The AI industry is undergoing a major rebranding as tech giants abandon the once-ubiquitous term “artificial general intelligence,” or AGI, in favor of a new wave of catchy, less ominous labels. Once the holy grail of AI development, AGI—defined as artificial intelligence that matches or exceeds human cognitive abilities—has fallen out of favor among top executives, who now see it as vague, overhyped, and increasingly problematic. CEOs across the industry are distancing themselves from the term. Dario Amodei of Anthropic calls AGI a “marketing term,” while OpenAI’s Sam Altman has dismissed it as “not a super useful term.” Google’s Jeff Dean avoids the conversation altogether, and Microsoft’s Satya Nadella has criticized the “AGI hype,” calling claims of achieving it “nonsensical benchmark hacking.” Even the original concept, which was first proposed in 1997 by researcher Mark Gubrud, has become a moving target—its definition varies wildly depending on who you ask, making it nearly impossible to measure. The shift is driven by several factors. First, the term has become entangled in legal and contractual confusion. The original 2019 OpenAI-Microsoft deal included an AGI clause that granted Microsoft rights to OpenAI’s technology upon AGI’s achievement. But when the contract was renewed in October, the terms changed: AGI would now require verification by an independent expert panel, and Microsoft wouldn’t lose access to the technology even after AGI was declared. The ambiguity has made the term increasingly impractical. Second, AGI has accumulated negative connotations. Years of doomsday rhetoric—books, headlines, and public warnings about AI destroying civilization—have led to public fatigue. What once seemed like responsible caution now feels like fear-mongering. As a result, companies are pivoting to branding that feels more grounded, human-centered, and less apocalyptic. Meta now promotes “personal superintelligence,” a vision where AI helps individuals achieve their goals, create, connect, and grow. Mark Zuckerberg framed it as a democratic alternative to centralized AI systems that could replace human labor. Microsoft, meanwhile, has introduced “humanist superintelligence,” emphasizing AI that serves people, solves real-world problems, and operates with purpose and restraint. The rebrand includes a new website with a warm, natural aesthetic and a focus on “approachable intelligence.” Amazon has adopted “useful general intelligence,” positioning its AI as practical, productive, and empowering—focused on making customers more capable rather than creating godlike systems. The company’s move to acquire Adept’s agentic AI technology underscores its commitment to building AI that can act autonomously on complex tasks. Anthropic, under Amodei, has settled on “powerful AI”—a term that describes a system smarter than any human across multiple domains, capable of writing novels, solving math problems, and coding at scale. It’s not just about answering questions; it’s about planning, executing, and learning over time, at speeds far beyond human capacity. The result is a new alphabet soup of AI visions: PSI, HSI, UGI, PI. While the core idea remains similar—systems that think, act, and adapt like humans—the language has shifted from existential dread to aspirational utility. Whether this rebranding will stick or simply delay the inevitable conversation about what true general intelligence really means remains to be seen. But one thing is clear: the era of AGI as a marketing term is over.