Topic
Explore the latest artificial intelligence news and research.
Gartner forecasts global AI spending will reach $2.59 trillion in 2026. Drawing on primary sources, this article examines how those costs are flowing through to Microsoft 365 price increases, cloud usage-based billing, and device prices.
NVIDIA has reportedly been in discussions over guarantees of roughly $250 billion to back OpenAI's lease and construction financing for a 10GW Ohio site. We examine the possibility that GPU demand and credit risk are becoming concentrated in the company.
NVIDIA and more than 50 companies have launched an AI defense alliance to protect everything from model weights to identity, permissions, harnesses, and logs. The initiative aims to fill gaps in defensive capability exposed by the Hugging Face breach using existing open technologies.
A TrendForce survey finds that NVIDIA's CPO switches have entered limited shipment. While practical deployment of the 409.6Tbps class is progressing, optical engine yield and advanced packaging capacity will determine the pace of production expansion.
As the AI boom intensifies the scramble for mobile memory, Apple's bargaining power has weakened, pushing prices on flagship products like the Mac Studio up by as much as $1,300. This piece examines the mechanics behind the hike, including Micron CEO's remarks and the rise of CXMT, and what they reveal about shifts in the supply chain.
Claude Opus 5 blocked all 1,290 browser injection attacks when combined with Auto Mode. The model alone scored 3.70%, showing the significant effect of the product-level defenses including permission management and dual-layer inspection.
Micron and Meta used DCPerf to examine how LPDDR5X speed and capacity affect AI preprocessing, web response, and Spark SQL. Running out of capacity and spilling to SSD produced the largest performance gap.
Shopify has released a developer preview that moves theme page structure away from JSON splitting and back toward Liquid-centric design. The approach aims to let both AI and humans read the same code, editing through typed contracts and inspection.
A claim of a universal jailbreak spanning GPT-5.6 Sol, Claude Opus 5, and Fable has surfaced. We check the undisclosed method against each company's safety evaluations, examining the reproduction conditions needed to prove universality and the gaps that remain.
On July 23, 2026, the UK AI Security Institute and the US CAISI jointly evaluated Kimi K3's cyberattack capabilities. It scored only 32% on ExploitBench (compared to roughly 76% for the average US model), but asymmetric testing conditions and the timing of the announcement complicate how the results should be interpreted.
