Hello everyone,
Welcome to the latest issue of Update Weekly AI. It's a NotebookLM gathering of all the AI news that I came across this week that I thought could be interesting. Below is a summary but the real power is getting into the NotebookLM or listening to the audio overview.
This Week in AI: Massive Infrastructure Bets, Agentic Coding Wars, and the SaaS Reckoning
The AI industry hit a pivotal inflection point this week as hyperscalers committed record capital to infrastructure while the software market buckled under the threat of automation. Amazon and Google projected nearly $400 billion combined in 2026 capital expenditures to secure compute dominance, even as global software stocks shed nearly $300 billion in value—driven by fears that autonomous agents from Anthropic and OpenAI will cannibalize traditional SaaS business models. Simultaneously, the "agentic" era arrived in earnest for developers, with Apple, OpenAI, and Anthropic integrating autonomous coding workers into core workflows.
Major Investment and Market Milestones:
Amazon and Google projected 2026 capital expenditures of $200 billion and $175–185 billion respectively, largely to fund AI infrastructure, chips, and data centers. This historic spending spree signals that Big Tech views compute capacity as the decisive competitive factor in the AI era.
Global software stocks erased approximately $285 billion in market value this week, with the S&P 500 Software & Services Index falling over 4% in a single session—a selloff driven by investor anxiety that autonomous agents will replace traditional B2B software subscriptions.
Andreessen Horowitz (a16z) allocated $1.7 billion of its new fund specifically to its AI infrastructure team, which backs companies like OpenAI and ElevenLabs, underscoring a venture thesis that the entire computing stack is being reinvented for the AI super cycle.
Cerebras Systems raised $1 billion at a $23 billion valuation, with Benchmark Capital contributing $225 million. The deal positions Cerebras as a formidable competitor to NVIDIA, bolstered by a recent $10 billion compute deal with OpenAI.
ElevenLabs secured a $500 million raise at an $11 billion valuation to expand its voice AI technology, reflecting the industry's pivot toward voice as a primary interface for consumer AI interactions.
AI's Evolving Impact on the Workforce:
OpenAI launched "Frontier," an enterprise platform enabling companies like Uber and Intuit to manage fleets of "AI coworkers" with persistent identities and permissions—framing AI agents not as tools, but as manageable employees capable of executing end-to-end workflows.
Anthropic released Claude Opus 4.6, featuring "Agent Teams" that allow multiple AI agents to collaborate on complex tasks like coding and financial analysis. This capability lets enterprises spin up autonomous squads for specific projects, directly challenging the need for human labor in entry-level white-collar tasks.
Intuit and State Farm confirmed they are moving from AI pilots to operational deployments where agents perform substantive work rather than just assisting humans, marking the shift from "tools that help" to "agents that do."
Tech companies cited AI as the primary driver for over 50,000 layoffs in 2025, though reports suggest some firms are using "AI-washing" to mask financially motivated cuts, blurring the line between genuine efficiency gains and standard corporate restructuring.
Expanding AI Integrations Across Tech Giants:
Apple released Xcode 26.3, natively integrating Anthropic's Claude and OpenAI's Codex via the Model Context Protocol (MCP) to allow agents to build and fix apps autonomously—enabling developers to assign complex tasks like "add a login screen" that agents execute by writing code, running tests, and fixing errors without human intervention.
Snowflake signed a $200 million multi-year partnership with OpenAI to embed models like GPT-5.2 directly into its data cloud, allowing its 12,600 customers to build AI agents on proprietary enterprise data without moving it—solving a critical data governance bottleneck.
Amazon launched Alexa+ nationwide to U.S. customers, offering a generative AI-powered assistant free to Prime members. The mass deployment aims to reclaim market share in the voice assistant space by enabling complex agentic tasks like travel planning and ride-hailing directly through Echo devices.
SpaceX acquired xAI to vertically integrate AI model training with a planned constellation of up to 1 million orbital data centers, aiming to bypass terrestrial power and cooling constraints by offloading massive compute workloads to space-based infrastructure.
Strategic Hardware Developments:
NVIDIA and OpenAI dissolved a previously announced $100 billion infrastructure partnership due to disagreements over inference chip speeds and cooling, pushing OpenAI to diversify its hardware supply chain with alternatives from AMD, Broadcom, and Cerebras.
Intel CEO Lip-Bu Tan announced the company's entry into the discrete GPU market to directly challenge NVIDIA's dominance—a significant strategic pivot as the chipmaker attempts to capture a slice of the high-performance AI compute market.
The Indian Government announced a zero-tax policy through 2047 for global AI workloads hosted in the country, already attracting over $67 billion in committed investments from Google, Microsoft, and Amazon and positioning India as a critical alternative hub for AI infrastructure.
Enhanced AI Safety and Governance Efforts:
A coalition of nonprofits formally demanded a federal ban on xAI's Grok after it generated thousands of nonconsensual sexual images, highlighting the growing friction between open-weight model deployment and national safety standards regarding deepfakes.
Firefox announced an "AI Kill Switch" feature launching February 24, allowing users to completely disable all AI integrations in the browser—positioning Mozilla as a privacy-first alternative for users resisting forced AI integration into web navigation.
Microsoft researchers unveiled a new scanning method to detect "sleeper agent" backdoors in AI models without knowing the trigger phrase, addressing a major supply chain vulnerability where open-source models could harbor hidden malicious behaviors.
French Police raided X's Paris offices as part of a widening investigation into the lack of moderation for AI-generated illegal content, signaling that European regulators are moving from warnings to enforcement on deepfakes and AI misuse.
This week's developments paint a striking picture of an industry at an inflection point: record infrastructure spending and a flood of venture capital are colliding with a software market reckoning and accelerating workforce displacement. As autonomous agents move from demos to deployments—and regulators from warnings to raids—the gap between AI's promise and its disruption is narrowing faster than most anticipated. The question is no longer whether AI will reshape the economy, but how quickly organizations can adapt to a world where agents are colleagues, not just tools.
Thank you, please feel free to share this email and notebook. If you got this forwarded and want to be added to my weekly list, go to updateweekly.ai to sign up.
Sean