DigiTimes: 2027 DRAM/HBM capacity already sold out According to a DigiTimes report, annual 2027 DRAM and HBM capacity has been fully booked out ahead of schedule. NAND flash supply isn't as tight as DRAM, but bookings are expected to be completely filled by the end of this month. Per sources cited by DigiTimes, companies aren't openly acknowledging that they need to lock in memory by this month — because they worry that if more players pile into the scramble for supply, their own allocations will shrink. Manufacturers also typically deliver only 60–70% of the volumes buyers originally requested, as memory makers prioritize demand from CSPs and AI majors. As a result, the DRAM volumes smartphone and PC makers can secure in 2027 are expected to fall markedly versus 2026. DigiTimes, citing industry sources, noted that with capacity now effectively almost entirely sold out, companies that have yet to finalize their volumes could face an even more acute "can't buy it even if you want to" situation in 2027. The report added that elevated memory prices are set to become the norm. While the sharp price spikes driven by the fight for allocation may subside — meaning the magnitude of 2027 price increases could be more moderate — the overall supply shortfall and the cost burden on the finished-goods industry are unlikely to be resolved in the near term.

$AMD Helios networking explained: Scale-up: inside the rack - 72 MI455X GPUs - 12 Broadcom Tomahawk 6 ASICs - 36 parallel network planes - 3.6 TB/s bidirectional per GPU - 260 TB/s aggregate The network is based on two standards: > UALoE defines how GPUs communicate, including memory reads and writes, transaction ordering, and how traffic is distributed across the 36 network planes > ESUN provides the Ethernet fabric underneath, handling forwarding, flow control, congestion management, link-level retries, and lower-overhead transmission for small messages The path is: MI455X → UALoE → ESUN Ethernet → Tomahawk 6 → MI455X The 31 TB of HBM4 stays physically distributed across 72 GPUs. UALoE lets any GPU read or write any other GPU's memory directly, making the whole rack act as a single computer Scale-out: between racks The racks are connected with Ethernet scale-out through 800G Pensando Vulcano AI NICs into a UEC-ready fabric The path is: Helios rack → Vulcano NIC → UEC Ethernet → another Helios rack

China's CXMT is seeing strong DRAM demand, with its production reportedly reserved through 2027. 🔗 Dell, HP, Lenovo, and Apple are reportedly among the companies expected to receive priority allocations.

has completed construction of a giant data center that houses only Chinese-made chips, a big step in Beijing’s efforts to replace restricted Nvidia silicon for future AI development

This is significant. $STM is the market leader in microcontrollers and they are having problems getting allocations at TSMC for their mature node chips. Reason is that TSMC is allocating this capacity to Nvidia ecosystem and asing STM to move to 40nm of below. As a result, STM is asking its customers to input orders for whole 2027. Companies with their own fabs will benefit from this situation.

I've been discussing Huawei's τ scaling (temporal scaling) with people recently, and noticed the conversation tends to stay at the surface level without reaching its substance — likely because many participants don't come from an EE background and aren't familiar with the classical meaning of τ in circuit theory. The very first time constant you learn in a circuits course is τ = RC: the resistance of a wire multiplied by its capacitance gives the order of magnitude of the time a signal needs to traverse that wire. The longer the wire, the greater the resistance and capacitance, and the slower the signal. Within this framework, the past sixty years of geometric scaling are reinterpreted as one particular implementation of temporal scaling. Transistors were shrunk to shorten switching delay; circuits were packed more tightly to shorten metal interconnects and reduce signal propagation delay. Geometric scaling was only ever the means — compressing delay was always the end. Huawei's thesis is that once geometric scaling stalls, you find other ways to keep compressing delay. As it happens, He Tingbo's τ scaling paper released its v2 a couple of days ago, expanding from 16 to 23 pages. I compared the two versions: the data and conclusions are unchanged. The additions are essentially responses to several points of criticism the industry raised about v1. Three are worth discussing. The most important addition is the test evidence now backing the previously bare claim of "41% energy efficiency improvement." In v1, that number had no baseline and no test conditions — the most obvious target for scrutiny. V2 supplies a full comparison table. The baseline is the 2025 Kirin 9030 Pro. Both chips use the same mature process node; the key difference is that the baseline uses a conventional planar design, while Kirin 2026 folds critical paths across two vertically bonded wafers. Folding shortens interconnects and reduces interconnect delay. The timing margin freed up on the critical path translates directly into a higher maximum clock frequency: 3.1 GHz at 1.1 V supply, 13% above the baseline. The "41% energy efficiency improvement" comes from a separate operating point specifically configured for an iso-performance comparison: voltage scaled down to 0.9 V, frequency scaled down to 2.5 GHz, with measured power at 25°C coming in at 0.59× the baseline. A back-of-the-envelope estimate checks out: dynamic power scales roughly with the square of supply voltage, so an 18% voltage reduction contributes about one-third of the power drop from the square term alone. Factor in the 9% frequency reduction and the interconnect capacitance eliminated by folding, and you land right around 0.59×. So the precise meaning of "41% energy efficiency improvement" is power reduction at iso-performance. In essence, the timing margin gained from folding is traded for lower power consumption; the efficiency gain comes from logic folding. As a side note, v2 also reports that power density after dual-layer stacking is actually 5.6% lower than the baseline. The second addition addresses the question peers are most likely to ask: 3D stacking has been around for years — AMD's 3D V-Cache and Intel's Foveros are both in volume production — so what's new about LogicFolding? To understand the paper's answer, you first need to know how two layers of silicon communicate. They rely on inter-layer bond pads, which function like elevators connecting the upper and lower floors. In prior production 3D stacking, bond pad pitch ranges from 9 μm to tens of micrometers, yielding roughly ten thousand connections per square millimeter — enough to attach a bus to an entire cache block. So the established design approach has been to move complete functional blocks wholesale onto the upper tier. AMD, for example, stacks an entire cache die on top of a processor die; the two tiers are designed independently and connected through an interface. But inside a chip, a single square millimeter contains hundreds of millions of transistors. If you want adjacent logic gates to sit on different tiers — one on top, one on the bottom — that connection density falls far short. Kirin 2026 brings bond pad pitch down to 1.5 μm, yielding 440,000 connections per square millimeter. That approaches the density of the top-level metal wiring inside a chip. Routing a signal across tiers costs roughly the same as routing it across metal layers within a single die. At this point, the two silicon layers merge into a single entity in the circuit sense. EDA tools can decide at the individual logic-gate level which gate goes on which tier, handing the problem to algorithms for global optimization — a completely different degree of design freedom from what came before. The paper also explains why they didn't take the more aggressive route of fabricating a second device layer directly on top of the first. That approach offers the finest inter-layer connectivity, but manufacturing the second layer requires high temperatures that damage the already-completed first layer. It isn't production-viable today. The third addition is thermal management. Vertical stacking significantly increases thermal density per unit area, and the lower die's heat dissipation path is blocked by the upper die. This is the first objection anyone raises about 3D stacking, and v1 did not address it in depth. V2 openly acknowledges that thermal management remains a key challenge for the LogicFolding architecture. The countermeasure is thermally-aware partitioning and floorplanning: during the design phase, high-power circuits are excluded from folding candidates, and the floorplan avoids placing high-power blocks in vertical adjacency to prevent hotspot superposition. Whether this strategy is a set of manually imposed engineering constraints or has already been codified into an automated flow within their internal EDA tools, the paper does not say. It only identifies a multi-physics tool chain as the single most important investment for the next decade. Combined with the measured data showing power density 5.6% below the baseline at the iso-performance operating point, the thermal concern has at least received a direct response. That said, this approach is fundamentally avoidance-based. As stacking grows to three or four tiers, the design space eligible for folding will be progressively squeezed by thermal constraints — a boundary the paper does not explore. Additionally, v2 includes a cross-sectional micrograph of the bond interface between the two wafers and explicitly states that wafer-on-wafer hybrid bonding is used. This spec is worth benchmarking against the industry: 1.5 μm pitch wafer-to-wafer hybrid bonding on a production logic chip has no precedent. TSMC's SoIC is currently in production at 6 μm pitch; Intel's Foveros Direct is at 9 μm. Impressive, to say the least. After comparing the two versions, I'm left with two questions. One is about equipment: who supplied the bonding tools capable of this spec? The paper says only that it is the result of years of process development across a multi-vendor ecosystem. The other is about EDA: designing two wafers as a single chip is beyond what any commercially available EDA tool can do today. The paper acknowledges this, stating only that methodological details will be "published within months." Yet the frequency table shows that the 2027-generation Kirin at 3.39 GHz is already tagged as having physical silicon, meaning this toolchain was up and running inside Huawei long ago — and has been validated on at least two product generations. My personal guess is that this EDA capability was built in-house by Huawei. If anyone has insight on this, I'd welcome the discussion.

WFE = 📈 The angstrom era of semis will demand chiplets thus all or most silicon will utilize advanced packaging. Huge need for new tools. Key nuggets from $AMAT Master Class on Advanced Packaging: - AMAT is increasingly an advanced packaging story, not just a front-end WFE story. - Five of six new platforms were aimed at back-end / packaging processes, showing where the innovation roadmap is shifting. - Advanced Packaging should exceed $2B in 2026E and grow >50% Y/Y, driven by HBM, panel-level packaging, hybrid bonding, and new integration schemes. - DRAM WFE remains a major growth lever, with AMAT arguing DRAM spend should stay well over 2x NAND WFE. - DRAM complexity lifts AMAT revenue per wafer: roughly +10% from 6F² to 4F², and another +15% moving toward 3D DRAM. -NEXX gives AMAT a more complete panel-level packaging flow, including lithography, deposition, and electrochemical deposition. - AMAT is pushing into some Lam-adjacent sockets, especially ECD and PE-CVD tools tied to advanced packaging and HBM.

Pretty much. But it’s not just one product line, it’s a demand stack. Tencent, Alibaba, ByteDance are all scaling inference clusters at the same time. $3B LTA from Tencent alone is just one buyer, one contract. And there’s a new variable: agentic AI is taking off fast, and Zhipu’s GLM-5.2 just hit frontier-level performance on coding/agent benchmarks while being open-weight and self-hostable at roughly 1/6 the cost of closed APIs. That changes the deployment math completely. Every major Chinese enterprise can now run frontier-grade agents on their own infra. Memory per AI server is already multiples of traditional cloud. Now multiply that by the kind of enterprise rollout these companies are gearing up for.

🚨 Japan’s power semiconductor integration still a long way to go? #Rohm# reportedly says talks with #Toshiba# and #MitsubishiElectric# are taking longer, as negotiations over fab integration, R&D resources, and strategic control prove complex.💡More: 🔗