OTT, SVOD, AVOD & FAST Streaming Platform News

YouTube TV has integrated ESPN's app-exclusive live events and library content directly into its linear channel guide, removing the requirement for subscribers to use the standalone ESPN app. This integration shifts viewership data and engagement from Disney's ecosystem to the YouTube TV platform, marking a significant change in their distribution agreement.

Cricket Australia has launched a global D2C streaming service, Cricket TV, powered by Eluvio’s Content Fabric and AI Video Intelligence. The platform supports live and archival content in 4K, utilizing AI for automated vertical video generation and dynamic transcoding without the need for file copies.

Fox has agreed to acquire Roku for $22 billion, a move intended to consolidate its position in the FAST market by combining The Roku Channel with Tubi. This acquisition occurs as major streamers like Disney+, Netflix, and Paramount+ increasingly explore free, ad-supported tiers to combat subscriber churn and expand their addressable markets.

Broadcom is promoting VMware Cloud Foundation 9.1 as a foundational infrastructure for enterprise agentic AI workloads. The company emphasizes the shift toward private cloud environments to address security, governance, and cost-efficiency challenges associated with scaling AI inference.

Adobe Connect 12.12.1 introduces a WebSocket-based fallback for audio and video to bypass restrictive network ports that block WebRTC. The update also includes improved noise suppression and new administrative controls for participant document annotations.

A report from Caretta Research, sponsored by Quickplay, claims that broadcast production teams spend 75% of their time on manual technical tasks. Quickplay uses these findings to promote its Content to Value Operating System as a solution for automating multi-platform distribution.

NASCAR is expanding its digital distribution strategy by leveraging FAST channels on platforms like Tubi and Prime Video alongside gaming integrations on Roblox and iRacing. The organization is shifting its focus toward digital-first content and interactive experiences to reach younger and international audiences beyond traditional linear television.

Sony Electronics and VITEC have announced a strategic collaboration to integrate VITEC's Avedia IPTV and digital signage platform directly into Sony's BRAVIA professional displays. The integration utilizes a system-on-a-chip architecture to eliminate the need for external media players in corporate, hospitality, and entertainment environments.

The global data center switch market is projected to grow from $19.16 billion in 2025 to $43.72 billion by 2035, driven by AI workload demands and 800 GbE infrastructure refreshes. This expansion reflects a broader industry shift toward high-bandwidth, low-latency networking required for hyperscale cloud and edge-compute architectures.

The AI infrastructure supply chain is facing a structural bottleneck due to limited capacity for advanced packaging and high-bandwidth memory (HBM) at TSMC and SK Hynix. This constraint forces streaming and tech firms to choose between optimizing inference efficiency or pursuing vertical integration to secure hardware allocations.

Amazon is rolling out its Alexa+ AI assistant to compatible Fire TV devices in the U.S., enabling natural language search, scene-specific navigation, and smart home integration. The update is being deployed automatically to existing hardware, with additional features available for Prime members or via a standalone subscription.

Sony has launched 'Live TV on PS5,' a FAST service featuring over 100 channels for U.S. PlayStation 5 users. The platform utilizes ad-tech from Publica and PubMatic to monetize content from partners including Lionsgate, Fox, and Crunchyroll.

Rypplzz and Castanet Corporation have announced a partnership to integrate spatial intelligence into hybrid ATSC 3.0 and 5G Broadcast networks. The collaboration aims to provide precise three-axis positioning and timing data for physical AI applications by leveraging high-power UHF spectrum.

Think has launched Think Grid, a managed AI infrastructure service featuring NVIDIA Blackwell accelerators and proprietary ILM orchestration. The service is hosted in Riyadh and targets enterprise and government customers with dedicated bare-metal nodes, aiming to provide a lower-cost alternative to hyperscaler AI compute.

This article provides a technical comparison between FastRTC and Open WebUI, highlighting their distinct roles in the AI stack for streaming professionals. FastRTC is positioned as a framework for real-time WebRTC media transport, while Open WebUI is identified as a platform for model management and AI workspace functionality.

Scalstrm is marketing its cloud-native infrastructure platform, which utilizes just-in-time transcoding and origin shielding to manage high-concurrency live events like the 2026 FIFA World Cup. The vendor claims its architecture reduces hardware requirements by 40x and storage costs by 65% compared to traditional streaming deployments.

AWS has released Glue 6.0, featuring a new Spark Real-Time Mode for sub-second event processing and native support for Apache Iceberg v3. The update also introduces Variant columns for heterogeneous schema handling and Arrow-native UDFs to improve performance in Python-based analytical pipelines.

Resilio has released an early preview of its MCP Server, which allows users to connect their management consoles to AI clients like Claude using the Model Context Protocol. The tool enables natural language troubleshooting, configuration auditing, and custom reporting for large-scale file transfer deployments.

Nvidia is investing $3.5 billion in MediaTek to integrate its NVLink Fusion technology into custom AI chips for hyperscalers. This strategic partnership aims to maintain Nvidia's dominance in data center infrastructure as cloud providers increasingly develop their own proprietary silicon.

AWS has launched its new R9g and R9gd memory-optimized instances powered by Graviton5 processors. These instances are designed for data-intensive workloads, offering significant improvements in compute performance, machine learning inference, and database processing speeds compared to previous generations.

Red Hat has detailed the use of GPU-pruner, an open-source tool designed to monitor NVIDIA DCGM metrics via Prometheus to identify and scale down idle GPU workloads in Kubernetes clusters. The tool aims to optimize hardware utilization for ML platforms like OpenShift AI by automatically scaling parent resources to zero without deleting workload metadata.

Adobe platform engineers have developed a Bi-LSTM predictive autoscaler for Kubernetes to mitigate 20% error rates in GPU workloads by forecasting demand 10 minutes in advance. The system, which runs inference on Prometheus metrics, successfully passed hackathon validation by catching simulated traffic spikes before they occurred.

Neocloud providers are increasingly competing with hyperscalers for AI-native workloads by offering specialized, cost-effective GPU infrastructure. While hyperscalers maintain dominance through enterprise-grade services and custom silicon, neoclouds are carving out a niche for high-performance training and inference tasks.

The European Union's Cyber Resilience Act (CRA) will mandate that manufacturers of products with digital elements report actively exploited vulnerabilities and severe security incidents starting September 11, 2026. Streaming hardware and software providers must establish 24-hour early warning processes to comply with these new reporting requirements ahead of the full regulation rollout in 2027.

llm-d 0.8 introduces multimodal serving capabilities, including token estimation for media assets and cache-aware routing. The update supports disaggregated topologies that separate visual encoding from prefill and decode stages to improve throughput and reduce latency in vision-language model inference.

Norsk and Oracle Cloud Infrastructure will show cloud-based Time-Addressable Media Store workflows at IBC 2026 in Amsterdam. The companies will demo the setup at Norsk stand 5.A53 and Oracle stand 5.A21, and present a joint session on September 11 at Content Everywhere Stage 2.

Norsk is bringing three new products to IBC 2026 at Stand 5.A53 in Amsterdam, including Norsk Playout, Norsk Probe and Norsk Remote Commentary. The company is also introducing Norsk Turnkey, a service for building custom live video workflows, alongside browser-based production tools and software monitoring features.

Datadog is expanding its observability and AI-driven monitoring tools in the ANZ region to help enterprises manage complex cloud and on-premises infrastructure. The platform is being utilized by companies like Flowstate to automate performance monitoring and identify bottlenecks in high-resolution 6K video processing workflows.

Germany currently holds the second-highest density of data centers globally, with 529 locations, primarily centered in the Frankfurt hub. The industry is shifting from traditional CPU-based infrastructure to high-density, liquid-cooled AI facilities, necessitating significant upgrades in power capacity and grid connectivity to maintain competitiveness.

Latvian communications operator Balticom experienced a cyber attack on an interactive TV server hosted in Bulgaria, resulting in the unauthorized broadcast of Russian propaganda to 5% of its subscribers. Investigations by Cert.lv confirmed the breach occurred at a third-party content delivery partner rather than within Balticom's internal infrastructure.

Akamai's 2026 Enterprise AI Usage Risk Report indicates that 5% of power users generate 12 times the AI activity of average employees, frequently utilizing unvetted browser and IDE extensions. The report highlights significant security risks, including the use of personal identities for corporate AI tasks and the presence of known vulnerabilities in nearly 16% of AI-related extensions.

IBM has announced a dual-ISA processor design for its next-generation z/Architecture, enabling native support for both z/Architecture and Arm instruction sets to improve software ecosystem compatibility. Additionally, the company is integrating HBM into its Spyre chips to better support agentic AI workflows and large language models in enterprise environments.

PlayBox Technology has launched Cosmos, a centralized media orchestration platform designed to manage linear programming, playout, and distribution for FAST and OTT channels. The platform aims to help broadcasters and startup operators scale multi-channel portfolios by automating workflows and centralizing control.

A software bug in the Roku OS 15.3 update is causing various Roku streaming players and smart TVs to spontaneously power on and initiate content playback during the night. While Roku has not yet provided a fix, users are reporting that disabling features like Fast TV Start and HDMI-CEC may mitigate the issue.

SpaceX is establishing an in-house foundry in Bastrop, Texas, to manufacture gas turbine blades and vanes, aiming to bypass global supply chain bottlenecks. This initiative is intended to accelerate the deployment of natural gas power plants required to support the growing energy demands of AI data centers.

fal has released H3 Max, a video generation model capable of producing 5-second clips in under 3 seconds. The model utilizes optimized kernels and post-training to achieve high-speed inference, targeting interactive and rapid-iteration creative workflows.

Sister Group has entered into a production and distribution partnership with broadcaster Amol Rajan to support his transition to a direct-to-audience creator model. The deal provides Rajan with infrastructure for YouTube and podcast distribution, including production services and audience data management.

Okta has launched Agent SSO, a new feature that registers AI agents as first-class identities using the Cross App Access (XAA) standard. The initiative aims to replace static credentials with short-lived tokens to improve security for non-human identities in enterprise workflows.

Salad and Vast.ai are competing in the distributed GPU cloud market, offering compute resources for AI and streaming workloads. The article provides a comparative overview of their pricing models, GPU availability, and service offerings for enterprise and developer use cases.

AWS has implemented stricter credit approval processes, including mandatory vice-president-level sign-offs and identity verification, following a massive fraud scheme involving the exploitation of Bedrock inference token discounts. The fraud, primarily orchestrated by proxy operators in the Greater China and Asia-Pacific regions, resulted in hundreds of millions of dollars in bad debt.

ESPN is launching a new streaming feature called RedZone at the U.S. Open, which provides live whip-around coverage across 16 courts simultaneously. The feature leverages the RedZone brand acquired from the NFL to deliver curated, high-leverage match highlights to viewers.

Groundcover provides a technical analysis of eBPF monitoring overhead in Kubernetes environments, detailing how event rates and hook types influence CPU and memory consumption. The guide offers platform engineers methodologies to measure in-kernel costs and optimize performance using kernel-level filtering and efficient data transport.

Independent 2026 benchmarks indicate that serverless cold-start latencies for major cloud providers have dropped significantly due to snapshot-based restore techniques. These improvements, alongside new billing models for initialization phases, are making event-driven functions increasingly viable for latency-sensitive streaming workloads.

A developer reports reducing infrastructure costs for a live scoreboard application from $1,100 to $16 per month by migrating from a managed AWS EKS and ElastiCache stack to a single 4-core server running Go, SQLite, and Caddy. The article highlights the efficiency gains of simplifying architecture for read-heavy workloads with 10,000 concurrent users.

Nvidia is shifting its strategic focus toward data orchestration and system-level efficiency with its new Vera Rubin architecture, moving beyond standalone GPU dominance. This approach addresses memory bottlenecks and data movement challenges in megascale data centers, positioning the company to compete with integrated chip designs from hyperscalers like OpenAI.

Factor House outlines five pillars for Kafka data governance, focusing on schema evolution, topic lifecycle management, PII guardrails, access control, and lineage. The guide emphasizes implementing these policies through automated configuration and code rather than manual cluster topology to ensure scalability and security.

This article provides a comparative cost analysis of AWS and Google Cloud compute, storage, and Kubernetes services as of August 2026. It highlights that while Google Cloud offers lower on-demand compute prices, AWS often provides better value for committed-use scenarios and ARM-based workloads.