When will any company achieve AGI?

The key reason to take this outcome seriously is that the market only requires a corporate announcement, not an objective scientific consensus. If one frontier lab believes it has built a system that meets its internal AGI threshold, or wants to position a model as such for strategic reasons, a declaration before the deadline is entirely plausible. Recent public timelines from major AI figures have moved into the 2026–2027 range, which materially raises the odds that at least one company will feel confident enough to use AGI language within the market window.

Against that, AGI remains a moving target, and firms have strong incentives to avoid making a claim that could trigger backlash, regulation, or embarrassment if experts disagree. The more capable the systems become, the more likely companies are to use softer labels such as “frontier,” “general-purpose,” or “powerful AI” rather than the loaded term AGI. That makes the event less about raw capability progress and more about corporate messaging discipline, which keeps the probability below a coin flip.

At 34% yes, the market appears somewhat too low relative to how near-term the leading-lab rhetoric has become. The market may be overweighting older expert surveys and underweighting the fact that a single self-interested announcement from any major lab is enough to resolve yes. Still, because the definition problem is real and the reputational downside of overclaiming is large, I would not push the probability much above the low-to-mid 40s.