📨 India Deploys, China's Robots Fight, Pentagon Threatens Anthropic

This week, the Pentagon threatened to designate Anthropic a "supply chain risk" despite its recent $30B Series G funding round, which valued the company at $380B. Meanwhile, India committed $1.7B to AI infrastructure, China's humanoid robots showcased martial arts at the Spring Festival Gala ahead of anticipated IPOs, and Korean investors shifted from US stocks to domestic chip giants.

India's AI summit has yielded significant commitments, including a $1.1B state-backed venture capital fund and Blackstone's $600M investment in GPU firm Neysa [TechCrunch]. Additionally, Peak XV invested $15M in a data center power startup. With 100 million weekly ChatGPT users and $68B in committed AI/cloud investment through 2030 [Reuters], India is positioning itself as a key market for AI deployment, even as AI advancements pose a potential threat to its $283B IT sector.

Chinese AI companies, such as Alibaba and DeepSeek, are releasing free, open-source models that compete with Western counterparts while offering lower operational costs. Despite limited access to advanced chips due to export controls, Chinese firms are achieving strong results [Reuters]. Concurrently, humanoid robot manufacturers like AgiBot and Unitree are preparing for IPOs after demonstrating choreographed routines and martial arts at China's Spring Festival Gala. These robots, though currently priced above $17,000, signify a growing trend in the sector [Reuters].

The Pentagon is considering designating Anthropic a "supply chain risk," a classification typically reserved for foreign adversaries. Such a designation would compel all U.S. defense contractors to sever ties with the company [Axios]. The core of the dispute lies in Anthropic's restrictions on military use, which conflict with the military's demand for access for "all lawful purposes." Anthropic specifically opposes the use of its AI for domestic surveillance and autonomous weapons. Claude, Anthropic's AI, was reportedly used via Palantir in a recent operation involving Maduro [WSJ].

Anthropic successfully closed a $30B Series G funding round, achieving a $380B valuation. This marks the 2^nd-largest private tech fundraising on record, following OpenAI's $40+B raise. Anthropic's annualized revenue has reached $14B, with Claude Code contributing $2.5B in run-rate revenue [CNBC].

Meanwhile, Meta plans to deploy millions of NVIDIA Blackwell and Rubin GPUs, along with NVIDIA CPUs, across its hyperscale data centers for AI training and inference. This partnership includes Spectrum-X Ethernet networking and confidential computing for WhatsApp. It also represents the first large-scale deployment of NVIDIA Grace CPUs, with a potential Vera CPU rollout in 2027 [Nvidia].

US stocks experienced a sharp decline on Thursday, driven by concerns over AI-driven disruption, leading to a rotation out of tech stocks. Software, trucking, logistics, and real estate sectors saw declines due to fears regarding AI's impact [YahooFinance].

Cisco's stock dropped 12% amid margin pressure caused by rising memory prices, a direct consequence of increased demand from AI data centers. Other major tech companies, including Nvidia, Meta, Amazon, and Apple, also recorded significant losses [CNBC].

This week featured three notable acquisitions focused on AI infrastructure funding, cloud capabilities, and data security:

University of Texas spinout Apptronik raised $520M, tripling its valuation to $5.3B in an extended Series A round totaling $935M. The company supplies robots to Google DeepMind, Mercedes-Benz, and logistics firms [TechCrunch].

Budapest-based Allonic secured a $7.2M pre-seed round for its biomimetic robotics platform, which utilizes soft tendons woven around 3D-printed skeletons [Allonic].

Stanford spinout Simile emerged from stealth with $100M in funding to develop AI models capable of anticipating human behavior and simulating decision outcomes [TFN].

As US AI stocks face sell-offs due to high capital spending and disruption fears, South Korean retail investors are increasingly active, with record margin loans and brokerage accounts. They are now investing in homegrown tech giants like Samsung and SK Hynix, anticipating a memory chip supercycle through 2027. Korean households hold $75B in local cash deposits and $170B in US equities. The government plans to offer capital gains tax exemptions to encourage the repatriation of domestic investment [Bloomberg].

OpenAI has introduced Lockdown Mode for ChatGPT, an optional setting designed for security-conscious users. This mode disables features vulnerable to prompt injection attacks that could lead to data leaks [OpenAI]. It restricts web browsing to cached content, preventing live network requests. As AI evolves from chatbots to agents executing complex tasks, security risks escalate. This update acknowledges the necessity of deterministic "hard blocks" to address AI vulnerabilities.

The European Defense Fund has allocated €7.3B through 2027 for research and development in AI, quantum computing, cybersecurity, drones, and biotech [European Commission]. Concurrently, the European Investment Bank has quadrupled its defense financing to over €4 billion by 2025, targeting military mobility, industrial capacity (including drone manufacturers), and SME liquidity [EIB]. This combined public funding creates a robust pipeline for defense tech startups and private investment.

👀 India's $68B AI infrastructure Pipeline: Will committed capital translate to operational GPU capacity and sustained deployment momentum through 2027?

👀 Pentagon-Anthropic Clash: A DoD "supply chain risk" designation could force defense contractors to abandon Claude, potentially reshaping enterprise AI adoption patterns.

👀 Korean investor Repatriation: Can tax incentives move $170B from US tech stocks into domestic AI chip leaders Samsung and SK Hynix?

📅 25-26 February 2026 | 🌍 Brussels & Online

Europe's premier digital policy forum, hosted by DIGITALEUROPE, will convene high-level policymakers, industry leaders, and innovators. The event will debate AI, cybersecurity, defense tech, and regulatory reform under the theme "Redesigning Europe's Digital Power," focusing on AI for industrial competitiveness, digital security, and regulatory simplification. Register

"The most dangerous data point is the one you never see.” Sahil Bloom

This sentence captures the essence of Survivorship Bias: our biggest blind spots arise from systematically overlooking failures, casualties, and negative outcomes that remain unseen. It reminds us to actively seek out what's invisible in our analysis before making important decisions. 

That’s it for this week.

Until next time,

The CLNM Capital

[Edited with the assistance of LLMs]

Want to share this newsletter with a friend? Use this link: https://clnm.beehiiv.com/