Rajasthan Government to Introduce Policies on AI, Gaming, and Data Centers

The government of Rajasthan in India is set to launch three new policies focused on artificial intelligence, gaming, and data centers. The Rajasthan AI Policy 2025 will promote the ethical and inclusive use of AI and establish a Centre of Excellence to support AI startups and research. An AVGC-XR Policy will support the animation, visual effects, gaming, and extended reality sectors, with plans for four Atal Innovation Studios backed by a Rs 1,000 crore budget. Additionally, a Data Centre Policy 2025 aims to create a secure ecosystem for private data centers to attract investment.

MIT Researchers Develop Efficient Machine Learning Method for Symmetric Data

Researchers at MIT have developed the first provably efficient method for machine learning with symmetric data, which could lead to more powerful and less resource-intensive AI models. The new algorithm requires fewer data samples for training, which can improve a model’s accuracy and adaptability. This approach is particularly useful for applications like drug discovery and materials science, where understanding the symmetry of molecular structures is crucial. The findings provide a foundational understanding and could aid in the development of new neural network architectures.

Generative AI Continues to Produce Vulnerable Code, Study Finds

A report from application security company Veracode has found that large language models introduce vulnerabilities in nearly half of the test cases when tasked with completing secure coding assignments. The study indicates that while AI models are improving at writing syntactically correct code, they are not getting better at writing secure code. Veracode’s research involved analyzing over 100 LLMs across 80 coding tasks related to common weaknesses like SQL injection and cryptographic flaws. The findings suggest that the size and complexity of the model do not necessarily correlate with more secure code output.

Sources:

Microsoft and Singapore’s DISG Launch AI Accelerator Program

Microsoft and Digital Industry Singapore (DISG) have launched the Agentic AI Accelerator program to assist up to 300 Singapore-based businesses in their AI transformation. As part of the Singapore Government’s Enterprise Compute Initiative, the program will provide businesses with up to S$250,000 in Azure cloud credits, along with AI training and tools over the next year. Microsoft will also offer additional funding for select businesses to co-develop their agentic AI roadmaps with Microsoft technology partners. The initiative builds on Microsoft Singapore’s AI Pinnacle Program, which aims to transform various industries through AI.

EHang Conducts Autonomous Cargo Flight Trial with VT-20 eVTOL

Chinese eVTOL manufacturer EHang has successfully completed a cargo delivery trial using its uncrewed VT-20 aircraft. The autonomous vehicle transported fresh seafood from Zhuhai to Guangzhou, a distance of just over 45 nautical miles, in 55 minutes on July 30. The flight was a partnership with Wanyi Tianxia (Zhuhai) Aviation. On its return journey, the aircraft carried medical test samples back to Zhuhai. According to the companies involved, the flight was significantly faster than ground transportation, especially during peak traffic hours. The VT-20 is a follow-up to EHang’s EH216-S air taxi, which is already in operation in several Chinese cities.

Swarm Robotics Poised to Replace the Assembly Line in Aerospace Manufacturing

The manufacturing industry is exploring swarm robotics as a more efficient and innovative alternative to the traditional assembly line, particularly in the aerospace sector. This approach utilizes autonomous mobile robots that can coordinate their actions, much like a swarm of insects. Companies like Arrival have already demonstrated the potential of this technology by using self-aware robots in their Microfactories, replacing fixed tooling and conveyors. The advancement of generative AI is seen as a key enabler for achieving the level of self-learning required for complex swarm robotic systems. Proponents believe that swarm robotics will offer greater speed, cost-effectiveness, safety, and precision compared to current assembly line methods.

Tesla is encountering significant legal and regulatory challenges regarding its autonomous driving technology. A recent Florida court case resulted in a $329 million verdict that assigned 33% liability to Tesla for a driver’s overreliance on its Autopilot system, highlighting a shift towards stricter accountability for semi-autonomous systems. In response to liability concerns, competitors such as Waymo and Mercedes-Benz are focusing on Level 4 autonomy within controlled environments. These legal and regulatory pressures, combined with declining EV sales, are causing concern among investors and could lead to a realignment of power in the autonomous driving sector.

Google Launches Gemini 2.5 Deep Think for Complex AI Reasoning

Google has launched Gemini 2.5 Deep Think, its most advanced AI model to date, which is designed to solve complex problems by utilizing multiple AI agents that work in parallel. This multi-agent approach allows the model to explore various lines of reasoning simultaneously before combining them for a more robust answer. Starting August 1, 2025, the model is available to subscribers of Google’s $250/month AI Ultra plan. A version of this model achieved a gold medal at the 2025 International Math Olympiad. Google is also making that version available to a select group of mathematicians and researchers to aid in new discoveries. Gemini 2.5 Deep Think leverages tools like Google Search for real-time information and code execution for problem-solving. The model card for Gemini 2.5 Deep Think was published on August 1, 2025, and notes that while its cybersecurity evaluation has improved, it still struggles with the most difficult real-world scenarios.

Sources:

KubeSphere Discontinues Open Source Edition, Citing License Violations and Market Shifts

The team behind KubeSphere, a popular open-source management platform for Kubernetes, has announced the immediate suspension of download links and free technical support for its open source version. The project, described as a “distributed operating system for cloud-native application management, using Kubernetes as its kernel,” simplifies Kubernetes setup and configuration. The decision has been met with criticism from the open source community, with Peter Smalls of SUSE stating that such moves “undermine the vital trust essential for a healthy open source ecosystem”. A former founding member of the KubeSphere team cited repeated license violations by third parties and the rapid adoption of AI as potential reasons for the change. The KubeSphere team’s official announcement on GitHub pointed to the need to adapt to the new era of AI and focus resources on commercial-grade services.

Compliance Automation Startup Comp AI Raises $2.6M in Pre-Seed Funding

Comp AI, a startup in the compliance automation sector, has raised $2.6 million in a pre-seed funding round co-led by OSS Capital and Grand Ventures. The company aims to disrupt the market for compliance with frameworks like SOC 2 and HIPAA, which it describes as costly and labor-intensive. Comp AI claims its AI-powered platform can automate up to 90% of the compliance process. The new funding will be used to expand the company’s operations over the next three months. Investors have praised the company’s rapid execution and its approach to a market in need of innovation.

Sources:

Global Cloud Computing Market Projected to Reach $1.6 Trillion by 2030

A new report from BCC Research projects that the global cloud computing market will grow from $738.2 billion in 2025 to $1.6 trillion by 2030, at a compound annual growth rate (CAGR) of 17.2%. The study highlights several growth drivers, including the increasing number of data centers, advancements in cloud networking, and the integration of AI with cloud platforms. The trend towards multi-cloud environments is also a significant factor, allowing businesses to use services from various providers to enhance flexibility and reduce costs. The report provides a detailed analysis of the market, broken down by service models, deployment models, enterprise sizes, and end-users across different regions.

Fujitsu Announces Development of Superconducting Quantum Computer with Over 10,000 Qubits

Fujitsu has initiated research and development for a superconducting quantum computer that will exceed 10,000 qubits, with a targeted completion by fiscal year 2030. This project is part of a NEDO initiative to advance the industrialization of quantum computers and will be a joint research effort with Japan’s National Institute of Advanced Industrial Science and Technology (AIST) and RIKEN until fiscal year 2027. The new computer will leverage Fujitsu’s “STAR architecture,” an early-stage fault-tolerant quantum computing architecture. The company aims to achieve 250 logical qubits by fiscal 2030 and 1,000 logical qubits by fiscal 2035. Following the development of the 10,000+ qubit machine, Fujitsu intends to research the integration of superconducting and diamond spin-based qubits.

Sources:

Study Details a Scalable Design for Photonic Quantum Computing

A recent study presents a detailed blueprint for constructing a scalable, fault-tolerant photonic quantum computer utilizing deterministic quantum dot emitters. This research, a collaboration between the University of Copenhagen, Sparrow Quantum, and the University of Bristol, aims to overcome significant obstacles in photonic quantum computing, including photon loss and the complexity of large-scale optical circuits. The proposed architecture integrates time-bin encoded photons and adaptive fusion gates within a low-depth structure to minimize optical complexity and facilitate real-time error correction. The design is based on a fusion-based quantum computing model that processes information through entangling measurements on small, entangled resource states produced by deterministic quantum dot-based emitters.

Sources:

Google Unveils Opal, a No-Code Platform for Building AI Mini-Apps

Google has introduced an experimental AI tool named Opal, which enables users to build small applications using AI prompts without writing any code. Opal allows for the creation of these mini AI apps by linking together AI prompts, models, and various tools through natural language and a visual editing interface. This new tool is designed to make AI application development more accessible to a wider audience.

Sources:

Proton Releases Cross-Platform, Open-Source Authenticator App

Proton has launched a new authenticator application that is both cross-platform and open-source. The app is designed with a focus on privacy and provides users with a secure method for two-factor authentication across different devices.

Sources: