Discover the latest breakthroughs in artificial intelligence, machine learning, and emerging AI technologies shaping our future.
A newly detailed flaw in ServiceNow’s Now Assist AI platform, dubbed ‘BodySnatcher’ (CVE-2025-12420), shows how weak authentication and over-privileged AI agents could have let attackers impersonate any user — including admins — across thousands of enterprises.
Caterpillar has expanded its partnership with Nvidia and unveiled the Cat AI Assistant at CES 2026, a multimodal AI copilot designed to guide operators, technicians, and fleet managers across millions of connected machines.
Taiwan’s TSMC has quietly confirmed it has begun volume production of its 2nm (N2) process, promising double‑digit performance and efficiency gains that will underpin the next wave of AI and mobile processors.
OpenAI has rolled out GPT-5.2, its most capable ChatGPT model to date, with sharper reasoning, fewer hallucinations, and new agent-focused variants designed to compete directly with Google’s Gemini 3.
Micron Technology plans to invest 1.5 trillion yen ($9.6 billion) in a new high-bandwidth memory plant in Hiroshima, deepening Japan’s bid to become a key hub for AI-era semiconductors.
Viral posts accusing Google of auto-opting Gmail users into training its Gemini AI model have sparked a privacy backlash, prompting official denials and renewed scrutiny of how ‘smart features’ and AI access user data.
Microsoft signed a five-year, $9.7 billion agreement with data-center operator IREN to secure Nvidia GB300 GPUs, adding up to 200 MW of liquid‑cooled AI capacity at IREN’s Texas campus by 2026.
Apple posted $102.5 billion in revenue and $1.85 EPS for fiscal Q4 2025, with iPhone and Services leading growth. Tim Cook forecast 10–12% year-over-year revenue growth for the December quarter.
Google and Reliance Jio are giving eligible Jio users no-cost access to Gemini 2.5 Pro and 2 TB of storage for 18 months, a ₹35,100 value—escalating the race to lock in everyday AI usage at massive scale.
Announced at GTC Washington, D.C., NVIDIA and Oracle will deliver the DOE’s ‘Solstice’ system with 100,000 Blackwell GPUs—plus a 10,000‑GPU ‘Equinox’—for a combined 2,200 exaflops of AI performance at Argonne National Laboratory.
Following its Washington, D.C. GTC keynote, Nvidia briefly crossed a $5 trillion market cap on October 29, driven by CEO Jensen Huang’s disclosure of $500 billion in AI chip orders and plans for seven new U.S. government AI supercomputers.
Qualcomm unveiled AI200 and AI250 rack-scale inference systems, marking a major push into data-center AI. Saudi-based HUMAIN plans a 200MW rollout starting in 2026.
SoftBank has approved the remaining $22.5 billion of its planned $30 billion investment in OpenAI, pending the AI firm’s transition to a for‑profit structure—an agreement tied to a broader $41 billion round announced in April.
OpenAI introduced ChatGPT Pulse, a preview feature that delivers once‑a‑day, personalized update cards generated from overnight research using a user’s chats, memory, and optional app connectors. The rollout starts on iOS and Android for Pro subscribers, with Plus support to follow.
CoreWeave signed a fresh $6.5 billion deal with OpenAI, lifting their 2025 total to $22.4 billion and underscoring a rapid buildout of AI infrastructure targeting 10 GW of capacity.
Microsoft now lets Microsoft 365 Copilot customers opt for Anthropic’s Claude Sonnet 4 and Opus 4.1 alongside OpenAI models, starting with the Researcher agent and Copilot Studio, with opt-in rollout beginning September 24.
The Stargate consortium announced five new U.S. AI data center sites—advancing toward a $500B, 10 GW target and bringing planned capacity to nearly 7 GW over the next three years.
Nvidia will invest up to $100 billion in OpenAI and partner to deploy at least 10 gigawatts of AI infrastructure, with the first 1GW slated for the second half of 2026.
Apple’s iPhone 17 lineup — including the ultra‑thin iPhone Air — went on sale September 19, boosting AAPL shares as Apple reportedly asks suppliers to increase production of the base model.
Nvidia will buy a $5 billion stake in Intel and jointly build custom x86 CPUs for AI systems and new PC SoCs that integrate Nvidia RTX GPU chiplets, sending Intel shares up as much as 30%.
OpenAI and Apollo Research unveiled new evaluations showing deceptive 'scheming' behaviors in frontier models and an early countermeasure, 'deliberative alignment,' that cut covert actions by ~30× in controlled tests.
Google and PayPal announced a multiyear strategic partnership to embed PayPal payments across Google products and co-develop AI-driven shopping and agentic commerce standards.
Oracle launched Java 25 (JDK 25) on September 16, adding 18 JEPs, long‑term support for at least eight years, and upgrades aimed at AI workloads and enterprise security.
Apple introduced the iPhone Air as its thinnest iPhone to date, pairing a 5.6mm titanium build with a 6.5-inch 120Hz display, A19 Pro silicon, and eSIM‑only connectivity starting at $999.
Apple has officially rolled out iOS 26 and iPadOS 26, introducing a stunning new design language called 'Liquid Glass' and a suite of powerful AI features under the 'Apple Intelligence' umbrella.
The popular note-taking app Goodnotes has officially launched on Android, bringing its acclaimed digital ink experience to a new ecosystem. However, early adopters find a promising but incomplete version compared to its iOS counterpart.
Google's DeepMind and Isomorphic Labs are leveraging advanced AI to dramatically accelerate pharmaceutical research, promising to bring new treatments to patients years faster than traditional methods.
Cornami and DESILO have deployed a large language model using Fully Homomorphic Encryption (FHE), enabling AI to analyze sensitive data without ever decrypting it, a landmark achievement for data security.
DJI's new Mini 5 Pro packs a groundbreaking 1-inch CMOS sensor and advanced LiDAR technology into its signature sub-250g frame, setting a new benchmark for professional-quality aerial photography in a compact form.
Researchers have developed a revolutionary photonic chip that uses light instead of electricity for core AI calculations, promising up to 100 times greater energy efficiency and paving the way for more sustainable AI.
A new generation of cybersecurity solutions is leveraging Artificial Intelligence and dedicated hardware to proactively detect and neutralize cyber threats before they can cause damage, marking a major shift from reactive to predictive security.
A breakthrough in FPGA architecture by Cornell University promises to make AI systems more powerful and sustainable by significantly reducing energy consumption and boosting performance.
Engineers have developed a revolutionary computer chip that uses light instead of electricity for AI computations, achieving up to 100 times greater energy efficiency.
OpenAI has officially launched GPT-5, a significant leap forward in AI that operates as a unified, intelligent system with groundbreaking performance in complex problem-solving, coding, and multimodal understanding.