The Hardware & Infrastructure Build-Out

The financial and physical scaffolding for the AI boom continues to expand at a staggering pace. Amazon is making a fresh $13B AI infrastructure investment in India, joining the global race to build out compute capacity in key markets. At the chip level, a deep dive into the math behind the OpenAI Jalapeño chip reveals the intense economic pressure driving custom silicon development to escape Nvidia’s margins. However, not all hardware news is positive: Cerebras stock plunged after its first earnings report as investors balked at a forecast for narrower gross margins. On the networking side, startup Netris raised a $15M Series A from a16z to provide software that helps “neocloud” AI infrastructure providers get their networks live faster. Meanwhile, the memory chip crunch is paying off handsomely for at least one US company, which saw its profit soar from $1.88 billion to $28.2 billion year-over-year.

AI Agents & Productivity Tools

The promise of AI to transform work is moving from theory to practice. OpenAI published a new research paper on how AI agents are transforming work, demonstrating their ability to handle longer, more complex tasks. Google’s DeepMind announced it is introducing computer use in Gemini 3.5 Flash, allowing the model to take actions like moving a cursor and clicking on a user’s screen—a key step toward practical agentic workflows. To make these multi-step agents more efficient, MIT researchers unveiled a new system called Murakkab that optimizes the speed and energy-efficiency of AI agent workflows. On the product side, Meta has revived its Facebook Creator Studio as an AI companion app, centered around an AI Creator Assistant to help users grow their audience. Separately, companies are scrambling to stop employees from maxing out AI budgets with small tasks, signaling a shift from unchecked experimentation to managed, cost-effective deployment.

Startups, Acquisitions & Policy

The corporate landscape is shifting rapidly. In a major acquisition for creative professionals, Adobe has acquired image and video enhancement tool maker Topaz Labs, planning to integrate its AI-powered tools across the Creative Cloud. A notable new challenger emerged as former Infosys chief Vishal Sikka launched a new startup, backed by Mayfield and Aramco Ventures, aiming to disrupt the traditional IT services world with AI. In robotics, Agility Robotics plans to go public via SPAC in a $2.5B deal, a major milestone for the humanoid robotics sector. On the policy front, Europe is pushing back on Washington’s chip war with China, concerned about the broad reach of proposed U.S. export controls on legacy chipmaking equipment.

The Human Factor

Amidst the technological sprint, human expertise and oversight remain critical. Ford had to hire back former engineers to fix mistakes made by its automated systems, a cautionary tale about over-reliance on AI and automation in complex manufacturing. In the job market, contrary to doom-laden predictions, new data suggests engineering jobs are the most resilient to AI disruption, making up a larger share of new hires. However, the competition for top AI talent is fierce: AI researchers continue to leave Google for its rivals, with the latest departures headed to Anthropic.

---

Editorial Take: Today’s stories paint a picture of an industry in a massive, multi-front investment phase. The billions flowing into chips, clouds, and networks (Hardware) are the bets on the table, while the work on agents and productivity tools (Research, Products) represents the race to build the most valuable things on top of that infrastructure. The underlying theme is cost and control: from custom chips to budget rationing, the era of blank-check AI spending is giving way to a more calculated, efficiency-driven chapter. The Ford story is a crucial reminder that for all this sophistication, judicious human oversight remains irreplaceable.