Gartner’s 2025 Hype Cycle: Generative AI Enters ‘Trough of Disillusionment’
According to Gartner’s 2025 Hype Cycle for Artificial Intelligence, Generative AI has officially moved into the ‘Trough of Disillusionment’. This shift indicates that the initial peak of inflated expectations is cooling as organizations now grapple with the technology’s limitations and seek tangible returns on investment (ROI). Gartner notes that while GenAI remains a potentially transformational technology, the primary challenge for many companies is proving its business value. Consequently, the industry’s focus is shifting towards foundational elements like AI-ready data and the development of AI agents, both of which are currently at the ‘Peak of Inflated Expectations’. This trend underscores a maturing understanding that successful enterprise AI relies on strong data governance, intelligent workflows, and robust operational frameworks. Despite cooling expectations for GenAI, overall IT spending is still projected to increase in 2025, driven partly by ongoing investments in AI software and infrastructure.
Sources:
Mount Sinai Tackles AI Bias in Healthcare with New ‘AEquity’ Tool
Researchers at the Icahn School of Medicine at Mount Sinai have developed AEquity, a novel tool designed to detect and mitigate biases within datasets used to train healthcare-focused machine-learning algorithms. This innovation addresses the critical concern that AI models trained on biased data can perpetuate and even worsen health disparities by leading to inaccurate diagnoses and treatments. AEquity was successfully tested on diverse health data, including medical images and patient records, where it identified both known and previously undiscovered biases. A key feature of the tool is that it is model-agnostic, meaning it can be applied to a wide range of machine-learning models, from simple algorithms to complex large language models (LLMs). The ultimate goal of AEquity is to provide a practical solution for developers and healthcare systems to build and deploy AI tools that are fair, equitable, and effective for all patient populations.
Sources:
SailPoint and HCLTech Partner on AI-Driven Identity Security for Enterprises
SailPoint has announced a strategic partnership with HCLTech to deliver advanced identity security solutions tailored for enterprises leveraging artificial intelligence. The collaboration integrates SailPoint’s AI-driven Identity Security Cloud with HCLTech’s comprehensive cybersecurity services. This partnership aims to help organizations manage the growing complexity of digital identities in AI-rich environments. As companies increasingly deploy non-human identities—such as machine accounts and autonomous AI agents—they face new governance challenges that traditional security models cannot handle. The joint solution will provide continuous, risk-aware governance and automate access controls for both human and digital identities, helping enterprises enforce zero-trust principles and ensure AI agents operate within secure, compliant boundaries.
Sources:
WeRide Launches World’s First 24/7 Fully Driverless Robotaxi Service in Guangzhou
Autonomous driving company WeRide has launched the world’s first 24/7, fully driverless ride-hailing service in the Huangpu District of Guangzhou, China. The service features the Robotaxi GXR, a vehicle specifically designed for mass production with Level 4 autonomous driving capabilities. Passengers can hail the robotaxis at any time for point-to-point travel within the designated district using the WeRide Go app. The Robotaxi GXR is equipped with WeRide’s advanced Sensor Suite 5.6, engineered to operate reliably in adverse weather conditions, and includes a full redundancy system for enhanced safety. This commercial launch follows an intensive 11-month period of public road testing and the successful acquisition of all necessary licenses to transition to a completely driverless operation.
OpenAI Begins Rollout of GPT-5, Its ‘Smartest, Fastest, Most Useful Model Yet’
OpenAI has started the broad rollout of GPT-5, its first major model upgrade since the release of GPT-4. Described by the company as its “smartest, fastest, most useful model yet,” GPT-5 introduces a novel “thinking” mode designed for longer, more complex reasoning tasks. This allows it to provide expert-level responses in demanding fields like coding, mathematics, and creative writing. GPT-5 is a unified system that seamlessly handles both quick, conversational replies and deeper, more reasoned analysis. A key goal of the new model is to significantly improve factual accuracy and reduce the instances of “hallucinations” that have affected previous models.
Sources:
Swiss Consortium Launches Apertus: A Fully Transparent, Open-Source AI Model
A consortium of leading Swiss institutions, including EPFL, ETH Zurich, and the Swiss National Supercomputing Centre (CSCS), has released Apertus, a new open-source AI language model. The name, derived from the Latin word for “open,” reflects the project’s core commitment to transparency. Its architecture, model weights, and training data are all publicly accessible. Apertus is designed as a foundational model, enabling developers and organizations to build a wide range of applications, such as custom chatbots and advanced translation tools. The model is available for download on Hugging Face and can be accessed via Swisscom, a strategic partner in the initiative. Future plans include expanding the Apertus model family and developing domain-specific versions for sectors like law and healthcare.
Sources:
Microsoft Unveils In-House AI Models to Reduce OpenAI Dependency
In a strategic move to reduce its reliance on partner OpenAI, Microsoft has introduced its first proprietary AI models for text and speech, named MAI-Voice-1 and MAI-1-preview. Microsoft’s AI chief, Mustafa Suleyman, emphasized the necessity of developing in-house expertise to create the industry’s strongest models. The company claims its MAI models can achieve top-tier performance at a significantly lower computational cost by utilizing carefully curated training data and advanced efficiency techniques.
Sources:
OpenAI Enhances Real-Time AI with gpt-realtime and General API Availability
OpenAI has announced the general availability of its Realtime API and the launch of gpt-realtime, its most advanced speech-to-speech model. The new model is engineered to deliver more natural, fluid, and multilingual conversations for real-world applications like interactive customer support. By integrating speech-to-text and text-to-speech capabilities into a single, unified model, gpt-realtime significantly reduces latency. Platform updates also include native support for SIP for direct phone call integration and the ability to process image inputs during conversations.
Sources:
Kubernetes 1.34 Released with Long-Term Support for Telco Clouds via Canonical
The latest version of Kubernetes, v1.34 “Of Wind & Will”, has been released, introducing 58 enhancements. Among these, 23 features have graduated to stable, including the General Availability of Dynamic Resource Allocation (DRA), which offers a more flexible and powerful way to manage specialized resources like GPUs. In related news, Canonical announced that its long-term support (LTS) version of Kubernetes 1.32 is now integrated with Sylva 1.5, a major European telco cloud initiative. This integration provides telecom operators with a production-grade, validated platform backed by Canonical’s 12-year security and maintenance guarantee. The collaboration aims to create a telco-friendly, cloud-native infrastructure stack that supports Europe’s technology sovereignty and security goals.
Sources:
Atlassian to Acquire The Browser Company for $610M in a Major DevOps Workflow Shift
Atlassian has entered into an agreement to acquire The Browser Company, creator of the Arc browser, for $610 million. This strategic acquisition signals a major shift by the software collaboration giant to deeply integrate AI-powered browser capabilities into its ecosystem of developer tools, including Jira, Confluence, and Bitbucket. The deal suggests a future where the browser evolves from a neutral interface into an active agent that orchestrates complex development workflows. For DevOps and platform engineering teams, this could lead to a more unified experience, breaking down silos between project management, source control, and CI/CD pipelines. For example, an AI-driven browser could detect a CI/CD pipeline failure, automatically surface the relevant Jira ticket, and suggest a code fix based on previous commits, all within a single, intelligent interface.
Sources:
Bosch and Alibaba Expand Partnership for AI and Cloud-Driven Innovation
Bosch and Alibaba Group have deepened their strategic partnership to accelerate digital transformation through advanced cloud computing and AI technologies. The expanded collaboration will focus on three key areas: cloud-based enterprise operations, AI-driven business innovations, and global e-commerce expansion. Under the agreement, Alibaba Cloud will support Bosch’s core technology stack. The two companies will also explore leveraging Alibaba’s AI models to enhance Bosch’s business innovations, such as developing a more intuitive smart cockpit for vehicles using multimodal AI. The partnership also extends to e-commerce, with Bosch planning to expand its presence on Alibaba’s global platforms across Southeast Asia, Spain, and Latin America.
Sources:
MOTH Launches Quantum Brush: A New Tool for Quantum-Powered Digital Art
The company MOTH has introduced Quantum Brush, an open-source digital painting tool that uses quantum computing to unlock new forms of artistic expression. The tool features four unique brushes, each designed to translate an artist’s strokes into distinct quantum algorithms, generating unique aesthetics based on quantum effects. These brushes are specifically designed to be compatible with today’s noisy intermediate-scale quantum (NISQ) devices. MOTH’s initiative is part of a broader effort to bridge the gap between quantum computing and the creative industries, including music and gaming, by empowering artists with tools to explore the technology’s artistic potential.
Sources:
Qubic Awarded Grant to Overcome Quantum Computing Heat Barriers with New Amplifiers
Quantum technology company Qubic has been awarded a CAD $925,000 grant from the Canadian government to develop advanced cryogenic amplifiers. This funding supports a larger $2.5 million project aimed at creating amplifiers from quantum materials to solve a critical thermal barrier in scaling quantum computers. Because most quantum computers operate in a cryogenic environment, heat generated by system electronics can interfere with qubit performance and limit scalability. Qubic’s quantum amplifier technology is designed to significantly reduce heat dissipation, potentially removing a key obstacle to building larger, more powerful quantum computers. The project is a collaboration with the University of Waterloo, the Institute for Quantum Computing, and the Quantum Nanofabrication and Characterization Facility.
Sources:
Software Engineering Brief: Chainguard’s Zero-CVE VMs, xAI’s Grok Code Fast 1, and AGENTS.md
In software engineering and open-source news this week:
- Chainguard has expanded its offerings of zero-CVE (Common Vulnerabilities and Exposures) virtual machines.
- xAI has released Grok Code Fast 1, a new AI model optimized for coding tasks.
- AGENTS.md, a new open standard for configuring AI coding agents, is gaining significant traction in the developer community.
- Tietoevry and Inmanta have deepened their partnership to advance automated telecom solutions.
Sources: