AI news on Monday, June 15, 2026

SIGNIFICANCE

★ Top story · Industry2d ago

Sarvam raises $300M Series B

Bengaluru-based Sarvam AI has secured $300 million in a Series B funding round, pushing the company to unicorn status. The investment includes a $150 million contribution from IT services firm HCLTech. This capital infusion highlights growing corporate interest in specialized Indian artificial intelligence development, providing the startup with the resources to expand its domestic enterprise solutions. The partnership with HCLTech marks a significant integration between India’s established IT infrastructure and emerging generative AI providers.

SSarvam AI TTechCrunch AI ZZamin.uz Read original ↗

Monday — June 15, 2026

Sarvam raises $300M Series B

From Chatbots to Collaborators: AI’s Next Era

Salesforce to Buy AI Customer Service Firm Fin for $3.6 Billion

Nvidia Set to Sell $25 Billion of Bonds, Joining AI Debt Boom

We’re strengthening our presence in Alabama through new investments and community support.

Adani, Jabil Alliance to Build AI Data Center Gear in India

Tencent-Backed Enflame Heads to IPO as China AI Chip Wave Grows

Pokémon Go data helped train AI now linked to military drones

Microsoft CEO Satya Nadella warns of "a small number of AI systems capturing all the economic returns"

The Download: the first brain implant power user and South Korea’s AI obsession

The US government may be asking Anthropic the impossible by demanding unhackable LLMs

Anthropic Set to Meet Trump Officials Over AI Security Concerns

Anthropic Export Ban Deepens Fears About US Stranglehold on AI

Anthropic shutdown sparks sovereignty debate across Europe

Anthropic Block Marks US Reversal, Warning to Silicon Valley

Meta’s new ‘AI Mode’ on Facebook pulls from public info across its platforms

STaR-DRO: Stateful Tsallis Reweighting for Group-Robust Structured Prediction

Stream3D: Sequential Multi-View 3D Generation via Evidential Memory

Did You Forget What I Asked? Prospective Memory Failures in Large Language Models

LiAuto-GeoX: Efficient Grounded Driving Transformer

Elastic Queries Reinforcement Learning: Self-Aware Policy Execution for VLA Models

Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal

LEPO: Latent Reasoning Policy Optimization for Large Language Models

Regional Climate Model Emulation with Diffusion Approaches: What is the Added Value of Generative Machine Learning?

Silent Failures in Federated Personalization of Foundation Models

Optimizing the Cost-Quality Tradeoff of Agentic Theorem Provers in Lean

Efficient On-Device Diffusion LLM Inference with Mobile NPU

EmoMind: Decoding Affective Captions from Human Brain fMRI

Exact Linear Attention

ClaimFlow: Tracing the Evolution of Scientific Claims in NLP

AcceRL: A Distributed Asynchronous Reinforcement Learning and World Model Framework for Vision-Language-Action Models

Squeeze-Release: Iterative Pruning with Exact Structural Minimization

Orchestra-o1: Omnimodal Agent Orchestration

A Virtuous AI is an Existential Risk

Curvature-Guided Geometric Representation for Protein-Ligand Binding Affinity Prediction

Robustness without Wrinkles: Parallel Simulation and Robust MPC for Certified Deformable Manipulation

Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments

Hidden in Plain Sight: Benchmarking Agent Safety Against Decomposition Attacks with DECOMPBENCH

Pano3D: Unified 3D Reconstruction and Panoptic Segmentation

One Layer's Trash is Another Layer's Treasure: Adaptive Layer-wise Visual Token Selection in LVLMs

SIMMER: Benchmarking Latent Failures in LLM Executable Planning with a World Model

SkillAudit: Ground-Truth-Free Skill Evolution via Paired Trajectory Auditing

WAM4D: Fast 4D World Action Model via Spatial Register Tokens

Benchmarking Web Agent Safety under E-commerce Deceptive Interfaces

NeST: Neuron Selective Tuning for LLM Safety

Can Deep Neural Networks Improve Compression of Very Large Scientific Data?

Refusal Beyond a Single Direction: A Preliminary Comparison of Diff-in-Means and INLP

I'm Sorry Driver, I'm Afraid I Can't Do That: Appraising the Safety of LLMs within Automotive Contexts

Beyond the Training Distribution: Evaluating Predictions Under Distribution Shift and Selection Bias

Closing the Reflection Gap: A Free Calibration Bonus for Agentic RL

TA-RAG: Tone-Aware Retrieval-Augmented Generation for Peer-Support Health Communication

Poker Arena: Multi-Axis Profiling of Strategic Reasoning and Memory in LLMs

Why do South Koreans love AI so much?

VikingMem: A Memory Base Management System for Stateful LLM-based Applications

The Race to Build Data Centers in Space

Top Banks Rush to Fill Chief AI Roles as Talent Jumps to Rivals

Retrospective Progress-Aware Self-Refinement for LLM Agent Training

A Multi-Agent AI System for Automated High School Transcript Processing: Collaborative Document Analysis at Scale

The Culture Funnel: You Can't Align What isn't in the Data

SpikF-GO: Spiking Fourier Graph Operators for Multivariate Time Series Forecasting

ORCA: A Platform for Open-Source Dexterity Research

Independent-Component-Based Encoding Models of Brain Activity During Story Comprehension

HiLo-Token: Input-Adaptive High-Low Frequency Token Compression for Efficient Image Editing

When and How Severely: Scenario-Specific Safety Envelopes for Driving VLAs

Claude Code Guide 2026: 25 Features with Examples + Demo

X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMs

Conformal calibration and look-elsewhere effect in anomaly detection for new-physics searches

When Sample Selection Bias Precipitates Model Collapse

Learning the Context of Errors: Black-Box Online Adaptation of Time Series Foundation Models

Can Editing 1 Neuron Fix Repetition Loops in LLMs?

Beyond a Single Explanation of the Adam--SGD Gap

Decompose Sparsely Where You Should, Absorb Densely Where You Should No

Formalizing Numerical Analysis: An Agent Pipeline and Quality Audit Beyond Kernel Acceptance

DLawBench: Evaluating LLMs Through Multi-Turn Legal Consultation

MAD: Manifold Attracted Diffusion

The Risk Shadow of Principal Component Analysis: When 99.9999% Variance Preservation Causes Catastrophic Decision Errors

CoRe: A Continuously Reward-Finetuned LLM Query Rewriter for Multi-Stage Context-Aware Relevance in Web-Scale Video Search

EvoTrainer: Co-Evolving LLM Policies and Training Harnesses for Autonomous Agentic Reinforcement Learning

$\mu_0$: A Scalable 3D Interaction-Trace World Model