Sequoia Capital has redefined Artificial General Intelligence (AGI) as “the ability to figure things out,” a shift that coincides with a broader industry reassessment of AGI timelines and capabilities. The venture capital firm’s January 2026 essay, “2026: This is AGI,” argues that AGI is no longer a distant prospect but a present reality, functionally achieved through the development of “long-horizon agents.”
This new definition, described by Sequoia partners Pat Grady and Sonya Huang as a “functional” rather than “technical” one, marks a departure from previous, more elusive attempts to define AGI. Historically, researchers have struggled to establish a concrete definition, often responding to inquiries with the sentiment that they “will know it when we see it,” according to the Sequoia essay. The firm’s framing prioritizes practical application over strict adherence to human-level cognition.
The claim that AGI has arrived rests on three key components, according to Sequoia: pre-training, which established baseline knowledge as demonstrated by the emergence of large language models in 2022; inference-time compute, exemplified by OpenAI’s o1 series in late 2024, enabling reasoning over that knowledge; and the development of long-horizon agents capable of sustained, multi-step perform. Coding agents, such as Claude Code, are cited as the first tangible example of this functional AGI, demonstrating the ability to work autonomously for approximately 30 minutes.
The shift in definition has prompted discussion about the value proposition of AGI and a potential industry pivot toward “hybrid AI,” according to a Forbes report published on the same day as Sequoia’s essay. The vagueness of the new AGI definition is seen by some as a move away from unrealistic hype and toward more feasible applications of artificial intelligence.
Sequoia’s assessment is based, in part, on data from METR, a research organization tracking AI agent capabilities through task completion time horizons. METR’s research indicates a doubling time of roughly seven months over the past six years in the duration of tasks AI agents can complete with 50% reliability. This trend suggests a rapid acceleration in AI capabilities.
The firm’s argument is also framed as a pitch to portfolio companies, signaling a belief that the technology has matured enough to warrant significant investment and development. Examples of AI applications that Sequoia identifies as functioning as specialists include OpenEvidence for medicine, Harvey for law, XBOW for penetration testing, and Day AI for sales.
Despite the firm’s confidence, the definition of AGI remains a subject of debate. Sequoia acknowledges that its definition is “imprecise” and will not resolve philosophical arguments about consciousness or human-level intelligence. The firm’s focus is on the business implications of AI systems that can independently solve complex problems.
