Rakuten Kobo rejected 45% of titles submitted to its Kobo Writing Life self-publishing platform in 2025, with CEO Michael Tamblyn attributing more than 80% of those rejections to books he deemed AI-generated. The move signals mounting pressure to filter synthetic content and reignites the debate around quality control, whether on a global storefront or within an organization’s own LLM pipeline.
Samsung SDI invests $20 million in Forge Nano to build a local battery supply chain in the United States. This move reduces supply risks and underscores the importance of industrial sovereignty, with direct implications for the reliability of on-premise tech infrastructure.
Alibaba.com is launching Accio Work, an agentic AI suite that automates export operations for Malaysian SMEs. The system autonomously executes instructions, from market research to cross-border logistics, operating 24/7. The launch includes a RM500,000 pitch competition to drive local adoption. However, removing human oversight can amplify pricing errors, inventory mismatches, and supplier risks. Alibaba acknowledges the risk, positioning automation to free teams for strategic decisions rather than eliminate control.
Licenses bought, tools distributed, yet adoption remains low. Atheni AI, founded by two unconventional entrepreneurs, focuses on personalized coaching to integrate AI into real workflows. A lesson that also applies to on-premise deployments, where the risk of shelfware is extremely high.
Acer's distribution arm Weblink is betting on the new MacBook Neo and software bundles to offset weak PC demand. A clear signal of how the channel is recalibrating its strategies, with one eye on margins and the other on value-added services.
Taiwan-based IPC maker IEI Integration saw its margin fall to 28% in Q1, but a rebound is expected. The figure highlights supply dynamics critical for enterprises relying on edge hardware for on-premise AI deployments.
The exploding intelligent eyewear market, highlighted by DIGITIMES, forces a rethink of architectures that bring LLMs directly onto devices, balancing privacy, latency, and data sovereignty.
Z.ai's valuation surge reignites the debate over China's AI bubble. Between revolutionary model promises and well-founded skepticism, those evaluating on-premise deployment must separate technical substance from market noise.
Groundhog Technologies marks a world first with a commercial deal for generative AI in telecom, and now targets LEO satellite network optimization. The move highlights rising demand for on-premise and edge AI in critical sectors, where latency and data sovereignty drive self-hosted architectures.
OpenAI has chosen a phased release for its new LLM GPT 5.6, officially citing U.S. regulatory uncertainty. The move signals a strategic shift with direct implications for on-premise deployment, data sovereignty, and the architectural choices of enterprises pursuing self-hosting. AI-RADAR analyzes the potential scenarios.
Memory shortages are tightening supply chains, and for Apple waiting is no longer viable. With M-series chips and growing AI ambitions, access to advanced memory is critical. Postponing investments means losing ground to competitors. For those designing on-premise AI infrastructure, the shortage makes hardware costs and availability more unpredictable, demanding careful TCO assessments.
Japan's government has earmarked 101.6 trillion yen for AI chip development, placing it at the centre of a 370 trillion yen growth roadmap. The move resets industrial priorities and raises practical questions for those assessing on-premise deployment: hardware availability, data sovereignty, and shifting total cost of ownership dynamics.
With the Dragonfly initiative, Qualcomm is doubling down on cloud AI, moving beyond its mobile chip heritage. The push could reshape the data center accelerator landscape, offering potential alternatives to NVIDIA and new options for organizations considering on-premise deployment.
Alibaba’s chip unit strengthens its financial base to accelerate AI hardware development. The move highlights a rush for technological sovereignty at the intersection of RISC-V, on-premise scenarios, and the need to bypass geopolitical semiconductor restrictions.
Accumulated deferred fixes, misconfigurations, and operational inefficiencies cost US enterprises $2.41 trillion annually, with a $1.52 trillion price tag to remediate. Jay Roland, founder of Varex Solutions, decries industry complacency. For on-premise AI adoption, this debt stalls infrastructure modernization and data sovereignty efforts, making long-term costs unsustainable.
Founded by former Meta researchers, the startup raised a significant round to meet nearly insatiable demand for testing autonomous agents. Its simulated environments promise to reduce risk before deployment, but for on-premise teams, agent reliability demands validation strategies adapted to local constraints and data sovereignty.
Notion will shut down Notion Mail on September 22, less than 18 months after launch. The reason reflects a broader shift: AI agents handle triage, responses, and scheduling without users ever opening an inbox. This move raises questions about data privacy and control, especially for those evaluating on-premise deployment to avoid emails being processed by external models.
The Trump administration proposes eliminating the brake pedal requirement for vehicles designed exclusively for autonomous driving. The regulatory change would remove a major barrier, accelerating designs with no human controls. For on-prem AI, the news triggers reflections on liability, safety, and the disappearing manual override in automated systems.
Indagari's analysis of credit card transactions from 28 million US consumers shows a 75% jump in Claude's paying user base and revenue since January, reshaping the rivalry with ChatGPT and raising questions for those evaluating LLMs for on-premise deployment.
After acquiring Skiff and launching Notion Mail, the company shelves the client and bets on AI agents. A shutdown that reflects a deeper shift: email automation moves from clients to artificial intelligence, with implications for organizations managing their own data.