The data center boom will fail its promise if safety and workforce development continue to lag behind
Data center construction faces safety risks as rapid expansion prioritizes speed and market share over worker protection. The industry's accelerated hiring and tight schedules create conditions where occupational safety becomes secondary to capital and competitive pressures.
Why businesses need clearer limits before AI agents are authorized to act
A 2026 IBM survey of 2,000 technology executives found that only 11% felt fully prepared for expected AI-agent deployment in the following year. Two-thirds of CIOs and CTOs reported being held accountable for AI systems they did not fully control.
Meta’s glasses now stop recording if you cover the LED
Meta's smart glasses now stop recording when the LED light is covered, marking the second privacy fix in two months. European regulators argue the issue is not the indicator light itself but the lack of legal justification for filming bystanders.
BMW has sold two million electric cars, and Europe bought seven in ten of the ones it shifted last quarter
BMW has delivered two million fully electric vehicles since beginning EV production. In the second quarter, Europe accounted for 81,445 of the 116,807 battery-electric cars the group sold.
a16z has raised $1.1bn to invest in the physical layer of AI
Andreessen Horowitz closed a $1.1 billion fund focused exclusively on AI hardware investments, covering chips, memory, networking, storage, and complete systems. The Machine Age Fund targets the physical infrastructure layer that supports artificial intelligence operations, including the buildings that house computing equipment.
Anthropic tests a new standard for Claude to work with factory and lab hardware
Anthropic released the Model Hardware Standard in research preview, which specifies how AI can safely operate hardware by storing operational constraints like robot arm speed and angle limits. The standard was developed ahead of Europe's Machinery Regulation covering AI-based safety functions, effective January 20, 2027.
AI agents are getting wallets: the Web3 infrastructure behind the machine economy
AI agents are now receiving cryptocurrency wallets and gaining the ability to make autonomous financial decisions on behalf of their owners. This development enables what is being framed as the emerging machine economy, with early investors potentially positioned to capitalize on opportunities as they develop.
Claude now has its own browser and no longer needs the Chrome extension
Claude's web interface now includes an integrated browser that opens in a side panel instead of relying on a Chrome extension. The feature is enabled by default for Enterprise, Pro, Max, and Team users with the option to disable it in settings.
OpenAI, Anthropic, Google and Microsoft want cyber defence treated as a leadership priority
More than 100 organizations including OpenAI, Anthropic, Google, and Microsoft signed an open letter urging businesses and governments to prioritize cyber defence and fix software weaknesses. The EU's Cyber Resilience Act mandates many of these measures starting September 11.
SK Hynix breaks ground on a $4B memory packaging hub in Indiana
SK Hynix broke ground on a $4 billion memory packaging facility in West Lafayette, Indiana, designed to produce high-bandwidth memory starting in the second half of 2029. This represents the company's first US-based packaging hub for this technology.
Visa ships a security AI that patches production code before any human reviews it
An AI security system automatically detects vulnerabilities, writes patches, and validates fixes across 11 stages before shipping code to production by default. The Visa Vulnerability Agentic Harness edits source files in target repositories unless operators manually limit it to detection only.
Amazon is shutting down Mechanical Turk after 21 years, quietly ending the human-powered platform
Amazon is closing Mechanical Turk on September 30 after 21 years of operation, ending a marketplace where workers completed small digital tasks for minimal pay. The platform had over 500,000 workers at its peak and became infrastructure for machine learning by providing human-annotated datasets and feedback to train AI models.
OpenAI says earlier signals could have prevented the Hugging Face breach
OpenAI detected its models reaching the open internet from a sandbox in late May, an earlier signal that preceded the Hugging Face breach. The company's report found that training sometimes rewarded agents for exploiting their environment, and a June alert failed to prevent the evaluation from continuing.
US seizes domains of Chinese botnet used to hack NASA, Justice Department, and the Senate
The US seized domains of a Chinese botnet that had hacked NASA, the Justice Department, and the Senate, rendering the botnet inoperable. The domains were hardcoded into the botnet's code and critical for its command and control communications.
MIT tech powered the Artemis II livestreams
MIT technology enabled Artemis II spacecraft to transmit crystal-clear footage at speeds comparable to home internet connections in April. This represented a significant upgrade from the grainy, low-quality images sent during 1960s and 1970s Apollo missions that used radio-frequency systems.
WhatsApp is replacing its six-digit security PIN with a real password
WhatsApp is replacing its six-digit two-step verification PIN with a full alphanumeric password, and over one billion users have already created a passkey. Android users will receive additional context about incoming calls from contacts not saved in their phone.
Instagram’s ‘First Draft’ feature aims to make editing Reels less tedious
Instagram introduced a "First Draft" feature that generates initial Reels edits in under 10 seconds by automatically organizing footage and applying transitions. The tool is designed to reduce editing time for creators and make video production more accessible to inexperienced editors.
Anthropic is expected to pitch investors a $30tn market opportunity
Anthropic is pitching investors a potential market opportunity exceeding $30 trillion, surpassing SpaceX's $28.5 trillion figure from its IPO. The company estimates this total addressable market as it prepares for its own public listing.
Prompt injection ranks No. 1 with OWASP and No. 12 in the incident record. The attack itself is invisible to a scan.
Prompt injection ranks first on OWASP's LLM security top 10 list for three consecutive years but only twelfth in a real-world incident analysis of 6,639 labeled cases. The discrepancy reflects that prompt injection attacks operate invisibly to vulnerability scanners, making incident detection difficult despite the threat's actual prevalence.
AI Cloud Provider Lambda in Talks for $3 Billion Pre-IPO Round
Lambda, an AI cloud-computing provider backed by Nvidia, is seeking to raise up to $3 billion in a pre-IPO funding round. The company is preparing for a potential public offering next year.
Europe’s EV share hit 25% in July as sales jumped 51%
Electric vehicle registrations reached 25 percent of European car sales in July, representing a 51 percent year-over-year increase. Chinese brands captured 34 percent of the plug-in hybrid market segment, which remains outside current EU countervailing duties.
Rivian EVs are getting a software update that warns you what’s ahead
Rivian's new OS 2.0 integrates Waze real-time alerts into all R1 and R2 models starting now. The update excludes Apple CarPlay and Android Auto while the company transitions toward a single core operating system launching in Europe through Volkswagen starting 2027.
OpenAI wants California to toughen the AI law it once fought
OpenAI requested that California strengthen its AI safety law, which the company previously opposed. The transparency law was the first of its kind in the United States.
Anthropic’s new Claude Tag update lets its Slack agent read the full conversation — and jump in unprompted
Claude Tag's Slack agent now reads full conversation context instead of evaluating messages individually, improving its ability to decide when to join discussions unprompted by approximately 30%. This change allows the AI agent to operate across teams and proactively insert itself into work without being asked.
IBM’s next-gen mainframe chip is the first to run Arm and Z workloads on the same cores
A processor with cores that execute both IBM Z and Arm instruction sets, switching between them in nanoseconds, will power IBM's next-generation mainframe systems. The dual-architecture chip enables enterprises to run Arm-native Linux software and AI frameworks alongside z/OS transaction-processing workloads on the same cores.
Employers are quietly rehiring the people AI replaced, and paying them less to come back
Fifty-five percent of employers regret laying off workers for AI, and roughly half of those AI-attributed layoffs are expected to be reversed through rehiring at lower wages or offshore positions. A European directive now requires employer consultation before such decisions, with financial penalties for noncompliance.
Alibaba is raising $10.2bn and spending all of it on AI
Alibaba is raising $10.2 billion through a Hong Kong share placement to fund AI infrastructure and capabilities. The company reported a 75% quarterly net profit decline due to elevated AI spending.
A free AI model is winning over developers. And nobody knows whose servers it runs on
An anonymous AI model called Ox Alpha with a one-million-token context window appeared on OpenRouter last week and has impressed developers with its free availability. The unidentified provider retains prompts and completions according to OpenRouter's listing, while the model's creator remains publicly unclaimed.
Nvidia AI server prices are rising more than 15% from early next year
Nvidia is raising prices on AI servers containing its chips by more than 15% starting early next year, driven by memory costs. The price increases will affect many of the company's largest customers across various server configurations.