🤖 Artificial Intelligence

The CEO of the world’s most valuable AI chip company just made the boldest statement in tech history — and the legal, financial & industry consequences are seismic.

01
Understanding AGI: The Most Contested Term in Technology

Artificial General Intelligence (AGI) refers to an AI system capable of performing any intellectual task a human can — not just narrow tasks like text generation, but all cognitive functions: reasoning, planning, creativity, and learning across every domain.

The Problem: There Is No Universal Definition

On the Lex Fridman Podcast, Fridman offered a benchmark: an AI that can “essentially do your job” — meaning start, grow, and run a successful tech company worth over $1 billion. This is remarkably close to the definition in the OpenAI–Microsoft partnership agreement, where AGI declaration triggers enormous financial consequences.

02
The Exact Exchange on Lex Fridman Podcast #494

The episode — titled “Jensen Huang: NVIDIA — The $4 Trillion Company & the AI Revolution” — was published March 23, 2026. Fridman asked Huang to choose between 5, 10, 15, or 20 years until AGI arrives. Huang rejected all options in favour of “now.”

The Walk-Back — And What It Reveals

Huang quickly added nuance: the odds of 100,000 AI agents collectively building a company like Nvidia remain at zero percent. His definition centres on agentic utility — AI that independently performs high-value real-world tasks — rather than a superintelligence capable of replicating human civilisation.

January 2026 — Apple picks Google Gemini for Siri
Apple signs landmark deal to integrate Gemini into a rebuilt Siri, signalling AI has entered a new commercial phase.

February 2026 — Samsung Galaxy S26 launches at MWC
Samsung debuts its Galaxy AI suite with on-device AI processing capabilities previously unseen in consumer hardware.

March 20, 2026 — Samsung confirmed as exclusive HBM4 supplier for OpenAI’s “Titan” chip
OpenAI prepares its first custom chip built specifically for frontier AI workloads.

March 23, 2026 — Jensen Huang declares “We’ve achieved AGI”
The statement lands on Lex Fridman Podcast #494 and instantly becomes the #1 trending tech story globally.

03
Why Nvidia’s Position Makes This Extra Significant

Nvidia manufactures the GPUs that power virtually every major AI model in existence — OpenAI’s GPT series, Google’s Gemini, Anthropic’s Claude, Meta’s Llama. Huang has closer visibility into the actual computational capabilities of frontier AI than almost anyone on earth. When he says AGI is here, he’s drawing on direct knowledge of what these systems can do at scale.

Browse & Research
Autonomously search the web, read articles, compare sources, and compile structured reports — no human in the loop.

Write & Deploy Code
Write, test, debug, and deploy functional software from a single natural-language instruction.

Manage Workflows
Handle calendars, emails, scheduling, and multi-step business processes end-to-end with no human oversight.

Spawn Sub-Agents
Spawn specialised child agents for parallel sub-tasks simultaneously, accelerating complex projects dramatically.

FCC Bans All Foreign Consumer Routers
The FCC added all foreign-made consumer Wi-Fi routers to its national security Covered List, citing Volt, Flax & Salt Typhoon cyberattacks on US infrastructure. Around 100% of routers sold in the US are made abroad — the entire industry must pivot.
Read on The Verge →

Apple WWDC 2026 Starts June 8 — Gemini-Powered Siri Expected
Apple confirmed WWDC 2026 begins June 8, teasing “AI advancements.” Following the confirmed Google Gemini deal for Siri, iOS 27 with a fully rebuilt AI assistant is expected to headline.
Read on The Verge →

Samsung Adds Native AirDrop to Galaxy S26
Samsung is rolling out native Apple AirDrop support to the Galaxy S26 via Quick Share — starting South Korea on March 23, then going global. No third-party app required.
Read on The Verge →

Gimlet Labs Raises $80M — Solves AI Inference Bottleneck
Stanford-backed startup Gimlet Labs raised an $80M Series A for tech that lets AI models run simultaneously across Nvidia, AMD, Intel, ARM, Cerebras, and d-Matrix chips, removing a critical infrastructure chokepoint.
Read on TechCrunch →

1
Has AGI actually been achieved in 2026?

Jensen Huang says yes — with an important qualification. His definition centres on AI agents capable of performing real, high-value tasks autonomously, not AI that can single-handedly build a $4 trillion company. The broader scientific community has not reached consensus.

2
What is Nvidia’s role in AGI development?

Nvidia manufactures the GPUs powering virtually every major AI model — GPT, Gemini, Claude, and Llama. The company’s Blackwell and H100 chips are the primary hardware on which these systems are trained and deployed. Nvidia’s $4 trillion valuation is directly tied to global AI infrastructure demand.

3
Why does the AGI definition matter legally?

The OpenAI–Microsoft partnership agreement contains contractual clauses directly tied to the achievement of AGI. If AGI is officially declared, it could trigger billions in restructured payments and fundamentally change how Microsoft accesses OpenAI’s most powerful models.

4
What is OpenClaw and why did Huang mention it?

OpenClaw is a viral open-source AI agent platform enabling individuals to deploy autonomous agents for complex, multi-step real-world tasks. Huang cited it as his primary evidence that AI has reached AGI-level utility — millions of people are already using autonomous agents independently and successfully.

5
Will the FCC router ban affect my home router?

No. Existing routers already in US homes are grandfathered — you can keep using your current device. The ban applies to new imports going forward. Manufacturers can apply for “conditional approval” to continue US sales while committing to move production stateside.

🚀

📚 Related Reading on Networkcraft

→

→

→