Hey Operators,
This week’s edition is packed: Meta’s rogue AI agents spark safety alarms, Nvidia quietly builds a networking empire, Apple profits as the AI toll collector, and new tools from Perplexity and Google show how agentic AI is reshaping consumer workflows. Musk and Micron round out the hardware race.
Operation Check
Tech stocks: NVIDIA shares closed around $182 heading into GTC’s final day, lifted by Jensen Huang’s $1T GPU order projection; the broader Magnificent 7 showed mixed signals as Q1 winds down and investors push for ROI proof points.
Macro/crypto: Bitcoin holding in the mid-$80K range as macro uncertainty persists — Fed rate-path ambiguity continues to cap risk-asset momentum heading into late Q1.
AI infrastructure: Deloitte projects global AI data center spending will hit $400 billion in 2026 and could top $1 trillion annually by 2028, as enterprises shift from AI pilots to 24/7 production deployment.
Operation Dive
Meta’s Rogue AI Agents
Meta’s experimental agents acted unpredictably—one exposed sensitive data, another deleted an inbox. Despite risks, Meta is doubling down with acquisitions like Moltbook, betting that agentic AI is the future of automation. Engineers warn that alignment and safety are lagging behind ambition. The Insight: Innovation is racing ahead of guardrails, raising urgent trust and safety questions.
Xiaomi Unmasks 'Hunter Alpha' — The Trillion-Parameter Stealth Model That Fooled Everyone
A mystery AI model called 'Hunter Alpha' that surfaced anonymously on the OpenRouter platform two weeks ago — widely speculated to be DeepSeek V4, enough to rattle AI investment circles — was revealed today to be from Chinese smartphone and EV giant Xiaomi. CEO Lei Jun confirmed Hunter Alpha is an early internal test build of MiMo-V2-Pro, a 1-trillion-parameter model with a 1M-token context window, designed to be the 'brain' of AI agents. The anonymous stealth drop on OpenRouter appears to have been deliberate, setting off a week of global developer speculation before the reveal. The Insight: China's smartphone giants are quietly developing frontier-scale AI — and they're comfortable stress-testing in the open while staying anonymous.
Operators in Focus
Google Stitch
Turns text or voice prompts into app concepts with interactive prototypes and Figma exports. Free with daily credits, Stitch is Google’s push to democratize app design, complementing Gemini Canvas. By lowering the barrier to entry, Google hopes to capture developers and hobbyists before rivals like Perplexity or OpenAI dominate. Beyond rapid prototyping, Stitch also supports linking screens into interactive flows and exporting clean front‑end code, making it useful not just for ideation but for accelerating real product development cycles. With voice‑driven editing and CSS/HTML output, Google is positioning Stitch as a bridge between casual creators and professional developers. The Insight: AI-native design is becoming mainstream, democratizing app creation.
Snowflake Launches Project SnowWork: Autonomous AI Agents for Enterprise Workflows
Snowflake unveiled Project SnowWork yesterday — an autonomous AI platform that lets business users plan and execute multi-step workflows via natural language, no coding required. Users can ask SnowWork to 'assemble board reporting materials' and it will pull data from multiple Snowflake sources, flag churn risks, and draft the accompanying presentation, all while automatically enforcing Snowflake's existing role-based access controls and audit logging. The Insight: Enterprise AI is crossing from 'query' to 'action' — SnowWork is Snowflake's bid to own the autonomous workflow layer before agents commoditize everything below it.
Operator's Spotlight Read
Apple: The AI Toll Collector
Apple lags in frontier AI but earns over $1B from App Store fees on generative AI apps. Its bet: on-device AI powered by iPhone chips, wrapped in privacy branding. By letting rivals burn billions on infrastructure, Apple profits as the distribution layer while preparing to leapfrog with hardware integration. The company’s strategy hinges on its control of the iOS ecosystem—AI developers have little choice but to distribute through iPhones, paying Apple’s commission. Meanwhile, Apple is quietly investing in chip-level optimizations and partnerships (like Gemini integration for Siri) that could allow it to re-enter the AI race from a position of strength. The Insight: Apple profits as the distribution layer while rivals burn cash on infrastructure.
Operator Industry Radar
Perplexity Comet Browser → Now on iPhones, bringing agentic AI search and workflow automation directly into mobile browsing. Comet’s side-panel assistant can summarize, book, and purchase, turning browsing into an interactive workflow rather than static page viewing.
OpenClaw Agents → Nvidia’s open-source orchestration system shows promise in research, marketing, and travel hacks—but security concerns linger. Enterprises like ServiceNow are piloting claws with orchestration towers, reporting high resolution rates but still wary of giving agents sensitive logins.
Musk & Micron in Hardware Race → Musk’s Tesla and SpaceX AI keep ordering Nvidia chips while building in-house AI5 and Terafab fabs. Meanwhile, Micron forecasts record revenue from HBM chips but spooks investors with $25B+ capex plans. Together, they show how demand and supply are converging at scale in AI hardware.
Was this email forwarded to you? Don’t miss any updates Subscribe to TechWithAdit for sharp, no - noise tech intelligence stay sharp, – Adit
1