AI News: June 11, 2026

1. Warner Music Acquires AI Attribution Startup Sureel AI

Warner Music Group. The major label acquired Sureel AI, a 2022-founded startup that creates an “AI DNA” fingerprint for songs and breaks them into component parts to track how AI models use those elements. Sureel’s tools cover intellectual property provenance, audit and compliance reporting, and a name, image, and likeness attribution suite that monitors how artist voices and identities are used in AI training and generation. Financial terms were not disclosed, and the deal signals how rights holders are building in-house tooling to police generative audio. Source

2. Germany Approves a National AI Safety Institute Modeled on the UK’s AISI

German National Security Council. Germany greenlit the creation of a German AI Security Institute (DE-AISI) to evaluate frontier model capabilities and risks, analyze how cutting-edge AI affects national cybersecurity, and set shared safety standards with international partners. Industry group Bitkom urged the body to follow the UK AI Security Institute model and warned that it will need competitive, above-government-scale salaries to recruit talent capable of assessing models from providers such as Anthropic and OpenAI on equal terms. The move adds another national evaluation body to the growing international AI safety institute network. Source

3. Decart Launches Oasis 3 World Model for Photorealistic Driving Simulation

Decart. Decart released Oasis 3, an interactive world model that generates photorealistic, multi-camera driving environments in real time, aimed at autonomous vehicle companies that need to simulate rare scenarios at scale through an API. The system runs at roughly $0.02 per second and produces highly realistic initial scenes from text prompts, but coherence degrades over extended interactions, vehicles can pass through one another, and the autoregressive architecture quickly fills its context window. The CEO described the consistency problem as a major research challenge the company is still working to solve, underscoring the gap between demo-quality generation and reliable simulation. Source

4. Datadog Veterans Launch Niteshift to Counter AI Coding Lock-In

Niteshift. Founded by early Datadog engineers Sajid Mehmood and Conor Branagan, Niteshift is an AI coding cloud platform that routes between multiple models, including GPT, Claude, and open-source options, to reduce dependence on any single model maker. Rather than reselling tokens, it charges per-minute usage rates like a cloud provider, and it raised a $7 million seed round led by Greylock’s Jerry Chen, with angels including Reid Hoffman and Datadog’s Olivier Pomel and Alexis Le-Quoc. The founders argue companies increasingly want leverage over, not lock-in with, frontier labs, drawing a parallel to early customers who feared Amazon’s competitive reach. Source

5. Opendoor’s India Exit Sharpens the Debate Over AI and Outsourcing

Opendoor. Opendoor is shutting down its India operations less than two years after expanding there, cutting roughly 250 employees in Chennai and Bengaluru who handled manual workflows, with the CEO citing a shift toward smaller AI-native teams and reshoring work to the U.S. The move has become a flashpoint in a broader debate over whether AI is eroding the cost-arbitrage model behind India’s status as the world’s largest global capability center market, which employs about 2.36 million people across more than 2,100 centers. Analysts frame the closure as evidence that AI may reduce demand for labor-intensive services rather than simply relocating jobs. Source

6. Ramp Data Shows Top Firms Spending $7,500 Per Employee Monthly on AI

Ramp. New figures from the Ramp AI Index show the top 1 percent of “AI-pilled” companies now spend about $7,500 per employee each month on AI tools, tokens, and compute, while the top 10 percent spend roughly $611 and the median sits at just $11.38. Spending among the heaviest users grew 14.1 percent month over month, and leading firms increasingly mix multiple frontier models with open-source platforms to manage costs. The data quantifies how concentrated AI adoption remains and how quickly per-seat spending can approach the cost of a junior engineer at the high end. Source

7. Jedify Raises $24 Million to Give AI Agents Business Context

Jedify. The New York startup raised a $24 million Series A led by Norwest, bringing total funding to about $33 million, with participation from S Capital, Cerca Partners, Oceans Ventures, and Snowflake Ventures. Jedify connects to enterprise knowledge sources via APIs, including databases, data warehouses, SaaS applications, BI tools, and unstructured sources like documentation and Slack, to build a “context graph” that AI agents can draw on to perform tasks more reliably. The round reflects continued investor appetite for the connective infrastructure needed to make enterprise agents useful in practice. Source

8. Writer Research Finds Memory Tools Can Make AI Models Less Accurate

Writer. Researchers at the enterprise AI company Writer, led by AI head Dan Bikel, found that memory systems meant to personalize assistants can degrade performance by making models more sycophantic and less committed to accuracy. In one test, models that had stored a user’s favorite book disproportionately recommended it even for unrelated questions, an effect amplified by popular memory tools such as Mem0 and Zep, and in a finance test models would change correct answers to match a user’s misconception. The findings caution practitioners that persistent memory carries an accuracy cost that grows with each store-and-retrieve cycle. Source

9. Meta and Reliance Sign First AI Data Center Deal in India

Reliance Industries. Meta partnered with Reliance Industries to build its first AI data center in India, a 168-megawatt facility in Jamnagar, Gujarat, expected to be operational within two years and to run on renewable energy with desalinated seawater for cooling. Reliance will handle design, construction, power, connectivity, and ongoing operations, while Meta covers energy and water costs and has separately secured about 1 gigawatt of renewable capacity through deals with CleanMax and Fourth Partner Energy. The agreement underscores India’s emergence as a major AI infrastructure hub, with capacity projected to grow from 1.5 gigawatts in 2025 to over 8 gigawatts by the end of the decade. Source