The Greensheet | June 2026
-
Memory pricing accelerates as allocation shifts. DDR5 RDIMM pricing has moved beyond May projections, while hyperscaler long-term agreements are limiting allocation for non-hyperscale buyers. DDR4 demand is also strengthening as industrial customers pay premiums to avoid DDR5 platform transitions.
- MLCC and passive constraints intensify. Murata, Samsung, and Taiyo Yuden are facing severe allocation, with lead times reaching 36 to 60 weeks on high-capacitance and automotive-grade lines. AI server and EV demand are consuming specialized capacity, while tiered pricing and broker restrictions point to structural shortage conditions.
- Texas Instruments disruption adds supply chain pressure. TI’s July 1 price increases, extended lead times, stricter quotation protocols, and expedite fee requirements are pushing automotive and industrial buyers toward open-market procurement and buffer buying.
- Nexperia fragmentation extends uncertainty. Sanctions tied to Yangzhou Yangjie and MCC are driving customers toward On Semi and Diodes alternatives, while Nexperia’s Malaysia ramp delay to Q4 2026 keeps allocation risk elevated. Verification protocols are also adding 2 to 4 weeks to procurement timelines.
-
Enterprise storage and HDD allocation tighten. SSD pricing has surged quarter over quarter, with weekly manufacturer resets now common. HDD capacity is reportedly sold out for 2026, while WD price increases and low-capacity drive exits are pushing customers toward higher-margin, high-capacity SKUs.
-
Texas Instruments' July 1 price adjustments range from 10% to 35% for standard portfolios and up to 85% for legacy and power management families. Booking lead times have extended to 180 days and 52+ weeks. The SN74 logic series is experiencing widespread decommits, and manufacturers are requiring detailed end-customer validation before releasing pricing.
-
Analog Devices portfolio-wide increases of 15% to 30% are fully implemented. Lead times remain at 32 to 40+ weeks, with T-Glass substrate constraints cited as the primary bottleneck for LT-series op amps and precision references. Automotive and industrial customers continue reporting decommits and expedite fee requests.
-
Renesas supply instability persists, with 52+ week lead times for power stage ICs such as ISL99390FRZ. Price increases of 5% to 50% take effect July 1. The timing of business divestment and recent system upgrade disruptions continues to drive design-out activity across NPI programs.
-
On Semi and Diodes Incorporated are absorbing overflow demand from Nexperia and MCC qualification shifts. Lead times have stretched to 40 to 52+ weeks, with On Semi announcing July pricing adjustments of approximately 15%+ and Diodes implementing backlog compression for commercial-grade orders.
-
STMicroelectronics' late June price increases of approximately 15% for general lines and 15% to 30% for automotive-grade products are now active. MCU lead times remain at 20+ weeks, while sensor series are seeing delivery pushes into Q2 2027. Mouser and other distributors have begun restricting PMIC and power switch allocations.
-
Lattice and Altera/Xilinx lead times have expanded to 42 to 56 weeks. Expedite fees are mandatory for orders requesting delivery in under 44 weeks. Decommits are being reported across XC7A50T and MachXO families as hyperscaler AI board demand consumes advanced packaging capacity.
-
Qorvo shortages continue across industrial and automotive PMIC portfolios, with manufacturers reporting significant decommits and authorized distribution unable to meet demand. Spot market pricing is rising quickly as brokers compete for the limited available stock.
-
Sony CMOS sensor shortages are impacting industrial and automotive segments, with specific MPNs at risk, including IMX250LLR, IMX178LQJ, IMX264LLR-C, IMX253LLR, and IMX546-AAQJ. A global Nittobo glass substrate shortage is limiting production to approximately 70% of demand through 2027, with glass cloth capacity heavily prioritized for Nvidia AI server programs.
-
Intel Server Granite Rapids and Sierra Forest remain tightly allocated, with lead times exceeding 6 months. Hyperscalers and cloud providers are absorbing more than 70% of available capacity. U.S. and APAC enterprise customers report needing strong project justification to secure even partial allocations.
-
AMD Server EPYC Genoa and Turin series remain sold out through 2026. High-frequency “F” models face 12 to 16+ week official lead times, with open-market premiums hovering around $1,000 per unit. No meaningful easing in AI or hyperscaler prioritization has been observed.
-
Intel Desktop and Mobile Alder Lake 12th Gen are officially EOL with limited manufacturer support, driving elevated open-market premiums. Raptor Lake Refresh and Arrow Lake remain constrained through August. Lunar Lake and Meteor Lake mobile processors remain tight through Q3. Panther Lake U-series shows improved availability for 12XE and 4XE iGPU variants in H1, while other SKUs remain constrained.
-
Small-core CPUs, including N-series N97, N100, N150, N305, and J-series J6412, remain critically constrained through Q3. Intel is deprioritizing low-margin SKUs to focus on Panther Lake and the 18A AI PC ramp, with portfolio-wide price increases of 5% to 25% effective Q2 2026.
-
Server CPUs and DCAI 10nm shortages persist through Q3 2026, impacting 4th and 5th Gen Xeon supply. Granite Rapids Xeon 6 allocation is prioritized to U.S. regions, while PRC customers require strong project justification to secure supply. Constraints are expected to continue through year-end.
-
Intel Atom A39xx has been notified for end-of-life, with a last order date of October 15, 2026, and a last shipment date of June 9, 2028. Elkhart Lake small-core processors across J, N, and X series face severe allocation constraints, with some configurations officially booked out to 2028.
-
NVIDIA Blackwell RTX 6000 and 5090 server editions have increased approximately 25% in price. RTX PRO 6000 Blackwell is trading at $11,000 to $12,500+, with lead times averaging 12+ weeks. Consumer RTX 5090 remains constrained by GDDR7 memory shortages, with production heavily prioritized toward higher-margin AI server demand.
-
RTX PRO 6000 Blackwell demand continues to spike, with an additional 15% to 20% pricing increase expected. Allocation remains highly restricted, and distributors are reporting no firm delivery schedules for workstation variants through Q4 2026.
-
NVIDIA RTX 4000 and 5000 Ada lead times remain extended at 48 to 52 weeks. Production continues to be curtailed to free capacity for Blackwell, with distributors reporting inventory reallocations to hyperscaler orders.
-
NVIDIA Jetson Modules are expected to see pricing increases of approximately 15% across the portfolio. Legacy modules including Nano and TX2 NX continue phased EOL transitions, with LTB windows closing in July 2026. Updated part numbers for T5000 and T4000 modules address acoustic noise issues.
-
DDR5 RDIMM contract pricing has jumped 75% to 125% quarter over quarter. Spot transactions for 64GB modules are consistently clearing above $2,800, while 96GB and 128GB modules are exceeding $3,800. Asia allocation has been reduced as manufacturers prioritize North American hyperscaler demand and stronger pricing acceptance.
-
DDR4 8Gb and 16Gb pricing continues to climb as industrial and networking customers finalize PPV approvals. Clean-pull and legacy inventory are being swept by brokers anticipating further capacity reallocation toward HBM and DDR5.
-
LPDDR4/5 and eMMC shortages for 8GB and 16GB automotive grades persist. Samsung and Micron have largely halted new quotations, with lead times extending into 2027. ISSI has confirmed that new eMMC orders will not ship until 2028, forcing aggressive dual-sourcing initiatives.
-
SanDisk low-density eMMC demand remains elevated for SDINBDG4-8G-ZA2 and SDINBDA6-64G-ZA1, with manufacturers reporting demand at 4 to 5 times available supply. Open-market pricing has tripled as production is deprioritized in favor of HBM3E and HBM4 output.
-
Enterprise SSDs have increased 50% to 90% quarter over quarter. Solidigm and Samsung are implementing weekly price resets. Low-end SATA models face 50% to 60% increases, while high-end SATA and NVMe models face 80% to 90% increases. Gen5 enterprise SSDs are projected to rise another 20% to 30% in Q3.
-
Solidigm SATA D3-S4520 and S4620 series face a final LTB date of September 30, 2026, driving secondary market volatility and premium open-market quoting.
-
Seagate and WD report 100% capacity utilization for 2026, with agreements extending into 2028 to 2030. WD implemented a 25% increase on June 10, with a 40% increase projected for July 6. Manufacturers are aggressively EOL-ing 1TB to 5TB 2.5" drives to force migration toward 12TB+ higher-margin SKUs.
-
Micron NOR Flash supply remains critically short. Lead times are stretched through 2H 2026 as wafer capacity is prioritized for enterprise memory and AI accelerator components.
-
Mellanox CX6 and CX7 100G, 200G, and 400G NICs face 30 to 52-week lead times as NVIDIA prioritizes CX8 production for next-generation GB-series GPU bundles. Bulk procurement by Chinese hyperscalers, including ByteDance and Tencent, has further tightened remaining CX7 availability.
-
Broadcom 5nm PCIe switch chips, including SS24 and SS26, are severely constrained. TSMC capacity limitations are pushing expected deliveries to January 2027, with no expedite options available. Baseline lead times remain at 52+ weeks.
-
Coherent and Finisar have pushed 400G and 800G transceiver lead times out by 6 months. NVIDIA is enforcing mandatory transceiver bundling with Q3400 switches, while H200 export restrictions continue creating volatility in spot markets for compliant interconnects.
- Murata, Samsung, and Taiyo Yuden are operating under severe allocation. High-capacitance and high-temperature lines, particularly those tied to Japan-exclusive production, face 36 to 60-week lead times. AI server and automotive demand are consuming more than 80% of specialized capacity. Distributors have implemented broker restrictions and tiered pricing.
- Panasonic and Kemet have announced 30% to 40% price increases effective in July. Lead times exceed 40 weeks. Low-margin commercial lines face potential discontinuation as manufacturers prioritize automotive and industrial grades.
- T-Glass substrate remains the primary cross-category bottleneck for ADI, Intel, Lattice, and Sony. Glass cloth fabric lead times have extended from 4 weeks to 20 weeks, causing decommits and forcing foundry diversions. Nittobo remains the sole qualified supplier for high-performance CMOS applications.
- Middle East tensions continue to impact PCB testing, capacitor production, and optical component manufacturing. Japanese and Korean manufacturers report activating emergency procurement protocols, with no meaningful capacity relief expected until late 2027.
- Lattice lead times remain extremely tight at 42 to 56 weeks. Orders requesting lead times under 44 weeks are subject to expedite fees. Decommit reports continue across XC7A50T and MachXO series.
- Taiwan memory chipmaker Winbond has reportedly entered Nvidia’s next-generation Vera Rubin platform supply chain with its NOR Flash products. Volume shipments are expected to begin in the second half of the year.. [Source: Digitimes]
- Intel is reportedly targeting Q1 2028 for CPUs featuring integrated Nvidia RTX GPU technology. If the schedule remains unchanged, the chips could debut at CES 2028. [Source: Digitimes]
- China has warned that global supply chains may become increasingly disconnected as key trading partners consider new measures to address trade imbalances and reduce rare earth dependencies. [Source: Bloomberg]
- Apple’s long-rumored foldable iPhone appears to be moving closer to reality, with multiple Chinese media outlets reporting that parts of the supply chain have begun small-volume shipments ahead of a planned launch next year. [Source: Digitimes]
-
Memory chipmaker Kioxia Holdings replaced Toyota Motor as Japan’s largest company by market value, underscoring how the global AI boom is reshaping Japan’s corporate landscape.
-
Contract manufacturer Flex outlined plans for a strategic breakup to monetize its exposure to AI. The company said it would spin off its cloud and power infrastructure business into a separate publicly traded company by early 2027.
-
Advantech unveiled its future strategy for edge AI and physical AI during COMPUTEX 2026. The company outlined plans to expand AI adoption in industrial environments by connecting AI agents, digital twins, and edge computing through its industrial AI software platform, WEDA.