June 22, 2026

AI News (06/22) : AI threats loom as drug discovery and infrastructure deals reshape tech landscape.

The AI world is a whirlwind of innovation and concern today. Intelligence agencies warn of devastating AI-powered cyber attacks just months away, while the race for compute power heats up with multi-billion dollar deals. From breakthroughs in healthcare diagnostics to massive infrastructure projects designed to fuel the AI boom, the technology continues its rapid, multifaceted expansion, forcing both caution and acceleration across industries.

AI Models Capable of Devastating Attacks Months Away, Five Eyes Warns

A rare joint statement from the Five Eyes intelligence alliance (Australia, US, UK, New Zealand, Canada) has sounded a stark alarm: powerful AI models capable of new, devastating cyber attacks are "months, not years" away. The warning follows the Trump administration's decision to block foreign nationals from using Anthropic's advanced Fable AI model, a move reflecting heightened national security concerns. The agencies emphasize AI's dual nature, capable of both bolstering and fundamentally transforming cyber threats, lowering barriers for bad actors and increasing attack sophistication.

Why it matters: This unprecedented public intervention highlights the urgent need for a "whole-of-society response" to escalating AI-driven cyber risks, framing it as a core business and leadership responsibility rather than a purely technical one.

Insilico Medicine, SK Biopharmaceuticals Strike $2.5B+ AI Drug Discovery Deal

Hong Kong-listed Insilico Medicine has sealed a partnership with Korea's SK Biopharmaceuticals, valued at over $2.5 billion, to develop neuroimmune therapies. Insilico will leverage its Pharma.AI platform for drug candidate design, while SK Biopharmaceuticals handles late-stage development and commercialization. This marks Insilico’s largest Asia-Pacific tie-up and its second multi-billion dollar deal, following a $2.75 billion agreement with Eli Lilly in March.

Why it matters: The deal underscores the accelerating trend of AI's integration into pharmaceutical R&D, promising to revolutionize drug discovery speed and efficiency, particularly in complex areas like neuroimmunology.

SpaceX Signs $6.3 Billion Computing Power Deal with Open-Source AI Startup Reflection

SpaceX has inked a substantial computing power agreement with Reflection AI, an open-source AI startup. The deal grants Reflection immediate access to Nvidia GB300 chips housed in SpaceX's Colossus infrastructure, with payments totaling approximately $6.3 billion through 2029. This move expands SpaceX’s compute-selling ventures, positioning it as a key provider alongside cloud giants, and notably supports a company dedicated to open-source AI models amidst growing calls for transparency and control.

Why it matters: This massive deal signifies compute power as a strategic currency in the AI race, and SpaceX's engagement with an open-source player reflects a critical shift towards decentralized AI access and development, especially after recent restrictions on closed models.

NVIDIA Unveils Halos for Robotics: Full-Stack Safety System for Physical AI

NVIDIA has introduced Halos for Robotics, the industry's first full-stack, comprehensive safety system for robotics and physical AI. Leveraging 18,600+ engineering years of autonomous vehicle safety development, Halos unifies AI compute and safety across hardware, software, and sensor data. Humanoid robotics leader Agility is the first to integrate Halos into its Digit robots, designed for factories, warehouses, and logistics.

Why it matters: As AI-enabled robots increasingly operate alongside humans, standardized, certifiable safety architectures like Halos are crucial for scaling autonomous systems into dynamic industrial environments with confidence and reliability.

Chevron Plans Massive Texas AI Power Plant for Microsoft

Chevron announced plans to develop a 2.67 gigawatt power plant in Reeves County, West Texas, specifically to fuel a new Microsoft data center. The facility will be co-located with the data center, allowing Chevron to bypass traditional grid connections and directly supply electricity generated onsite using Permian Basin natural gas. This project is Chevron’s first power plant for AI, leveraging its energy expertise for the surging power demands of data centers.

Why it matters: This venture highlights the energy sector's direct engagement with the AI boom, showing how traditional oil and gas companies are pivoting to meet the massive, localized power demands of AI infrastructure, while also addressing environmental concerns through non-potable water use.

Microsoft & Chevron Formalize 20-Year Gas Deal for AI Data Centers

Following the announcement of Chevron's West Texas AI power plant, Microsoft has formalized a 20-year gas deal with the energy giant. This long-term agreement secures the natural gas supply critical for powering Microsoft's new data center in Reeves County. The collaboration ensures a stable, dedicated energy source for the compute-intensive operations of AI, bypassing reliance on the conventional energy grid and addressing the escalating energy needs of large-scale AI facilities.

Why it matters: A 20-year commitment for gas supply underscores the immense and long-term energy requirements of AI data centers, forging new direct utility-like partnerships between tech and energy sectors.

Insilico's Pharma.AI Platform Secures Major Korean Biotech Partnership

Insilico Medicine's Pharma.AI platform continues to attract major biopharma partners, most recently securing a global R&D collaboration with Korea’s SK Biopharmaceuticals, valued at over $2.5 billion. The deal enables SK Bio to leverage Pharma.AI’s generative AI and automation programs for target validation, molecule optimization, and candidate optimization in neuroimmune disorders. This latest partnership builds on Insilico's track record, including a significant deal with Eli Lilly and collaborations with Qilu Pharmaceutical and Servier.

Why it matters: The repeated high-value deals for Insilico's Pharma.AI platform demonstrate the industry's growing confidence in generative AI as a transformative force, capable of accelerating and improving the quality of the entire drug development pipeline.

AI Tool Reveals Hidden Organ Damage from High Blood Pressure

Researchers at the University of Oxford have developed an AI tool called "HyperScore" that uses machine learning to analyze imaging and clinical data, revealing hidden organ damage caused by high blood pressure. Studying over 27,000 individuals, the AI identified six distinct "HyperTrajectories" of hypertension-related disease, showing how damage manifests differently in various organs. This tool could enable more personalized treatment and earlier identification of risks for stroke, heart failure, or kidney disease.

Why it matters: Moving beyond simple blood pressure readings, this AI innovation offers a personalized, holistic view of hypertension's impact, paving the way for targeted interventions and better patient outcomes by understanding individual disease progression.

AI Helps Diagnose Rare Childhood Illnesses After Years of Uncertainty

OpenAI’s o3 model has proven to be a "game changer" at Boston Children’s Hospital, helping diagnose 18 children whose rare illnesses had stumped doctors for years. The AI analyzed genomes and clinical data, identifying new diagnoses, including rare neurodevelopmental and neuromuscular disorders. One patient, Kyra Benton, received a diagnosis of myofibrillar myopathy just before her 20th birthday after struggling with unexplained symptoms for years. Human experts then reviewed the AI's findings for final confirmation.

Why it matters: This application showcases AI's profound potential to cut through diagnostic complexity in rare diseases, offering hope and answers to families by accelerating the identification of genetic links and enabling timely treatment.

NVIDIA Unveils Vera Rubin Platform for AI & HPC Infrastructure

NVIDIA has launched the Vera Rubin platform, a rack-scale supercomputer combining Vera CPUs, Rubin GPUs, advanced networking, and a full software stack. Announced at ISC High Performance 2026, the platform is designed for scientific research and large data center operators, offering over seven exaflops of AI performance and five petaflops of native double-precision computing. It promises up to four times performance boosts for memory-bound fluid dynamics applications and aims to meet the escalating demands of "agentic AI."

Why it matters: The Vera Rubin platform represents a significant leap in the convergence of AI and high-performance computing, providing unprecedented power for scientific discovery and signaling NVIDIA's continued dominance in delivering integrated solutions for frontier AI and research.

To stay up to date on everything going on in AI, check out the tracker at the500feed.com

Sources

View the live feed — updated daily from 500+ sources →← All posts