2026-04-01
By Vadym · Generated with AI, curated by me
The bill is coming due. Oracle fires 30,000 people to pay for data centers, the public turns skeptical just as adoption peaks, and the protocol that quietly wired every AI agent together just crossed 97 million installs. The money is moving — but so is the backlash.
Oracle began executing what may be the largest layoff in its history on March 31, with employees across the US, India, Canada, and Mexico receiving termination emails at 6 a.m. with no prior warning. TD Cowen estimates the cuts will hit 18% of Oracle’s 162,000-person workforce — up to 30,000 jobs — freeing $8–10 billion in cash flow. The money goes straight into an AI infrastructure buildout that requires an estimated $156 billion in capital spending. Oracle disclosed a $2.1 billion restructuring plan in its March 2026 SEC filing.
Physical Intelligence, the two-year-old robotics startup founded by former Google DeepMind researchers, is in talks to raise approximately $1 billion at a valuation exceeding $11 billion, according to Bloomberg. That’s a near-doubling from its $5.6 billion valuation just four months ago. Founders Fund is set to participate, with Lightspeed Venture Partners and returning backers Thrive Capital and Lux Capital also in talks. The 80-person company is building general-purpose AI models that can power robots to perform tasks from folding laundry to peeling vegetables. Co-founder Sergey Levine described it simply: “Think of it like ChatGPT, but for robots.”
The Model Context Protocol hit 97 million monthly SDK downloads in March 2026, up from roughly 2 million at launch in November 2024. Every major AI provider — including the teams behind GPT, Gemini, and most agent frameworks — now ships MCP-compatible tooling. The ecosystem has grown from a handful of reference implementations to over 5,800 community and enterprise servers covering databases, CRMs, cloud providers, and developer tools. The White House’s national AI policy framework released March 20 specifically identified agentic AI infrastructure as a priority area for investment and governance.
There’s a tension running through today’s stories that won’t resolve itself quietly. The industry is spending at a scale that requires sacrificing its own workforce. The public is using AI tools while increasingly believing those tools will hurt them. And the infrastructure layer — protocols, models, datasets — is maturing faster than the trust and governance needed to manage it. Oracle fires 30,000 to build AI. Physical Intelligence raises a billion to teach robots to fold shirts. MCP wires 97 million agent connections with no security standard in sight. The technology is accelerating. The social contract around it is fraying. 2026 isn’t the year AI arrives — it’s the year we find out whether anyone planned for what happens after it does.
— Boba
Curated by Vadym