Wall Street CEOs Summoned to Discuss Anthropic AI Risks | Bloomberg Tech 4/10/2026

Key Concepts

• AI Infrastructure & Compute: The critical shortage of physical data center capacity and high-performance computing (HPC) hardware (Nvidia GPUs) required to train and run advanced AI models.
• Systemic Cyber Risk: Concerns regarding the potential for advanced AI models (like Anthropic’s "Mythos") to identify and exploit vulnerabilities in financial systems.
• Project Glass Wing: A collaborative initiative involving major financial institutions and AI labs to stress-test models and establish security guardrails.
• Capital Expenditure (CapEx) Supercycle: The massive, multi-billion dollar investment in AI hardware, energy, and data centers, driving a "tale of two markets" (hardware outperformance vs. software volatility).
• Geopolitical Overhang: The impact of Middle Eastern conflicts and trade restrictions (e.g., bans on H100/H200 chips to China) on global supply chains and market stability.

---

1. AI and Financial Systemic Risk

Treasury Secretary Scott Bessent and Fed Chair Jerome Powell held an urgent, private meeting with Wall Street leaders to address the cyber risks posed by Anthropic’s new AI model, Mythos.

• Key Argument: While AI offers significant upside for efficiency, there is a fear that these models could be weaponized to identify and exploit vulnerabilities in the U.S. financial infrastructure.
• Methodology: Anthropic has implemented a "limited rollout" strategy, providing access to only ~40 partners (including JP Morgan and Palo Alto Networks) to conduct stress tests and establish "guardrails" before wider dissemination.
• Perspective: Experts note that this is a precautionary measure; the goal is to balance innovation with the protection of sensitive consumer data.

2. The Compute Crunch and CoreWeave

CoreWeave has emerged as a "luxury builder" for AI infrastructure, securing multi-year, multi-billion dollar contracts with Meta and Anthropic.

• Business Model: CoreWeave acts as a specialized provider of Nvidia compute. They utilize a "land and expand" strategy, securing long-term, committed contracts that allow them to raise capital at lower costs.
• Financials: The company recently executed a massive $21 billion deal with Meta and issued significant convertible debt. Analysts highlight a $90 billion contracted backlog as a signal of long-term revenue stability.
• Market Insight: Analysts argue that the industry is nowhere near "built out." With Microsoft and other tech giants struggling to find enough physical space for compute, hardware and semiconductor companies are seeing massive revenue growth, while software companies face volatility due to high CapEx requirements.

3. Market Dynamics: Hardware vs. Software

The market is currently bifurcated, with hardware/semiconductor stocks rallying while software names struggle.

• TSMC Performance: TSMC reported a 35% revenue increase, defying geopolitical tensions in the Middle East and signaling that demand for AI infrastructure remains robust despite macroeconomic uncertainty.
• Software Sell-off: Software companies are experiencing a "resumption of AI disruption fears." Investors are concerned about the ROI of AI software and the high costs of energy and infrastructure required to sustain these platforms.
• Energy Constraints: Energy prices are becoming a bottleneck. OpenAI has reportedly paused its "Stargate" data center project in the UK due to high energy costs, illustrating that the "chokehold" on AI growth is increasingly tied to power availability.

4. Artemis 2 and Space Innovation

The Artemis 2 mission represents a significant milestone in human spaceflight, with the crew returning from a historic lunar flyby.

• Technical Challenge: The splashdown is a critical test for the Orion capsule’s heat shield, which faced degradation issues during the Artemis 1 mission.
• Methodology: NASA and Lockheed Martin have engineered a new re-entry profile to minimize exposure to extreme temperatures (up to 5,000°F).
• Future Outlook: Professor Lori Lechner (ASU) emphasized that the data gathered from this mission is vital for accelerating the Artemis 3 and 4 timelines, with the goal of putting boots on the moon by 2028.

5. Corporate and Labor Developments

• Apple Unionization: Apple is closing its first unionized retail store in Towson, Maryland. While Apple cites "declining conditions at the mall," the union characterizes the move as "union busting." Legal experts suggest this will trigger a long-term battle to determine if the closure constitutes illegal retaliation under labor laws.
• Export Controls: US charges against a Super Micro co-founder regarding the smuggling of banned Nvidia H100/H200 chips to China highlight the ongoing tension between AI development and national security/trade policy.

---

Synthesis

The current tech landscape is defined by a massive, capital-intensive buildout of AI infrastructure. While hardware providers like CoreWeave and TSMC are thriving on the back of insatiable demand for compute, the broader market is grappling with the "angst" of high energy costs, geopolitical instability, and the systemic cyber risks inherent in deploying powerful, unproven AI models. The successful return of the Artemis 2 crew serves as a symbolic counterpoint, highlighting the potential for human-led innovation to overcome extreme technical and physical challenges.