AI News
Page 10 of 15 · 443 articles
-
HUMAIN makes first Saudi investment, backing enterprise AI firm MOZN
PIF-backed HUMAIN made its first investment in a Saudi company, backing enterprise AI firm MOZN to co-build production-scale AI for financial crime prevention and compliance.
-
Hassabis steps back as DeepMind CEO in major Google AI reshuffle
Demis Hassabis is ceding his DeepMind CEO role to become chairman and Alphabet chief scientist, while Jeff Dean and other senior researchers depart to launch a new company.
-
EU begins enforcing AI Act transparency rules for chatbots
The EU's AI Act transparency rules took effect August 2, requiring chatbots to disclose themselves and AI-generated content to carry clear labels, with fines up to 15 million euros.
-
MCP ships its largest spec update ever, moves to stateless core
The Model Context Protocol released its 2026-07-28 spec, the largest revision since launch, removing session state from the protocol layer so servers scale on ordinary HTTP infrastructure.
-
Abu Dhabi launches world's first integrated AI judicial platform
The Abu Dhabi Judicial Department will roll out an AI platform covering the full litigation workflow starting September, with judges retaining sole authority over final rulings.
-
FCC bans new Chinese humanoid and quadruped robot imports
The FCC barred imports of new Chinese-made humanoid and quadruped robots and connected power inverters, citing national security and supply-chain risk to the US AI buildout.
-
Claude Opus 5 takes FrontierBench lead over GPT-5.6 Sol
Anthropic's newly shipped Claude Opus 5 scored 43.3% on FrontierBench v0.1 at maximum effort, ahead of GPT-5.6 Sol's 37.5%, though results are benchmark-specific.
-
Nvidia and SK Group sign $500B-plus AI infrastructure deal
Nvidia and South Korea's SK Group signed letters of intent for a partnership exceeding $500 billion, spanning a 2-gigawatt AI factory and a long-term HBM4 memory supply deal.
-
Kimi K3 open weights go live, largest release ever
Moonshot AI released Kimi K3 full open weights, a 2.8-trillion-parameter model, a day ahead of schedule and the largest open-weight release to date.
-
Report: Saudi Arabia strengthens position among leading AI nations
A Microsoft-Accenture report highlighted Saudi AI leadership, citing Stanford ranking the Kingdom first globally in AI security and women empowerment in AI, plus a new Aramco-Pasqal quantum deal.
-
AMD unveils Instinct MI455X GPU, its answer to Nvidia Rubin
AMD unveiled Instinct MI455X with 432GB of HBM4 memory and up to 40.26 PFLOPS of compute, paired with the 72-GPU Helios rack, directly targeting Nvidia Vera Rubin.
-
Alphabet raises 2026 capex to $205B as Cloud revenue jumps 82%
Alphabet reported Q2 revenue of $119.8B, up 24% YoY, and raised 2026 capex guidance to $195-205B. Shares fell over 6% despite the beat, as free cash flow turned negative.
-
Bipartisan bill would force AI firms to keep a shutdown "kill switch"
Reps. Ted Lieu and Nathaniel Moran introduced the AI Kill Switch Act, requiring developers of powerful AI systems to retain the ability to throttle, suspend, or shut them down.
-
Microsoft and Mistral expand partnership on sovereign AI in Europe
Microsoft and Mistral announced an expanded partnership, with Microsoft funding Mistral European GPU infrastructure using thousands of Nvidia Vera Rubin chips.
-
Fireworks AI raises $1.5B Series D at $17.5B valuation
AI inference cloud Fireworks AI raised a $1.505B Series D at a $17.5B valuation, up 4.4x from its $4B Series C seven months earlier.
-
e& UAE and Core42 launch UAE's first sovereign GPU compute platform
e& UAE and Core42 launched Sovereign AI Compute, giving in-country GPU access with no upfront cost and zero egress fees.
-
Saudi Arabia's AI tool adoption hits 45% as new internet report shows rapid digital growth
Saudi Arabia's Communications, Space and Technology Commission (CST) released its Saudi Internet Report 2025, finding that 45.2% of Saudi internet users now use AI tools. The report also found average mobile data consumption of 53GB per person monthly — roughly three times the global average — and a median mobile download speed of 216Mbps, placing the Kingdom among the top five G20 countries for mobile internet performance.
-
DeepSeek is quietly building its own AI inference chip to cut reliance on Nvidia and Huawei
Chinese AI lab DeepSeek is developing its own custom chip designed for inference rather than training, according to Reuters sources, in an effort to reduce dependence on both Nvidia and Huawei processors. The roughly year-old project remains early-stage, with DeepSeek in talks with chip-design, foundry, and memory partners while facing US and Dutch export controls that block access to the most advanced overseas fabs.
-
TSMC posts record $40.2B quarter, raises AI chip outlook past 40% growth
TSMC reported record Q2 2026 revenue of $40.2 billion, up 36% year over year, with net profit surging 77.4% to a new single-quarter high. The company raised its full-year AI chip growth outlook above 40%, pledged an additional $100 billion to its Arizona fabs, and confirmed CoWoS advanced packaging capacity is sold out through year-end.
-
Google ships Gemini 3.6 Flash, cuts prices, and confirms Gemini 4 pretraining has begun
Google released three new Flash-tier models on July 21: Gemini 3.6 Flash, a cheaper and more token-efficient default workhorse; Gemini 3.5 Flash-Lite at $0.30/$2.50 per million tokens; and Gemini 3.5 Flash Cyber, a restricted vulnerability-research model for governments and trusted partners. Google also confirmed it has begun pretraining Gemini 4, while flagship Gemini 3.5 Pro remains stuck in partner testing, months past its promised June launch.
-
AI drug discovery investment surges past $2B as development timelines compress sharply
AI-driven drug discovery attracted more than $2 billion in recent investment, according to a new BCC Research analysis, as AI tools compress traditional preclinical discovery timelines of 4-5 years down to 12-18 months and pharmaceutical executives report AI is already reducing time to produce medicines for clinical testing by roughly 20-30%.
-
Saudi Arabia deepens AI and space cooperation with China in Beijing talks
Saudi Communications and IT Minister Abdullah Alswaha met with senior Chinese officials in Beijing, including China's Minister of Science and Technology Yin Hejun, to expand cooperation on artificial intelligence, the digital economy, technology transfer, and talent exchange programs. Alswaha was accompanied by CST Governor Haytham AlOhali and Saudi Ambassador to China Abdulrahman Alharbi.
-
Huawei debuts Atlas 950 SuperPoD, claims 6.7x Nvidia's compute using zero US components
Huawei publicly displayed its Atlas 950 SuperPoD for the first time at the World AI Conference in Shanghai, linking up to 8,192 domestically-made Ascend NPUs into a single computing node and claiming 6.7 times the computing power of Nvidia's next-generation NVL144 system. The most significant detail isn't the performance claim, which is unverified, but that the system contains zero US-origin components.
-
AI chip startup Etched in talks for $20B valuation, betting against general-purpose GPUs
AI chip startup Etched is negotiating two funding rounds simultaneously — one led by existing investor Jane Street at a $20 billion valuation, another led by Sequoia Capital at $10 billion — quadrupling its prior valuation as investors seek alternatives to Nvidia. Neither deal has closed. Etched builds Sohu, a chip designed exclusively for transformer-based inference rather than general-purpose computing.
-
SK Group chair warns of 60-100% jump in AI memory demand as capacity lags
Speaking at the Korea Chamber of Commerce and Industry's Jeju Forum on July 19, SK Group Chair Chey Tae-won said foreign governments now treat AI memory access as a matter of economic security, and warned that customers are asking for 60-100% more AI memory capacity in 2027 while, in his words, no company currently has meaningful new capacity coming online to meet it.
-
Egypt's Kuadra raises funding to launch the country's first unified AI platform for construction
Egyptian construction-tech startup Kuadra has closed a funding round led by Cairo-based Edafa Venture and is preparing to launch what it calls the first unified AI-powered platform managing the entire project lifecycle for contracting companies, automating tender analysis, work allocation, and bid evaluation.
-
Nonprofit Current AI reaches $400M in commitments to build a public alternative to Big Tech AI
Current AI, a nonprofit founded in 2025 by Martin Tisne and now led by CEO Ayah Bdeir, has secured roughly $400 million in commitments from the French government, the Ford Foundation, the MacArthur Foundation, DeepMind, and Salesforce to build open, public AI infrastructure for communities and languages underserved by commercial AI systems. Its first project, Suno Sutra, is an offline device running AI in 22 Indian languages.
-
Japan and Nvidia launch $6.2B national physical AI factory, aiming for a ‘third option’
Nvidia, Japanese startup Noetra, and Japan's Ministry of Economy, Trade and Industry launched what they call the world's first national AI infrastructure for physical AI — a $6.2 billion facility with 27,500 Nvidia Rubin GPUs and 13,750 Vera CPUs at 140 megawatts. Noetra CEO Hironobu Tamba called it a bid for "a genuine third option" beyond US and Chinese AI stacks.
-
EU forces Google to open Android and Search data to rival AI assistants
The European Commission issued two binding decisions under the Digital Markets Act on July 16, ordering Google to give rival AI assistants the same system-level access to 11 Android feature groups that Gemini enjoys, and to share anonymized Google Search data with competing search and AI providers on fair terms. Android interoperability is due by July 2027; search data sharing begins January 2027.
-
Kimi K3 adds Standard and High reasoning tiers, easing its max-only cost problem
Two days after launching Kimi K3, the largest open-weight model released to date, Moonshot AI has begun rolling out Standard and High reasoning-effort tiers alongside the existing Max-only mode. The update targets developers who saw per-task costs spike since K3 launched with only maximum-effort reasoning available.