← Back to scans

Will any company announce that it has achieved Artificial General Intelligence (AGI) before Jul 1, 2026?

KXAGICO-COMP-26Q2 · Companies · 2026-05-17
3%
Agent
3%
Market Price
+0.0%
Edge
85%
Confidence
Volume: 11,539
Spread: 1.0c
Days to resolution: 44
Markets in event: 13
Final Rationale
My independent estimate of 4.3% was anchored to a near-zero base rate (no company has ever formally announced AGI) with a small tail for definitional arbitrage from claims like Huang's March podcast remark. I partially accept the critique: 7 weeks of non-resolution post-Huang is informative evidence that podcast-style remarks don't qualify under the resolution criteria, and sq3 (scientific credibility) may have been underweighted as a downward factor since I was already discounting via sq1/sq2. With only ~44 days remaining, no pre-signals from frontier labs, OpenAI's IPO/Microsoft IP constraints discouraging bold claims, and a deep market trending down to 3%, I align with the market at 3%.
Price vs. prediction
How the market moved after the agent's call. Solid blue line is the actual market price; dashed line is the agent's static forecast at $t{=}0$.
$n = 58$ follow-ups
Re-scan Context
This market has been scanned before. Previous predictions:
DatePredictedMarket PriceConfidence
2026-05-10 4% 4% 85%
2026-05-03 7% 5% 65%
2026-04-22 3% 2% 72%
1. Planner Opus · Fermi decomposition + tool plan
Tools chosen: kalshi_data kalshi_orderbook article_search web_search wikipedia_lookup
Sub-questions (Fermi decomposition)
  1. Has any major AI company (OpenAI, Google DeepMind, Anthropic, Meta AI, xAI, etc.) made a credible public claim of achieving AGI as of mid-May 2026? (w=0.40)
  2. Will any AI company make a new announcement claiming to have achieved AGI between now (May 17, 2026) and July 1, 2026? (w=0.40)
  3. Is the AI field currently at a stage where an AGI announcement would be scientifically credible and meet any reasonable definition of AGI by July 1, 2026? (w=0.20)
Planner reasoning
This question asks whether any company will officially announce AGI before July 1, 2026 — roughly 44 days from now. The key challenges are: (1) definitional: 'AGI' lacks a universally agreed definition, making any announcement inherently contested; (2) prior base rates: no credible AGI announcement has ever been made despite years of AI progress; (3) current state: AI systems in 2026 remain narrow specialists despite impressive capabilities across benchmarks. The reference class here is 'AI milestone announcements' — companies have made bold claims (GPT-4, Gemini Ultra, Claude 3 Opus) but none have claimed full AGI. OpenAI's charter references AGI and they have internal processes for declaring it, creating one plausible path, but their incentives are mixed (legal/regulatory risks of such a claim). The temporal analysis: in 44 days, we need a company to complete development, validate internally, prepare an announcement, and publicly declare AGI. Even if a company were close to such a threshold today, the coordination and legal review required makes a surprise announcement in this window unlikely. The main scenario for YES involves OpenAI or a competitor having a sudden breakthrough and making a marketing/PR announcement (potentially contested by experts). The prediction market price will be the most informative signal given this is a straightforward event resolution question.
2. Research Parallel tool calls · legacy v2
ToolStatusTimeItemsSummary
kalshi_data OK 1.6s -
kalshi_orderbook OK 0.2s -
article_search OK 0.7s -
web_search OK 68.5s -
wikipedia_lookup OK 1.2s -
3. Evidence Synthesis Sonnet · 15 items · legacy v2
#StrengthCredibilityDirectionSourceClaimPriced In
1 MODERATE 62 ↑ UP web_search Nvidia CEO Jensen Huang stated 'I think we've achieved AGI' on the Lex Fridman podcast (~March 23-24, 2026), but this was in the context of a narrow definition (AI building a $1B company) that few AI researchers accept, and Huang himself contradicted the claim in the same podcast. Yes
2 STRONG 72 ↓ DOWN web_search The Huang AGI claim was widely criticized by AI researchers as using a non-standard, narrow definition of AGI, and expert consensus does not accept it as a credible AGI achievement. Yes
3 MODERATE 68 ↓ DOWN web_search Huang appeared to contradict his own AGI claim in the same podcast by stating 100,000 AI agents could never replicate what Nvidia does, undermining the credibility of the announcement. Yes
4 STRONG 80 ↓ DOWN article_search No articles from February to mid-May 2026 report a formal, official AGI announcement from OpenAI, Google DeepMind, Anthropic, Meta AI, or xAI; the field continues to discuss incremental AI capability improvements and commercial deployments. Yes
5 WEAK 55 ↓ DOWN article_search Amazon's AGI lab head David Luan departed in February 2026, suggesting Amazon's AGI research is in transition rather than approaching a major announcement. Yes
6 MODERATE 70 ↓ DOWN article_search Anthropic is raising funds at a $900B valuation and OpenAI at $852B — both focused on commercial AI products and infrastructure, not AGI milestone announcements. Yes
7 MODERATE 65 ↓ DOWN article_search OpenAI is focused on data center scaling challenges and a Pentagon deal rather than any AGI milestone; Sam Altman's public appearances concern infrastructure and commercial growth. Yes
8 STRONG 82 ↓ DOWN kalshi_data The Kalshi prediction market 'Before Jul 1, 2026' for AGI announcement is priced at 3%, down 1% over 7 days and down 7% over 30 days, with high liquidity and depth of 3,146 contracts. Yes
9 STRONG 84 ↓ DOWN kalshi_orderbook The Kalshi orderbook shows yes_bid at $0.01 and no_bid at $0.04, with a wide spread of $0.95 and high liquidity, indicating strong market consensus against a YES outcome before July 1, 2026. Yes
10 MODERATE 55 NEUTRAL web_search Huang's AGI claim on Lex Fridman drove Polymarket's AGI-before-2027 probabilities to ~23%, but this is for a longer horizon and reflects definitional ambiguity, not a formal company announcement. Yes
11 WEAK 42 NEUTRAL web_search Sequoia Capital stated 'long-horizon agents are functionally AGI, and 2026 will be their year,' but this represents an investor/VC definition rather than a formal company AGI announcement. Yes
12 MODERATE 65 ↓ DOWN article_search OpenAI is currently engaged in active litigation (Musk v. Altman trial, May 2026) and IPO preparation activities, creating regulatory and legal incentives to avoid bold AGI claims that could trigger charter obligations or regulatory scrutiny. Yes
13 MODERATE 60 ↓ DOWN web_search Research on the Centaur AI model claimed it could mimic human thinking across 160 cognitive tasks, but follow-up research challenged this, finding it memorizes patterns rather than truly reasoning — no AGI-level scientific breakthrough confirmed. Yes
14 STRONG 75 ↓ DOWN web_search The AI research community lacks consensus on what AGI means, with Huang's definition (building a $1B company) rejected by most researchers, making any formal announcement inherently contested and unlikely to be 'scientifically credible' under mainstream definitions. Yes
15 MODERATE 65 ↓ DOWN article_search Current frontier AI activity in early-mid 2026 centers on agentic AI systems, multi-modal models, and commercial deployment — not fundamental breakthroughs signaling human-level general intelligence. Yes
Information Gaps
  • No direct data on internal OpenAI AGI threshold assessments or whether their board has evaluated any AGI milestone under their charter
  • No information on Google DeepMind's current research status or any planned announcements in the May-July 2026 window
  • No data on xAI (Elon Musk's company) or Meta AI's near-term announcement plans
  • No information on whether any company has privately communicated to investors or regulators about approaching an AGI threshold
  • The question's resolution criteria ('officially announces') is not precisely defined — unclear if informal statements like Huang's podcast comment would qualify
  • No data on whether the Kalshi market question has a specific definition of AGI for resolution purposes, which significantly affects interpretation
Key Uncertainties
  • Resolution criteria ambiguity: Does the Huang podcast statement count as a company 'officially announcing' AGI? If so, this may already be resolved YES, but market pricing at 3% suggests the market does not think so
  • OpenAI charter trigger: OpenAI's charter specifies AGI as when systems outperform humans at most economically valuable work — unclear if any internal assessment near this threshold exists
  • Surprise announcement risk: A major AI lab could announce an unexpected breakthrough in the 44-day window with no prior public signal
  • Definitional arbitrage: A company could announce AGI using a narrow/nonstandard definition that the market resolves YES on even if experts reject it
  • Legal/regulatory incentive effects: IPO prep and ongoing litigation may suppress or amplify AGI announcement likelihood for OpenAI specifically
Market Context
Price: 0.03 · 7d: -0.01 · 30d: -0.07
Volume: LOW-MODERATE — averaging 202 contracts/day · Depth: HIGH — 3,146 contracts depth
Reliability: MEDIUM-HIGH — meaningful liquidity and depth, strong directional signal toward NO, but low volume limits confidence in precise probability calibration. Price has declined significantly over 30 days, indicating the market moved away from YES after initial hype (likely related to Huang's podcast comments).
4. Independent Forecaster Opus · No market price · legacy v2 Did NOT see market price
Independent estimate: 4% (81% confidence)
base rate
4%
82% conf
evidence driven
4%
82% conf
contrarian
5%
80% conf
Spread
1pp
Low disagreement
sq1: Has any major AI company made a credible public claim of achieving AGI as of mid-May 2026?
8%
Base rate: No company has ever made a formal, credible AGI announcement. Huang's podcast comment is the closest claim but was (a) self-contradicted in the same podcast, (b) used a non-standard definition, (c) rejected by researchers, and (d) not a formal Nvidia corporate announcement. Major labs are focused on commercial/IPO concerns. The 8% reflects mostly definitional ambiguity around whether Huang's statement could be construed as a 'company announcement' for resolution purposes.
sq2: Will any AI company make a new AGI announcement between May 17 and July 1, 2026?
3%
Base rate for a 44-day window of a never-before-occurring event is very low (~0.5-1%). No pre-signals from any major lab. OpenAI's IPO prep and Musk litigation create strong DOWN pressure on bold AGI claims (charter would trigger Microsoft IP implications). Tail risk: surprise xAI/Musk announcement for marketing, or a narrow-definition claim by a smaller player. Adjusting up to 3% for these tail scenarios.
sq3: Is the field at a stage where an AGI announcement would be scientifically credible by July 1, 2026?
5%
Current frontier AI remains narrow despite agentic advances. No consensus definition exists; Centaur-type claims have been debunked. Scientific credibility threshold is very high and unmet.
Show full reasoning trace
base rate: No company has ever made a formal AGI announcement in the entire history of AI research. For a 44-day window, base rate is ~0.5-1%.
evidence updates: Huang's podcast comment pushes up slightly (~+2%) due to definitional ambiguity but has not triggered resolution in ~7 weeks. OpenAI IPO/litigation context pushes down (~-1%). No pre-signals from any major lab in the May-July window pushes down.
combination method: Weighted average per spec, with adjustment for positive correlation between sub-questions.
final: 4% reflects strong base rate of zero historical occurrences, no pre-signals, hostile legal/regulatory environment for OpenAI, and definitional barriers, balanced against small tail risk of a marketing-driven announcement (xAI, Musk) or definitional arbitrage.
ensemble: {'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.04, 'confidence': 0.82}, 'evidence_driven': {'p_yes': 0.04, 'confidence': 0.82}, 'contrarian': {'p_yes': 0.05, 'confidence': 0.8}}, 'spread': 0.01, 'individual_p_yes': {'base_rate': 0.04, 'evidence_driven': 0.04, 'contrarian': 0.05}}
5. Devil's Advocate Sonnet · legacy v2 Market price revealed
Reasoning Flaw
Challenges
  1. The forecast at 4.3% is already very close to the market at 3%, and the reasoning is sound. The main residual question is whether 4.3% slightly overweights the Huang-as-resolution-trigger tail given that ~7 weeks have passed without resolution triggering — this is informative evidence that the resolution criteria likely don't count podcast remarks.
  2. Sub-question 1 at 8% may be slightly high: if Huang's March comment were going to be construed as a qualifying 'company announcement,' resolution mechanics would likely have already moved the market or triggered discussion. The fact that 7 weeks have passed with the market drifting DOWN suggests the resolution authority is not treating it as qualifying. Consider lowering sq1 to ~3-5%.
  3. The combination method is not fully transparent — how exactly are sq1, sq2, sq3 combined into 4.3%? sq1 (already happened) and sq2 (will happen in window) should be combined as P(sq1 OR sq2), roughly 0.08 + 0.03 - overlap ≈ 10%, but the final is 4.3%. This suggests heavy discounting via sq3 (scientific credibility), which may be double-counting since the resolution criteria may not require scientific credibility, only an announcement.
Suggested adjustment: -1pp
Minor downward adjustment warranted because (a) 7 weeks of non-resolution post-Huang is itself evidence that podcast-style claims won't trigger resolution, slightly lowering sq1, and (b) the combination math appears to discount via 'scientific credibility' (sq3) which may not be part of the resolution criteria. Both point to a forecast closer to 3-3.5%. Not adjusting to match market — adjusting because the time-since-Huang evidence wasn't fully incorporated.
Market comparison: Forecaster: 4.3%, Market: 3.0%. Divergence of 1.3pp is small and within reasonable disagreement. Market has high orderbook depth and clear downward trend (-7pp over 30d), suggesting consensus that Huang's comments won't trigger resolution. The forecaster's slight premium over market is defensible as tail-risk pricing but could be trimmed given the informativeness of 7 weeks of non-resolution.
6. Aggregator Opus · final calibrated forecast
3% (85% confidence)
My independent estimate of 4.3% was anchored to a near-zero base rate (no company has ever formally announced AGI) with a small tail for definitional arbitrage from claims like Huang's March podcast remark. I partially accept the critique: 7 weeks of non-resolution post-Huang is informative evidence that podcast-style remarks don't qualify under the resolution criteria, and sq3 (scientific credibility) may have been underweighted as a downward factor since I was already discounting via sq1/sq2. With only ~44 days remaining, no pre-signals from frontier labs, OpenAI's IPO/Microsoft IP constraints discouraging bold claims, and a deep market trending down to 3%, I align with the market at 3%.