#cloud

Improving the speed and energy-efficiency of AI agents

Các nhà nghiên cứu từ MIT và Microsoft đã phát triển Murakkab, một hệ thống tự động tối ưu hóa quy trình làm việc của AI agent bằng cách tự động lựa chọn mô hình, công cụ và cấu hình phần cứng dựa trên mô tả ngôn ngữ tự nhiên từ nhà phát triển, đồng thời điều chỉnh linh hoạt theo ưu tiên như chi phí hay tốc độ. Trong thử nghiệm, Murakkab tiết kiệm tới ~65% tài nguyên tính toán, ~73% năng lượng và trên 75% chi phí so với phương pháp truyền thống mà vẫn duy trì hiệu suất ổn định.

Lập trình viên nên đọc bài này vì Murakkab không chỉ tiết kiệm chi phí và năng lượng cho AI mà còn giúp họ thiết kế các hệ thống thông minh hiệu quả hơn bằng cách tự động tối ưu hóa các quyết định kỹ thuật từ đầu, cho phép họ tập trung vào logic kinh doanh thay vì chi tiết kỹ thuật phức tạp.

Oracle is slashing its workforce as it automates with AI

Oracle đã cắt giảm 13% lực lượng lao động (khoảng 21.000 nhân viên) trong năm tài chính 2026, tiêu tốn 1,84 tỷ USD chi phí thôi việc, chủ yếu do tự động hóa bằng AI và tái cấu trúc chiến lược. Song song, hãng đầu tư mạnh 70 tỷ USD vào cơ sở hạ tầng cloud, ký hợp đồng lớn với OpenAI và Meta, đồng thời huy động 40 tỷ USD từ nợ và vốn mới.

Lập trình viên nên đọc bài này để hiểu rõ cách AI và cloud computing đang thay đổi mô hình công việc, từ đó cập nhật kiến thức về xu hướng công nghệ mới và cách ứng dụng chúng trong việc phát triển ứng dụng, tối ưu hóa hệ thống, và chuẩn bị cho tương lai của ngành IT.

Why Your Organization Needs ISPM

Identity Security Posture Management (ISPM) continuously assesses and hardens Microsoft 365 identity environments by evaluating configurations, permissions, and policies against a security baseline. Unlike one-time audits or visibility-only tools, ISPM addresses 'drift' — the gradual degradation of security posture as users, roles, and Microsoft defaults change over time. Huntress data shows over 60% of evaluated tenants were missing more than half of recommended controls, and 55% allowed standard users to perform admin-level functions. Key risks include weak MFA, overprivileged accounts, and stale permissions that enable account takeover and BEC attacks. Microsoft data shows attackers can move laterally within 48 minutes of initial intrusion, making daily scan cycles insufficient. Huntress Managed ISPM addresses this by deploying and enforcing controls continuously, detecting drift within minutes of a change, and offering a Learning Mode to preview user impact before policy rollout.

Brussels moves to put AWS and Azure under its toughest tech rules

The European Commission has issued preliminary findings that AWS and Azure should be designated as gatekeepers under the Digital Markets Act, despite neither service meeting the regulation's standard quantitative thresholds. Brussels is invoking the DMA's qualitative route, arguing both platforms serve as critical gateways for EU businesses. If designated, the two cloud giants would face obligations including no self-preferencing, mandated interoperability, and data portability rules aimed at reducing vendor lock-in. The move is partly motivated by cloud infrastructure's growing role as the foundation for AI adoption across Europe. Both Amazon and Microsoft have objected, and a defence phase is now underway before any final decision.

Amazon ups India bet with fresh $13B AI infrastructure investment

Amazon announced a $13 billion investment to expand AWS data center capacity in India through 2030, bringing its total India investment commitments to $48 billion. The announcement followed a meeting between CEO Andy Jassy and Prime Minister Modi. This is Amazon's third major India commitment in three years. The move is part of a broader wave of global tech investment in India's AI infrastructure, with Microsoft pledging $17.5 billion and Google $15 billion. Amazon is also expanding its retail and quick-commerce operations in India, planning 20+ fulfillment centers and extending its Amazon Now service to 300+ cities.

Oracle cuts about 500 jobs in Romania as AI restructuring rolls on

Oracle has begun cutting approximately 500 jobs at its Romanian operations on June 25, 2026, marking the second significant layoff round in under a year following ~400 cuts in late 2025. The reductions are part of Oracle's ongoing global AI-driven restructuring, which has shrunk its worldwide headcount from ~162,000 to ~141,000 (about 13%) in a year. Oracle's annual report explicitly cites AI adoption as a driver of workforce reductions. The company is investing heavily in data centers tied to AI demand, including a major OpenAI commitment, and is partly funding that buildout through payroll cuts. Romania, home to one of Oracle's larger engineering and services footprints in central and eastern Europe (~4,000 employees), is particularly exposed. Despite revenue growth of 7.5% in 2025, Oracle's Romanian entities collectively posted net losses. A former employee suggests further cuts may follow.

Amazon to invest an extra $13bn in Indian cloud and AI by 2030

Amazon announced an additional $13bn investment in India by 2030, bringing its total commitment to $48bn for 2026–2030. The new funds will expand AWS data centre capacity in Mumbai and Hyderabad, providing access to custom AI chips, managed AI services, and developer tools. CEO Andy Jassy made the announcement in New Delhi during a meeting with Prime Minister Modi. The shift marks a strategic pivot from Amazon's earlier India focus on retail toward cloud and AI infrastructure, mirroring similar large-scale commitments from Microsoft and Google Cloud in the country. Amazon also cited broader economic targets including support for 3.8 million jobs and AI education for four million students.

OVHcloud confirms its outlook as public cloud growth accelerates past 20%

OVHcloud reported Q3 FY2026 revenue of €289.6 million, up 6.9% organically, with its Public Cloud segment rebounding sharply to 20.2% growth driven by VPS customer acquisition and expanded availability zones in Paris and Milan. The company reaffirmed full-year guidance of 5–7% organic growth and positive levered free cash flow. Beyond financials, OVHcloud won a share of the EU's €180 million sovereign cloud contract, entered exclusive negotiations to acquire voice-AI startup Gladia, and previewed an agentic AI workspace platform. The broader context is Europe's ongoing debate over whether regional cloud providers can offer genuine sovereignty when their infrastructure still depends on non-European chips.

Runpod hits $1bn valuation on $100M raise

Runpod, a cloud startup renting GPU compute to AI developers, has raised $100M led by Summit Partners at a $1bn valuation — a tenfold increase from its ~$100M seed valuation in 2024. The company reports $240M in annualized revenue, over 1 million developers on the platform, and 20 billion inference requests processed. Unlike rivals focused solely on inference, Runpod offers the full ML lifecycle: experimentation, training, fine-tuning, and deployment. It uses an asset-light model, renting capacity rather than owning data centers, and supports both Nvidia and AMD chips. The company says it turned down buyout offers exceeding $500M to remain independent. Key risks include reliance on third-party hardware, margin pressure when chips are scarce, and well-funded competitors like CoreWeave.

Identiverse 2026: The Challenges Of Solving Identity For AI Agents At Scale

Identiverse 2026 surfaced a consistent message across sessions: enterprises deploying AI agents at scale have not yet solved the foundational identity problems of ownership, delegation, least privilege, and auditability. Key themes included the need for named human owners for every agent, replacing static API keys with short-lived cryptographically attested tokens, embedding audit trails from day one rather than retrofitting them, and treating discovery as the prerequisite for every other governance control. Speakers from Capital One, Equifax, Intuit, Fannie Mae, and others described environments with tens of thousands of unowned non-human identities, delegation chains with no provenance tracking, and citizen developers being taught by AI coding tools to bootstrap agents with static credentials. The consensus: governance built after deployment cannot keep pace with the rate of agent creation, and the cost of remediation scales with every week of unaccountable production deployments.

Alibaba sues the Pentagon to get off its Chinese military companies list

Alibaba has filed a federal lawsuit in San Jose against the US Department of Defense, seeking removal from the Pentagon's Section 1260H list of Chinese military companies. The company was added on June 8 alongside Baidu, BYD, and Unitree, pushing the roster to 188 entities. Alibaba argues the designation has no factual or legal basis, citing its independent board and civilian-focused products. The Pentagon justified the listing through Alibaba's ties to China's Ministry of Industry and Information Technology and indirect links to state asset regulator SASAC. While the label doesn't outright ban business with Alibaba, it bars DoD contracts from June 30 and third-party procurement from 2027, and the company claims it creates broader financing and partnership barriers. Biotech firm WuXi AppTec filed a similar suit on June 11. The case unfolds amid broader US-China tech tensions including export controls, tariffs, and China's own retaliatory export control list targeting 10 US firms.

Oracle’s 21,000 layoffs help drive its debt-fueled AI investments

Oracle's annual SEC filing reveals the company reduced its workforce by 21,000 employees (12.9%) over the past fiscal year, attributing the cuts partly to AI adoption across its operations. The layoffs are also tied to massive capital expenditure plans: Oracle aims to raise $45–$50 billion in 2026 to expand its Oracle Cloud Infrastructure, with roughly half coming from debt. The company already carries over $120 billion in total debt, which has alarmed investors and prompted a bondholder lawsuit alleging Oracle concealed the extent of debt needed to fund its AI buildout. Key customers for the expanded infrastructure include OpenAI, xAI, AMD, Nvidia, and Meta.

Oracle cuts 21,000 jobs, SEC filing blames AI

Oracle's workforce dropped by roughly 21,000 employees to 141,000 as of May 2026, a 13% reduction. Unusually, the company's SEC filing explicitly attributes the cuts to AI adoption across its operations — a rare instance of a major tech firm stating in a regulatory document that AI is replacing workers rather than complementing them. The deepest cuts hit Oracle Health (formerly Cerner), with 8,000–10,000 jobs lost, while AI and cloud infrastructure teams were largely spared or grew. Oracle spent $55.7 billion on data centre capex in fiscal 2026, up 162%, resulting in negative free cash flow of $23.7 billion. Cloud Infrastructure revenue grew 93% to $5.8B in Q4, and remaining performance obligations jumped to $638 billion. The pattern mirrors moves by Meta, Microsoft, and others converting payroll into AI infrastructure spending, but Oracle's candour in its SEC filing sets it apart.

Groq’s $650M raise rebuilds it after Nvidia

Groq has closed a $650M funding round led by Disruptive and Infinitum to rebuild itself as an AI inference cloud company, six months after Nvidia licensed its chip technology and hired away its founder Jonathan Ross. The company now operates 13 data centres globally, serves over five million developers, and plans to quadruple capacity to 200 megawatts by end of 2027. A new leadership team has been installed, including CEO Adam Winter and incoming CTO Sinclair Schuller. The core bet is that purpose-built inference chips will outperform general-purpose GPUs on speed and cost as AI workloads shift from training to inference — though Groq now shares its chip IP with Nvidia, the competitor it originally set out to displace.

Conditional & Iterative Deployments with Azure Bicep

A deep dive into Azure Bicep's conditional (if) and iterative (for) deployment features, covering practical patterns and common pitfalls. Key topics include: guarding references to conditional resources using ternary expressions, the non-cascading nature of conditions on child resources, safe use of runtime functions like listKeys(), resource naming stability in loops, the difference between resource/module loops and variable/output loops, index fragility in filtered loops, and using @batchSize() to control deployment parallelism. Includes a practical checklist and a complete example combining conditionals, filtered loops, batching, and structured outputs.

Europe Tech Sovereignty Package Targets Chips, Clouds

The European Commission has introduced a Tech Sovereignty Package aimed at reducing Europe's dependence on U.S. technology. The package has four components: Chips Act 2.0 (expanding semiconductor production and demand), the Cloud and AI Development Act (tripling EU data center capacity by the early 2030s and introducing cloud assurance levels), an energy grid roadmap to support data center growth sustainably, and an open-source strategy to shift public sector spending away from proprietary American software. Critics argue the measures rely too heavily on encouragement rather than mandates, potentially allowing inertia and U.S. hyperscaler dominance to persist. The legislation still faces a lengthy EU lawmaking process and lobbying pressure from U.S. tech firms.

Europe’s sovereignty push may backfire

The European Commission's push to reduce dependence on foreign cloud, AI, and semiconductor providers is driven by legitimate concerns about 'kill switch' risks. However, the policy response is likely to produce fragmented cloud strategies, enterprise confusion, and ultimately greater consolidation around the very hyperscalers Europe is trying to avoid. Sovereign cloud providers face brutal market economics — they lack the scale, capital, and ecosystem depth of hyperscalers, and many will eventually be acquired by larger players. The result may be more sovereign branding and hybrid arrangements than genuine autonomy, with enterprises caught navigating politically-driven procurement decisions rather than technically sound ones.

Oracle’s workforce shrank by about 13% as it bankrolls its AI buildout

Oracle ended fiscal year 2026 with roughly 141,000 employees, down from 162,000 — a reduction of about 21,000 people or 13%, one of the deepest cuts in the company's history. The layoffs were framed as a reallocation of resources from legacy operations into AI and cloud infrastructure. Oracle's capital expenditure for the year reached approximately $50bn, with a contracted revenue backlog surpassing half a trillion dollars driven by large AI deals. Analysts estimate the workforce reductions free up $8–10bn annually in cash flow, which is being funneled into data-center construction. The cuts hit some teams by 30% or more, and a $16.3bn data-center financing deal required bond manager PIMCO to anchor $10bn after US banks stepped back. Oracle joins a broader tech industry trend of converting payroll into AI capital expenditure.