Enterprise IT spending is forecast to grow 5% in 2027, the strongest spending outlook since October 2022, as businesses continue to invest in AI despite rising hardware costs and pressure on tech budgets. But 53% of enterprises exceeded their AI budgets in the previous six months, and the money has to come from somewhere, with 60% cutting contractors and consultants and 57% delaying hardware upgrades. Read the full article on where the budget is actually going: buff.ly/sr0xpnS #ITSpending #AI #EnterpriseIT #TechBudgets

Databricks acquired Row Zero to bring enterprise spreadsheets into its AI platform, letting users analyze massive datasets while maintaining centralized governance. Read more: buff.ly/1efsF3j #Databricks #EnterpriseAI #DataAnalytics #Spreadsheets #AIAgents

VAST Data extended DataEnclave to self-hosted and on-premises environments, using NVIDIA Confidential Computing to protect sensitive data accessed by AI workloads. Read more: buff.ly/dSPCB47 #VASTData #ConfidentialComputing #AISecurity #EnterpriseAI #DataSecurity

AI infrastructure spending has hit an economic milestone in the U.S., with inflation-adjusted spending on information-processing equipment reaching $752 billion in the second quarter of 2026 and surpassing residential investment at $748 billion. Economist Stijn van Nieuwerburgh estimates data center and AI infrastructure spending will total $10.3 trillion between 2025 and 2032, even as S&P Global Ratings projects all six major tech companies will generate negative free operating cash flow in 2026 and 2027. Read the full article on the shift and whether the spending can justify itself: buff.ly/E0yt567 #AgenticAI #AIInfrastructure #DataCenters #TechEconomy

Oracle issued a force majeure notice to the developer of a massive New Mexico data center, positioning itself to defer payments if regulatory and construction hurdles delay the planned 2028 launch, while both Oracle and Blue Owl publicly insist the project remains on schedule. The debt tied to the $18 billion loan is already trading at stressed levels, making Project Jupiter a high-stakes test of the financial risks embedded in the race to build AI compute. Read the full article on the force majeure filing and what it signals for AI infrastructure: buff.ly/GW8RoUV #AI #Oracle #DataCenters #AIInfrastructure

IBM has integrated its mainframe-based digital asset platform with permissioned blockchain networks, including a SWIFT shared ledger launched with support from 40 financial institutions, using ISO 20022 messages so banks can process tokenized deposits through payment workflows that already exist on the mainframe. An on-premises edition is now in beta for firms in highly regulated environments that need digital sovereignty, with IBM Crypto Express HSMs handling security. Read the full article on how mainframes are becoming part of the digital asset equation: buff.ly/v3mbNlo #IBM #Mainframe #Blockchain #DigitalAssets

Tax compliance is one of those functions people outside the business tend to underestimate, until you consider what is being sold, where the buyer is, which exemptions apply and how obligations change as the business grows. Avalara processed more than 54 billion AvaTax API calls last year, and after a day at its CRUSH conference I came away thinking the company is putting that established position to work, developing more than 100 agents and a hub to coordinate them while keeping its deterministic tax engine responsible for the calculations. Read the full article on what Avalara is building and the lesson for other software companies: buff.ly/SSd2cJS #AI #AgenticAI #Avalara #TaxCompliance

Meta used its annual Connect conference to expand well beyond the smartphone, unveiling a keychain-sized AI device called Muse Charm, 100-gram VR glasses, and a camera-free pair of Ray-Ban Audio glasses, all tied to its new Muse personal AI agent. The agent now integrates with Walmart, Best Buy, Sephora and PayPal, with Meta planning to distribute Muse free and monetize it by taking a fee on transactions. Read more on Meta's push to build the next dominant consumer computing platform: buff.ly/CNBUs0O #Meta #AI #AIAgents #SmartGlasses

China has launched a review of Broadcom networking switches in state-backed data centers after finding they account for up to 90% of the equipment in some State-owned facilities, part of Beijing's "domestic chips for domestic use" push. The shift would open a multi-billion-dollar market for Huawei, H3C and Ruijie, though experts note Huawei's switches remain less energy-efficient, which could raise operating costs for data center operators. Read the full article on the review and what it means for the global AI hardware supply chain: buff.ly/POZwNUF #AI #Broadcom #China #Semiconductors

NVIDIA is trading near its lowest valuation in more than a decade despite surging financial performance, as market skepticism grows over the longevity of the AI boom. The chipmaker sits below 17 times forward earnings even as analysts project revenue and net income jumping roughly 90% and 99%, with gross margins set to dip below 72% as memory costs climb and hyperscalers build their own AI hardware. Read the full article on the disconnect between NVIDIA's earnings and its valuation: buff.ly/yr4ZmgO #NVIDIA #AI #Semiconductors #StockMarket

Nutanix has acquired French software developer Ryax Technologies to improve GPU utilization and simplify the use of agentic AI across hybrid cloud environments, folding its compute orchestration technology into future releases of Nutanix Kubernetes Platform and Nutanix Enterprise AI. This addresses an efficiency challenge: expensive GPUs are often underutilized even as businesses struggle to obtain sufficient computing capacity, and Ryax's fractional GPU support lets multiple AI applications share a single processor rather than each reserving an entire one. See what the acquisition means for agentic AI deployments: buff.ly/jgxkydT #Nutanix #GPU #AgenticAI #HybridCloud

The days of relying solely on static dashboards to monitor IT environments are starting to fade, and Riverbed is embedding AI agents that let IT teams launch natural language queries to spin up the views they need on demand. It is not clear yet how far this transforms IT management, since many organizations have already found themselves rehiring personnel to manage the hundreds of agents they deployed. Get the details on Riverbed's agentic observability push: buff.ly/vwts8Pb #Observability #AIAgents #ITOperations #AIOps

IT talent is still relatively constrained, so when one professional takes on a new role, it tends to create opportunities for others. This week's ten postings lean heavily on site reliability and infrastructure work, from Ironclad and Snowflake to Anduril and Shield AI, with pay ranging from $100,000 to $270,000. Check out the openings and the salary ranges: buff.ly/QroX8fJ #ITJobs #Jobs #TechJobs #Hiring

Virginia hosts the world's largest concentration of data centers, and Governor Abigail Spanberger just unveiled a framework to curb their environmental and economic costs, with new oversight of power generation, water use and noise pollution plus restrictions on the non-disclosure agreements that have kept the public in the dark. Executive actions can begin immediately, while limits on on-site natural gas generation, a 25-megawatt threshold for local approval and a temporary electricity consumption tax will need approval in the 2027 legislative session. Get the full breakdown of Virginia's policy shift and the pushback from industry groups: buff.ly/vyqEk8u #DataCenters #AI #Virginia #AIPolicy

Oregon's data centers already consume about 23% of the state's retail electricity, a share researchers expect to reach as high as 32% by 2030 as AI drives demand for computing infrastructure. The report also finds the industry's direct employment is modest, property tax exemptions outweigh what operating data centers paid, and gaps in water and community impact data make the full costs hard to judge. Get the full findings on Oregon's data center boom and what it means for the grid: buff.ly/8ZkdPvG #DataCenters #AI #Energy #Oregon

Enterprise IT funding this week went where the AI buildout actually hurts, into power, workflow reliability, physical security, energy-efficient silicon and the networking layers that keep accelerators from sitting idle. Crusoe's oversubscribed $3.9 billion Series F led the way alongside $550 million for Temporal, $270 million for Exein, over €200 million for EUCLYD and $205 million for Cornelis, with the common thread being systems design rather than another model. Read the full article on the five deals to watch and what they signal for the enterprise AI stack: buff.ly/8OS2RPD #AI #Cloud #Cybersecurity #Semiconductors

The House passed the Ratepayer Protection Act 417-3, and the Senate killed it a day later when Sen. Jon Husted sought unanimous consent and Sen. Martin Heinrich blocked the request, arguing the bill only asks states to consider standards rather than requiring large power users to pay for the grid upgrades they create. Heinrich's own GRID Savings Act, which would give FERC a larger role and require financial commitments before upgrades proceed, then ran into the same obstacle, leaving the question of who pays for data center infrastructure unresolved. Read the full article on the Senate clash and what it means for the AI boom's electricity bill: buff.ly/MVJQlJf #AI #DataCenters #Energy #AIPolicy

The DOE has launched a quantum computing competition with up to $215M in planned funding to advance fault-tolerant systems for scientific workloads. Read the full article: buff.ly/yLs30Xd #QuantumComputing #DOE #HPC #QuantumTechnology

Tuning quantum circuits gets dramatically more expensive as the subproblems grow, and a DQAOA-GPT framework from Oak Ridge, IonQ, Nvidia and the University of Tennessee held runtime nearly flat at about 28 seconds while the conventional method climbed from roughly 34 seconds to more than 11 minutes. The catch is that every circuit was simulated on a single Nvidia H200 GPU with no quantum processor involved, so the work is benchmark-scale validation rather than a demonstration of quantum advantage. Read the full article on how generative AI changes the cost curve for quantum optimization: buff.ly/Ist5coD #QuantumComputing #GenAI #IonQ #NVIDIA

The biggest thing standing between AI and more capacity isn't chips, it’s the grid, and Google, NVIDIA, and Emerald AI just launched a coalition arguing data centers should shift or cut power draw when the grid strains in exchange for faster connections. Eighteen launch partners signed on, from Anthropic to Constellation and National Grid, and the pitch landed the same week the House passed the Ratepayer Protection Act 417-3 to make data centers cover the added cost of the power they demand. Read the full article on the flexible load push and the energy fight in Washington: buff.ly/RqjayOh #AgenticAI #DataCenters #Energy #Google