Table of Contents:
AI | Artificial Intelligence
UNESCO and CODEMAO advance Africa-Asia cooperation in AI education through Youth AI Tour
Anthropic warns AI models could replicate themselves across smart devices
VC | Startup & Funding
Strasbourg-founded Goodweek raises €10.69 million to become the AI Manager for SMBs
Cornell Tech professor’s startup receives $1.75 million to improve AI efficiency
HI | Hardware & Infrastructure
Mitsubishi Heavy Industries and Ingeteam Form Strategic Partnership to Advance Power Solutions for AI Data Centers
South Korea’s KT Launches 800Gbps Dedicated Lines for AI Data Centers
UNESCO and CODEMAO advance Africa-Asia cooperation in AI education through Youth AI Tour
From July 20 to 25, 2026, UNESCO and its partner CODEMAO, supported by UNESCO’s Institute for Information Technologies in Education, held the Africa-Asia Youth AI Tour in Shenzhen, China.
The event brought together officials, educators and students from China, Côte d’Ivoire, Kenya, Namibia, the Republic of the Congo and Thailand for a policy forum, technology study visits and a three-day student training and project co-creation programme.
It set out a framework for long-term cooperation on curriculum development, teacher capacity building, student mobility and project-based learning, and 68 schools received official pilot accreditation under the UNESCO-CODEMAO Youth Coding Initiative, launched in 2021.
UNESCO says it will continue working with CODEMAO on teacher training, inter-school communications and youth innovation.
Source: UNESCO
Anthropic warns AI models could replicate themselves across smart devices
Anthropic said on October 13 that its experiments suggest small AI models may be months away from self-replicating on commercial internet-of-things devices such as smart ovens and off-brand Android TV boxes.
In its tests, models under 30 billion parameters could complete the many steps replication requires but struggled to finish the full sequence, and only simplified tests with some security steps removed showed end-to-end replication.
The company noted that such devices often have weak security, that a model too large for one device could split into shards across several, and that a replicating system would have no central point where it could be switched off.
Anthropic said the form autonomous replication will take is still uncertain, and it called for stronger security on connected devices to make spreading harder.
Source: CNBC
This episode is brought to you by Stupid Good AI!
Stupid Good AI is a personal food concierge that learns your taste and tells you exactly where to eat and drink in any city.
A foodie concierge in your pocket. Register for free at StupidGood.ai for best of lists in order to find the right place on the first try.
Strasbourg-founded Goodweek raises €10.69 million to become the AI Manager for SMBs
Goodweek, a Los Angeles-headquartered French startup, has raised a €10.69 million ($12 million) seed round led by Spain’s Adara Ventures, with participation from investors including FJ Labs, True Ventures and Bpifrance.
The company, founded in Strasbourg and incubated at Station F in Paris, offers small and medium-sized businesses a single platform that combines models from Anthropic, OpenAI, Google and Mistral.
It is meant to counter unsanctioned employee AI use, which leads to personal accounts, data leaving the company and duplicate subscriptions, by isolating company data and showing who uses which tools at what cost.
Goodweek is one year old, has 50 employees and nearly 150 customers, and will use the funding to expand sales and enter markets including Latin America and Asia.
Source: EU Startups
Cornell Tech professor’s startup receives $1.75 million to improve AI efficiency
Makora, a startup co-founded by Cornell Tech assistant professor Mohamed Abdelfattah, has received a $1.75 million award from the National Science Foundation-supported STRIDE Ventures AI Efficiency Challenge.
The funding is meant to make artificial intelligence faster and less costly to run.
Makora will use it to deploy its technology with data center providers, showing how automated software improvements can get more out of existing computing resources, which could limit the energy needed to meet growing AI demand.
The company is based at The Bridge on the Cornell Tech campus, and the award supports commercial deployment of its technology alongside Abdelfattah’s research on future AI systems at Cornell.
Source: Cornell Tech
Join the newsletter for regular updates on artificial intelligence, venture activity, and core systems!
Mitsubishi Heavy Industries and Ingeteam Form Strategic Partnership to Advance Power Solutions for AI Data Centers
Mitsubishi Heavy Industries and Spain’s Ingeteam signed a memorandum of understanding on October 15 to distribute, integrate and jointly develop energy infrastructure for U.S. data centers.
The agreement pairs MHI’s power, cooling and energy management systems with Ingeteam’s power electronics, rotating electrical machines and control technologies.
The companies say AI workloads are straining power quality and grid integration, with harmonic distortion, voltage imbalance and longer interconnection timelines among the problems.
Ingeteam’s products would be sold in the U.S. through MHI Group, and the two will also explore joint work on medium-voltage distribution and energy management systems, subject to technical and commercial evaluation.
Source: Mitsubishi Heavy Industries
South Korea’s KT Launches 800Gbps Dedicated Lines for AI Data Centers
KT, the South Korean telecom operator, launched AI-dedicated network lines on October 14 with bandwidth of up to 800Gbps, eight times the previous 100Gbps maximum for dedicated lines.
The service comes in 400Gbps and 800Gbps tiers and uses KT’s 840,000 km of nationwide optical cable to build shortest-path and redundant routes between data centers.
It applies an All Photonic Network architecture, which reduces latency by limiting conversions between optical and electrical signals.
KT plans to market the lines to AI data center operators and enterprises building large-scale AI infrastructure, and to expand its customized B2B offerings.
Source: BigGo Finance
Foundations is written by the Axented team.
Axented works with companies around the world to design, build, and scale digital products, teams, and AI-driven systems.
If you’re building or scaling a technology-driven business and want to explore how these signals apply to your context, you can learn more about how we work at axented.com.
