foundry committed a substantial $100 billion investment—eventually leading to a greater production footprint in Phoenix, Arizona. Reports suggest that ground has already been broken, in a low-key manner—as of yesterday (April 29)—at a planned third location, only hours after TSMC's receiving of permits—aka a "thumbs up" from the US government. According to local news outlets, key administrative representatives were in attendance to witness the initiation of construction work. TSMC's third plant is destined to pump out cutting-edge products via a 2 nm (N2) process technology, with Apple, NVIDIA and AMD confirmed as "front of the queue" customers. Despite recent fanfare and celebrations, industry analysts reckon that it will take up to a decade for the foundry's North American operation to solidify a dependable supply chain. In the interim, certain elements will require shipping to overseas locations—for packaging and finalization purposes.

A series of April leaks have suggested that Apple's mixed reality headset engineering team is concocting two distinct next-gen solutions. Mid-month, leakers shared alleged early shots of "Vision Air"-related connectors and external parts—hinting at a potential dark blue colorway. Combined with a selection of fairly legitimate-sounding predictions from a notorious industry watcher, so-called Vision Pro sequels are on the way. Apple's Chinese manufacturing partners are reportedly deep into mass production of crucial "Vision Pro 2" components. Mark Gurman's "Power On" newsletter has provided plenty of inside knowledge stories over the past couple of months—his latest article included a section dedicated to fresh VR/AR insights: "I reported earlier this month that Apple is full steam ahead on two new successors to the Vision Pro (2023): a lighter version at a cheaper price point, and a Mac-tethered model aimed at applications that need maximum responsiveness."

He continued: "all signs point to the lighter model arriving between the end of this year and the first half of 2026. Despite the first version selling poorly, the company isn't abandoning ship here. The main uncertainty is whether the lighter version will be considered a replacement for the Vision Pro or a cheaper alternative." In theory, Apple could test "more mainstream" gaming waters with an initial rollout of the claimed cheaper + lightweight "Vision Air" model—perhaps set to do battle with readily available rival devices; e.g. Meta's dominant Quest 3 range. A full-blown Vision Pro follow-up could launch later on in 2026—likely reserved for upper-crust customers, with an increased focus on productivity applications.

Global PC shipments grew 6.7% YoY in Q1 2025 to reach 61.4 million units, according to Counterpoint Research's preliminary data. The growth was mainly driven by PC vendors accelerating shipments ahead of US tariffs and the increasing adoption of AI-enabled PCs amid the end of Windows 10 support. However, this surge may be short-lived, as inventory levels are likely to stabilize in the next few weeks. The impact of the US tariffs is expected to dampen the growth momentum in 2025.

Apple and Lenovo delivered strong performances in the quarter, largely due to new product launches and market dynamics. Apple experienced 17% YoY growth in shipments, driven by its AI-capable M4-based MacBook series. Lenovo's 11% growth reflected its expansion into AI-enabled PCs and its diversified product portfolio. Lenovo remained the brand with the largest market share during the quarter. HP and Dell, on the other hand, benefited from the US market pull-ins during the quarter, with 6% and 4% YoY growth respectively, and maintained their second and third places in Q1. We also found that the pull-ins happened for other major brands too ahead of the tariff uncertainty, leading to the market share further consolidating around major brands.

Mid-way through April, we heard about sanctions affecting shipments of NVIDIA's H20 AI chips into China. Despite (rumored) best efforts made by Jensen Huang and colleagues, the US government has banned the export of Team Green's formerly sanction-conformant design. Similarly, NVIDIA prepared a slightly less potent GPU for gaming applications—exclusively for the Chinese market. Despite sporting a restricted GB202 "Blackwell" GPU die, the GeForce RTX 5090D 32 GB is still a monstrous prospect. According to Chinese PC hardware news sources, Team Green representatives have sent alerts to "all" of its Chinese add-in-card partners (AICs)—early warning signs have indicated a possible cut-off of GB202 GPUs in the near future. A member of the Chiphell forum disclosed some insider knowledge and dismissed unfounded speculation about RTX 5090D cards being replaced by "full fat" RTX 5090 options.

sthuasheng commented on Team Green's alleged bulletin—distributed at some point last week: "the notice only said that the supply of RTX 5090D was suspended, ...this did not mean any sales or transportation ban; it urged everyone not to make any speculations or judgments unless there was an official notice issued at a later date. After this notice was issued, each AIC began to notify agents to suspend sales, because the inventory of 5090D has always been very small, so it is necessary to keep these stocks to observe the subsequent situation and deal with the subsequent after-sales. At the same time, we might as well speculate that each AIC and dealer may also have the intention to stockpile 5090D units and then sell them at an elevated price." BenchLife.info decided to reach out to industry moles, following an absorption of various Chiphell whispers.

Up until very recently, the microSD Express format was viewed as a commercial disappointment for involved manufacturers and suppliers. Potential buyers were not attracted to premium price points; as demanded by designs that can leverage (up to) 900 MB/s read speeds. Hermitage Akihabara—a Japanese electronics and computer hardware chain—has tracked an alarmingly sharp climb in demand for microSD Express products. The popular retailer documented this recent trend across several articles, with last week's investigation diving deep into units selling: "at a rate of 337.62 cards per hour!" The upcoming launch of a highly-anticipated mainstream gaming device has seemingly sharply driven up sales of a largely forgotten storage format. Hermitage Akihabara explored this unprecedented turnaround in fortunes: "Nintendo's influence is amazing. The 'microSD Express card,' which had hardly any sales due to lack of use, sold out in an instant as soon as it was announced that it would be used in the Nintendo Switch 2. It was sold out not only in (our) Akihabara shops but also on EC sites nationwide, and it caused such a stir that it was even covered in the general news."

Hermitage Akihabara expects to replenish its stock within a non-specific timeframe—but prices could remain "quite high," due to manufacturers (reportedly) being fairly conservative with production output. The store's "behind the scenes of the microSD Express card sellout" news piece outlined some compelling data points: "our mail order service received a large number of orders. According to the calculations made by the mail order staff, they were selling at a rate of 5.63 units per minute, or 337.62 units per hour, which was also a surprise. Of course, we didn't have that much stock, but it seems that we were selling as much as we had. One interesting thing is that most of the buyers ordered only one copy. At most, three copies were purchased, and more than 80% of people ordered one copy. Most of the buyers were pure users who wanted to save one to use with the Switch 2." As reported earlier in the month, Lexar has readied the world's first 1 TB microSD Express card. Their $199.99 (MSRP) 1 TB PLAY PRO microSDXC Express model is already available to purchase, months in advance of Switch 2's global launch date (June 5). Industry insiders believe that Samsung is collaborating with Nintendo on an "official" range of microSD Express cards.

Not too long after delivering its in-depth Switch 2 showcase, Nintendo of America and retail partners postponed pre-launch reservations for the forthcoming hybrid gaming console. An official announcement stated: "pre-orders for Nintendo Switch 2 in the U.S. will not start April 9, 2025 in order to assess the potential impact of tariffs and evolving market conditions. Nintendo will update timing at a later date. The launch date of June 5, 2025 is unchanged." The vast majority of industry watchdogs reckon that company will be assessing an adjustment of launch day product MSRPs. Last week's North American presentation outlined price points of $449.99 and $499.99; for a standard package and a Mario Kart World Pack (respectively). Theorized adjustments could extend to physical copies of games, as well as accessories; e.g. the Switch 2 Pro Controller.

Gamers further north—in the neighboring nation of Canada—were relieved about last week's alert being non-applicable to their market region. Up until yesterday, pre-orders were on track at CA$629.99 (base console) and CA$699.99 (with Mario Kart World). Unfortunately, the House of Mario's Canuck branch has confirmed that recent events have had a knock-on effect for local customers. As disclosed in a statement to MobileSyrup, Nintendo of Canada explained: "pre-orders for Nintendo Switch 2 in Canada will not start on April 9, 2025 in order to align with the timing of pre-orders to be determined in the U.S....Nintendo will provide updated information at a later date. The launch date of June 5, 2025 is unchanged." MobileSyrup's report was updated yesterday evening; with links to social media bulletins posted by the regional arms of Walmart and GameStop. Both retail outlets repeated Nintendo of Canada's disappointing message.

Last week, industry reports pointed to evidence of NVIDIA H20 AI GPU shortages in China—supply chain insiders expressed frustration about limited availability, and alleged price hikes. Days later, local media outlets have disclosed staggering sales figures. Two unnamed sources opine that the likes of Tencent, Alibaba and ByteDance have spent roughly US$16 billion on H20 purchases, across the first three months of 2025. Back in February, Reuters noted an extraordinary surge in orders for: "(Team Green's) H20 model—the most advanced AI processor legally available in China under U.S. export controls—driven by booming demand for Chinese startup DeepSeek's low-cost AI models." The unprecedented rush—to secure precious AI-crunching hardware—was likely motivated by whispers of elevated restrictions; coming from across the Pacific Ocean.

Curiously, local government bodies have allegedly "advised" a stoppage of H20 orders—a recent Financial Times article suggested that this message was directed at the nation's largest AI players (mentioned above). A few industry moles believe that NVIDIA's engineering department is working on another Chinese market exclusive AI chip, although it is not clear whether a new entrant will be designed to conform to recently introduced "not very strict" environmental regulations. Anonymous sources have made noise about an upgraded H20 variant; sporting fancy HBM3E modules.

Yesterday's global launch of Radeon RX 9070 XT and RX 9070 was a relatively successful affair; day one stock was swiftly snapped up. Naturally, buying conditions were not ideal for every customer. Despite a prominent UK retailer's teasing of plentiful supply (pre-launch), online feedback pointed to apparent limited supplies of RDNA 4 cards across European regions. Fresh reports suggest that anticipated fluid price conditions have caused a larger than expected rush at retail; hence the rapid depletion of opening day stock. As reported yesterday, a handful of PC hardware stores have alluded to forthcoming upward shifts in pricing for the lower-end of an all-board partner-built Radeon RX 9070 series graphics card lineup. VideoCardz has highlighted a disgruntled customer's experience with Ebuyer UK. Bran180s—a Radeon subreddit member—managed to snag a baseline MSRP conformant Sapphire PULSE RX 9070 XT model for the ideal launch price (£569.99, including VAT), but the webstore reneged this transaction.

A screenshot was uploaded to Reddit, alongside a short story: "was on the website ready for the launch of the RX 9070 XT, got one straight away and paid no issues. Ebuyer emailed me today to cancel, and now the price is £150 more." The British e-tailer has issued apologies, following the absorption of online criticism (see relevant screenshot below). The "normal price" of Sapphire's basic Pulse card was eventually adjusted to a mere £664.98, but Ebuyer has de-listed this SKU (at the time of writing). Other UK webshops—Scan, AWD-IT, CCL, Box etc.—have similarly implemented price hikes across low, mid and premium card tiers. Australia's Hardware Unboxed managed to extract an official response from AMD—their social media post quoted Frank Azor. The Team Red exec indicated that his team is ready to intervene: "it is inaccurate that $549/$599 MSRP is launch-only pricing. We expect cards to be available from multiple vendors at $549/$599 (excluding region specific tariffs and/or taxes) based on the work we have done with our AIB partners, and more are coming. At the same time, the AIBs have different premium configurations at higher price points and those will also continue."

A South Korea-based hardware enthusiast—Harukaze5719—has discovered a curious listing of ZOTAC's GeForce RTX 5090 SOLID OC model on a Chinese auction site (asking price: ~$4175 USD). This finding was shared with their audience, via a social media post—a screenshot was accompanied by harukaze5719's short and succinct message: "OMG..." Officially, NVIDIA and its board partners cater to the Chinese hardware market with a restricted variant of the flagship "Blackwell" GPU—GeForce RTX 5090D. Despite under-the-hood nerfing, this region-exclusive model still offers enviable performance (when pushed).

The "full fat" GeForce RTX 5090 GPU appears to be an even hotter commodity—with demand (at launch) exceeding far exceeding supply. ZOTAC's South Korean branch warned potential customers about difficult conditions almost two weeks ahead of Blackwell's market debut on January 30. Harukaze5719 has likely expressed semi-personal frustration over apparent South Korea market-destined ZOTAC stock turning up for sale in China. So-called "Chinese GPU shopping tourists" could have made the journey to a South Korean PC hardware store, with their sole objective being the acquisition of GeForce RTX 5090 graphics cards. Tom's Hardware has gathered various related tidbits from Japanese and Taiwanese news sources—where local TV coverage explored the events of last week's launch event.

Earlier this week, MSI's China office admitted that it had only a limited supply of GeForce RTX 50 series graphics cards available for launch day. A VideoCardz report from last night points out similar problems affecting retail channels in North America. MSI's official US store has delayed shipments of GeForce RTX 5090 and RTX 5080 cards. This announcement was extracted from a conversation conducted via Discord—between a company representative (MSI Spence) and a VideoCardz staffer (WhyCry). In full, MSI stated: "we are going to allow GeForce RTX 5090 pre-orders on January 30, at 6 AM PST—with a ship date of February 6. The GeForce RTX 5080 will be available for purchase on January 30 at 6 AM PST, with an immediate ship date of (same day) January 30. GeForce RTX 5090 pre-orders will open on January 30 at 6 AM PST, with a ship date of February 6."

Additionally, StinceBuilt has corroborated reports of newly adjusted timeframes—the custom PC builder publicly disclosed its own dealings with the NVIDIA board partner: "we're being told by the MSI Team that the MSI RTX 5090 Variants are now being pushed back until February 6th for launch. RTX 5090 Day of Launch is going to be very rough, both on the consumer retail/ecommerce side, and to all system integrators. Stock is minimal at best. Please be patient. We've confirmed this with other shops like ourselves."

TrendForce reports that a recovery in demand for general servers—coupled with an increased production share of HBM by DRAM suppliers—has led suppliers to maintain their stance on hiking prices. As a result, the ASP of DRAM in the third quarter is expected to continue rising, with an anticipated increase of 8-13%. The price of conventional DRAM is expected to rise by 5-10%, showing a slight contraction compared to the increase in the second quarter.

TrendForce notes that buyers were more conservative about restocking in the second, and inventory levels on both the supplier and buyer sides did not show significant changes. Looking ahead to the third quarter, there is still room for inventory replenishment for smartphones and CSPs, and the peak season for production is soon to commence. Consequently, it is expected that smartphones and servers will drive an increase in memory shipments in the third quarter.

TrendForce's latest forecasts reveal contract prices for DRAM in the second quarter are expected to increase by 13-18%, while NAND Flash contract prices have been adjusted to a 15-20% Only eMMC/UFS will be seeing a smaller price increase of about 10%.

Before the 4/03 earthquake, TrendForce had initially predicted that DRAM contract prices would see a seasonal rise of 3-8% and NAND Flash 13-18%, significantly tapering from Q1 as seen from spot price indicators which showed weakening price momentum and reduced transaction volumes. This was primarily due to subdued demand outside of AI applications, particularly with no signs of recovery in demand for notebooks and smartphones. Inventory levels were gradually increasing, especially among PC OEMs. Additionally, with DRAM and NAND Flash prices having risen for 2-3 consecutive quarters, the willingness of buyers to accept further substantial price increases had diminished.

Two months ago, NVIDIA introduced its GeForce RTX 40 SUPER series to the market, bringing a trio of models: RTX 4070 SUPER, RTX 4070 Ti SUPER, and RTX 4080 SUPER. Today, according to the report from ComputerBase, NVIDIA's latest trio has recorded a drop in pricing recently, and it now retails under MSRP in German stores. The RTX 4070 SUPER started with an MSRP of 659 Euros ($599 in the US) and is now available from 589 Euros. Its older brother, the GeForce RTX 4070 Ti SUPER, started with an MSRP listing of 889 Euros ($799 in the US) and is now retailing from 840 Euros. Lastly, the NVIDIA GeForce RTX 4080 SUPER has been listed at 1,109 Euros ($999 in the US) and is now retailing with a small discount at 1,092 Euros.

Once NVIDIA launched a new GPU generation, it became a custom for these cards to be retailed over their MSRP long before prices were adjusted and settled. However, with the latest SUPER refresh, this seems to be one of the fastest price adjustments. This could be caused by either an improvement in the supply chain or leveled supply and demand, making it so that these cards are finally trading below their launch-level MSRPs.

TrendForce reports a substantial 24.5% QoQ increase in NAND Flash industry revenue, hitting US$11.49 billion in 4Q23. This surge is attributed to a stabilization in end-demand spurred by year-end promotions, along with an expansion in component market orders driven by price chasing, leading to robust bit shipments compared to the same period last year. Additionally, the corporate sector's continued positive outlook for 2024 demand—compared to 2023—and strategic stockpiling have further fueled this growth.

Looking ahead to 1Q24, despite it traditionally being an off-season, the NAND Flash industry is expected to see a continued increase in revenue by another 20%. This anticipation is underpinned by significant improvements in supply chain inventory levels and ongoing price rises, with clients ramping up their orders to sidestep potential supply shortages and escalating costs. The ongoing expansion of order sizes is expected to drive NAND Flash contract prices up by an average of 25%.

The NVIDIA H100 Tensor Core GPU was last year's hot item for HPC and AI industry segments—the largest purchasers were reported to have acquired up to 150,000 units each. Demand grew so much that lead times of 36 to 52 weeks became the norm for H100-based server equipment. The latest rumblings indicate that things have stabilized—so much so that some organizations are "offloading chips" as the supply crunch cools off. Apparently it is more cost-effective to rent AI processing sessions through cloud service providers (CSPs)—the big three being Amazon Web Services, Google Cloud, and Microsoft Azure.

According to a mid-February Seeking Alpha report, wait times for the NVIDIA H100 80 GB GPU model have been reduced down to around three to four months. The Information believes that some companies have already reduced their order counts, while others have hardware sitting around, completely unused. Maintenance complexity and costs are reportedly cited as a main factors in "offloading" unneeded equipment, and turning to renting server time from CSPs. Despite improved supply conditions, AI GPU demand is still growing—driven mainly by organizations dealing with LLM models. A prime example being Open AI—as pointed out by The Information—insider murmurings have Sam Altman & Co. seeking out alternative solutions and production avenues.

According to a recent MyDrivers news piece, graphics card models based on AMD's China-exclusive Radeon RX 6750 GRE GPU are "selling like hotcakes" in the region. Team Red's Radeon RX 6750 GRE was released last October, and made available in two memory configurations: 10 GB with a recommended starting price of 2219 RMB (~$312), and 12 GB going for 2379 RMB (~$335). Industry experts were surprised to hear about the mid-range Navi 22 (RDNA 2) GRE card becoming a popular choice in mainstream and budget-conscious gaming circles—as of early January, AMD has taken the unusual step of delaying its Radeon RX 7600 XT launch in China.

Chinese hardware news outlets are reporting that the 12 GB variant is selling for as low as 2239 RMB (~$315) at select retail sites—customers are reported to be snapping these cards up due to an irresistible "price-to-performance ratio," despite onboard technology being almost three years old. AMD is reported to be implementing a "strict" price control policy to prevent unbalanced supply conditions—Wccftech suggests that adjustment "instructions" have been sent out to AIB partners. The MyDrivers article points out that the policy adjustments will take time to stabilize Radeon RX 6750 GRE prices—the aforementioned 2239 RMB shop offers were still live late last week (report is dated January 26).

OpenAI CEO, Sam Altman, had a turbulent winter 2023 career moment, but appears to be going all in with his company's future interests. A Bloomberg report suggests that the tech visionary has initiated a major fundraising initiative for the construction of OpenAI-specific semiconductor production plants. The AI evangelist reckons that his industry will become prevalent enough to demand a dedicated network of manufacturing facilities—the U.S. based artificial intelligence (AI) research organization is (reportedly) exploring custom artificial intelligence chip designs. Proprietary AI-focused GPUs and accelerators are not novelties at this stage in time—many top tech companies rely on NVIDIA solutions, but are keen to deploy custom-built hardware in the near future.

OpenAI's popular ChatGPT system is reliant on NVIDIA H100 and A100 GPUs, but tailor-made alternatives seem to be the desired route for Altman & Co. The "on their own terms" pathway seemingly skips an expected/traditional chip manufacturing process—the big foundries could struggle to keep up with demand for AI-oriented silicon. G42 (an Abu Dhabi-based AI development holding company) and SoftBank Group are mentioned as prime investment partners in OpenAI's fledgling scheme—Bloomberg proposes that Altman's team is negotiating a $8 to 10 billion deal with top brass at G42. OpenAI's planned creation of its own foundry network is certainly a lofty and costly goal—the report does not specify whether existing facilities will be purchased and overhauled, or new plants being constructed entirely from scratch.

TrendForce has issued its latest findings, indicating that the global SSD market has rectified its supply and demand dynamics in 2022, following a resolution in the shortage of master control ICs that had hampered the market in 2021. Despite the normalization of supply, global SSD shipments witnessed a decline, with only 114 million units shipped in 2022—a 10.7% decrease from the prior year.

The top three SSD shipment leaders of 2022 were Kingston, ADATA, and Lexar, with Kingston and ADATA maintaining solid advantages and experiencing growth in market share over 2021. Lexar's growth was attributed to an aggressive push for revenue in anticipation of going public. Kimtigo, in 2022, made significant strides in expanding into industrial control and OEM markets, which in turn boosted its shipment volume and market share. Netac maintained its competitive edge in the SSD market alongside securing several government orders in the enterprise SSD sector, keeping its market share and ranking consistent with the previous year.

TSMC Chairman Mark Liu was asked to comment on all things artificial intelligence-related at the SEMICON Taiwan 2023 industry event. According to a Nikkei Asia report, he foresees supply constraints lasting until the tail end of 2024: "It's not the shortage of AI chips. It's the shortage of our chip-on-wafer-on-substrate (COWOS) capacity...Currently, we can't fulfill 100% of our customers' needs, but we try to support about 80%. We think this is a temporary phenomenon. After our expansion of advanced chip packaging capacity, it should be alleviated in one and a half years." He cites a recent and very "sudden" spike in demand for COWOS, with numbers tripling within the span of a year. Market leader NVIDIA relies on TSMC's advanced packaging system—most notably with the production of highly-prized A100 and H100 series Tensor Core compute GPUs.

These issues are deemed a "temporary" problem—it could take around 18 months to eliminate production output "bottlenecks." TSMC is racing to bolster its native activities with new facilities—plans for a new $2.9 billion advanced chip packaging plant (in Miaoli County) were disclosed during summer time. Liu reckons that industry-wide innovation is necessary to meet growing demand through new methods to "connect, package and stack chips." Liu elaborated: "We are now putting together many chips into a tightly integrated massive interconnect system. This is a paradigm shift in semiconductor technology integration." The TSMC boss reckons that processing units fielding over one trillion transistors are viable within the next decade: "it's through packaging with multiple chips that this could be possible.".

NVIDIA could be adjusting its enterprise-grade GPU production strategies for the Chinese market, according to an article published by MyDriver—despite major sanctions placed on semiconductor imports, Team Green is doing plenty of business with tech firms operating in the region thanks to an uptick in AI-related activities. NVIDIA offers two market specific accelerator models that have been cut down to conform to rules and regulations—the more powerful and expensive (250K RMB/~$35K) H800 is an adaptation of the western H100 GPU, while the A800 is a legal market alternative to the older A100.

The report proposes that NVIDIA is considering plans to reduce factory output of the A800 (sold for 100K RMB/~$14K per unit), so clients will be semi-forced into purchasing the higher-end H800 model instead (if they require a significant number of GPUs). The A800 seems to be the more popular choice for the majority of companies at the moment, with the heavy hitters—Alibaba, Baidu, Tencent, Jitwei and ByteDance—flexing their spending muscles and splurging on mixed shipments of the two accelerators. By limiting supplies of the lesser A800, Team Green could be generating more profit by prioritizing the more expensive (and readily available) model.

TrendForce reports that OEMs have continued making concerted efforts to scale back production. However, given that the trajectory of market demand is still unclear, it's expected that the NAND Flash market will continue to be in a state of oversupply in 3Q23. Cautious inventory management by buyers is preventing a stabilization in NAND Flash prices even with an anticipated seasonal surge in demand for 2H23. TrendForce predicts that NAND Flash wafers will be the first to see a price hike in 3Q23 as prices for module products such as SSDs, eMMCs, and UFS will likely continue to fall due to tepid downstream demand. Consequently, the overall ASP of NAND Flash is forecast to continue dropping by about 3~8% in 3Q23, though a possibility exists prices may recover in 4Q23.

Client SSD: Although notebook shipments are expected to gradually recover in 3Q23, reversing an oversupply of SSD will continue to be challenging. Furthermore, a portion of suppliers have implemented aggressive promotions to secure customer orders and hit shipping targets in light of weakened demand and less-than-satisfactory order volumes from major clients, putting pressure on other suppliers. TrendForce estimates that the ASP of client SSDs will fall by 8~13% in the third quarter.

The top technology companies in China have been ordering a lot of NVIDIA enterprise-grade GPUs, even though U.S. sanctions have prevented the shipment of A100 and H100 models (plus AMD's MI250 Instinct accelerator) to the nation in recent times. ByteDance - best known for developing TikTok - managed to grab plenty of Ampere enterprise units prior to last Autumn's cutoff period, and has continued to purchase Team Green's H800 GPU, which is a cut-down version of the H100 flagship. Smaller outfits are relying on less direct sources to acquire HBC GPUs—according to a Reuters investigative article, international trade restrictions have created a thriving black market for "top-end NVIDIA AI chips."

Their reporters carried out some on-site sleuthing: "Visiting the famed Huaqiangbei electronics area in the southern Chinese city of Shenzhen is a good bet - in particular, the SEG Plaza skyscraper whose first 10 floors are crammed with shops selling everything from camera parts to drones. The chips are not advertised but asking discreetly works...They don't come cheap. Two vendors there, who spoke with Reuters in person on condition of anonymity, said they could provide small numbers of A100 artificial intelligence chips made by the U.S. chip designer, pricing them at $20,000 a piece - double the usual price."