AI pressure shifts to regulation as China models and IBM slide hit enterprise tech

AI companies and their critics spent the week moving the fight from product roadmaps to policy, as executives and economists called for new rules while New York imposed a data center moratorium. The shift matters for enterprise buyers and infrastructure operators because AI demand is still driving multibillion-dollar buildouts, even as local opposition and geopolitical competition become harder to treat as side issues.

A group of AI leaders and economists called for a new regulatory approach to artificial intelligence, citing concerns including job losses, according to SiliconANGLE. Google DeepMind Chief Executive Demis Hassabis separately called for an AI standards body. The proposals follow earlier efforts from OpenAI, Anthropic and others to shape how governments regulate the technology, though no concrete legislative outcome was disclosed.

The public backlash is also becoming more physical and local. The Wall Street Journal reported that some hard-line anti-AI activists have discussed potential violence, and Futurism reported that some technology executives have hired armed bodyguards. Data centers have become a more immediate target for opposition because of their electricity, water and pollution effects. New York Gov. Kathy Hochul signed an executive order creating what SiliconANGLE described as the first U.S. state moratorium on data centers.

That has not stopped capital spending. Meta Platforms nearly doubled its planned investment in the Hyperion data center campus in Louisiana to more than $50 billion, according to SiliconANGLE. Data center operator Switch is considering an initial public offering that could raise as much as $10 billion, RuntimeWire reported. QumulusAI, described as a neocloud provider, went public through a direct listing.

China and sovereign AI move up the agenda

China’s Moonshot AI released Kimi K3, which SiliconANGLE described as the world’s largest open-weights model and a system moving closer to U.S. leaders across several capabilities. The release adds pressure to the assumption that the most important AI models and cloud infrastructure will remain concentrated among U.S. providers.

SiliconANGLE contributors Amit Ayal Govrin, John Furrier and Dave Vellante argued that sovereign AI is becoming a central fault line, with deployment environments splitting by region, regulation and governance requirements. The practical implication for vendors is that model access, data residency, security controls and cloud partnerships may become procurement constraints rather than policy talking points.

The U.S. still has a major infrastructure advantage through Nvidia. Vellante and theCUBE Research’s Bob Laliberte interviewed Nvidia networking executive Gilad Shainer and concluded that Nvidia’s networking position is a real part of its AI factory advantage, while industry concerns about lock-in remain unresolved.

IBM drops, funding keeps flowing

IBM shares fell more than 26% Tuesday after the company projected an earnings shortfall tied to large mainframe orders that did not arrive, according to SiliconANGLE. The drop came despite some AI-related gains at the company. IBM, Alphabet, Intel, SAP, ServiceNow and Tesla are among the enterprise names scheduled to report earnings next week.

AI and infrastructure financing remained active. The Wall Street Journal reported that Databricks is set to reach a $188 billion valuation with new investment from Coatue. Fireworks closed a $1.5 billion round at a $17.5 billion valuation, SiliconANGLE reported. Other disclosed rounds included Emergent at $130 million, TerraFirma at $115 million, AGI at $70 million, InstaLILY at $60 million, Microagi at $55 million and Valarian at $50 million.

Outside core AI software, defense technology startup Helsing raised $1.8 billion at an $18 billion valuation, while Walden Robotics launched with $300 million and Singularity raised $80 million at a $400 million valuation. The week’s financing data points show that investors are still underwriting AI-adjacent infrastructure, robotics and defense bets, even as the policy and power constraints around AI become more visible.

This story draws on original reporting from SiliconANGLE.