DataNews · Section
Enphase Energy
Enphase has begun building 4 kW power modules in Texas for its IQ Solid-State Transformer platform, designed to deliver up to 5 MW per rack and convert medium-voltage utility power directly into the 800V DC architecture emerging inside AI data centers.
Corning
Verizon and Corning have signed a multi-billion-dollar supply agreement for more than 80 million miles of high-density optical fiber through 2032, supporting both nationwide broadband expansion and long-haul connectivity between AI data centers.
Qualcomm
Amazon and Qualcomm have entered a multigenerational collaboration covering custom AI-inference silicon and optical networking, while an SEC filing reveals an Amazon equity warrant whose vesting milestones are tied to Qualcomm purchases reaching as much as $60 billion.
Arm
Arm has introduced Neoverse CSS N4, expanded its AGI CPU ecosystem and launched an 80-plus-company Physical AI initiative as it attempts to make its architecture the common compute layer connecting agentic AI from cloud data centers to robots and autonomous systems.
Pixxel
Pixxel has raised $100 million in a Series C led by Temasek and Seraphim to expand its satellite, Earth-intelligence, sovereign-space and manufacturing platforms.
iPronics
Silicon-photonics startup iPronics has raised $125 million with participation from Nvidia to scale optical circuit switches designed to dynamically reconnect increasingly large AI clusters while reducing power, space and networking bottlenecks.
NVIDIA
Nvidia has agreed to acquire Hugging Face for $12.93 billion, extending the chipmaker deeper into the software and open-model ecosystem while pledging that the platform will remain open, multi-cloud and compute-agnostic.
Bloomberg
Apple was reported to be eliminating more than 200 positions across Siri, Vision Pro and software teams.
Starcloud
Starcloud raised $250 million at a $2.3 billion valuation, with Nvidia and Cisco Investments joining as new investors.
Mitsubishi Electric
Mitsubishi Electric agreed to acquire PCI Energy Solutions for $1.4 billion, adding software used across power-market and grid operations.
Castelion
Castelion announced more than $1 billion in financing at a $13 billion valuation to scale advanced defense manufacturing.
Micron
Micron committed $10 billion over a decade to a new Boise research center focused on advanced memory and computing systems.
Amazon
Amazon says Prime Air will expand dramatically from its current launch sites toward service in nearly 500 U.S. cities and towns.
Reuters
Stripe agreed to acquire AI-model routing platform OpenRouter in a transaction reported at slightly more than $8 billion.
Reuters
AI chip startup Etched raised $700 million at a $21 billion valuation and says it has more than $1 billion in customer contracts.
Reuters
Google is changing enforcement of its site reputation abuse policy in the EEA after EU Digital Markets Act scrutiny of publisher demotions.
Reuters
Anthropic explored acquiring AI-chip startup MatX for about $7 billion before talks shifted toward a possible partnership, underscoring its push into custom silicon.
Reuters
Internal documents reviewed by Reuters show Meta explored shrinking some teams by up to 60% under Project OT while AI agents were expected to absorb work.
Reuters
Anthropic will spend $45 billion over six years on roughly 460 MW of Nscale AI compute capacity in West Virginia using Nvidia Vera Rubin systems.
TechCrunch
Robotics startup Generalist reportedly reached a $3 billion valuation after raising nearly $200 million in additional financing led by 8VC.
Louisiana Economic Development
SpaceX plans a $100 billion launch campus in Vermilion Parish, Louisiana, designed for extremely high Starship launch cadence.
Reuters
Moonshot AI is reportedly negotiating revenue-sharing arrangements with Microsoft, Amazon and Google for hosting Kimi K3, seeking as much as 30% of related revenue.
Gatik
Gatik raised $200 million as its autonomous middle-mile freight business moves toward larger commercial deployment.
Emerald AI
Emerald AI raised $150 million at a $1.05 billion valuation to expand software that helps AI data centers operate more flexibly around grid constraints.
Infineon Technologies
Infineon is acquiring C2i Semiconductors to expand its AI data-center power-management portfolio, including software-defined multiphase control and vertical power delivery.
Business Insider
Hugging Face is exploring a potential sale that could value the AI developer platform at $13 billion or more, nearly triple its 2023 valuation.
Quintessent via Business Wire
Quintessent raised a $40 million oversubscribed Series A and began customer sampling of a quantum-dot DWDM comb laser designed for AI optical interconnects.
nVent
nVent agreed to acquire Maverick Power for $1.75 billion, with up to $550 million of additional consideration tied to performance, expanding its data-center power-distribution portfolio.
Reuters
Xiaomi unveiled its second-generation Xring O3 smartphone processor as it expands a multibillion-dollar effort to control more of its own silicon stack.
XPeng
XPeng's robotics business raised more than $900 million at a post-money valuation above $6.3 billion, backed by IDG Capital, Tencent, Alibaba and others.
Reuters
Nvidia is reportedly discussing an investment in Perplexity that would value the AI search and agent startup above $30 billion as its annualized revenue exceeds $750 million.
Times of AI
OpenAI is now urging California to strengthen SB 53 with monitoring during model training and stronger cybersecurity protections.
Waymo
Waymo has detailed a purpose-built 5nm ASIC delivering more than 1,000 TOPS for front-end sensor processing in its robotaxis.
Satya Nadella / NVIDIA
Microsoft has received its first production Nvidia Vera Rubin systems, marking the transition of Nvidia’s next-generation rack-scale AI platform into hyperscale deployment.
Axios
CrowdStrike global CTO Elia Zaitsev is leaving to co-found Cognition, a venture firm targeting a $170 million fund focused on cybersecurity for the AI-agent era.
OpenAI
OpenAI cut GPT-5.6 Sol API pricing by more than 20% for three months, extending model-price competition to its flagship tier.
DeepSeek
DeepSeek released an experimental vision-enabled V4 Flash model that adds image and screenshot understanding to its low-cost agent platform.
Kakao
Kakao plans to split into KakaoAI and KakaoX, isolating KakaoTalk, AI, advertising and commerce in a separately valued public company.
Nvidia
Nvidia made a minority investment in Cloverleaf Infrastructure, a developer focused on securing land and power for large AI data-center projects.
Reuters
Muon Space raised $250 million as it expands satellite manufacturing and spacecraft architectures intended to support high-power orbital computing workloads.
Binance
Binance launched Agent OS, allowing permissioned AI applications to access trading, wallets, payments and market data through MCP-compatible interfaces.
Salesforce
Salesforce launched Slack Code, creating shared project channels where teams can supervise, review and approve work performed by AI coding agents.
TechCrunch
Rivian spinout Also raised $150 million to expand autonomous-driving technology and commercial delivery vehicle programs.
Fortune
Rillet raised $100 million at a $1 billion valuation as its agent-first accounting platform targets legacy ERP and finance systems.
Reuters
Major global banks are adopting Ant International's Falcon time-series model for FX exposure and cash-flow forecasting.
Reuters
Munich Re agreed to acquire cyber-insurance provider At-Bay for $575 million, deepening the convergence of security telemetry and underwriting.
OpenAI
OpenAI launched a teen-specific ChatGPT experience with automatic age estimation, stronger safety restrictions and parental controls.
Reuters
Velaura AI raised $110 million at a valuation above $1 billion for chip and software technology designed to reduce data-center electricity consumption.
Reuters
Albertsons says customers using its more comprehensive conversational AI shopping assistants generate up to a 26% increase in average order value.
Reuters
Deutsche Telekom agreed to acquire Fiberhost and Inea for about €1 billion, expanding T-Mobile Polska into owned fixed-fiber infrastructure.
Atlassian
Atlassian's new data-contribution rules govern how eligible Jira, Confluence and related cloud data can be used to improve AI and product experiences.
Business Insider
Amazon is redistributing internal workloads across more AWS regions under Region Flex as electricity and data-center capacity constrain expansion.
The Wall Street Journal
New reporting links Meta to a proposed 2.2-GW hyperscale data-center development near Maysville, Kentucky, amid local land and power disputes.
Bloomberg
Nvidia was reported to be in early discussions with South Korean AI-chip company Rebellions about a partnership, investment or possible acquisition.
Newcomer / The Information
Nvidia was reported to have agreed to a $6 billion non-exclusive Poolside technology license plus a $1 billion equity investment.
Reuters
Cerebras unveiled the CS-4 inference system and said it is targeting 600 MW of deployed compute capacity by the end of 2027.
Business Insider
EliseAI was reported to be discussing a roughly $300 million financing at a $3.7 billion valuation.
Reuters
Pony.ai reported a 691.2% increase in robotaxi revenue and said its overseas contracted deployment pipeline exceeds 4,000 vehicles.
Reuters
AI-content startup Higgsfield raised $400 million at a $5.4 billion valuation while reporting a $700 million annualized revenue run rate.
Reuters
Uber and Zipline announced a U.S. autonomous-drone delivery partnership alongside an undisclosed strategic investment by Uber.
Reuters
Smack Technologies raised $61 million to accelerate battlefield AI systems and a wearable display as Pentagon buyers seek a broader military-AI supplier base.
Navitas Semiconductor / GlobeNewswire
Navitas Semiconductor agreed to acquire Claros for up to approximately $232.8 million, adding vertical power-delivery and integrated-voltage-regulator technology for AI processors.
NVIDIA
NVIDIA says Groq 3 LPX is now in full production as a specialized inference accelerator for the Vera Rubin platform, targeting low-latency token generation for agentic AI workloads.
Thomson Reuters
Thomson Reuters launched Thomson, its first proprietary large language model, after investing about $40 million in talent and compute to specialize an open-source foundation with proprietary professional content.
Snowflake
Snowflake opened CoCo automations to public preview, allowing recurring unattended AI-agent runs in Snowflake-managed sandboxes with auditable Cortex threads.
The Wall Street Journal
Applied Materials reported $9.12 billion in quarterly revenue and is expanding manufacturing capacity as AI-driven semiconductor demand creates multi-year visibility into advanced packaging and chip-equipment requirements.
Reuters
SMIC reported second-quarter profit of $479.2 million as AI-related semiconductor demand helped drive stronger foundry utilization and capacity expansion.
Reuters
Lenovo reported a 43% jump in quarterly revenue while its AI-server pipeline reached $54 billion, up 157% sequentially.
Associated Press
Google unveiled the Pixel 11 family, Pixel Watch 5 and Pixel Tag, using the launch to deepen Gemini AI integration across its hardware ecosystem.
Reuters
Meta Platforms is entering a landmark federal trial in Oakland over claims that Facebook and Instagram were designed in ways that harmed or addicted younger users and violated children’s privacy laws.
The Wall Street Journal
Form Energy raised $750 million as its 100-hour iron-air battery technology gains traction for grids supporting AI data centers.
Reuters
Foxconn reported a 35% rise in second-quarter profit as cloud and networking products, including AI servers, reached 51% of quarterly revenue for the first time, while the manufacturer prepared Nvidia Vera Rubin server racks for fourth-quarter shipments.
Reuters
Alpha Compute signed a binding term sheet for a 200MW Pennsylvania AI data-center campus that could expand to 1GW and include on-site gas generation tied to Marcellus shale rights.
Reuters
Longtime OpenAI executive Brad Lightcap is leaving the company to start a new venture, adding another prominent founder candidate to the frontier-AI ecosystem.
NVIDIA
Nvidia introduced Nemotron 3.5 Lightning and NeMo Switchyard, an experimental routing layer designed to send AI requests to different models based on quality, latency and cost.
River AI
River AI, founded by xAI co-founder Igor Babuschkin, raised $1.1 billion with backing from General Catalyst, Nvidia, AMD Ventures and others to build enterprise infrastructure for customizable, customer-owned AI models.
Riot Platforms
Riot Platforms disclosed a 20-year, 191 MW Rockdale compute agreement expected to generate about $9.1 billion through 2048, underscoring the shift of power-rich Bitcoin mining sites toward AI infrastructure.
Reuters
IBM and Together AI signed a $240 million multiyear agreement to build a large AI inference cluster on IBM Cloud using Nvidia HGX B300 systems, Blackwell 300 processors and Spectrum-X networking.
Reuters
French media groups are asking the country's competition regulator to intervene against Google's AI-generated search summaries, arguing that the feature diverts traffic from original publishers.
Axios
Cambridge Aerospace raised $300 million at a $3.4 billion valuation to scale lower-cost interceptors for drones, cruise missiles and other air-defense threats.
Reuters
Archer Aviation agreed to acquire Boeing's Wisk Aero, Insitu and SkyGrid businesses, expanding into autonomous flight, defense drones and airspace software.
Reuters
Meta released Muse Glimmer, a compact open-weight model designed to run agentic workloads locally on a Mac or PC with a single GPU, while expanding its downloadable Muse model strategy.
Reuters
Meta is creating a $1 billion fund for communities affected by its expanding data-center footprint as the company accelerates AI infrastructure spending.
Financial Times
King’s Cross is becoming one of Europe’s densest AI clusters as frontier-model companies, startups, researchers and capital concentrate in central London.
The Australian
OpenAI and South Australia signed an MOU to explore AI skills, investment, government adoption and infrastructure aligned with the state's data-center strategy.
Chron / Houston Chronicle
Amazon is planning a major data-center campus at GW Ranch in Pecos County, Texas, using new on-site generation initially rather than relying on the ERCOT grid as AI infrastructure demand accelerates.
The Wall Street Journal
Lumilens raised more than $700 million at a $5.5 billion valuation as demand grows for optical interconnects used to connect large AI accelerator clusters.
Financial Times
Moonshot AI is restructuring its ownership and corporate organization as it prepares for a potential Hong Kong listing and adds major Chinese state-backed investors.
Reuters
Nvidia is preparing to invest as much as $3 billion in Lancium, the power-infrastructure developer behind the first operational Stargate AI campus in Abilene, Texas.
TechCrunch
Rippling says AI token spending briefly approached 40% of its R&D payroll equivalent before model routing and governance reduced the cost burden while keeping usage high.
Cloudflare
Cloudflare launched Kitesurf, a browser engine designed for autonomous AI agents, with lower CPU and memory use than Chromium in Cloudflare’s internal benchmark.
Reuters
A New Mexico court ordered Meta to contribute $567 million to a teen mental-health fund after an earlier $375 million jury award, bringing the combined judgment to $942 million, while imposing youth-safety product requirements in the state.
The Wall Street Journal
Nielsen agreed to acquire DoubleVerify for approximately $2.15 billion, combining audience measurement with digital-ad verification, fraud detection and media-quality technology.
Reuters
Only 7% of executives surveyed by KPMG said their organizations had established measurable ROI from AI, while nearly half had questioned, delayed, or reduced AI-agent deployments after costs exceeded expected benefits.
Barron’s
Atlassian beat fiscal fourth-quarter expectations and reported continued adoption of Rovo, its AI search, chat and agent platform embedded across Jira, Confluence and other enterprise workflows.
The Australian
Australian AI-infrastructure company Firmus has secured large new equity commitments as investors keep funding power, cooling and GPU infrastructure behind AI factories.
Reuters
Alibaba plans to require some large commercial users of its next open-weight Qwen model to share revenue, testing a freemium licensing model for open-weight AI.
Reuters
AMD agreed to acquire Toronto-based Taalas, a startup developing specialized silicon intended to reduce compute and memory bottlenecks during AI inference.
Wall Street Journal
SpaceX and Tesla plan to initially invest $16.8 billion in Terafab, a Texas semiconductor complex intended to support AI, robotics, autonomous vehicles and space-based data centers.
Reuters
DeepSeek has resumed a funding round seeking nearly $8 billion at a valuation close to 500 billion yuan, or about $74 billion, with Monolith Management reportedly in talks to participate.
Reuters
Unitree priced its Shanghai IPO at 150.8 yuan per share, valuing the humanoid-robotics company at about $9.04 billion, while DeepSeek invested roughly $20.8 million and agreed to collaborate on AI models for humanoid machines.
Reuters
Microsoft has opened its largest data-center hub in India in Hyderabad, adding a fourth cloud region as it expands roughly $20.5 billion of committed investment in India.
Reuters
Deutsche Telekom expanded its 2026 buyback after stronger earnings and high demand for its Munich AI computing facility, which is fully utilized.
Reuters
Meta launched Muse Code, a beta AI coding agent powered by Muse Spark 1.2 that can write, debug and verify code while managing longer software-development assignments.
Reuters
Anthropic is hiring an internal semiconductor team to co-design custom chips and Claude models as AI computing costs and capacity constraints rise.
Reuters
Alphabet reorganized Google DeepMind leadership as Demis Hassabis moved to a new chief scientist role and several senior AI researchers left to form Discovery Loop.
Reuters
Recent results from SAP, Capgemini, Sopra Steria and OVHcloud suggest enterprise AI spending is moving from experimentation toward integration, governance, modernization and controlled infrastructure.
Reuters
CoreWeave plans to build and operate three Indonesian AI data centers with 360 megawatts of contracted power, marking its first physical expansion into Asia-Pacific.
Reuters
Volta Infra raised capital at a $2.4 billion valuation and announced a six-year European AI cloud agreement reportedly worth $10 billion.
Reuters
Apple asked a federal judge to restrict OpenAI and former Apple employees from using alleged trade secrets in a consumer-hardware dispute.
Reuters
ArcelorMittal expanded its Microsoft cloud partnership after a separate AWS industrial-AI collaboration, showing a practical multicloud manufacturing architecture.
Orange
Orange and infrastructure investor Morrison are planning a jointly controlled French data-center platform targeting AI and cloud demand.
The Verge
Google paused a Google Earth generative-AI image feature one day after launch after users shared realistic synthetic scenes that raised policy and misinformation concerns.
Reuters
A Manhattan federal judge rejected most of Perplexity AI's bid to dismiss Reddit's lawsuit over alleged data scraping for AI-powered search.
Reuters
NXP Semiconductors is reportedly discussing a possible acquisition of Ambarella, a chip designer focused on low-power AI processors for cameras, vehicles and robotics.
Reuters
A German court ruled that AI music company Suno violated copyright protections by processing works represented by GEMA without authorization.
Reuters
Reddit's second-quarter revenue rose 61% to $805 million and its advertiser count increased 70%, while management said search-algorithm changes and SEO headwinds reduced user acquisition.
Reuters
AI cloud provider Nscale plans to acquire Anyscale, bringing together GPU infrastructure and software used to build and operate distributed AI workloads.
Reuters
Intel has reportedly provided RosaicLabs with access to Atom processor technology, including register-transfer-level code that could allow the startup to develop custom x86-based chip designs.
Reuters
Eliyan raised $145 million at a $1 billion valuation to license chip-interconnect technology and sell chiplets that move data between AI processors, memory and networking components.
Reuters
Coursera is investing $100 million in LearnVector, a new AI-education company founded by Andrew Ng that plans to build personalized training for white-collar workers, enterprises, governments and universities.
S&P Global
S&P Global has agreed to acquire datacenterHawk, combining asset-level data-center and fiber intelligence with power-market, grid, energy and technology research. Financial terms were not disclosed.
Lenovo
Lenovo has published a validated on-premises generative-AI design combining ThinkAgile MX650a V4, NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and Azure Arc-enabled Foundry Local. The architecture keeps models and prompts on customer-owned infrastructure while extending enterprise governance into the data center and edge.
OpenAI
OpenAI is rolling out Health in ChatGPT to U.S. users aged 18 and older on web and iOS, allowing people to connect supported medical records and Apple Health information for more contextual health conversations.
OpenAI
OpenAI has introduced Presence, an enterprise deployment product designed to help organizations operate voice and chat agents that can answer questions, use company systems, take approved actions and escalate to people.
OpenAI
OpenAI has documented a broad set of publisher deployments spanning document analysis, verification, transcription, translation, personalized briefings, archive search, audience tools and advertising workflows.
OpenAI
OpenAI has launched a small-business program combining virtual training, in-person AI academies, practical guides and partner integrations intended to help lean teams adopt ChatGPT Work and GPT-5.6.
ABB
ABB says its SACE Infinitus solid-state circuit breaker is available at 2,500 amps and certified to IEC 60947-2. The company links growing demand for the technology to a broader shift from AC toward DC distribution in data centers, marine systems and microgrids.
Siemens
Siemens has launched its SENTRON ECPD electronic circuit protection device in the UK, saying it can respond up to 1,000 times faster than conventional electromechanical protection and reduce short-circuit energy release by 99.5%. The product combines protection, switching, measurement and communications for low-voltage distribution systems.
NVIDIA
NVIDIA has introduced Blackwell-powered Jetson Thor T3000 and T2000 modules aimed at robotics, visual AI and edge workloads that need to run foundation models on compact, power-efficient systems. The launch expands the hardware available for moving advanced inference away from centralized data centers.
IBM
IBM has introduced the compact Power S1112, a one-socket Power11 system designed for on-premises deployment and capable of running AI inference locally using on-chip Matrix Math Acceleration. The launch reinforces local inference as a mainstream enterprise-server workload.
Siemens
Siemens and Infineon are collaborating on semiconductor circuit breakers for data centers, factories and battery systems. Infineon will supply silicon-carbide power modules for Siemens SENTRON 3QD2 breakers, which use semiconductor switching and protection algorithms for microsecond-scale interruption.
NVIDIA
NVIDIA’s DGX Station for Windows and DGX Spark push advanced AI development, inference and autonomous-agent workloads onto deskside systems built around Grace Blackwell processors. The products illustrate how local AI compute is emerging as an owned enterprise hardware category alongside cloud AI.
Eaton
Eaton is investing more than $30 million in a new 370,000-square-foot Nebraska manufacturing facility to expand U.S. medium-voltage switchgear production. The company explicitly links the expansion to AI data-center demand and expects production to begin in the first half of 2027.
Lenovo
Lenovo is expanding its Hybrid AI Advantage with NVIDIA across workstations, edge systems, data centers and larger AI factories, using Blackwell-based RTX PRO systems and an expanded services collaboration with IBM Technology Lifecycle Services. The portfolio reflects a broader move toward mixed local, edge and data-center inference.