As the OECD warns global growth will moderate to 2.9% in 2026 amid trade tensions and rising debt, AI continues its multi-year run as the dominant investment theme — but where the value lies is shifting fast. MIT's Iceberg Index reveals $1.2 trillion in cognitive automation potential hidden across administrative and professional services, 5x larger than visible tech impacts. Meanwhile, NVIDIA + University of Hong Kong research indicate that small AI models can outperform giants at 30% of the cost.
🌍 Strategic Landscape
OECD Forecasts Moderating Global Growth Amid Trade & Debt Risks
The OECD projects global GDP growth will moderate from 3.2% this year to 2.9% in 2026, citing higher tariffs and policy uncertainty. While growth is currently supported by strong AI investment, the outlook is fragile due to rising public debt from defense and aging, and high valuations in AI and crypto assets. The OECD urges co-operation on trade, fiscal discipline, and structural reforms. 🔗 OECD Economic Outlook
EU and Singapore Strengthen Ties on Safe AI, Online Child Protection and Digital Trade
On December 1, the EU and Singapore hosted the second session of their Digital Partnership Council in Brussels. They agreed to deepen cooperation on safe AI, protecting children online, fighting scams, digital IDs, cybersecurity, freer data flows, semiconductors, and quantum tech. They welcomed the new EU-Singapore Digital Trade Agreement from May 2025 that removes trade barriers and builds trust for consumers and businesses. 🔗 European Comission
The Iceberg Index — Skills-centered KPI for the AI economy: How Cognitive Automation Creates $1.2T Investment Hotspots
An MIT-associated study, Project Iceberg, utilizes Large Population Models to simulate 151 million US workers and 32,000 skills, addressing how AI is reshaping the $9.4 trillion labor market. It introduces the Iceberg Index, a skills-based metric measuring the wage value of tasks AI can perform. Analysis shows the true exposure: $1.2 trillion across administrative and professional services, is five times greater and more distributed than the visible tech-sector impact. 🔗 Project Iceberg Report
Source: The Iceberg Index Report
📈 Tailwind of the Week
At AWS re:Invent 2025, Amazon introduced the Trainium3 AI chip and UltraServer, delivering up to 4× the performance and 40% better energy efficiency for training and inference compared to Trainium2. The new UltraServer platform scales to clusters of up to 1 million Trainium3 chips. AWS also revealed Trainium4, the next-generation chip designed from the ground up to natively support Nvidia’s ecosystem, standards, and software stack — making it easier for customers running large-scale AI workloads to adopt or mix AWS silicon alongside Nvidia GPUs. 🔗 TechCrunch
📉 Headwind of the Week
Global markets turned risk-off as Bitcoin plunged 6.4% in its sharpest drop since March, dragging crypto-linked stocks down and pushing the Dow over 400 points lower. The sell-off comes amid renewed fears of an AI bubble, rising bond yields, and uncertainty over U.S. monetary policy and Fed leadership, tightening financial conditions for speculative tech and AI assets. 🔗 The Wall Street Journal
✨ M&A + Acquisitions’ Highlights
This week's deals highlight the strategic race to secure AI infrastructure capabilities, with Nvidia deepening its chip-design ecosystem while enterprise players move to capture AI-native customer engagement:
Category | Notable Companies | Highlights |
|---|---|---|
Chip / AI-infrastructure acquisition | Marvell Technology → Celestial AI 🔗 Reuters | Marvell agreed to acquire Celestial AI for US$ 3.25B (cash + stock, with earn-outs) |
Enterprise software / SaaS (AI customer-experience) | Tata Communications → Commotion Inc. | Tata Communications acquired a 51% stake in Commotion, a US-based AI-native enterprise SaaS firm focused on real-time, personalized customer engagement |
AI-ecosystem / tools investment — strategic minority stake | Nvidia → Synopsys 🔗 Reuters | Nvidia announced a US$ 2B investment into Synopsys’ chip-design software business as part of a broader collaboration. |
🌱 Early-Stage Investment Opportunities
This week's selection highlights ventures building critical AI infrastructure to capture the $1.2T cognitive automation opportunity revealed by MIT's Iceberg Index, focusing on energy-efficient computing, expert-verified training data, and autonomous workflow automation:
🚀 Growth-Stage Investment Opportunities
Our growth-stage picks focus on cognitive automation across high-value services, AI-driven clinical advancement, and the critical need for secure, open-source infrastructure to power the next generation of applications:
🔮 What to Watch
Trade Tensions Surge: Monitor escalating U.S. tariffs and retaliatory measures, which could reduce global growth by 0.5% by 2026, per OECD warnings.
AI Bubble Risk: Watch for corrections if AI investments underdeliver, potentially eroding tech-sector resilience that's offsetting tariff impacts.
EU-Singapore Digital Integration: Track roll-out of joint AI safety standards, digital ID pilots, and age verification rules. Expect accelerated semiconductor and quantum research partnerships in 2026.
Cognitive Automation Investment Shift: Monitor capital flowing toward administrative, financial, and professional services beyond traditional tech hubs, guided by MIT's Iceberg Index findings.
+ News on AI
AI Spending Surge in 2026: Investment Broadens Beyond Tech Giants
Global enterprise AI investment is expected to surge in 2026 as adoption broadens beyond tech giants. Market analysts note spending is decentralizing, with a wider range of businesses investing in AI infrastructure, software, and devices, signaling mainstream adoption. 🔗 Fortune
Small AI Models May Outperform Giants at Fraction of the Cost
According to a study by NVIDIA and the University of Hong Kong, smaller AI models can outperform frontier systems through intelligent coordination. Their ToolOrchestra system trains an 8B orchestrator model using reinforcement learning to manage specialized tools efficiently. On challenging reasoning benchmarks, it scored 37.1% versus GPT-5's 35.1% while being 2.5 times more efficient. On other tests, it surpassed GPT-5 using only 30% of the cost. The research challenges AI's "bigger is better" paradigm, suggesting lightweight coordination of diverse tools may define next-generation systems. 🔗 Read the paper on arXiv by NVIDIA + Univ. of Hong Kong
AI-Generated “Slop” Videos Hooking Babies on Youtube
Low-quality AI-generated videos are proliferating on YouTube, particularly targeting children under two. Creators use tools like ChatGPT and video generators to mass-produce simple, repetitive content for quick profit. With over 60% of U.S. parents reporting daily YouTube use by toddlers (per Pew Research), child development experts warn of potential developmental impacts. This growing scrutiny could trigger stricter content moderation requirements and age-verification regulations, adding compliance costs for platforms like YouTube and AI content-generation companies. 🔗 Bloomberg
🎓AI-Learning
FOTC/Google Webinar: The Future of AI in Tech
📅 Jan 23, 2025 | ⏰ 10:00 to 11:30 CEST | 🌐 Online
Online Free webinar featuring experts from Google, FOTC, and Laurens Coster demonstrating how AI and cloud computing enhance operational efficiency and drive innovation in organizations — with practical applications for data management, automation, and AI business strategies through real-world case studies. 🔗 Register
That’s it for this week.
Until next time,
The CLNM Capital
Want to share this newsletter with a friend? Use this link: https://clnm.beehiiv.com/