RobinsonLabs homelab · ai · ops
NEWS WIRE · LIVE
// News · the wire

The wire, auto-curated daily.

updated June 28, 2026 · archive →
// still developing

Fable 5 and Mythos 5 remain offline under US export controls

No detail available.

// 1 source
Build Fast with AIbuildfastwithai.com ↗
// ai · tech

AI & Technology

3 stories
// ai · tech

Meta commits $14.3 billion for 49% stake in Scale AI as founder leads new superintelligence lab

A landmark investment secures high-quality training data and reshapes the AI data landscape.

2 sources
★ Latest
on the wire

Meta commits $14.3 billion for 49% stake in Scale AI as founder leads new superintelligence lab

A landmark investment secures high-quality training data and reshapes the AI data landscape.

Meta Platforms has finalized a $14.3 billion investment in Scale AI, acquiring a 49% non-voting stake in the data-labeling startup and valuing it at approximately $29 billion. The deal brings Scale AI founder Alexandr Wang, 28, to Meta to head a new superintelligence research lab, addressing the critical constraint of high-quality training data for large language models. The partnership structure keeps Scale AI independent to avoid antitrust scrutiny while granting Meta significant operational control. Competitors including Google, OpenAI, and xAI have reportedly paused or wound down their relationships with Scale AI following the announcement.

// 2 sources
Reutersreuters.com ↗
Forbesforbes.com ↗
// ai · tech
DeepMind launches AlphaGenome, a unified AI model that processes one million DNA base pairs to predict gene regulation
The breakthrough model outperforms specialized predecessors across 22 of 24 genomic benchmarks.
2 sources →

DeepMind launches AlphaGenome, a unified AI model that processes one million DNA base pairs to predict gene regulation

The breakthrough model outperforms specialized predecessors across 22 of 24 genomic benchmarks.

Google DeepMind has introduced AlphaGenome, a new AI model capable of analyzing up to one million DNA base pairs at single-nucleotide resolution to predict how genetic variations impact gene regulation. Available via API for non-commercial research, the model outperformed the best external models on 22 of 24 evaluations for single-sequence predictions and matched or exceeded top models on 24 of 26 variant-effect predictions. AlphaGenome unifies the prediction of diverse modalities including gene expression, splicing, and protein binding, eliminating the need for multiple specialized models and offering a foundation for future adaptation in genomics research.

// 2 sources
DeepMind Blogdeepmind.google ↗
Naturenature.com ↗
// ai · tech
Nvidia's Vera Rubin AI supercomputer racks now cost $7.8 million each, driven by a 485% surge in memory costs
Memory now comprises a quarter of the total system cost as demand for high-bandwidth memory spikes.
2 sources →

Nvidia's Vera Rubin AI supercomputer racks now cost $7.8 million each, driven by a 485% surge in memory costs

Memory now comprises a quarter of the total system cost as demand for high-bandwidth memory spikes.

Nvidia's next-generation Vera Rubin NVL72 rack-scale AI systems are estimated to cost $7.8 million each, a significant increase from the previous generation's $4 million. The primary driver is the cost of memory, which has risen by approximately 485% and now accounts for 25% of the total system cost, or about $2 million per rack. The system features a threefold increase in LPDDR5X capacity compared to the prior generation, alongside the introduction of significant 3D NAND storage. These rising hardware costs add financial pressure to data center projects as the industry continues to scale AI infrastructure.

// 2 sources
Tom's Hardwaretomshardware.com ↗
NVIDIA Newsnvidianews.nvidia.com ↗
// servicenow

ServiceNow & ITSM

3 stories
// servicenow
ServiceNow launches autonomous AI specialists across every major business function
The company is shifting from advisory AI to AI that senses, decides, and acts, debuting specialized agents for IT, CRM, Employee Services, and Security/Risk that handle end-to-end workflows without human hand-off.
2 sources →

ServiceNow launches autonomous AI specialists across every major business function

The company is shifting from advisory AI to AI that senses, decides, and acts, debuting specialized agents for IT, CRM, Employee Services, and Security/Risk that handle end-to-end workflows without human hand-off.

ServiceNow announced at Knowledge 2026 that its Autonomous Workforce now includes dedicated AI specialists for IT (AIOps, SRE), CRM (sales, quoting, orders), Employee Services (HR, Finance, Legal), and Security/Risk. These agents run on the unified ServiceNow platform, leveraging the Context Engine and Workflow Data Fabric to resolve millions of cases monthly—over 100 million customer cases and 40 million employee cases—while resolving 91% without reassignment. The AI Control Tower governs these agents, ensuring they operate within enterprise guardrails with full audit trails.

// 2 sources
ServiceNow Newsroomnewsroom.servicenow.com ↗
Fortunefortune.com ↗
// servicenow
AI Control Tower expands to govern all enterprise AI agents, including those from partners like Microsoft and NVIDIA
ServiceNow positions its Control Tower as the essential governance layer for the entire AI workforce, capable of discovering, observing, governing, securing, and measuring any agent across the enterprise.
2 sources →

AI Control Tower expands to govern all enterprise AI agents, including those from partners like Microsoft and NVIDIA

ServiceNow positions its Control Tower as the essential governance layer for the entire AI workforce, capable of discovering, observing, governing, securing, and measuring any agent across the enterprise.

Originally introduced as a visibility tool, the AI Control Tower has evolved into a comprehensive governance platform that can detect rogue agent behavior and automatically shut down agents acting outside permissions (the "Kill Switch"). It now integrates deeply with Microsoft’s Agent 365 ecosystem and NVIDIA’s agent frameworks, managing agents in Outlook, Word, and PowerPoint, while tracking usage and business impact at the token level. This expansion reinforces ServiceNow’s strategy as the "AI control tower for business reinvention," ensuring that as AI agents proliferate across functions and partners, they remain secure, governed, and valuable.

// 2 sources
ServiceNow Newsroomnewsroom.servicenow.com ↗
cxfoundationcxfoundation.com ↗
// servicenow
ServiceNow deepens strategic partnerships with Microsoft, NVIDIA, and Lenovo to embed AI workflows across the enterprise
New integrations enable AI agents to operate headlessly in tools like Teams and Zoom, while device intelligence from Lenovo drives proactive IT resolution.
2 sources →

ServiceNow deepens strategic partnerships with Microsoft, NVIDIA, and Lenovo to embed AI workflows across the enterprise

New integrations enable AI agents to operate headlessly in tools like Teams and Zoom, while device intelligence from Lenovo drives proactive IT resolution.

ServiceNow announced expanded collaborations with Microsoft, NVIDIA, and Lenovo to embed its AI workflows deeper into the enterprise. With Microsoft, ServiceNow specialists are available on the Agent 365 Marketplace, and governance extends across the ecosystem. With NVIDIA, the partnership delivers autonomous agents that navigate a desktop like a human, supported by the OpenShell runtime. With Lenovo, real-time device intelligence triggers automatic workflows for proactive IT issue resolution, cutting support costs by 30% and resolving 40% of issues before users notice them. These partnerships highlight ServiceNow’s role as the connecting layer for AI across diverse platforms.

// 2 sources
cxfoundationcxfoundation.com ↗
Fortunefortune.com ↗