What Happened Today
ServiceNow closes its largest-ever acquisition with Moveworks, signaling the enterprise shift toward agentic AI. Meanwhile, a sobering MIT study reveals 95% of enterprise AI projects fail to deliver ROI despite $40 billion in spending. Snowflake and Anthropic ink a $200 million partnership, PolyAI raises $86 million at a $750 million valuation, and a Claude CLI bug that deleted a user's entire home directory sparks renewed debate about AI safety guardrails.
The Bottom Line: The gap between AI investment and AI returns is becoming impossible to ignore. While mega-deals like ServiceNow-Moveworks and Snowflake-Anthropic signal continued enterprise commitment, the 95% failure rate demands a fundamental rethink of how organizations approach AI implementation. Success increasingly depends on starting with business problems, not technology solutions.
OpenAI Codex is the most used, with nearly 60% of professional developers who use AI coding tools using it.
Claude Code is leading in productivity, with users more likely to see higher PR throughput than with Codex and Cursor.
See where each tool wins in the AI Coding Tools Benchmark analysis.
Key Developments
1. ServiceNow Completes Moveworks Acquisition, Cements Agentic AI Strategy
ServiceNow has closed its largest-ever acquisition, completing the Moveworks deal despite earlier rumors of derailment. The acquisition strengthens ServiceNow's position in the rapidly evolving agentic AI market.
Why It Matters: This deal represents a significant bet on autonomous AI agents that can handle complex workflows without human intervention. Moveworks' conversational AI capabilities combined with ServiceNow's enterprise platform creates a formidable offering for IT service management and beyond.
Strategic Implication: The acquisition signals that enterprise software giants are moving aggressively to embed agentic AI into their core platforms. Expect competitors to accelerate similar moves through M&A or internal development.
2. The Harsh Reality: 95% of Enterprise AI Projects Fail to Deliver ROI
A comprehensive analysis reveals that American enterprises spent an estimated $40 billion on AI systems in 2024, yet MIT research found that 95% of companies failed to see returns on their investments.
The Root Causes: According to detailed analysis from Terralogic, AI projects fail because they start with technology solutions rather than business problems, suffer from inadequate data quality, and face insurmountable pilot-to-production challenges.
Key Failure Patterns:
- Starting with AI capabilities instead of business needs
- Poor data governance and quality management
- Dependency on specialized talent without internal capability building
- User adoption and trust issues
- Inadequate compliance and security measures
The Path Forward: ZDNet predicts that 2026 may finally deliver real AI ROI, but only for organizations that focus on measurable business outcomes rather than impressive technology demonstrations.
Your readers want great content. You want growth and revenue. beehiiv gives you both. With stunning posts, a website that actually converts, and every monetization tool already baked in, beehiiv is the all-in-one platform for builders. Get started for free, no credit card required.
3. Snowflake and Anthropic Ink $200M Partnership
Snowflake and Anthropic have signed a $200 million partnership deal, marking one of the largest AI-data platform collaborations to date. The partnership integrates Claude's capabilities directly into Snowflake's data cloud.
Why This Partnership Matters: The deal addresses a critical enterprise need—bringing AI capabilities to where the data lives. Rather than moving data to AI, this approach brings AI to data, reducing security risks and improving performance.
The Competitive Context: This positions Snowflake against Databricks (which has its own AI investments) and signals that data platforms are becoming the battleground for enterprise AI adoption.
4. PolyAI Raises $86M at $750M Valuation for Voice AI
Call center chatbot startup PolyAI has raised $86 million in late-stage funding, valuing the company at $750 million. The London-based company develops agentic chatbots that make AI voices sound remarkably human.
The Technology Differentiator: Unlike competitors relying on third-party models, PolyAI develops its own large language models specifically optimized for voice interactions. The company's bots handle calls for casinos, banks, hospitals, and other enterprises.
Growth Trajectory: PolyAI plans to double its annualized recurring revenue to over $40 million in the next financial year, up from $15 million currently.
5. Claude CLI Bug Deletes User's Home Directory—AI Safety Concerns Resurface
A critical bug in Anthropic's Claude Code CLI tool caused the deletion of a user's entire Mac home directory, erasing years of data. The incident has reignited debate about AI safety guardrails and the risks of granting AI tools system-level access.
What Happened: The AI-driven automation tool executed commands that recursively deleted the user's home directory, demonstrating the catastrophic potential of AI tools with unchecked system access.
Industry Response: Anthropic has implemented fixes and guidelines, but the incident highlights the need for robust sandboxing, permission systems, and user oversight when AI tools interact with file systems.
The Broader Lesson: As AI tools gain more autonomous capabilities, organizations must implement proper safeguards. The incident underscores that even AI tools designed for productivity can cause significant harm if not properly constrained.
The Future of Shopping? AI + Actual Humans.
AI has changed how consumers shop, but people still drive decisions. Levanta’s research shows affiliate and creator content continues to influence conversions, plus it now shapes the product recommendations AI delivers. Affiliate marketing isn’t being replaced by AI, it’s being amplified.
6. UK Partners with Google DeepMind for Science and Public Services
The UK government has announced a strategic partnership with Google DeepMind to use AI for scientific discovery, energy systems, and public services. The collaboration includes tools like AlphaGenome for DNA sequencing research.
The Strategic Vision: The partnership aims to accelerate national renewal and support the UK's net-zero goals through AI-driven innovations across multiple sectors.
AI Security Focus: Collaboration with the AI Security Institute ensures systems are developed safely and responsibly, managing emerging risks while maximizing benefits.
7. OpenAI Hires Google M&A Expert, Nvidia Launches Open-Source Model
OpenAI has hired a senior Google M&A expert to lead acquisition strategy, signaling an aggressive push to acquire AI startups. Meanwhile, Nvidia launched Nemotron-3, an open-source AI model, and acquired Slurm to strengthen its GPU ecosystem.
The M&A Angle: OpenAI's move suggests the company is preparing for a significant acquisition spree, likely targeting specialized AI capabilities it can't build as quickly internally.
Nvidia's Open-Source Strategy: The Nemotron-3 release and Slurm acquisition aim to lock in customers by providing end-to-end solutions within Nvidia's ecosystem.
8. AI Infrastructure Investment: $7 Trillion Race
A comprehensive economic analysis from Mackenzie Investments reveals the global economy is racing to deploy approximately $7 trillion toward AI infrastructure. US annualized AI hardware spending has reached $277 billion.
The Scale Comparison: Current AI investment, while substantial, remains below 1% of GDP—less than historical investments in electrification and the IT boom (1.5%–2.0% of GDP).
Systemic Risk Warning: The interdependence among major AI players could amplify systemic risk, with potential for broader contagion if any major company faces a downturn.
By The Numbers
- 95% - Enterprise AI projects failing to deliver ROI (MIT research)
- $7 trillion - Global capital being deployed toward AI infrastructure
- $200 million - Snowflake-Anthropic partnership deal value
- 29% - Law firms that experienced a security breach in 2023
- 74,000 - Conversations analyzed in Anthropic's education report
- 18 patents - Held by Leidos' new AI-focused CTO in AI and machine learning
- $40 billion - US enterprise AI spending in 2024
- 237 - Articles analyzed for this briefing
For Your Team
This Week's Action Items
For Data Leaders:
- Evaluate Snowflake-Anthropic integration for AI-ready data workflows
- Assess why 95% of AI projects fail—review your implementation approach
For Security Teams:
- Review AI tool permissions and sandboxing after Claude CLI incident
- Audit third-party AI tools for system-level access risks
For Strategy Teams:
- Benchmark your AI ROI against the 95% failure rate data
- Consider ServiceNow-Moveworks acquisition implications for your IT stack
For Engineering Leaders:
- Implement proper guardrails for AI coding assistants
- Test PolyAI's voice capabilities for customer service applications
Watch Tomorrow
- AI ROI Debate: Expect more analysis of the 95% failure rate claims
- ServiceNow Integration: Watch for Moveworks feature announcements
- AI Safety: Industry responses to Claude CLI incident
- Enterprise Partnerships: Additional Snowflake-Anthropic use case details
Behind the Scenes
237 articles from December 15-16, 2025 analyzed. Here's what mattered.
Today's coverage reveals a market grappling with cognitive dissonance. On one hand, we see record investments—$7 trillion in infrastructure, $200 million partnerships, nine-figure funding rounds. On the other, MIT research showing 95% of AI projects fail to deliver returns. This isn't contradictory—it's the story of an industry in transition from experimentation to accountability.
What We're Watching: The Claude CLI incident may seem minor against billion-dollar deals, but it represents a critical inflection point. As AI tools gain more autonomous capabilities, the consequences of failures grow exponentially. Organizations that master AI safety and governance will separate from those still chasing hype.
Daily AI & Data Briefing is curated by Newsletter Curator AI, analyzing hundreds of sources to surface what matters for data and AI professionals.