AWS

Together AI has secured a $240 million agreement with IBM Cloud to deploy large-scale Nvidia HGX B300 infrastructure. The systems, representing Nvidia's previous generation of hardware, are scheduled to launch in the first quarter of 2027. This partnership highlights a strategic move to secure substantial compute capacity despite the hardware not being the current flagship.

• Together AI locks in $240M for IBM Cloud-hosted Nvidia HGX B300 capacity
• Hardware is last-gen B300, not the latest AI accelerator models
• Deployment timeline targets Q1 2027 for system launch
• Signals continued demand for large-scale GPU clusters via cloud providers
• No immediate impact on existing database or cloud engineering fleets

The UK government has established a £14 billion cloud framework for public sector procurement, explicitly designed to increase opportunities for small and medium-sized enterprises. This move aims to counter the current market dominance of global cloud providers by encouraging buyers to diversify their supplier base. The initiative signals a policy shift toward supporting local tech suppliers within critical infrastructure contracts.

• £14B framework targets public sector cloud procurement across UK government bodies.
• Policy explicitly favors SMEs to challenge the dominance of global hyperscalers.
• Buyers are encouraged to diversify suppliers beyond the usual major cloud vendors.
• Local tech firms gain a structured pathway into high-value government contracts.
• Framework implementation begins August 2026 with immediate procurement opportunities.

Amazon OpenSearch Service now supports combining full-text, vector, and geospatial searches within a single query against JSON documents. The blog demonstrates this capability using a restaurant discovery application as a concrete working example. This approach allows developers to build complex discovery layers without managing separate search engines for each data modality.

• Unified query interface handles text, vectors, and location data in one request.
• Simplifies architecture by replacing multiple specialized search stores with OpenSearch.
• JSON-native support eliminates schema transformation overhead for document storage.
• Practical example provided for building composite search experiences.

OVH Cloud has announced significant price increases, with gaming server rates rising by up to 87%, to offset the financial impact of the RAMpocalypse. The cost adjustments also apply to conventional instances, affecting both legacy and newer hardware configurations. This move reflects the broader industry struggle with memory-related security vulnerabilities and their remediation expenses.

• Gaming workloads face the steepest 87% rate increases at OVH Cloud.
• Standard instances on both old and new hardware see price hikes.
• Costs are driven by the financial burden of the RAMpocalypse.
• Review current instance types for potential budget impact.

AWS Database Blog details how to configure connection pooling for the Advanced JDBC Wrapper on Aurora and RDS using a new assistant tool. The post clarifies the architectural differences between internal and external pooling modes and guides users in selecting the appropriate strategy. This assistant automates the generation of correct configuration parameters to optimize database connectivity.

• AWS provides a configuration assistant to automate JDBC wrapper pooling setup
• Clarifies distinctions between internal and external pooling modes in the wrapper
• Helps practitioners select optimal pooling strategy for Aurora and RDS workloads
• Reduces configuration errors by generating correct parameters via the assistant