When looking at prediction markets, crowdforecasting platforms, and institutional sentiment, the consensus on who will reach Artificial General Intelligence (AGI) first is no longer a one-horse race. While OpenAI held a commanding lead in market sentiment for years, a fierce three-way battle for the top spot has emerged between Anthropic, Alphabet (Google), and OpenAI.Data from prediction markets like Metaculus, alongside tracking on Kalshi and Polymarket, breaks down how the market views the frontrunners.
The Market Leaderboard
Aggregate forecasting metrics isolate the probability of which entity will debut the first “weakly general” AI system:
| Company / Entity | Market Probability (Consensus) | Core Market Thesis |
| Anthropic | ~29% to 30% | Favored for their rapid algorithmic scaling, strong positioning in end-to-end coding automation, and internal focus on building self-correcting AI research loops. |
| Alphabet (Google) | ~28% to 29% | Seen as a massive threat due to unparalleled compute infrastructure, data advantages, and the multi-modal integration seen across the Gemini ecosystem. |
| OpenAI | ~21% to 22% | The historical favorite; still heavily backed by retail sentiment and Kalshi macro-timelines (e.g., a 55% chance to hit AGI by 2030), though recent execution timelines have compressed their lead. |
| Other / Decentralized | ~10% to 11% | Includes wildcards like Meta (FAIR), xAI, or open-source consortiums. |
The Underlying Macro Views
The broader capital and prediction markets are trading on a few distinct shifts in how AGI is expected to manifest:
- Corporate vs. State Supremacy: Prediction markets place an 81% to 88% probability that the first AGI will be developed by a for-profit corporation rather than a nationalized government project or university group.
- The Shift to “Functional AGI”: Traditional prediction metrics for absolute human-level AGI across all domains skew out into the 2030s. However, the market is actively pricing in the arrival of “functional AGI”—systems characterized by long-horizon autonomous agents capable of independent, complex workflows (like software engineering or multi-step scientific synthesis)—much sooner, with significant volatility expected across labor and equity markets over the next 12 to 24 months.
- Algorithmic Efficiency Over Raw Compute: The market has recently upgraded Anthropic and Google because compute bottlenecks have made purely expanding cluster sizes less efficient. Markets are betting heavily on companies that can maximize “test-time compute” (reasoning at the point of output) and non-linguistic logic benchmarks, like ARC-AGI 2.
Are you tracking this from a capital allocation standpoint, or are you more curious about the specific technical benchmarks (like ARC-AGI or HLE) these companies are hitting?