Desk

Wire

Named accounts, then the replies. Open items wait for a sitting.

@Frenchie_·@smsehy·@dylan522p·@StockSavvyShay·@Jambooxyz·@MilkRoadAI

• 14 Oct · 08:29 · @Simne1coreSource · Open · @wireTrane and LiquidStack have launched a 2.5 MW CDU. It is the piece that takes fluid from the hall circuit and delivers it to the racks at a set temperature, flow and head. Shipments from Q2 2027.

  @Simne1core · Self-thread quotes the datasheet: 2.5 MW at 4 °C ΔT, 2.0 MW at 3 °C, 3750 l/min at 3.5 bar available head, facility inlet to 45 °C, 52 kW nominal. Notes dual-feed/ATS options and that a larger CDU moves the constraint to hall piping, manifolds, cold plates and circuit redundancy. No external replies.

  Trane and LiquidStack launch a 2.5 MW CDU with shipments starting in Q2 2027, rated 2.5 MW at 4 °C ΔT and 3750 l/min at 3.5 bar.

  Open the issue

• 14 Oct · 07:03 · @OptionKing666Source · Open · @wireHD Hyundai is expanding its HiMSEN land-based gas engine capacity by 3.3GW to capture 9GW of potential AI data center demand, according to J.P. Morgan. Hanwha Engine is simultaneously preparing 250MW to 450MW of four-stroke engine supply for data center operators seeking onsite generation to bypass five-to-ten-year utility grid interconnection delays. Continuous 24/7 data center operation compresses engine overhaul cycles from traditional marine intervals down to 3 years. This operational shift lifts aftermarket maintenance margins above the 30% to 35% levels typical in marine service contracts. Prolonged procurement delays in dual-fuel components remain the primary delivery risk to monitor across Korean engine builders.

  @OptionKing666 · Thread had no replies. Post of 14 Oct 2026 cites the J.P. Morgan note of 6 Oct. Continuous operation shortens overhaul cycles to three years and lifts aftermarket margins; dual-fuel component delays remain the delivery risk.

  HD Hyundai expands HiMSEN land-based gas engine capacity by 3.3 GW to capture 9 GW of potential AI data center demand; Hanwha prepares 250-450 MW four-stroke supply for onsite generation that bypasses multi-year grid interconnects.

• 14 Oct · 05:30 · @OptionKing666Source · Open · @wireBloom Energy's 800V DC-native fuel cell platform cuts non-compute capex by $3.6bn (27%) for a 1 GW AI data center facility compared to traditional AC setups, according to J.P. Morgan. The architecture also reduces five-year total cost of ownership by $5.5bn (9%) by eliminating power conversion layers, transformers, and copper cabling while delivering direct onsite electricity. Conventional gas turbines face lead times of 5 to 7 years and grid interconnects take up to a decade, whereas solid oxide fuel cell factories target ramp times of roughly 2 years. J.P. Morgan sees this architecture bypassing power equipment shortages, though high upfront stack costs and manufacturing scale-up across licensed partners remain key operational variables to monitor.

  @OptionKing666 · Thread had no replies. Post of 14 Oct 2026 cites the J.P. Morgan note of 6 Oct. The architecture eliminates conversion layers and copper for direct onsite 800V DC.

  Bloom Energy captures the 800V DC-native SOFC step that cuts non-compute capex 27 percent for a 1 GW AI facility per J.P. Morgan, with factories targeting roughly two-year ramps versus five-to-seven-year turbine lead times.

  Open the issue

• 13 Oct · 18:00 · @eetimesSource · Open · @wireAmperesand solid-state transformers use SiC semiconductors to stabilize AI data center power grids, prevent outages, and support 800V DC architectures.

  @timothywedell · A reply notes that soft trip with a visible warning beats a hard cut. EE Times of 13 Oct 2026 quotes CEO Brian Dow on battery-backed microgrid operation and 98.5 percent efficiency. Company site confirms first commercial units delivering 2026; Nevada plant opened January 2026 with Gen 3 testing.

  Amperesand captures the medium-voltage SST step for 800V DC AI factories, with first commercial deliveries beginning in 2026 and a Reno facility already running a production line toward more than 20 GW per year capacity.

  Open the issue

• 14 Oct · 04:16 · @Schulz_ResearchSource · Open · @wireLITEON told OCP that its 800V-to-50V power brick now delivers 10 kW at close to 99% peak efficiency, with a 12 kW version due in 2027. The talk walked the path from the 800V rack bus to the GPU: intermediate bus converter takes 50V down to 12V at 1.5 kW peaking above 98%, then the regulator next to the GPU. For NVIDIA Vera Rubin NVL72 there is a 95 kW DC/DC power shelf and a 280 kW in-rack CDU. LITEON also showed a 2.6 MW CDU from DCX. LITEON says its 800V power rack and higher-end power products finish validation and enter mass production in the second half of this year.

  @AIQuanting · A reply asks whether the 91% figure is the last step or the full path from 800V. The post of 14 Oct 2026 cites the OCP talk. LITEON press confirms the Grid-to-Chip portfolio including 800V bricks, IBC, and shelves for Vera Rubin and ORv3. Mass-production clock is company-stated for 2H 2026; monthly filings do not yet break out the 800V line.

  LITEON captures the 800V-to-50V brick at 10 kW near 99% efficiency and targets mass production of its 800V power rack in the second half of 2026, with a 12 kW version in 2027.

  Open the issue

• 14 Oct · 03:44 · @renewablemirrorSource · Open · @wireInfineon launches the XDP8HD40S-1 high-voltage hot-swap controller for 800 VDC and ±400 VDC AI servers. The device provides digital power-path protection so modules and subsystems can be connected or disconnected without disrupting operation, with integrated SOA control, SiC cascode support, PMBus, and black-box recording.

  @renewablemirror · Thread had no replies. The post of 14 Oct 2026 cites the launch. Infineon specifies digital SOA control, gate driver for SiC cascode, PMBus 1.5, and metering accuracy around 1% for voltage and current, targeting intermediate bus and GPU power systems.

  Infineon captures the hot-swap protection step for 800 VDC and ±400 VDC AI server power paths with the XDP8HD40S-1 controller.

  Open the issue

• 13 Oct · 11:24 · @NadirH22Source · Open · @wiresouth korea cleared its first test of 20 humanoid robots working next to people in real plants: a chip plant, a car plant, a shipyard and a logistics site. the reason it needed an exemption: the safety rules assume fixed robots behind fences.

  @NadirH22 · Thread had no replies. Seoul Economic Daily of 13 Oct 2026 names the regulatory sandbox exemption for Robros. Robots will clean circular components, position car body parts, stack shipyard materials and move packaging. Data will rewrite safety rules written for fenced fixed robots.

  South Korea grants Robros an exemption to deploy 20 humanoid robots alongside workers at Nepes chip plant, Sungwoo Hitech auto plant, HD Hyundai Samho shipyard and Delivery Lab logistics site.

  Open the issue

• 13 Oct · 10:52 · @droidhuntSource · Open · @wireSankai Mechatronics closed a nine-figure RMB Series B to scale humanoid joint reducers — its parts already power Beijing's Tiangong robots, and a 1M-units/yr plant is under construction. State industrial funds led.

  @droidhunt · Thread had no replies. Post of 13 Oct 2026 names the Series B close, state funds, existing Tiangong supply and the plant under construction. No further plant clock in the thread.

  Sankai Mechatronics closes a nine-figure RMB Series B and builds a 1 million unit per year plant for humanoid joint reducers that already power Beijing's Tiangong robots.

  Open the issue

• 13 Oct · 09:40 · @BRR_bull_runSource · Open · @wire$SMCI expanded its line of racks and servers built to Open Compute Project specs (an open hardware standard for data centers). The systems are available now. The specs: liquid-cooled $NVDA HGX B300 nodes at up to 18 per rack, or 144 GPUs. Rear-door heat exchangers from 10 to 120 kW. An in-rack coolant unit rated 250 kW in 4U. Racks with a certified 5,000 lb static load. $NVDA Vera Rubin NVL72 and $AMD Helios rack options are also on the list.

  @BRR_bull_run · Thread had no replies. Post of 13 Oct 2026 names systems available now. Specs cover liquid-cooled B300 nodes, 250 kW 4U CDU, high static load racks. Question in the post is how many existing halls can take a 144-GPU rack without new cooling.

  SMCI ships liquid-cooled OCP-spec racks now with up to 144 NVIDIA HGX B300 GPUs per rack, rear-door exchangers to 120 kW and a 250 kW in-rack coolant unit, plus Vera Rubin NVL72 and AMD Helios options.

  Open the issue

• 13 Oct · 10:30 · @everything_PE_Source · Open · @wireInfineon is launching a 27 kW three-phase PSU reference design for OCP ORV3 AI server racks, reaching over 98% peak efficiency and 116 W/in³ power density without a separate capacitor bank unit. The design targets 800 VDC or ±400 VDC architectures, uses CoolSiC and CoolMOS MOSFETs with an integrated energy buffer for 20 ms hold-up, and is showcased at OCP Booth E26.

  @everything_PE_ · Thread had no replies. everything PE post of 13 Oct 2026 cites the launch. Reference design uses CoolSiC, CoolMOS, EiceDRIVER and PSOC, with five-level ANPC PFC and three-level LLC, on the floor at OCP Booth E26 12-15 Oct.

  Infineon launches a 27 kW three-phase PSU reference design for OCP ORV3 AI server racks, achieving over 98 percent peak efficiency and 116 W/in³ density without a separate capacitor bank, for 800 VDC architectures.

  Open the issue

• 13 Oct · 11:00 · @financespotnewsSource · Open · @wireCisco launches Silicon One G302 51.2T top-of-rack switching silicon with 200G SerDes, 1.6T optics, and direct-to-chip liquid cooling at OCP; customer shipments targeted before end-2026, with P200 systems confirmed in production at Microsoft and Alibaba.

  @financespotnews · Thread had no replies. Post of 13 Oct 2026 cites the OCP launch. Cisco newsroom of 12 Oct confirms G302 planned to begin shipping before the end of 2026 and P200-powered systems in production at Microsoft and Alibaba.

  Cisco ships Silicon One G302 51.2T top-of-rack silicon with 200G SerDes, 1.6T optics and direct-to-chip liquid cooling before end-2026; P200 systems already in production at Microsoft and Alibaba.

• 13 Oct · 10:25 · @TheINIofficialSource · Open · @wireAdani Enterprises and Jabil sign a non-binding MoU for a proposed 50:50 joint venture to manufacture AI and data centre hardware in India, including AI racks, servers, storage, networking, power distribution and liquid-cooling systems. The multi-gigawatt platform targets domestic and global demand. The proposed JV is subject to definitive agreements and applicable approvals.

  @TheINIofficial · Thread had no replies. News post of 13 Oct 2026 cites the non-binding MoU. Scope covers racks, liquid cooling and power gear. Still subject to definitive agreements.

  Adani Enterprises and Jabil sign a 50:50 MoU to manufacture AI racks, servers, power distribution and liquid-cooling systems in India for a multi-GW platform.

• 13 Oct · 10:57 · @Awulakpakpa24Source · Open · @wireDelta Electronics is showcasing its 3.6MW liquid to liquid CDU at OCP with three hot-swappable 25kW pumps running on 800V DC. Huayuan expects next generation cooling solutions to start mass production in H2 2026. Delta is targeting Q1 2027 for mass production of its separate 2MW Project Deschutes CDU. Google already has a 2MW CDU design compatible with electronically driven canned motor pumps. GOOGL, MSFT and NVDA are working on common 800V DC interfaces through OCP.

  @Awulakpakpa24 · Thread had no replies. Post of 13 Oct 2026 names the OCP showcase and the two production clocks. The 3.6 MW unit is on the floor; volume shipments are separate.

  Delta Electronics showcases a 3.6 MW liquid-to-liquid CDU at OCP with three 25 kW hot-swappable pumps on 800V DC; Huayuan targets next-generation cooling mass production in H2 2026 and Delta targets Q1 2027 for its 2 MW Project Deschutes.

  Open the issue

• 13 Oct · 13:08 · @OptionKing666Source · Open · @wireHon Hai is projected to expand its NVIDIA NVL72 rack market share to 43% in 2027, up from 37%, according to Morgan Stanley. Total industry rack output will rise 10% YoY to 90.4K units, following a 209% surge to 82K units in 2026. Hon Hai volume will reach 38.8K racks, while Quanta holds a 26% share and Wistron takes 24%. This shift squeezes second-tier builders, whose combined volume share drops from 14% to 7% by 2027. Scale advantages in cooling and power distribution increasingly favor the top three assemblers. Slower-than-expected AI server penetration or delays in customer data center buildouts represent the key downside risk to these volume projections. Report date: 2026-10-12 · MS

  @OptionKing666 · Thread had no replies. The chart is the Morgan Stanley note of 12 Oct 2026. Hon Hai volume to 38.8K racks, Quanta 26 percent, Wistron 24 percent. Second-tier share falls from 14 to 7 percent by 2027.

  Morgan Stanley says Hon Hai expands NVIDIA NVL72 rack share to 43 percent in 2027 (38.8K units) as scale in cooling and power distribution favors the top three assemblers.

• 13 Oct · 02:56 · @CoolerMasterSource · Open · @wireLiquid cooling without rebuilding the facility. The Cooler Master 300 kW AALC In-Row CDU provides 300 kW nominal and up to 450 kW maximum cooling capacity for AI infrastructure, including GB200 and GB300 NVL72 systems.

  @CoolerMaster · Thread replies were non-technical. Company post of 13 Oct 2026 names the in-row CDU for existing halls, supporting NVL72 systems.

  Cooler Master ships a 300 kW AALC In-Row CDU (up to 450 kW) for GB200 and GB300 NVL72 racks without requiring a facility rebuild.

  Open the issue

• 13 Oct · 07:33 · @allday_stocksSource · Open · @wireSolarEdge showcases 800V DC AI Data Center Power Technology at Nvidia AI Factory MGX Ecosystem Showcase. Developing grid-to-rack power architecture using solid-state transformers and medium-voltage-to-800V DC conversion. Three-phase conversion stage already operating under load in engineering labs; full system validation underway. Working with Nvidia on protection and grounding standards for scalable, high-density AI infrastructure.

  @allday_stocks · Thread had no replies. Company announcement of 13 Oct 2026 names the SST cell and containerized architecture on the floor at OCP. Three-phase MV-to-800V stage runs under load; system validation continues. Joint protection framework with NVIDIA from September.

  SolarEdge showcases its MVAC-to-800 VDC solid-state transformer at the NVIDIA AI Factory MGX Ecosystem Showcase at OCP 12-15 Oct, with the three-phase conversion stage already operating under load in labs and full-path validation underway.

  Open the issue

Sittings

• 2026-10-14 · @desk

  +8

  LITEON targets 800V-to-50V mass production in 2H 2026

  LITEON told OCP its 800V-to-50V power brick delivers 10 kW at close to 99 percent peak efficiency, with a 12 kW version due in 2027. The path continues through an intermediate bus converter at 1.5 kW above 98 percent, then the regulator next to the GPU. For Vera Rubin NVL72 it showed a 95 kW DC/DC shelf and a 280 kW in-rack CDU. A reply asks whether the 91 percent figure is the last step or the full path from 800V. LITEON states the 800V power rack finishes validation and enters mass production in the second half of this year.

  LITEON captures the 800V-to-50V brick at 10 kW near 99 percent efficiency. Mass production of the 800V power rack is targeted for the second half of 2026, with a 12 kW version in 2027.

  Open the issue

• 2026-10-13 · @desk

  +8

  Delta captures Google and Nvidia DC/DC power into 2027

  J.P. Morgan says Delta Electronics DC/DC server power revenue rises about 60 percent year on year in 2027 on Google vertical power modules and Nvidia lateral products, driven mainly by TPU volume. DC/DC moves from about 6 percent of 2026 revenue to 7 percent in 2027. Vicor’s Section 337 complaint is viewed as manageable via licensing or redesign. A reply confirms the 7 percent share is the TPU-driven piece.

  Delta Electronics captures Google vertical and Nvidia lateral DC/DC server power. Revenue on that line rises about 60 percent year on year in 2027.

  Open the issue

• 2026-10-12 · @desk

  +8

  AWS Trainium 3 liquid-cooled servers slip to December

  J.P. Morgan says Amazon AWS Trainium 3 liquid-cooled server shipments delay from September to December 2026. The slip creates a temporary air pocket for liquid-cooling and chassis suppliers ahead of a larger rack buildout in 1H27. Wiwynn September revenue surged 100 percent year on year with AWS over 50 percent of that month. Accton grew 66 percent under consignment. Thread had no replies.

  Wiwynn and Accton capture Trainium 3 liquid-cooled server shipments. The delay moves volume from September to December 2026. Larger rack build sits in 1H27.

  Open the issue

• 2026-10-10 · @desk

  

  GlobalFoundries starts US silicon interposers for TSMC CoWoS in 1H 2028

  GlobalFoundries will manufacture silicon interposers at its Malta, New York fab for TSMC CoWoS under a multi-year $2 billion agreement announced 8 Oct 2026. The site is the first U.S. source of those interposers, including embedded deep-trench capacitor parts. Volume production is expected to ramp in the first half of 2028. The reply only restates the $2 billion New York figure.

  GlobalFoundries captures the first U.S. source of silicon interposers for TSMC CoWoS at Malta, New York. Volume production ramps in the first half of 2028.

Printed

• Issue 01 · Electrification+6The current is the chainNVIDIA captures the chip. The building captures the current. Price the phase, not the logo.
• Issue 02 · Physical AIThe body is a reducerAssembler is not the industry. The reducer is. Sit Tsu and Hotaka, then sit the qualification calendar.
• Issue 03 · Cross-book+8Watt and handDo not pick a winner in the logo layer. Pick the step that cannot be skipped, and the year it starts to print.

Next to sit

• Next

  The SST certification calendar

  After 2029 is a slide. The desk wants the cert, the SiC, and who eats the transformer.

• Queue

  Un-caging the cell

  Sourcing is step one. Thousands of hours under thermal load is the gate that ships volume.

• Queue

  The park that iterates overnight

  Harmonic, servo, and force sensor in one drive. That velocity is the second book on the body.

Method

1. 01

   Break on X

   The chart, the thread, the first map. That is the desk’s inbound.

2. 02

   Sit the comments

   The missing column usually lives under the post. Margin. Who. When.

3. 03

   Pull the sources

   SemiAnalysis, the vendor note, the GTC tape. No logo soup.

4. 04

   Write the view

   One sentence that can sit on the canvas. Price the phase, not the logo.

5. 05

   Draw the chain

   Steps, who captures, when it prints. The building, not the brand.

6. 06

   Cut the briefing

   A night voice note. Walk the chain in under four minutes.

7. 07

   Cut the short

   Nine by sixteen. One line. Then YouTube, then TikTok.