#AI

985 articles

Peter Smith Peter Smith Industry
1 month ago

NVIDIA Blackwell Ultra Dominates AI Training Benchmarks

NVIDIA's Blackwell Ultra GPUs delivered record-breaking MLPerf Training v5.1 results, showing massive performance jumps across all benchmarks. This comes as potential tax policy changes threaten to increase costs for AI infrastructure operators.

Saad Ullah Saad Ullah Industry
1 month ago

J.P. Morgan Says AI-Semiconductor Boom Will Run Through 2027

An analysis of J.P. Morgan's latest semiconductor outlook showing that the AI-driven expansion is still in early stages and projected to continue through 2027, challenging fears of an impending bubble.

Saad Ullah Saad Ullah Industry
1 month ago

AI Update: GPT-5.1 Improves Instruction Accuracy

GPT-5.1 now follows instructions more accurately and respects custom formatting preferences, including the option to avoid em-dashes when specified.

Eseandre Mordi Eseandre Mordi Benchmarks
1 month ago

Desktop vs. Mobile: How People Actually Use GenAI Tools

A look at traffic data revealing the growing divide between mobile-friendly and desktop-focused AI platforms.

Eseandre Mordi Eseandre Mordi Markets
1 month ago

Anthropic Reveals $50B Strategy to Expand AI Infrastructure

Anthropic unveiled a $50 billion investment to build AI data centers in Texas and New York, marking its first major move into self-operated infrastructure amid fierce AI industry competition.

Usman Salis Usman Salis AI models
1 month ago

Google (GOOGL) Unveils JAX-Privacy 1.0 for Differentially Private ML

Google Research launched JAX-Privacy 1.0, a library designed to make differentially private machine learning easier on the high-performance JAX framework. The release aims to help researchers and developers combine cutting-edge privacy algorithms with JAX's scalability.

Peter Smith Peter Smith AI models
1 month ago

OpenAI GPT-5.1 Uses Fewer Tokens on Easy Tasks, More on Hard

A chart comparing GPT-5 and GPT-5.1 reveals the newer model generates far fewer tokens on easy prompts and significantly more on hard ones—showing –57% tokens at the 10th percentile and +71% at the 90th percentile.

Saad Ullah Saad Ullah AI models
1 month ago

DeepMind Aligns Vision Models to See More Like Humans

DeepMind's new results show that "aligned" vision models—trained to interpret images more like humans do—outperform standard models on both human-alignment tasks and generalization benchmarks.

Saad Ullah Saad Ullah Markets
1 month ago

AMD Forecasts $1 Trillion Data Center Market by 2030

AMD's recent Analyst Day presentation reveals ambitious projections for AI and data center growth, forecasting a $1 trillion market opportunity by decade's end.

Saad Ullah Saad Ullah Industry
1 month ago

Google Launches Private AI Compute for Secure Gemini Cloud

Google introduces Private AI Compute — a Gemini-powered framework that processes AI requests in the cloud while guaranteeing end-to-end privacy, making user data inaccessible even to Google itself.

Saad Ullah Saad Ullah Industry
1 month ago

ChatGPT Introduces Real-Time Group Chats

OpenAI is rolling out a Group Chat feature in ChatGPT that lets people work together in real time, share files, and create images directly inside the AI platform.

Eseandre Mordi Eseandre Mordi AI models
1 month ago

Global Data Center Demand to Triple by 2030

McKinsey forecasts that worldwide data center capacity will nearly triple by decade's end, with AI workloads driving 70% of the expansion as computing infrastructure enters a transformative era.

Usman Salis Usman Salis Industry
1 month ago

Cursor Launches New GPT-5 Codex Fast Models

Cursor has released two new GPT-5 Codex variants — FAST and High FAST — that code twice as fast but cost double, targeting developers who value speed over savings.

Peter Smith Peter Smith Industry
1 month ago

Tech Leaders See 2035 as AGI Turning Point

A decade-long timeline is emerging — but the agreement itself might say more about us than the tech.

Saad Ullah Saad Ullah AI models
1 month ago

Kimi-K2 Becomes China's Best Open-Source AI Model

Kimi-K2 Thinking scored 42.1% on the WeirdML benchmark, making it China's top open-source model and outperforming both Claude 4.1 Opus and Grok-4

Eseandre Mordi Eseandre Mordi Markets
1 month ago

AMD Targets Over $20 EPS in New Growth Model

AMD announced aggressive long-term targets including $20+ earnings per share and 35%+ annual revenue growth — significantly above Wall Street's bullish estimates.

Peter Smith Peter Smith Industry
1 month ago

Grok 4 Fast Challenges GPT-5 With Equal Power at 95% Lower Cost

A new AI model delivers GPT-5-level performance while slashing costs by up to 95% and running 150% faster, potentially reshaping the economics of enterprise AI deployment.

Saad Ullah Saad Ullah Industry
1 month ago

Anthropic Projects Profitability Ahead of OpenAI

How Claude's enterprise focus and lean infrastructure strategy could make Anthropic the first major AI lab to turn a profit

Eseandre Mordi Eseandre Mordi Industry
1 month ago

Anthropic Forecasts Lower AI Compute Costs Than OpenAI

A closer look at how Anthropic plans to compete with OpenAI through efficiency rather than scale

Saad Ullah Saad Ullah Industry
1 month ago

Anthropic Projects Lower Infrastructure Costs Than OpenAI

Anthropic believes it can run AI operations more cost-effectively than OpenAI. Internal projections show their server costs relative to revenue will drop significantly, potentially falling below OpenAI's as early as next year.

Posts feed
ALLEX Robotic Hand Achieves 100-Gram Force Detection in Precision Task Demo
4 hours ago
GPT-5.2 Pro Tops FrontierMath Leaderboard With 29.2% Score
5 hours ago
GLM-4.7 Hits 1224 ELO: New Top Open-Weights Model Takes Lead
9 hours ago
MiniMax M2.1 Coding Model Scores 88.6% on Benchmark, Beats Claude and Gemini
21 hours ago
AI Models Now Handle 5-Hour Software Tasks Autonomously
21 hours ago
Longer AI Reasoning Chains Prove Easier to Monitor, OpenAI Study Reveals
22 hours ago
AI-Designed Protein Shows 53% Stronger Anti-Inflammatory Effect
22 hours ago
VinMotion Launches Motion 2 Humanoid Robot With 40kg Load Capacity
1 day ago
Google's Control of 4-Layer AI Stack Creates Cost Edge
1 day ago
META: Meta Platforms Expands AI Stack With Manus Acquisition
1 day ago
Follow us on Twitter

Disclaimer: Aigazine provides content for informational and educational purposes only. We do not offer financial, investment, or legal advice. Any decisions made based on the information published on this website are solely at the reader’s own discretion and risk. We encourage all readers to conduct their own research and seek professional guidance when necessary.

Aigazine is a news platform focused on artificial intelligence, covering global AI trends, technology, and innovation. Aigazine is part of TheTradable.com project and is based in Tbilisi (0179, Georgia, Tbilisi City, Vake District, 49 Besarion Zhghenti Street, VAT 305786600).

© 2025 aigazine.com

We use cookies to improve your experience on our site and to show you relevant advertising. To find our more, read our privacy policy and cookie policy