Anthropic signs $35B Lambda cloud deal, Nscale raises $3B - Silicon Brief #86

September 07, 2026 - Silicon Brief
Hi there,

AI infrastructure demand is shaping both power planning and data center expansion. A CERA Consulting study says targeted U.S. transmission investment could save electricity customers up to $15.3 billion by 2050 as AI increases power demand, while Equinix introduces Equinix Fabric One to connect enterprise, cloud, and AI environments.

Sharon AI has signed a five year agreement with Rafay Systems supporting up to 150,000 GPUs, and Nscale has committed $3.5 billion in computing resources to Figure, with potential deployments involving up to 100,000 Nvidia GPUs. Elsewhere, Dirac and Amazon Web Services (AWS) are working on AI manufacturing systems, while BYD expands in house development of automotive chips.
Nscale Secures $3 Billion for AI Infrastructure in Texas and North Carolina
Nscale closes approximately $3 billion in financing for AI infrastructure projects in Ward County, Texas, and Madison County, North Carolina. The funding supports a purpose built AI campus, a colocation site, GPU capacity and networking infrastructure. Read more
Lam Research begins construction on Oregon AI semiconductor lab
Lam Research begins building a $620 million semiconductor research lab in Tualatin, Oregon. The facility is part of the company's previously announced $3 billion global expansion of research and development labs. Read more
Salute to Acquire T5 Operations for AI and Cloud Data Centers
Salute announces a definitive agreement to acquire T5 Operations on September 1, 2026. The deal expands Salute's facility management and operations services for AI, high performance computing, and cloud data centers. Read more
CERA Study Says Transmission Investment Could Save Customers $15.3 Billion
A CERA Consulting study, commissioned by National Grid, the Electricity Customer Alliance, and Converge Strategies, finds that expanded U.S. transmission infrastructure could save electricity customers up to $15.3 billion by 2050. The analysis links interregional transmission to rising demand from AI and industrial activity, as well as grid reliability and national security. Read more
Dirac and AWS partner on AI manufacturing platform BuildOS
Dirac partners with Amazon Web Services to run its BuildOS AI process design platform on AWS infrastructure. The companies will support automated production workflows for defense, aerospace and industrial manufacturers, including sensitive workloads through AWS GovCloud. Read more
Equinix Launches Fabric One for Enterprise AI Connectivity
Equinix introduces Equinix Fabric One, a managed connectivity service for linking enterprise, cloud and AI environments. The service will enter beta testing later this year, with AWS and Google Cloud as partners, ahead of planned general availability in 2027. Read more
Broadcom reports 221% growth in AI semiconductor revenue
Broadcom reports a 221% increase in AI semiconductor revenue and projects total fourth quarter revenue of $21.7 billion. Read more
Sivers Semiconductors invests USD 30 million in Glasgow InP photonics facility
Sivers Semiconductors will invest USD 30 million to expand its indium phosphide photonics manufacturing facility in Glasgow, Scotland. The expansion will increase production capacity for lasers used in AI datacenters and optical networks, with operations expected to begin in the fourth quarter of 2027. Read more
Cango's EcoHash begins commercial GPU services in Georgia
Cango subsidiary EcoHash has activated its first GPU servers at a 3 megawatt AI compute section in Georgia. The company is now providing commercial GPU computing services while adding further server capacity. Read more
LITEON Invests $176 Million in DCX Liquid Cooling Systems
LITEON Technology will acquire about 25% of Poland-based DCX Liquid Cooling Systems for approximately $176 million. The companies will develop liquid cooling and integrated power-and-cooling systems for hyperscale AI data centers. Read more
Lyte Raises $165M Series C at $1.6B Valuation
Lyte, a Sunnyvale physical AI company, raises $165 million in Series C funding led by Maverick Silicon, valuing the company at $1.6 billion. The funding will support production of its custom silicon and LyteVision platform, alongside expansion in robotics and industrial deployments. Read more
Stäubli introduces power and liquid cooling connectivity for AI data centers
Stäubli introduces electrical connectors and liquid cooling couplings for high density AI data centers. The products support power distribution and cooling systems, including configurations compatible with Open Compute Project Open Rack v3 architectures. Read more
ASUS expands AI infrastructure partnerships from cloud to edge
ASUS announces partnerships with Intel, AMD, IBM, Samsung, and other companies to support AI infrastructure across cloud and industrial edge deployments. The company also presented servers, AI platforms, and rugged hardware for machine vision, automation, video analytics, and healthcare. Read more
NOVOSENSE reports 66.7% first half 2026 revenue growth
NOVOSENSE reports first half 2026 revenue of about $374.54 million, up 66.73% year on year, while shareholder attributable net profit reached about $9.97 million. The Chinese chipmaker cites growth in automotive, industrial and energy markets, and says several products are shipping for AI server power applications. Read more
D-Robotics powers home robots at IFA 2026 with Sunrise AI chips
D-Robotics says several home robots shown at IFA 2026, including TCL's hey AiMe, Vbot's SuperDog and xLean's TR1, use its Sunrise AI chips. The company supplies chips and software for robotics platforms spanning 5 to 560 TOPS. Read more
Ezurio announces Carbon AM67 edge AI module with 4 TOPS
Ezurio announces the Carbon AM67, an edge AI system on module built on Texas Instruments AM67x processors. The module provides up to 4 TOPS of AI performance and is available for sampling, with production modules and kits expected in October. Read more
Vobile and UNN sign deal for 25MW AI infrastructure in Brunei
Vobile and Brunei's Unified National Networks, or UNN, signed agreements to develop AI factories in Brunei Darussalam. UNN will provide 25MW of capacity for Vobile's DreamMaker platform, with a further 25MW reserved for expansion. Read more
BYD develops in house automotive chips
BYD expands its semiconductor operations to include power devices and smart driving chips, according to a DIGITIMES Intelligence report. Its Xuanji A3 chip is reportedly in scaled production on a 4 nm automotive process, as China seeks to localize more auto semiconductor supply. Read more
Anthropic signs $35 billion cloud deal with Lambda
Anthropic signs a $35 billion cloud computing agreement with Lambda, which is backed by Nvidia, for access to a Texas data center. The deal follows a separate $45 billion agreement with Nscale as AI developers compete for computing capacity. Read more
Kulicke & Soffa Advances Co-Packaged Optics for AI Infrastructure
Kulicke & Soffa announces advanced packaging and thermo-compression bonding solutions for co-packaged optics, targeting higher bandwidth and data processing efficiency in AI data centers and networking. The company will present the technologies at SEMICON Taiwan from September 2 to September 4, 2026. Read more
SLB acquires Kelvion for $4.1 billion to expand data center business
SLB will acquire thermal management provider Kelvion for $4.1 billion, including cash and assumed debt. The deal expands SLB's data center infrastructure business as AI investment increases demand for cooling systems, subject to regulatory approvals. Read more
BUUU Group to Acquire 60% of Brightray Science for AI Data Center Expansion
BUUU Group agrees to acquire a 60% stake in modular data center provider Brightray Science, which will become a consolidated subsidiary. The company also secures financing agreements that could provide more than US$60 million, subject to warrant exercises, closing conditions and regulatory approvals. Read more
Flex to acquire EPC Power for $4.4 billion
Flex agrees to acquire EPC Power for $4.4 billion, pending regulatory approval and other closing conditions. The deal would add power conversion technology for AI data centers and grid applications to Flex's Cloud and Power Infrastructure segment, which the company plans to spin off in 2027. Read more
MINISFORUM Unveils Ryzen AI MAX+ PRO 495 Systems at IFA 2026
MINISFORUM introduces an AI mini workstation and AI-focused NAS powered by AMD Ryzen AI MAX+ PRO 495 processors. The systems offer up to 131 TOPS of AI performance, while the NAS supports up to 200TB of local storage. Read more
Sharon AI signs five year Rafay Systems deal for 150,000 GPUs
Sharon AI will use Rafay Systems as the orchestration and operations platform for its AI Factory environments. The five year agreement covers infrastructure management for up to 150,000 GPUs across Asia Pacific and other markets. Read more
Nscale commits $3.5 billion in compute to Figure
Nscale signs a multi year agreement to provide Figure with $3.5 billion in computing resources for humanoid robotics development. The deal could exceed $6 billion and involve up to 100,000 Nvidia GPUs, with initial deployments planned in Texas in the second half of 2027. Read more
HUMAIN and Compal Electronics Partner on Cloud to Edge AI Infrastructure
HUMAIN and Compal Electronics partner to develop AI infrastructure and intelligent computing platforms, with work spanning cloud to edge deployments. The companies presented the collaboration at LEAP 2026, including plans for scalable systems for sectors such as healthcare. Read more
Nvidia invests $3.5 billion in MediaTek for custom AI chips
Nvidia is investing $3.5 billion in Taiwanese chipmaker MediaTek to support the design of custom AI chips for companies and cloud providers. The partnership gives MediaTek access to Nvidia technology and NVLink Fusion for its custom chip business. Read more
Light Source Communications Secures Second Tenant for Tulsa to St. Louis Fiber Route
Light Source Communications secures a second hyperscale tenant for its dark fiber route between Tulsa and St. Louis. The more than 500 mile network is designed for AI and high performance computing workloads, with construction scheduled for completion in the third quarter of 2027. Read more
European Union awards Bull €387.8 million contract for LUMI-AI supercomputer
The European Union awards French state owned Bull a €387.8 million contract to build the LUMI-AI supercomputer in Finland. The system is expected to begin operating in the second half of 2027 under the EuroHPC initiative. Read more
Cloover Reaches Profitability and Expands Across Europe
Cloover, an AI native energy platform, reports profitability and an annual revenue run rate above $350 million, three years after launch. The company will expand into the UK, France and Poland, backed by more than $1.3 billion in financing capacity for connected home energy systems. Read more
Aranya Raises $11M for AI Inference Infrastructure
Aranya raises $11 million across seed and pre-seed funding rounds to develop its multicluster operating system for AI workloads. The company says its platform can turn bare metal hardware into production-ready clusters within 48 hours and plans to expand its team. Read more
CollPlant Completes LightSolver Acquisition for Photonic Computing
CollPlant Biotechnologies completes its acquisition of Israeli deep tech company LightSolver, expanding into photonic and high performance computing while continuing its regenerative medicine operations. The company issued shares and pre funded warrants, and invested $5 million in LightSolver's laser based computing platform and commercialization. Read more
AFL says AI data centers strain long haul fiber capacity
AFL says growing AI workloads are increasing demand for capacity, low latency and resilient network connections. The company reports that long haul fiber routes face immediate pressure, with ribbon fiber lead times exceeding 60 weeks in many markets. Read more
Corvex Expands Data Center Capacity Fivefold for AI Workloads
Corvex plans to increase its data center capacity from about 1.5 MW to 8 MW by the end of 2026, supporting roughly 3,000 additional GPUs at two US facilities. The AI computing platform secured $33 million in funding and holds an option that could raise capacity above 20 MW. Read more
Together AI and HUMAIN Partner on Saudi AI Data Center
Together AI and HUMAIN partner to build a 250 megawatt data center in Saudi Arabia and combine their AI infrastructure and cloud services. The companies project more than $5 billion in first year revenue from the collaboration. Read more
Mixx Technologies launches SxC Connector for AI infrastructure
Mixx Technologies introduces the SxC Connector, an optical connector designed for high capacity AI infrastructure. The company says it can terminate up to 24,576 fibers in one rack unit and will demonstrate it at SEMICON Taiwan beginning September 2, 2026. Read more
stackArmor receives FedRAMP High authorization for The Armory
stackArmor, a Quantum Sky company, receives FedRAMP High authorization for The Armory, its secure cloud environment built on Google Cloud. The platform supports software deployments handling sensitive federal information for agencies and independent software vendors. Read more
SWI Group Joins NVIDIA Cloud Partner Program
SWI Group becomes a Preferred Partner in the NVIDIA Cloud Partner program. The agreement gives the investment firm access to NVIDIA resources for AI infrastructure deployments in Europe and the US. Read more
OpenAI Reportedly Buys Thousands of Mac Mini and Mac Studio Systems for AI Training
OpenAI reportedly purchased tens of thousands of Apple Mac mini and Mac Studio systems for AI training and autonomous computer use agents. The purchases highlight growing demand for Macs in local AI workloads, while memory chip shortages constrain supply. Read more
S&P Global Completes datacenterHawk Acquisition
S&P Global completed its acquisition of datacenterHawk on September 1, 2026. The deal combines datacenterHawk's data center market intelligence with S&P Global Energy's insights to support analysis of AI infrastructure. Read more
Acer Launches Veriton RI110 AI Mini Workstation
Acer introduces the Veriton RI110 AI Mini Workstation, a compact system built for running AI workloads, including models with up to 120 billion parameters. It includes an Intel Core Ultra X7 processor, up to 96 GB of memory, 4 TB of storage, and OCuLink connectivity, with regional availability beginning in late 2026. Read more
Tensordyne outlines Napier AI inference system
Tensordyne outlines its Napier AI inference system, which combines custom processors, on chip SRAM, HBM3e memory and a low latency interconnect for large models. Its TDN AIP chip is in fabrication at TSMC, while the company’s performance and cost claims are based on modeling and simulation. Read more
Chatsworth Products joins Digital Realty AI lab in London
Chatsworth Products joins the Digital Realty Innovation Lab in London, where organizations can test AI and hybrid cloud infrastructure using real workloads. The company will showcase its ZetaFrame Cabinet System and other equipment for high density AI deployments. Read more
Zendure unveils Agentic HEMS for autonomous home energy management
Zendure introduces Agentic HEMS at IFA 2026, an AI based system that forecasts and coordinates household energy assets including solar panels, batteries, electric vehicles, heat pumps, tariffs and home loads. The platform combines ZenPulse, ZENKI AI Agents and Zen+OS, with integrations for Home Assistant, Homey and MQTT. Read more
SEMI reports record $40.53B semiconductor equipment billings in Q2 2026
SEMI reports global semiconductor equipment billings reached a record $40.53 billion in the second quarter of 2026, up 23% from a year earlier and 11% from the prior quarter. The increase reflects investment in AI infrastructure and advanced chip manufacturing across major regions. Read more
Nvidia Expands AI Hardware With Vera Rubin Architecture
Nvidia is broadening its AI hardware portfolio beyond GPUs with the Vera Rubin architecture, including the Vera CPU and specialized data center systems. The expansion targets more efficient data orchestration in large AI data centers as competition from Amazon and Google grows. Read more
SpaceX Builds Texas Turbine Foundry for AI Power Demand
SpaceX is building a foundry in Bastrop, Texas, to produce gas turbine components and shorten turbine production timelines by up to 18 months. The effort responds to rising electricity demand from AI infrastructure, while adding to concerns about emissions from gas turbine projects, including disputes in Memphis. Read more
DoGo Power Integrates AI With Grid Forming Energy Storage
DoGo Power presented its AI integrated, grid forming energy storage technology at Intersolar South America 2026 in São Paulo. The company says the systems are designed to support renewable power grids, and its 4S architecture received the Top Innovation Award Brazil. Read more

We hope you enjoyed this article.

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

Subscribe to Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.