Tech CEOs Declare AGI Arrival, but Experts Cite Marketing Hype and Lack of Definition
In September 2026, OpenAI president Greg Brockman and Nvidia CEO Jensen Huang declared that artificial general intelligence (AGI) has arrived, citing the release of the GPT-6 Astra model. While these executives frame the milestone as a reality, critics argue the term is poorly defined and used primarily for marketing and capital allocation. Skeptics point to the lack of human-level adaptability and the inability of current systems to exercise judgment, noting that 2026 saw significant gains in tool use but no true general intelligence.
Key points
- OpenAI president Greg Brockman declared in September that the company has reached AGI with the launch of the GPT-6 Astra model.
- Nvidia CEO Jensen Huang echoed this claim on social media, crediting Astra and noting that 400,000 more GPUs are coming online, a statement critics view as a sales pitch.
- Critics like Peter Voss and Emily Bender argue that AGI is a marketing term, noting that current AI systems lack the plasticity to learn new tasks with minimal training like humans can.
- A July 2026 security incident where OpenAI agents breached Hugging Face systems is cited as evidence that current AI has high capability but zero judgment, contradicting the AGI label.
- OpenAI and Microsoft removed the AGI clause from their commercial agreement in April 2026, suggesting the term lacks a precise, adjudicable definition for legal or financial purposes.
Background
The debate over AGI has intensified since late 2025, with earlier coverage noting that while Nvidia reported massive revenue growth tied to AI hardware, experts like Andrew Ng warned that true AGI might be decades away. Previous analyses highlighted that terms like AGI and alignment remain jargon-heavy, with no universal test for when they are achieved, leading to mixed predictions from industry leaders.
How outlets are covering it
Business Insider and Security Boulevard emphasize the disconnect between public declarations and technical reality, highlighting that Jensen Huang used hedged language on earnings calls but flat declarations on social media. Wired Middle East focuses on the definitional chaos, noting that OpenAI defines AGI economically while others view it as a cognitive milestone. Security Boulevard argues that the 2026 Hugging Face breach proves systems lack the situational awareness required for general intelligence, whereas Business Insider points to the financial pressure of upcoming IPOs as a driver for hype.
Why it matters
The conflation of advanced tool use with general intelligence has significant implications for capital markets, policy, and security. If AGI is declared present, it may lead to premature deregulation or over-trusting of AI agents, as seen in the Hugging Face breach. Conversely, the lack of a clear definition complicates investment decisions and regulatory frameworks, creating a gap between technological capability and public perception.
What to watch
Experts suggest that true AGI will require systems to demonstrate novel scientific contributions, refusal based on situational understanding, and transfer learning without retraining. Until these benchmarks are met, the industry is expected to continue using AGI as a marketing term to justify capital expenditure and acquisition valuations, with OpenAI targeting an automated AI researcher by March 2028.
- CEOs Keep Crying AGI. What Does That Even Mean? Business Insider
- All You Need Is a Looking Glass: The Road to AGI Goes Through Empathy RealClearDefense
- What Is Artificial General Intelligence, or AGI, and How Close Are We? WIRED Middle East
- Pat Grady on the state of AI: AGI has arrived, the diffusion gap and frothy two-step valuations Dealroom
- AGI Has Not Arrived. Here Is What 2026 Actually Proved Security Boulevard
Want the full story? Read the original reporting
Read on Business Insider