In this week’s real-time analytics news: MIT focuses on making AI-generated code more accurate in any language.
Keeping pace with news and developments in the real-time analytics and AI market can be a daunting task. Fortunately, we have you covered with a summary of the items our staff comes across each week. And if you prefer it in your inbox, sign up here!
MIT researchers and others have developed a new approach that automatically guides an LLM to generate text that adheres to the rules of the relevant language, such as a particular programming language, and is also error-free. The method allows an LLM to allocate efforts toward outputs that are most likely to be valid and accurate while discarding unpromising outputs early in the process. This probabilistic approach boosts computational efficiency.
Due to these efficiency gains, the researchers’ architecture enabled small LLMs to outperform much larger models in generating accurate, properly structured outputs for several real-world use cases, including molecular biology and robotics. In the long run, this new architecture could help nonexperts control AI-generated content. For instance, it could allow businesspeople to write complex queries in SQL, a language for database manipulation, using only natural language prompts.
Real-time analytics news in brief
Astronomer announced the release of Apache Airflow 3, marking the most significant advancement in the project’s history. Airflow 3 offers a comprehensive redesign that delivers long-standing community requests while also expanding the platform’s capabilities to support complex AI, ML, and near real-time data workloads. Additionally, Airflow 3 contains a significant number of new features that make Airflow easier to use, more secure, and able to run any task, anywhere.
Buoyant announced the release of Linkerd 2.18. The release includes three significant changes based on our experience helping our customers run Linkerd at the boundaries of scale: protocol declarations, GitOps-compatible multi-cluster, and changes to Gateway API support. This release also introduces an experimental build of the proxy for Windows environments, an exciting new area for Linkerd.
Dataiku announced the launch of AI Agents with Dataiku. This new set of capabilities is designed to create and control AI agents at scale, delivering a new class of AI applications powered by analytics, predictive models, and agents. The platform addresses the challenges of ungoverned AI agent deployments by enabling centralized creation, governance, and integration into existing systems. Key features include centralized agent creation via no-code and code options, LLM Mesh for secure agent orchestration, observability and performance monitoring tools, and more.
Diligent announced Diligent AI Risk Essentials (AI Risk Essentials), a new solution designed to initiate and strengthen an organization’s enterprise risk management (ERM) program. Supported by benchmarking risk data from SEC 10-K filings, AI Risk Essentials can be implemented in less than a week, enabling GRC professionals to immediately prepare for board-level discussions, swiftly identify risks, conduct risk assessments, implement mitigation strategies, and monitor progress.
Endor Labs announced a major expansion of its AppSec platform. At the core of the platform are dedicated AI agents built specifically for application security. The new AI Security Code Review uses multiple AI agents to review every pull request (PR) for architectural changes that impact an organization’s security posture and fall outside the scope of legacy Static Application Security Testing (SAST) and vulnerability scanning tools.
Illumio released a new agentless container security solution that delivers deep, real-time visibility into Kubernetes workloads without using in-container agents. The solution is designed to run across any Kubernetes environment and allows customers to visualize intra-cluster traffic between pods, services, and namespaces; detect lateral movement risks and unauthorized communication paths; and enforce segmentation policies to contain potential breaches.
Lenovo introduced 21 newAI-optimized ThinkSystem and ThinkAgile solutions, expanding its enterprise storage and infrastructure portfolio to support IT modernization and AI workloads. The new ThinkSystem and ThinkAgile models are designed with accelerated computing and efficiency that companies need to kickstart their AI journey and maximize enterprise AI efforts. To that end, the new systems include next-generation storage arrays, hyperconverged infrastructure (HCI) appliances, and AI Starter Kits aimed at improving efficiency, scalability, and sustainability.
Lenses.io announced the release of Lenses 6.0, enabling organizations to modernize applications and systems with real-time data as AI adoption accelerates. Lenses 6.0 simplifies the complex task of connecting applications and systems by providing data visibility and accessibility so that developers, regardless of skill level, can autonomously build innovative AI applications seamlessly. Additionally, Lenses 6.0 helps businesses to transition from their systems sharing and processing data that is minutes or hours old to data that is real-time.
NVIDIA announced that NeMo microservices are now generally available for developers and enterprises to build AI agents that learn from experience with business intelligence and insights using inference data from reasoning models, including NVIDIA Llama Nemotron. The microservices provide an end-to-end developer platform for creating state-of-the-art agentic AI systems and continually optimizing them with data flywheels informed by inference and business data, as well as user preferences.
OneTrust announced multiple AI innovations to help teams scale their operations, increase efficiency, and respond faster. Two new OneTrust agents, the Privacy Breach Response Agent, and Privacy Risk Assessment Agent, automate simple tasks and entire privacy processes. AI Document Scanning, available with OneTrust’s spring platform release, and AI Inventory Analysis, currently in preview, auto-populate assessments by scanning documents, past assessments, and existing inventories to identify relevant information and pre-populate new assessments. And OneTrust Copilot gives privacy teams easy access to the exact regulatory, product, and program intelligence they need to keep up with regulatory updates.
PingCAP announced a series of product updates that make it easier for businesses to scale with TiDB, a unified database platform for transactions, real-time analytics, and AI applications. The new release delivers new AI features and expanded multi-cloud options. Key features include AI-ready SQL and multi-cloud freedom and flexibility. To that latter point, TiDB’s Cloud Dedicated option now supports Microsoft Azure alongside AWS and Google Cloud. Support for Alibaba Cloud International is scheduled for later this year.
SandboxAQ announced the general availability of AQtive Guard, a platform designed to manage and secure Non-Human Identities (NHIs) and other cryptographic assets used by AI agents. AQtive Guard’s Discover module enables organizations to maintain an accurate inventory and control over both NHIs and cryptographic assets such as keys, certificates, algorithms, and libraries. AQtive Guard’s Protect orchestrates automated remediation workflows and enforces protection policies such as credential rotation or certificate renewal.
Snyk launched API & Web, a dynamic application security testing (DAST) solution built to secure the APIs that power the next wave of AI-driven apps. The solution allows developers to test the security of all of their APIs and web apps, whether the code was written by a developer or AI. It also provides detailed recommendations on how to fix issues that are found.
Vanta announced new ways to help organizations confidently demonstrate AI security and evaluate AI risk across their ecosystem. Vanta’s new AI Security Assessment provides a standardized approach to evaluating AI-related security risks and enables companies to have a better understanding of how these risks impact their overall security program. AI Security Assessments feature an accessible and practical set of evaluative questions covering ten critical categories, from governance and organizational management, data privacy and security, bias, human oversight, and more.
Partnerships, collaborations, and more
Fujitsu Limited and Supermicro announced that they are deepening their strategic collaboration in AI computing to facilitate the secure application of generative AI within dedicated environments. To that end, the two companies, alongside the wider Fujitsu Group, announced the launch of the PRIMERGY GX2570 M8s, a Supermicro OEM server product featuring a high-performance GPU, SupportDesk, and Infrastructure Manager. Starting in July 2025, Fujitsu will combine these OEM products, maintenance services, and integrated management tools with its Takane large language model (LLM) for enterprises to offer managed services for building generative AI infrastructure.
ASAPP announced it has achieved the Amazon Web Services (AWS)Generative AI Competency. Recognition as an AWS partner in this category means the company drives the advancement of services, tools, data strategy, and infrastructure pivotal for implementing generative AI technologies across diverse industries.
Couchbase announced a strategic partnership with Arize AI to equip enterprises with a reliable solution for building and monitoring RAG and agent applications at scale. Via the partnership, organizations can leverage Couchbase Capella as a single data platform for LLM caching, agent memory, vector embeddings, and operational workloads while using Arize AI to ensure high reliability and transparency.
Datadog announced that it has acquired Metaplane. The acquisition will empower data teams to take action on insights and make a broader impact throughout their organizations and across the complete data stack. Additionally, the acquisition will help Datadog accelerate its expansion into data observability, building on launches of related products like Data Jobs Monitoring and Data Streams Monitoring.
OPAQUE Systems announced the availability of its secure AI solution on the Microsoft Azure Marketplace. By integrating confidential computing with popular data and AI tools, OPAQUE lets enterprises process sensitive data fully encrypted, from ingestion to inference, without costly code rewrites or specialized cryptographic skills.
Polyhedra unveiled the Marketplaceon EXPchain, enabling frictionless integration between blockchain applications and AI capabilities. Specifically, the new marketplace will offer developers seamless access to plug-and-play tools for content creation, automation, or data intelligence, allowing builders to supercharge their applications with cutting-edge AI tools.
Qumulo announced the availability of Pay-As-You-Go pricing for Cloud Native Qumulo (CNQ) in AWS Marketplace. The new pricing removes key friction points for customers, allowing them to use CNQ immediately, with costs that scale dynamically with usage.
If your company has real-time analytics news, send your announcements to [email protected].
In case you missed it, here are our most recent previous weekly real-time analytics news roundups:
- Real-time Analytics News for the Week Ending April 19
- Real-time Analytics News for the Week of April 12
- Real-time Analytics News for the Week Ending April 5
- Real-time Analytics News for the Week Ending March 29
- Real-time Analytics News for the Week Ending March 22
- Real-time Analytics News for the Week Ending March 15
- Real-time Analytics News for the Week Ending March 8
- Real-time Analytics News for the Week Ending March 1
- Real-time Analytics News for the Week Ending February 22
- Real-time Analytics News for the Week Ending February 15
