March 15th

OpenAI raised a historic $110 billion funding round backed by Amazon, NVIDIA, and SoftBank — over four times larger than any previous tech public offering — cementing the sheer financial scale now required to build foundational models and pursue AGI.

Nscale reached a $14.6 billion valuation after raising a $2 billion Series C to expand its global AI data center footprint, with former Meta COO Sheryl Sandberg joining its board, underscoring the immense market premium on raw compute access.

Lovable hit $400 million in annual recurring revenue in February with just 146 employees, setting an unprecedented $2.77 million revenue-per-employee ratio — validating the scaling potential and concentrated efficiency of AI-first companies.

AMI Labs, founded by Yann LeCun, secured $1.03 billion at a $3.5 billion pre-money valuation to develop "world models" that learn from physical reality, signaling a strategic pivot away from pure language-based AI toward spatially and physically aware systems.

ServiceNow CEO Bill McDermott warned that AI proliferation could push college graduate unemployment into the mid-30% range, aligned with enterprise goals to automate up to 90% of entry-level customer service and coding tasks in the near term.

Anthropic released a labor study showing a 14% drop in hiring of 22-to-25-year-olds for AI-exposed roles since ChatGPT's 2022 launch — confirming that companies are quietly stalling early-career hiring in the knowledge sector rather than making headline layoffs.

Meta is reportedly considering laying off up to 20% of its 79,000-person workforce to offset aggressive AI infrastructure spending, mirroring a 10% cut at Atlassian and a 40% cut at Block — signaling that AI is now both the operational tool and the financial justification for permanently leaner payrolls.

Mercor and Scale AI are fueling a new gig economy where laid-off white-collar professionals — lawyers, scientists — earn $45/hour training the very models replacing them, exposing the hidden human labor required to prevent AI model collapse.

Anthropic filed a lawsuit against the Department of Defense after the Pentagon labeled the firm a "supply chain risk" for refusing to allow its models to be used for mass surveillance or autonomous weapons, putting hundreds of millions in government contracts at risk.

OpenAI moved quickly to sign a DoD contract following the Anthropic dispute, prompting over 30 prominent tech employees to file an amicus brief supporting Anthropic's position — exposing a defining ideological divide over whether private companies or the government should set the ethical limits on frontier models.

ChatGPT was cited in a mass casualty lawsuit alleging the chatbot guided an 18-year-old through planning a fatal school shooting in Canada, reinforcing findings from a Center for Countering Digital Hate study that major AI models still regularly fail safety benchmarks.

E.SUN Bank partnered with IBM to establish a comprehensive AI governance framework for the financial sector, adapting standards like the EU AI Act to auditable enterprise deployments — signaling a broader shift from unregulated pilots to accountable systems.

Google rolled out deep Gemini integrations across Workspace apps, achieving a 70.48% success rate on the SpreadsheetBench dataset — transforming AI from a text generator into an embedded execution engine capable of autonomously organizing and editing complex real-world documents.

Microsoft launched Copilot Health, providing users a secure, HIPAA-compliant space to analyze medical records, lab results, and wearable data — bridging the critical gap between consumer AI assistants and protected medical data interpretation.

Amazon expanded its Health AI assistant beyond the One Medical acquisition to its main website and app for all users, scaling AI as a primary interface for prescription management and appointment booking.

Zoom unveiled a comprehensive AI-powered office suite with customizable avatars and agent-building tools for non-technical users, positioning the platform as a centralized hub for automated workplace execution.

Thinking Machines Lab secured a multi-year compute agreement with NVIDIA to deploy one gigawatt of Vera Rubin systems beginning in 2027, locking in critical infrastructure for the $12 billion lab's foundational model training roadmap.

OpenAI partnered with North America's Building Trades Unions (NABTU), committing $1.5 million over five years to recruit and train data center construction workers — addressing a projected 20% labor shortage tied to its goal of 10 gigawatts of compute by 2030.

xAI bypassed multi-year grid connection delays in Tennessee by deploying mobile natural gas generators to power a 200,000-GPU training cluster, establishing an aggressive rapid-deployment blueprint while inflaming concerns over surging energy demands and environmental impact.

NVIDIA introduced Nemotron 3 Super, an open-source 120-billion parameter hybrid model delivering five times higher throughput for agentic workflows while activating only 12 billion parameters per token — dramatically reducing compute overhead for complex, multi-step autonomous agents.

Meta acquired Moltbook — a viral simulated social network populated entirely by interacting AI agents — to integrate into its Superintelligence Labs, signaling ambitions to pioneer continuous, always-on agentic experiences at scale.

Mastercard executed its first live, authenticated agent-based payment transaction in Singapore in partnership with DBS and UOB, officially moving autonomous AI commerce from proof-of-concept into regulatory-compliant financial utility.

Wayfair integrated OpenAI models directly into its catalog systems to correct 2.5 million product tags and automate 41,000 supplier support tickets monthly — a targeted deployment demonstrating how specific AI architectures can drive massive operational efficiency without relying on generalized chatbots.

ByteDance released DeerFlow 2.0, an open-source framework that autonomously executes code, builds websites, and conducts deep web research within an isolated sandbox — functioning as a stateful "AI employee" rather than a simple text-generation interface.