GenAI Daily - January 19, 2026: ASEAN+ Enterprises Accelerate AI Investment, DeepSeek's Training Breakthrough, Rubin Platform Launch

Top Story

Enterprises across Asia Pacific are accelerating their shift from AI experimentation to execution, with 96% of organizations planning to increase AI investments over the next 12 months, according to the 4th edition of the Lenovo CIO Playbook 2026 - The Race for Enterprise AI, commissioned by Lenovo with insights from IDC. On average, organizations expect AI spending to grow by 15%, spanning GenAI and Agentic AI, public cloud AI services, on-prem AI infrastructure, and AI security tools.

The research reveals a decisive shift toward outcome-based deployments rather than pilot programs.

88% of AP organizations expect a positive ROI from AI in 2026, with an average anticipated return of 2.8x (US$2.85 for every US$1 invested). Yet, scaling AI beyond pilots remains a key challenge, reinforcing the importance of governance, operating models, and lifecycle management.

The Playbook finds that 86% of organizations across Asia Pacific now incorporate on-premises or edge environments as part of hybrid AI architectures, effectively making hybrid AI the default model for enterprise AI deployments.

Why it matters: This marks a fundamental shift from AI experimentation to production-grade deployment, with enterprises demanding measurable ROI rather than proof-of-concept demonstrations.

Key Developments

DeepSeek Unveils Breakthrough Training Architecture for Next-Gen Models

The Chinese AI startup published a research paper on Wednesday, describing a method to train large language models that could shape "the evolution of foundational models," it said. The paper, co-authored by its founder Liang Wenfeng, introduces what DeepSeek calls "Manifold-Constrained Hyper-Connections," or mHC, a training approach designed to scale models without them becoming unstable or breaking altogether.

Wei Sun, the principal analyst for AI at Counterpoint Research, told Business Insider on Friday that the approach is a "striking breakthrough." She added that even with a slight increase in cost, the new training method could yield much higher performance.

The timing suggests preparation for DeepSeek's rumored V4 model launch, expected mid-February.

Impact: This architecture could enable more efficient training of large models, potentially maintaining DeepSeek's cost advantage over Western competitors. (Source)

NVIDIA Launches Rubin Platform with Six New AI Chips

NVIDIA today kickstarted the next generation of AI with the launch of the NVIDIA Rubin platform, comprising six new chips designed to deliver one incredible AI supercomputer.

NVIDIA today announced new open models, frameworks and AI infrastructure for physical AI, and unveiled robots for every industry from global partners. Expanding the open model universe, NVIDIA today released new open models, data and tools to advance AI across every industry.

The Nemotron 3 family of open models - in Nano, Super and Ultra sizes - introduces the most efficient family of open models with leading accuracy for building agentic AI applications. Nemotron 3 Nano delivers 4x higher throughput than Nemotron 2 Nano and delivers the most tokens per second for multi-agent systems at scale through a breakthrough hybrid mixture-of-experts architecture.

Impact: The Rubin platform positions NVIDIA for the next wave of AI infrastructure demand while open-sourcing Nemotron 3 democratizes access to high-performance agent development. (Source)

Siemens and NVIDIA Expand Partnership for Industrial AI Operating System

Siemens and NVIDIA expand their partnership to build the Industrial AI Operating System, reinventing the entire end-to-end industrial value chain through AI - from design and engineering to manufacturing, production, operations, and into supply chains. Siemens' primary product launch at CES 2026 is the Digital Twin Composer, available on the Siemens Xcelerator Marketplace mid-2026.

Siemens and NVIDIA will work together to build AI-accelerated industrial solutions across the full lifecycle of products and production, enabling faster innovation, continuous optimization, and more resilient, sustainable manufacturing. The companies also aim to build the world's first fully AI-driven, adaptive manufacturing sites globally, starting in 2026 with the Siemens Electronics Factory in Erlangen, Germany, as the first blueprint.

Impact: This partnership signals the industrial sector's move toward AI-native manufacturing, potentially transforming how physical products are designed and produced. (Source)

Enterprise AI Enters "Accountability Phase" as Pragmatism Takes Hold

If 2025 was the year AI got a vibe check, 2026 will be the year the tech gets practical. The focus is already shifting away from building ever-larger language models and toward the harder work of making AI usable.

"Fine-tuned SLMs will be the big trend and become a staple used by mature AI enterprises in 2026, as the cost and performance advantages will drive usage over out-of-the-box LLMs," Andy Markus, AT&T's chief data officer, told TechCrunch.

January 2026 marks a turning point for AI. The focus has shifted from experimentation to accountability, reliability, and real-world impact.

Organizations are moving from "Can AI do this?" to "Can this AI be relied on when it matters?"

Impact: The shift toward smaller, specialized models and reliability-first deployments marks enterprise AI's evolution from experimental to mission-critical infrastructure. (Source)

Product Launches

• NVIDIA Nemotron 3 Family: Open-source agentic AI models in Nano, Super and Ultra sizes with 4x higher throughput for multi-agent systems - First models designed specifically for building AI agents at enterprise scale (NVIDIA)

• Siemens Digital Twin Composer: Industrial metaverse platform launching mid-2026 on Siemens Xcelerator Marketplace - Combines digital twins with real-time engineering data for AI-driven manufacturing (Siemens)

• LG CLOiD Smart Robot: AI-powered home robot using NVIDIA Jetson Thor platform and Isaac Sim - Simulates behavior in virtual environments before real-world deployment (AI Apps)

Funding & Deals

• xAI - $20B Series E: Elon Musk's AI company building foundation models to "understand the universe" - Led by Valor Equity Partners with strategic investment from NVIDIA and Cisco (xAI)

• Cyera - $400M Series F: AI-powered data security platform at $9B valuation, protecting data for 20% of Fortune 500 companies - Led by Blackstone alongside Accel and Sequoia Capital (PYMNTS)

• Cast AI - Strategic Investment: Cloud and Kubernetes automation platform surpassing $1B valuation with OMNI Compute GPU marketplace launch - Participation from Pacific Alliance Ventures and SoftBank Vision Fund 2 (PYMNTS)

• Protege - $30M Series A: Real-world data aggregation platform for AI training led by Andreessen Horowitz - Provides rights-protected datasets across healthcare, media, and audio sectors (PYMNTS)

Related reading: Check out this week's [Deep Insights analysis] for strategic context on these developments.