AI infrastructure providers

Nvidia has unveiled its next-generation co-packaged optics (CPO) technology, according to industry reporting. CPO integrates optical components directly with silicon to move data faster while cutting power consumption, a key consideration for AI data centers. Further details on specifications, products, and availability were not provided in the initial report.

News coverage in the semiconductor sector focuses on how companies supplying chips for artificial intelligence are shifting toward strategic partnerships to develop infrastructure. The discussion centres on collaborations between chipmakers, cloud providers and data centre operators, alongside ongoing innovation in AI accelerator hardware, as demand for computing power continues to reshape the industry's business models.

Nebius, the AI infrastructure provider listed on Nasdaq, is drawing investor attention as demand for computing capacity surges. The company, which emerged from Yandex and rebranded after divesting its Russian assets, is seen alongside peers like CoreWeave as a beneficiary of the AI data-center build-out. Market watchers are tracking its stock amid heavy spending on GPU capacity by hyperscalers and AI labs.

Chinese lidar maker Hesai has launched Kosmo, a 3D data platform designed to support robot training and the development of world models. The platform aims to provide spatial data infrastructure for robotics and embodied AI applications, extending Hesai's business beyond automotive lidar sensing into the growing robotics sector.

German hosting provider Hetzner has introduced a new dedicated GPU server, the GEX63, aimed at professional AI and machine learning workloads. The launch is drawing attention among developers and infrastructure enthusiasts, who are discussing the pricing and hardware configuration as an alternative to larger cloud providers for training and inference work.

A developer has released Talorys, an open-source personal AI agent that can be self-hosted entirely on Cloudflare's free tier. The project, hosted on GitHub, lets users run their own AI assistant without paying for infrastructure or relying on third-party hosted services. It is drawing attention among developers interested in low-cost, self-custodied AI tooling.

Russian telecom operator MTS has introduced a virtual secretary feature that answers missed calls using a synthetic copy of the subscriber's own voice. Coverage across Russian tech outlets highlights the AI voice-cloning capability, while MTS Exolve also announced expanded cloud infrastructure via colocation in MWS Cloud. Commentators are also weighing what 5G and 4G differences mean practically for regions like Bashkortostan, where MTS is a major provider.

A writer and self-described LLM critic is calling for artificial intelligence models that individuals can train themselves on their own machines, with no data centers required. The alternative proposed is server-based models that are highly specialized and trained only on open data. The appeal reflects a broader debate among AI watchers about user control, privacy, and concentration of AI infrastructure in large companies, resonating with people who want more transparency and independence from big tech providers.

An analysis in The Conversation examines whether the current AI investment boom rests on circular money flows, with major tech companies investing in, and simultaneously buying services from, the same AI providers. The piece asks who is really paying whom, and whether the arrangement reflects genuine economic value or a bubble vulnerable to correction.

AI cloud company Prime Intellect has acquired a startup focused on making model inference faster, according to a report by The Information. The deal signals continued consolidation in the AI infrastructure space, as providers race to cut the cost and latency of running large models. Details of the target and price were not disclosed.

Google has launched Gemini 4 Argon, its newest flagship artificial intelligence model, making it available to developers and business customers. The release continues the company's push to compete in the enterprise AI market, with the model expected to power applications and services built on Google's AI infrastructure. Further details on pricing, capabilities and availability were not immediately provided.

Mirantis announced that its k0rdent AI platform now integrates the open-source NVIDIA Run:ai scheduler, aiming to provide full-stack operations for NVIDIA DSX AI Factory environments. The move lets enterprises orchestrate AI workloads and GPU resources across Kubernetes clusters through a single management layer, positioning k0rdent as a control plane for AI infrastructure deployments built on NVIDIA's reference architecture.

GPU cloud provider RunPod has named a new chief financial officer whose stated priorities are scaling the company's finance operations and supporting its AI infrastructure buildout. The appointment, reported by CFO Dive, signals the fast-growing AI cloud firm is professionalizing its financial leadership as demand for compute capacity keeps rising.

French cloud provider Scaleway is set to build a new AI supercomputer in Sweden powered by AMD chips. The announcement puts AMD in direct competition with Nvidia in the AI infrastructure market and adds to Europe's growing capacity for training large AI models.

AI cloud provider Nscale has announced the acquisition of Anyscale, the company behind the Ray distributed computing framework. The deal is intended to strengthen Nscale's full-stack AI cloud platform, adding Anyscale's software and engineering talent to its infrastructure business. Further details on the transaction's value and terms have not been disclosed.

Microsoft plans to broaden its Azure artificial intelligence hardware lineup with the Maia 300 custom AI chip, slated for 2027. The move is part of the company's effort to reduce reliance on outside chip suppliers and scale its AI cloud infrastructure, though details on performance, manufacturing partner, or rollout timeline beyond the year remain limited.

Nebius has added NVIDIA Run:ai to its AI Cloud Applications catalog, making the open-source scheduling and orchestration layer available for customers running GPU workloads. The integration lets teams manage and optimize AI training and inference clusters directly through Nebius's cloud platform. The announcement points to tightening cooperation between Nebius and NVIDIA as demand for AI infrastructure keeps climbing.

Shpyrd, a new cloud computing service, is presenting itself as a cloud designed specifically for the AI era, promising infrastructure tailored to artificial intelligence workloads rather than traditional cloud hosting. Details remain sparse beyond its website, with early reactions focused on curiosity about how it differs from established providers like AWS, Google Cloud and Azure.

The French cloud provider Scaleway plans to build a supercomputer dedicated to artificial intelligence in Sweden. The announcement has circulated in Swedish tech discussions, highlighting growing Nordic investment in AI infrastructure and high-performance data centers, as European companies race to expand computing capacity for AI workloads.

Community College Daily publishes an argument that community colleges should build and own their own AI infrastructure rather than relying on licensed products from outside vendors. The piece frames ownership as a way for two-year institutions to control costs, protect student data, and tailor AI tools to their teaching missions. It adds to a wider debate in higher education over dependence on commercial AI providers.

SpaceX is reportedly in negotiations to provide the Pentagon with AI computing power under a contract worth billions of dollars. The talks would mark a significant expansion of Elon Musk's company beyond launch services into defense computing infrastructure. Details on the scale, timeline and scope of the potential deal have not been confirmed by either SpaceX or the Department of Defense.

Nscale, the AI cloud infrastructure company, is being discussed in connection with energy and power, reflecting growing attention to the electricity demands of large-scale data centres used for artificial intelligence workloads. Observers are weighing how much power Nscale's operations require and what that means for its expansion plans, though no specific announcement has been confirmed.

The Healthcare Sector Coordinating Council has published an AI clinical resilience playbook aimed at helping healthcare organizations deploy artificial intelligence safely. The guidance covers managing patient safety risks and cybersecurity threats that come with clinical AI systems. It is intended to give hospitals and other providers a practical framework for adopting AI while protecting patients and network infrastructure.

Microsoft is pushing forward with its own internally developed AI chip, a move aimed at reducing its reliance on Nvidia's dominant graphics processors for artificial intelligence workloads. The announcement underscores a broader race among major tech companies to build in-house silicon for data centers and control their AI infrastructure costs.

Marvell Technology says it will present its AI data center solutions at the OCP Global Summit 2026. The company, a major supplier of custom chips and networking components for artificial intelligence infrastructure, is expected to highlight technology aimed at hyperscale data centers. The announcement comes as chipmakers increasingly use industry events to promote their AI hardware roadmaps to cloud providers and investors.

Microsoft cloud chief Scott Guthrie discussed the company's new custom AI chip for Azure in an interview, saying demand for AI computing power is 'insatiable'. He also addressed the ongoing price war among AI model providers, as cloud and chip rivals compete aggressively on cost and capacity to serve booming generative AI workloads.

Analysts argue Australia can strengthen its AI sovereignty by hosting and adapting foreign open-weight models domestically rather than relying on them as-is. Running the models on local infrastructure would make them more secure and better suited to Australian needs, even though their parameters were trained abroad. The discussion centres on reducing dependence on foreign providers and mitigating risks tied to overseas-controlled AI systems.

A new report asks whether artificial intelligence can foresee internet outages before they occur. The piece explores how AI models might analyse network traffic patterns to detect early warning signs of disruptions, potentially allowing providers to fix problems proactively. Interest reflects growing concern about network reliability as connectivity becomes essential infrastructure for work, commerce and daily life.

Engineers and software commentators are discussing why token prices make up only a small fraction of the real cost of running AI features in production. The argument is that the total bill is an architecture problem: infrastructure, latency, fallbacks, monitoring, and engineering time around an AI feature often outweigh what providers charge per token.

A new policy discussion is examining how the European Union can use legislation to reduce its dependence on foreign cloud providers and secure digital infrastructure within Europe. The topic touches on regulation, data protection and AI readiness, with supporters arguing that legal tools are Europe's strongest lever to build an independent cloud sector amid geopolitical tensions and reliance on US-based technology companies.