Governance and Regulation
Microsoft has signalled its intent to join the European Union’s voluntary code of practice for general-purpose AI, committing to publish training content summaries and implement copyright-compliance policies aligned with the bloc’s upcoming AI Act. In contrast, Meta Platforms has declined to sign, citing legal uncertainties and concerns that the guidelines exceed the AI Act’s scope—joining 45 European firms in warning that the code could stifle innovation across the region. Read more
On the same day, the European Commission released detailed guidance for AI systems deemed to pose systemic risks, targeting models that could significantly impact health, safety, fundamental rights, or society. Companies including Google, OpenAI, Meta, Anthropic, and Mistral must conduct rigorous risk assessments, adversarial testing, incident reporting, and robust cybersecurity measures by August 2, 2026 to avoid fines of up to €35 million or 7% of global turnover. Read more
Meanwhile, the Judicial Council of California adopted landmark rules for generative AI use in courts, requiring either a full ban or custom policies by September 1 that address confidentiality, bias mitigation, oversight, transparency, and security. Led by Chief Justice Patricia Guerrero’s AI task force, the policy prohibits the input of confidential data into public AI tools and mandates human verification of AI outputs across 65 courts handling five million cases annually. Read more
Ethical Stewardship and Community Funding
OpenAI announced a $50 million fund to support nonprofits and community organisations in education, healthcare, economic opportunity, and community-led AI research. The initiative follows recommendations from OpenAI’s nonprofit commission, which consulted over 500 experts to align philanthropic efforts with the mission of advancing AI for public benefit. Read more
As part of its long-term stewardship strategy, OpenAI plans to convert its for-profit subsidiary into a public benefit corporation, with its nonprofit arm retaining ownership. This structural change aims to safeguard the company’s societal mission while enabling the investments needed for continued innovation. Read more
Defense and Agentic AI
The U.S. Department of Defense awarded contracts worth up to $200 million each to OpenAI, Google, Anthropic, and Elon Musk’s xAI to prototype advanced “agentic AI” systems designed to assist military decision-making and operational workflows. DoD Chief Digital and Artificial Intelligence Officer Doug Matty emphasised that these AI “agents” will transform the department’s ability to support warfighters and maintain strategic advantage. Read more
Next-Generation AI Agents and Market Moves
OpenAI’s new ChatGPT “agent” feature enables the chatbot to execute complex, multi-step tasks—such as browsing the web, using plugins, and making reservations—through its own virtual environment. Early testers praise its autonomous capabilities, though regulatory constraints in the EU have delayed broader rollout. Read more
Meta confirmed plans to expand its compute infrastructure with a 1-gigawatt data center in Ohio and a 5-gigawatt “Hyperion” facility in Louisiana slated for 2026. Google also secured AI talent by acquihiring startup Windsurf for roughly $2.4 billion after a bidding battle with OpenAI, underscoring the resource-hungry nature of the AI race. Read more
Talent Wars and Supercomputing Push
Meta Platforms has recruited two leading Apple AI researchers—Mark Lee and Tom Gunter—to its new Superintelligence Labs team, following the earlier poaching of their former supervisor. This aggressive talent acquisition is part of Meta’s strategy to accelerate its AI roadmap. Read more
The company has also increased its 2025 capital expenditure target to $64–$72 billion, deepening partnerships with firms like Scale AI and committing hundreds of billions to build mega-scale data centers and supercomputing clusters in pursuit of future breakthroughs. Read more
Open-Source Frontiers
Mistral launched Voxtral, its first open-source AI audio model capable of transcribing and understanding up to 40 minutes of multilingual speech—including English, Spanish, French, and Hindi—at just $0.001 per minute. The release highlights growing momentum behind community-driven AI innovation. Read more
Perplexity AI is in talks with smartphone manufacturers, including Apple and Samsung, to pre-install its Comet AI browser on mobile devices. Comet integrates AI into browsing for personal data queries, scheduling, and summarization, aiming to challenge entrenched defaults like Chrome. Read more
Corporate Restructuring and Efficiency
Amazon Web Services has cut “several hundred” roles across its division as part of CEO Andy Jassy’s vision for leaner, AI-augmented operations. The layoffs affect specialist and support teams, reflecting a strategic realignment to reduce bureaucracy and reallocate resources toward innovation—even as AWS posted a 17% year-over-year revenue increase. Read more
