Today's AI News — February 26, 2026

1. Nvidia Unveils Next-Gen AI System "Vera Rubin"

Santa Clara, California — Nvidia headquarters. On February 13, CEO Jensen Huang stood before a massive machine assembled from 1.3 million components. Named Vera Rubin, this next-generation AI supercomputer carries 72 Rubin GPUs and 36 Vera CPUs.

Performance has improved tenfold per watt over its predecessor, Grace Blackwell. Power consumption roughly doubled, yet processing efficiency rose by an overwhelming margin. It is also Nvidia's first system to adopt 100% liquid cooling, significantly cutting water usage in data centers.

More than 80 suppliers across 20 countries, including TSMC, contributed components. AWS, Google Cloud, Microsoft, and Oracle plan to offer Vera Rubin-based cloud services from the second half of 2026. Meta announced it will deploy the system in its own data centers by 2027.

Sources: CNBC, Nvidia Newsroom

2. OpenAI Launches GPT-5.3-Codex, Ushering in the Agent Era

OpenAI made two moves at once. The first is the release of GPT-5.3-Codex — a coding agent model that goes beyond writing and reviewing code to perform nearly every task a developer handles at a computer.

The second is the full rollout of ChatGPT Agent. Integrating the former Operator service, this agent autonomously handles everything from web browsing to executing real-world tasks. Booking flights and ordering groceries have moved from "showing you how" to "doing it for you."

In agent mode, endurance for repetitive, tedious work — like sifting through hundreds of emails to extract action items — has been dramatically enhanced. Meanwhile, older models such as GPT-4o, GPT-4.1, and o4-mini were retired from ChatGPT on February 13.

Sources: OpenAI, Releasebot

3. Privacy Authorities from 52 Countries Adopt Joint Declaration on AI Deepfakes

Deepfake images generated by services like xAI's Grok are spreading worldwide. Non-consensual sexual imagery and content targeting minors sit at the heart of the problem.

On February 23, the Global Privacy Assembly (GPA) adopted the "Joint Declaration on AI-Generated Content and Data Protection." Sixty-one supervisory authorities from 52 countries — including South Korea's Personal Information Protection Commission, along with counterparts from France, the UK, Singapore, Canada, and the EU — signed on.

The declaration calls for safeguards to prevent the creation of non-consensual sexual content, transparency about the scope of AI system use, swift reporting and takedown procedures, and enhanced protections for children and adolescents.

Sources: Korea Policy Briefing, AI Newspaper

4. South Korea Spotlighted as World's Most AI-Optimistic Nation

Politico, the U.S. political media outlet, described South Korea as "a nation in love with AI." The article spotlighted the tech optimism permeating Korean society.

The Ministry of Education announced the "2026 AI and Digital Intensive Program for Working Professionals" on February 20, followed by a plan for "University AI Core Curriculum Development Support" on February 24. AI education infrastructure is expanding rapidly, covering everyone from college students to working adults.

On February 25, the KAIST-hosted "AI Factory Connect Day 2026" took place in Seoul, sharing real-world cases of AI transformation (AX) in manufacturing. Starting February 28, the "Seoul AI Festival 2026" opens at Dongdaemun Design Plaza (DDP).

Sources: Newspim, Korea Policy Briefing, AI Newspaper

5. Apple Accelerates AI Wearable Strategy Centered on Visual Intelligence

CEO Tim Cook sent a signal. Apple's next big direction is AI wearable devices built around Visual Intelligence. According to Bloomberg, an AI-powered wearable strategy will take shape alongside a low-cost MacBook lineup set to be announced on March 4.

What Apple calls Visual Intelligence is the ability of cameras to understand and respond to what they see in real time. The company aims to move beyond text-based conversational AI toward a stage where AI interprets the physical world before your eyes.

A turning point is approaching as Apple's AI strategy — once confined to the iPhone and Mac — extends to devices worn on the wrist and face.

Source: Bloomberg

6. MIT Technology Review Names "Generative Coding" a Top 10 Technology for 2026

MIT Technology Review selected "Generative Coding" as one of its 10 breakthrough technologies for 2026. The evidence: executives at Microsoft and Google disclosed that AI now writes more than 30% of Microsoft's code and over 25% of Google's.

AI that generates code has already moved beyond the experimental stage. Workflows in which a developer states an intent and the AI writes, tests, and deploys the code are taking root in enterprise settings.

The trajectory of 2026 is clear. The center of gravity is shifting from the race to build ever-larger language models toward deploying smaller models in the right places and weaving AI seamlessly into human workflows.

Source: MIT Technology Review

7. NASA Perseverance Achieves Autonomous Mars Navigation via AI-Designed Routes

The Perseverance rover, rolling across the Martian surface, has made history. It successfully navigated along a route designed not by a human controller but by AI.

Signals from Earth take anywhere from 4 to 24 minutes to reach Mars. In an environment where real-time control is impossible, the AI analyzed the terrain, determined a safe path on its own, and moved the rover accordingly.

This marks a live deployment of AI-driven autonomous navigation in space exploration. If self-driving cars on Earth dodge obstacles on roads, the AI on Mars must find its way between boulders and sand dunes. The conditions are incomparably harsher.

Source: ScienceDaily

This newsletter is automatically generated and delivered every morning.
KJ Kim AI News Briefing

 

#ArtificialIntelligence #AI #Nvidia #VeraRubin #OpenAI #GPT5 #ChatGPT #CodingAgent #Deepfake #DataPrivacy #AIEducation #Apple #VisualIntelligence #GenerativeCoding #NASA #AutonomousDriving #AIBriefing #KimKyungJin #LawyerKimKJ #KimKJAI