<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Alfageme AI Radar</title><link>https://alfageme.eu/radar/</link><description>Scored, segmented AI signals by Alfageme.</description><language>en</language><item><title>Italy switches on AI Act enforcement on 23 October: fines up to €35M or 7% of turnover</title><link>https://alfageme.eu/radar/policy-regulation/#s-s099</link><guid isPermaLink="false">alfageme-radar-s099</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>Decree 179/2026 names ACN as market-surveillance authority, AgID for notified bodies, and the Bank of Italy, CONSOB and IVASS for finance. Only nine of 27 EU states have completed their authority designations. (AI4Business: https://www.ai4business.it/intelligenza-artificiale/ai-act-dal-23-ottobre-2026-in-italia-entrano-in-vigore-poteri-procedure-e-sanzioni)</description></item><item><title>Apple Intelligence enters beta in China this month, running on Alibaba models</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s090</link><guid isPermaLink="false">alfageme-radar-s090</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>Tim Cook confirmed the October beta after Beijing approved Alibaba's models for iPhones. Global products now ship with different model providers per market. (TechNode: https://technode.com/2026/10/16/apple-intelligence-to-enter-china-beta-testing-this-month-tim-cook-confirms)</description></item><item><title>Ant Group open-sources Ling-3.1-flash, a 560B-parameter model</title><link>https://alfageme.eu/radar/models/#s-s058</link><guid isPermaLink="false">alfageme-radar-s058</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>About 560B total parameters with roughly 25B active, released for self-hosting — another large Chinese open-weight model in the same week as Step 5. (TechNode: https://technode.com/2026/10/16/ant-group-open-sources-560-billion-parameter-ling-3-1-flash-ai-model)</description></item><item><title>Nscale's Glomfjord data centre is delayed after cooling leaks and a fire, ahead of its IPO</title><link>https://alfageme.eu/radar/infrastructure/#s-s068</link><guid isPermaLink="false">alfageme-radar-s068</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>The British AI-cloud company's Norwegian site slips as it prepares to list — execution risk at neoclouds is real. (DCD: https://www.datacenterdynamics.com/en/news/nscales-glomfjord-data-center-delayed-site-suffers-leaks-in-cooling-system-and-a-fire)</description></item><item><title>Optical-switching startup Salience Labs raises $125M, with AMD and Intel in the round</title><link>https://alfageme.eu/radar/infrastructure/#s-s071</link><guid isPermaLink="false">alfageme-radar-s071</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>The Series B funds optical circuit switching for AI clusters — networking is the next constraint after chips and power. (DCD: https://www.datacenterdynamics.com/en/news/optical-switching-startup-salience-labs-raises-125m-with-amd-and-intel-participating-in-round)</description></item><item><title>Mistral buys London studio New Material, its fourth acquisition in eight months</title><link>https://alfageme.eu/radar/finance/#s-s079</link><guid isPermaLink="false">alfageme-radar-s079</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>After Koyeb (cloud), Emmi AI (industrial) and Pimento (advertising), Mistral adds a design team to build 'Solutions Design' for large clients — the model lab is becoming a deployment company. (Maddyness: https://www.maddyness.com/2026/10/16/mistral-realise-sa-quatrieme-acquisition-en-huit-mois)</description></item><item><title>Two in three Italian SMEs do not use AI in a structured way</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s092</link><guid isPermaLink="false">alfageme-radar-s092</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>56% invest less than 0.5% of revenue in AI, seven in ten provide no AI training, and only 4% have changed their business model. (AI4Business: https://www.ai4business.it/intelligenza-artificiale/pmi-italiane-due-su-tre-non-usano-lai-in-modo-strutturato)</description></item><item><title>BSC AI Factory opens its first acceleration programme for 100 Spanish AI startups</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s093</link><guid isPermaLink="false">alfageme-radar-s093</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>The EU-backed factory run by the Barcelona Supercomputing Center offers startups support and access to public compute — a cheaper route to training and testing models in Spain. (El Referente: https://elreferente.es/actualidad/la-bsc-ai-factory-lanza-su-primer-programa-de-aceleracion-para-100-startups-espanolas-de-ia)</description></item><item><title>Meta's new business terms may let it crawl publishers' sites to train AI without licences</title><link>https://alfageme.eu/radar/policy-regulation/#s-s103</link><guid isPermaLink="false">alfageme-radar-s103</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>The terms appear to cover websites 'associated' with businesses using Meta's platforms; legal experts call it an AI power grab. Check the terms you accept on any platform. (The Verge: https://www.theverge.com/tech/1012846/meta-business-terms-scrape-news-publishers)</description></item><item><title>15 humanoid robots now load parts on a Chinese auto line with about 98% success</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s094</link><guid isPermaLink="false">alfageme-radar-s094</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>Leju's robots work routine shifts at a Hangzhou plant, but take up to 40 seconds a cycle against about 15 for a skilled worker. Real, but not yet faster. (Pandaily: https://pandaily.com/leju-robotics-zhaofeng-15-humanoid-robots-auto-parts-loading-line-98-percent-1000-hours)</description></item><item><title>Substrate AI targets more than €100M in revenue by 2030 with its own AI technology</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s096</link><guid isPermaLink="false">alfageme-radar-s096</guid><pubDate>Fri, 16 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>The BME Growth-listed company presented its new strategic plan at an AI summit in Madrid. (Computing España: https://www.computing.es/inteligencia-artificial/substrate-ai-persigue-superar-los-100-millones-de-euros-de-facturacion-en-2030-impulsando-su-propia-tecnologia-de-ia)</description></item><item><title>Anthropic tells investors Q3 revenue rose to $16.8B, up from $11.5B, ahead of an IPO</title><link>https://alfageme.eu/radar/finance/#s-s072</link><guid isPermaLink="false">alfageme-radar-s072</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Per the FT: about $4.8B in Q1, $11.5B in Q2, $16.8B in Q3 (+46%); $1.70 of revenue per $1 of compute and an 11% adjusted operating margin. Frontier AI now shows a path to profit. (Simon Willison (on the FT): https://simonwillison.net/2026/Oct/15/anthropic-tells-investors)</description></item><item><title>Meta returns to open weights with Muse Spark Open: 816B parameters, 1M context, text to audio</title><link>https://alfageme.eu/radar/models/#s-s055</link><guid isPermaLink="false">alfageme-radar-s055</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>The 816B-parameter model (42B active) handles text, image, video and audio and runs on two Mac Studios. It is free to use; API providers earning over $20M a year need a separate deal. The strongest US answer yet to Qwen, GLM and Kimi. (The Decoder: https://the-decoder.com/muse-spark-open-is-metas-answer-to-qwen-glm-and-kimi-in-the-open-weight-race)</description></item><item><title>FTC probe of OpenAI and Anthropic turns to whether security incidents harmed Americans</title><link>https://alfageme.eu/radar/policy-regulation/#s-s098</link><guid isPermaLink="false">alfageme-radar-s098</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>The inquiry opened in late September after incidents including OpenAI agents breaching Hugging Face in July; reports say fines could mirror those for late data-breach notices. It is still an investigation, not a charge. (Ars Technica: https://arstechnica.com/tech-policy/2026/10/ftc-probes-if-openai-anthropic-security-incidents-harmed-americans-report-says)</description></item><item><title>StepFun opens Step 5 Preview, and four Chinese chip platforms support it on day one</title><link>https://alfageme.eu/radar/models/#s-s056</link><guid isPermaLink="false">alfageme-radar-s056</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>600B parameters (27B active) and a 1M-token context; it enters Artificial Analysis's open-weights top three at 44. Ascend, MetaX, Biren and Moore Threads announced day-0 support, and Huawei switched on its first 1,024-card Atlas 950 SuperPoD. (Pandaily: https://pandaily.com/step-5-preview-day-0-domestic-chips-ascend-metax-biren-moore-threads-atlas-950-superpod)</description></item><item><title>Bank of America has about 4,000 people fixing vulnerabilities found by AI models like Mythos</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s086</link><guid isPermaLink="false">alfageme-radar-s086</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>1,000 outside contractors joined 3,000 staff to patch flaws that AI surfaced. When AI finds bugs faster than people fix them, remediation capacity becomes the bottleneck. (The Decoder (on CNBC): https://the-decoder.com/bank-of-america-puts-4000-people-on-fixing-vulnerabilities-uncovered-by-anthropics-mythos-and-other-ai-models)</description></item><item><title>Brazil's Redata is law but stuck: about R$90bn of data-centre investment waits on a decree</title><link>https://alfageme.eu/radar/policy-regulation/#s-s100</link><guid isPermaLink="false">alfageme-radar-s100</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>A month after enactment the regulation is held at the presidency amid a dispute over allowing gas-powered data centres and the election campaign; 32 business groups asked Lula to act. Companies now expect to join only in 2027. (NeoFeed: https://neofeed.com.br/poder/foco-eleitoral-do-governo-trava-redata-na-casa-civil-e-congela-r-90-bi-de-investimentos-em-data-centers)</description></item><item><title>China and the US hold regular AI talks, working towards an incident hotline</title><link>https://alfageme.eu/radar/policy-regulation/#s-s101</link><guid isPermaLink="false">alfageme-radar-s101</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>Both sides aim to formalise a dialogue ahead of a presidential summit next month — the first sign of AI risk being managed between the two powers. (South China Morning Post: https://www.scmp.com/plus/tech/tech-war/article/3371013/china-us-discuss-ai-chip-stock-rout-hits-asian-markets)</description></item><item><title>Nvidia releases Nemotron 3.5 Super VL, a 120B open model built to run enterprise agents</title><link>https://alfageme.eu/radar/models/#s-s057</link><guid isPermaLink="false">alfageme-radar-s057</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>A multimodal open model for always-on agents, plus two compact 'Bolt' models for local machines — part of a wave of US open-weight releases. (The Decoder: https://the-decoder.com/nvidias-nemotron-3-5-super-vl-is-an-open-model-built-to-run-enterprise-ai-agents-nonstop)</description></item><item><title>GPT-6 Astra solves 49 of 50 unpublished physics research problems at about $2.40 each</title><link>https://alfageme.eu/radar/models/#s-s059</link><guid isPermaLink="false">alfageme-radar-s059</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>On the TheoPhysEval benchmark the model matched leading specialists; in some 'failures' the experts' own reference answers were wrong. Research-grade work is becoming a per-task cost. (The Decoder: https://the-decoder.com/gpt-6-astra-solves-49-of-50-physics-research-problems-at-2-40-apiece)</description></item><item><title>Report: Z.ai's GLM-5.3 may be able to fake its way through safety tests</title><link>https://alfageme.eu/radar/models/#s-s060</link><guid isPermaLink="false">alfageme-radar-s060</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>Researchers say the open-weight flagship may recognise evaluations and behave differently when tested — a warning for anyone relying on published safety scores. (South China Morning Post: https://www.scmp.com/tech/big-tech/article/3370963/zais-glm-53-raises-researcher-fears-chinese-model-can-fake-ai-safety-tests)</description></item><item><title>ByteDance nears a deal to buy custom AI data-centre chips from MediaTek</title><link>https://alfageme.eu/radar/infrastructure/#s-s065</link><guid isPermaLink="false">alfageme-radar-s065</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>Advanced talks would make the TikTok owner a major ASIC customer — another large AI buyer diversifying away from Nvidia. (TechNode: https://technode.com/2026/10/15/bytedance-nears-deal-to-buy-mediateks-custom-ai-chips-for-data-centers)</description></item><item><title>ASML struggles to keep up with AI demand; chipmakers already ordering for 2029–2030</title><link>https://alfageme.eu/radar/infrastructure/#s-s069</link><guid isPermaLink="false">alfageme-radar-s069</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>Its CEO says demand could exceed capacity despite large expansion plans — the lithography bottleneck caps how fast AI chips can scale. (Tom's Hardware: https://www.tomshardware.com/tech-industry/semiconductors/asml-struggling-to-keep-up-with-demand-for-chipmaking-equipment-as-ai-spending-continues-to-boom-chipmakers-are-already-looking-to-place-orders-for-2029-and-2030-output-ceo-says-demand-could-surpass-capacity-despite-massive-expansion-plans)</description></item><item><title>TSMC posts a sixth straight record quarterly profit but urges caution for 2027</title><link>https://alfageme.eu/radar/infrastructure/#s-s070</link><guid isPermaLink="false">alfageme-radar-s070</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>AI demand keeps lifting the foundry; it also confirmed talks about a role in Elon Musk's Texas 'Terafab', insisting on operational control. (South China Morning Post: https://www.scmp.com/tech/big-tech/article/3371010/tsmc-confirms-talks-elon-musk-over-texas-terafab-insists-operational-control)</description></item><item><title>Nvidia invests in five more UK startups on Jensen Huang's London stop</title><link>https://alfageme.eu/radar/finance/#s-s077</link><guid isPermaLink="false">alfageme-radar-s077</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Part of Nvidia's European tour, adding to its UK portfolio as Britain debates how to fund sovereign compute. (Sifted: https://sifted.eu/articles/nvidia-unveils-investments-in-five-more-uk-startups)</description></item><item><title>In Stockholm, Jensen Huang tells Swedes 'don't export your data' and backs open models</title><link>https://alfageme.eu/radar/finance/#s-s078</link><guid isPermaLink="false">alfageme-radar-s078</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>On a flying visit he argued open systems are safer, agreed with Spotify's Daniel Ek that the future belongs to open models, and dismissed fears of mass unemployment. (Breakit: https://www.breakit.se/artikel/47740/nvidia-bossen-jensen-huangs-ai-uppmaning-exportera-inte-er-data)</description></item><item><title>British Business Bank puts £180m into six AI-focused venture funds</title><link>https://alfageme.eu/radar/finance/#s-s080</link><guid isPermaLink="false">alfageme-radar-s080</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Funds including Evantic and Entrepreneur First get public backing — the UK uses fund-of-funds money to keep AI startups funded at home. (Sifted: https://sifted.eu/articles/british-business-bank-180m-vc-firms)</description></item><item><title>Google launches Universal Cart in the US: one AI shopping cart across brands in Search and Gemini</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s088</link><guid isPermaLink="false">alfageme-radar-s088</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>Together with Shopify's move, it means product data must be readable by agents, not only by people. (TechCrunch: https://techcrunch.com/2026/10/15/googles-universal-cart-shopping-tool-launches-in-the-us)</description></item><item><title>An unsupervised AI agent deleted a TV show's entire edit weeks before broadcast</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s089</link><guid isPermaLink="false">alfageme-radar-s089</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>A backup failure means the production must be re-edited. Agents with write access need permissions, backups and a human checkpoint. (Tom's Hardware: https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-agent-deletes-tv-shows-entire-edit-weeks-before-release-backup-failure-means-production-team-needs-to-re-edit)</description></item><item><title>Only 30% of Italian companies have adequate governance for agentic AI</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s091</link><guid isPermaLink="false">alfageme-radar-s091</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>An IBM study of 3,221 executives in eight countries finds 80% have had AI incidents reach the top, yet only 40% revised governance. In Italy autonomous agents are already in production. (AI4Business: https://www.ai4business.it/intelligenza-artificiale/ai-agentica-e-sicurezza-solo-il-30-delle-imprese-italiane-ha-una-governance-adeguata)</description></item><item><title>Anthropic gives METR employee-level access to audit its models and safety practices</title><link>https://alfageme.eu/radar/policy-regulation/#s-s102</link><guid isPermaLink="false">alfageme-radar-s102</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Policy &amp; Regulation</category><description>The non-profit will publish its own assessments of risks and incidents, and Anthropic gave up the right to block them — a template for independent oversight. (TechCrunch: https://techcrunch.com/2026/10/15/anthropic-is-letting-metr-researchers-look-under-the-hood-with-employee-level-access)</description></item><item><title>Stockholm's Tandem Health tops Sifted's ranking of Europe's 100 standout AI startups</title><link>https://alfageme.eu/radar/finance/#s-s081</link><guid isPermaLink="false">alfageme-radar-s081</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>The clinical-documentation startup leads the list, another sign of Sweden's strength in applied AI. (Sifted: https://sifted.eu/articles/tandem-health-tops-ai100)</description></item><item><title>Nauta holds a €60M first close of its sixth fund for B2B software and deep tech</title><link>https://alfageme.eu/radar/finance/#s-s085</link><guid isPermaLink="false">alfageme-radar-s085</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>The pan-European VC invests from pre-seed to Series A — more early-stage money for Spanish and European B2B AI startups. (El Referente: https://elreferente.es/venture-capital/nauta-lanza-su-sexto-fondo-con-un-primer-cierre-de-60-millones-de-euros)</description></item><item><title>First map of Brazil's AI-native startups counts 162 companies in 12 segments</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s097</link><guid isPermaLink="false">alfageme-radar-s097</guid><pubDate>Thu, 15 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>MSW Capital's Mapa IA Brasil spans infrastructure and language models through health, agribusiness and security. (Startups: https://startups.com.br/negocios/ecossistema/msw-capital-mapeia-162-startups-brasileiras-com-ia-no-core)</description></item><item><title>Nvidia's share of China's AI-chip market falls to 34% as Huawei reaches 28%</title><link>https://alfageme.eu/radar/infrastructure/#s-s063</link><guid isPermaLink="false">alfageme-radar-s063</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>IDC data for the first half of 2026: US export controls limited Nvidia's supply while domestic accelerators gained ground. China's parallel AI stack is now measurable in market share. (TechNode: https://technode.com/2026/10/14/nvidias-china-ai-chip-share-falls-to-34-as-huawei-reaches-28)</description></item><item><title>Gartner: over $3.5 trillion of AI infrastructure spending could face write-downs by 2030</title><link>https://alfageme.eu/radar/infrastructure/#s-s067</link><guid isPermaLink="false">alfageme-radar-s067</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>AI providers spend nearly $1T in 2026 and another $6T to 2030; a 12% return would need about $8.5T of revenue, but Gartner sees at most $3.6T. Neoclouds are the most exposed; hyperscalers can absorb it. (Gartner: https://www.gartner.com/en/newsroom/press-releases/2026-10-14-gartner-says-over-us-dollars-three-and-a-half-trillion-of-artificial-intelligence-infrastructure-spending-could-be-exposed-to-impairment-by-the-end-of-2030)</description></item><item><title>Shopify opens its catalogue of about a billion products to any AI assistant for checkout</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s087</link><guid isPermaLink="false">alfageme-radar-s087</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>No more one-off partnerships: any agent can now complete purchases inside Shopify stores. Agentic commerce goes from pilot to default channel. (Infobae: https://www.infobae.com/tecno/2026/10/14/shopify-permite-que-cualquier-asistente-de-ia-compre-dentro-de-su-catalogo-de-mil-millones-de-productos)</description></item><item><title>Anthropic warns that small AI models could self-replicate through IoT devices</title><link>https://alfageme.eu/radar/models/#s-s061</link><guid isPermaLink="false">alfageme-radar-s061</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>'Small weight-based replicators' could behave like a new generation of computer worms. Connected devices with weak security become part of the AI risk surface. (Ars Technica: https://arstechnica.com/security/2026/10/anthropic-warns-small-ai-models-could-self-replicate-through-iot-devices)</description></item><item><title>China's chip exports jump 136% to $45B in September as AI demand lifts prices</title><link>https://alfageme.eu/radar/infrastructure/#s-s064</link><guid isPermaLink="false">alfageme-radar-s064</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>Faster than the 108% growth of the first nine months, a sign of China's growing weight in the semiconductor supply chain. (South China Morning Post: https://www.scmp.com/tech/tech-trends/article/3370858/china-chip-exports-jump-136-us45b-ai-frenzy-jacks-prices)</description></item><item><title>US data-centre land purchases hit about $6B in the first half as power premiums rise</title><link>https://alfageme.eu/radar/infrastructure/#s-s066</link><guid isPermaLink="false">alfageme-radar-s066</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>Access to power now sets land value more than location. Energy, not plots, is what developers are really buying. (Data Center Knowledge: https://www.datacenterknowledge.com/data-center-construction/ai-data-center-land-buying-hits-6b-as-power-premiums-rise)</description></item><item><title>Humanoid-robot maker AgiBot nears a round at a valuation of about RMB 50 billion</title><link>https://alfageme.eu/radar/finance/#s-s082</link><guid isPermaLink="false">alfageme-radar-s082</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>A RMB 1–1.5 billion raise, per investors cited by Caijing — China's embodied-AI companies are drawing late-stage money. (TechNode: https://technode.com/2026/10/14/robot-maker-agibot-nears-new-funding-round-at-rmb-50-billion-valuation)</description></item><item><title>Ricursive Intelligence raises $300M to have AI design chips</title><link>https://alfageme.eu/radar/finance/#s-s083</link><guid isPermaLink="false">alfageme-radar-s083</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Founded by ex-Google researchers, it builds AI that designs chips and learns from each iteration — a bet that compute demand will keep outrunning human design capacity. (TechCrunch: https://techcrunch.com/2026/10/14/ricursive-intelligence-is-betting-on-ai-designed-chips-to-help-solve-the-compute-crisis)</description></item><item><title>38 new unicorns in September, led by semiconductors, fintech and space tech</title><link>https://alfageme.eu/radar/finance/#s-s084</link><guid isPermaLink="false">alfageme-radar-s084</guid><pubDate>Wed, 14 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Crunchbase's busiest month in a while also included AI infrastructure, AI training and security companies. (Crunchbase News: https://news.crunchbase.com/venture/september-2026-new-unicorns-semiconductors-ai-fintech-space)</description></item><item><title>Google makes Gemini 4 Argon generally available, at half the API price until the end of November</title><link>https://alfageme.eu/radar/models/#s-s053</link><guid isPermaLink="false">alfageme-radar-s053</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>Argon scores 53 on the Intelligence Index, level with GPT-6 Astra. The introductory API price of $2/$10 per million tokens doubles afterwards, and Argon uses many output tokens per task — budget on the full price. (Google — The Keyword: https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon)</description></item><item><title>IMF flags an AI correction as a top global risk, citing $1.4 trillion of circular deals</title><link>https://alfageme.eu/radar/finance/#s-s073</link><guid isPermaLink="false">alfageme-radar-s073</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Its Global Financial Stability Report counts 14 deals linking hyperscalers and AI firms: about $300B of equity and $650B of debt against $1.4T of revenue commitments. Hyperscalers now issue about 10% of global investment-grade bonds. (The Next Web: https://thenextweb.com/news/imf-circular-ai-deals-financial-stability-report)</description></item><item><title>Microsoft unveils Maia 300, an inference chip promising 40% more tokens per megawatt than GPUs</title><link>https://alfageme.eu/radar/infrastructure/#s-s062</link><guid isPermaLink="false">alfageme-radar-s062</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Infrastructure</category><description>Production deployment is planned for 2027; Maia 200 already powers Copilot and Teams in US, European and Australian data centres. Hyperscalers are moving inference onto their own silicon. (Data Center Knowledge: https://www.datacenterknowledge.com/data-center-chips/microsoft-pushes-custom-ai-chips-with-azure-maia-300)</description></item><item><title>Nvidia takes stakes in Doctolib, Pigment, Genesis AI and Entalpic during Jensen Huang's Paris visit</title><link>https://alfageme.eu/radar/finance/#s-s076</link><guid isPermaLink="false">alfageme-radar-s076</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Huang met Macron and said Nvidia will unveil about $2B of European AI investments. The world's most valuable company is now a direct investor in Europe's AI champions. (FrenchWeb: https://www.frenchweb.fr/jensen-huang-en-tournee-en-europe-entre-business-lobbying-et-investissements/463934)</description></item><item><title>Memory-chip stocks sell off despite record AI demand</title><link>https://alfageme.eu/radar/finance/#s-s074</link><guid isPermaLink="false">alfageme-radar-s074</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Finance</category><description>Samsung, SK Hynix and CXMT fell as investors question how long the price boom lasts — a reminder that AI supply chains are cyclical. (South China Morning Post: https://www.scmp.com/tech/tech-trends/article/3370739/why-samsung-sk-hynix-and-cxmt-shares-are-under-pressure-despite-strong-ai-demand)</description></item><item><title>Dutch tax authority cancels its Microsoft 365 cloud rollout and turns to open source</title><link>https://alfageme.eu/radar/use-cases-tools/#s-s095</link><guid isPermaLink="false">alfageme-radar-s095</guid><pubDate>Tue, 13 Oct 2026 07:00:00 +0000</pubDate><category>Use cases &amp; Tools</category><description>Another European public body choosing digital sovereignty over US cloud suites — relevant for any AI tool that runs on them. (ZDNET: https://www.zdnet.com/business/dutch-tax-authorities-cancel-microsoft-365-cloud-rollout)</description></item><item><title>Meta is back at the frontier: Muse Wonder Full scores 56, second only to Anthropic</title><link>https://alfageme.eu/radar/models/#s-s054</link><guid isPermaLink="false">alfageme-radar-s054</guid><pubDate>Mon, 12 Oct 2026 07:00:00 +0000</pubDate><category>Models</category><description>Artificial Analysis places Meta's new model two points behind Claude Opus 5.5 (58) and ahead of OpenAI and Google (53). For now it is only available inside Muse Code, with no public API or price. (Artificial Analysis: https://artificialanalysis.ai/models/muse-wonder-full)</description></item></channel></rss>