Page 3 of 4 · 90 stories in this section
OpenAI models are now available on Amazon Bedrock, ending Microsoft exclusivity. AWS also announced a $50 billion investment deal and a cloud agreement worth over $100 billion over eight years.
Roblox slashed its 2026 bookings forecast to $7.33B-$7.6B from $8.28B-$8.55B after age-verification and safety features reduced user engagement and new sign-ups, sending shares down 18%.
Alphabet shares jumped 10% after Google Cloud revenue surged 63% to $20.03B, while Meta fell 8% on debt-funded capex concerns. The Magnificent Seven's Q1 earnings growth hit 57%, far outpacing the S&P 500 ex-tech.
SoftBank is creating Roze, an AI and robotics firm focused on automating data center construction, with plans for a U.S. IPO as early as 2026 at a $100 billion valuation.
Google Cloud revenue jumped 63% to $20.03B, beating estimates, while AWS grew 28% to $37.6B. AI demand drove growth, with hyperscaler capex projected to reach $725B in 2026.
Microsoft's AI business hits a $37B annual run rate as Copilot crosses 20 million paid enterprise seats, agent mode becomes default in Office apps, and Accenture commits 740,000 seats in the largest deal to date.
Meta reported Q1 revenue of $56.3 billion, beating estimates, but daily active people fell to 3.56 billion, missing expectations. The company raised its 2026 capex forecast to $145 billion, sending shares down 7%.
OpenAI is partnering with Qualcomm and MediaTek to develop a smartphone that relies on AI agents instead of apps, with mass production expected in 2028, according to analyst Ming-Chi Kuo.
Intel reported Q1 2026 revenue of $13.6B, crushing expectations by over $1B. Data center and AI revenue rose 22%, the foundry unit grew 16%, and the stock surged 16% after hours.
At Google Cloud Next, Alphabet unveiled TPU 8t for training and TPU 8i for inference — the first time Google has shipped purpose-built dies for each workload, claiming 80% better inference economics and a supply chain spanning Broadcom, MediaTek, and TSMC's 2nm node.
Tesla's $25B capex plan funds simultaneous launches of its Cybercab robotaxi, Optimus humanoid robot factory, and an AI compute doubling while FSD subscribers hit 1.28 million.
Apple unified its hardware engineering and silicon divisions under chip architect Johny Srouji, splitting the operation into five named groups days before John Ternus prepares to become CEO September 1.
Reuters says Nvidia wants full upfront payment for China H200 orders, shifting AI chip demand into a test of export controls, buyer liquidity and supply-chain leverage.
Nvidia's $5.5 billion H20 charge is more than an inventory write-down. It marks a tougher U.S. export regime moving from performance thresholds to direct product-by-product enforcement in China.
ASML's weaker-than-expected bookings suggest AI spending is not yet strong enough to lift every corner of the semiconductor market.
Nvidia’s NVLink Fusion turns a proprietary interconnect into a platform for third-party CPUs and accelerators, deepening its hold over how AI datacentres are designed.
CoreWeave's IPO filing shows AI infrastructure demand is real, but so are the losses, leverage and customer concentration behind the boom.
Both edge platforms made WebAssembly components their default deployment target in Q1 2025, abandoning container-based serverless patterns and claiming 3-5x faster cold starts for enterprise workloads.
TSMC lifted its 2026 outlook and spending plans after a sharp first-quarter jump, offering one of the clearest signals yet that AI demand is flowing into foundry revenue and factory expansion.
TSMC posted $1.8 billion in Q1 2026 net profit, up 58% year-over-year, as AI accelerator orders hit 35% of total revenue. The chipmaker raised full-year guidance to 25-29% growth.
The Bytecode Alliance ratified WIT IDL in September 2025, enabling Rust, Go, C++, and Python components to interoperate without JavaScript host shims for the first time.
Data centers retire AI accelerators every 3-5 years, generating 4 million metric tons of e-waste annually by 2026 — a hidden environmental liability the industry has largely avoided addressing.
Frontier AI reasoning providers have cut per-token costs by an order of magnitude in 18 months, making AI-augmented workflows economically viable for mainstream enterprises and threatening seat-based licensing models that have sustained enterprise software for decades.
Adobe unveils Firefly AI Assistant for cross-app creative workflows. Conversational agent orchestrates multi-step processes across Photoshop, Premiere, Illustrator, Lightroom, Express. Shifts from individual AI features to unified agent carrying context across Creative Cloud ecosystem.