Scan: When will nuclear fusion be achieved?

← Back to scans

When will nuclear fusion be achieved? — Before 2030

KXFUSION-30-JAN01 · Science and Technology · 2026-05-17

30%

Agent

28%

Market Price

+2.4%

Edge

35%

Confidence

Volume: 28,271

Spread: 0.4c

Days to resolution: 1324

Markets in event: 3

Final Rationale

My independent estimate of 28% reflected a ~75/25 weighting toward strict resolution (sustained/wall-plug net gain) over loose (NIF-style scientific ignition counting), combined with sub-forecasts showing SPARC and Helion as the only credible pre-2030 strict-interpretation candidates with historically slip-prone timelines. I partially accept Challenges 2 and 3 from the critique: the disjunctive combination of sq1/sq2/sq3 does mechanically imply a higher number, and my strict-interpretation weight was partly justified by circular market inference, so I nudge up ~2pp to 30%. I reject Challenge 4 as already incorporated (the contrarian ensemble member at 0.34 captures regime change from private capital/HTS magnets). The market at 27.6% aligns near-perfectly with my estimate, but given LOW-MEDIUM liquidity and the legitimate internal-consistency issue flagged, I sit slightly above market rather than anchoring exactly.

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:

Date	Predicted	Market Price	Confidence
2026-05-09	32%	32%	35%
2026-05-02	40%	38%	42%
2026-04-06	42%	40%	45%

1. Planner Opus · Fermi decomposition + tool plan

Tools chosen: article_search web_search kalshi_data wikipedia_lookup code_execution

Sub-questions (Fermi decomposition)

Has nuclear fusion already been 'achieved' (net energy gain / ignition) as of today, May 2026, under any credible scientific definition? (w=0.15)
Will a government lab or private company demonstrate sustained net energy gain fusion (Q>1 on a wall-plug or scientific basis) before January 1, 2030? (w=0.50)
Will ITER, NIF, Commonwealth Fusion, TAE Technologies, Helion, or another leading program reach a meaningful fusion milestone (ignition or Q≥1) by end of 2029? (w=0.35)

Planner reasoning

This question hinges critically on the definition of 'nuclear fusion achieved.' NIF already achieved scientific ignition (Q~1.5 on laser energy) in December 2022, but wall-plug efficiency is far below 1 (lasers used ~300x more grid energy than fusion output). The market likely requires a more robust standard — either scientific net energy gain at a meaningful scale, or commercial/practical net energy gain. Key causal mechanisms to assess: (1) NIF's path to higher yield shots and whether Q>1 on wall-plug is achievable by 2029 — very unlikely given current laser efficiency limits; (2) Commonwealth Fusion's SPARC tokamak timeline, targeting first plasma around 2025-2026 and net energy demonstration around 2027-2028 — this is the most credible near-term candidate but still faces substantial engineering risk; (3) Helion's ambitious 2028 commercial target (with Microsoft contract) — technologically aggressive; (4) ITER has slipped to 2033+ for first plasma. Reference class: Major physics milestones consistently slip 5-15 years from initial estimates. The NIF 2022 result shows progress is real but the gap between scientific ignition and practical fusion remains enormous. Prediction markets on fusion have historically priced this well below 10%. Temporal analysis: With 1,324 days (~43 months) remaining, the realistic sequence for YES would require: (a) a program like SPARC or Helion completing construction by mid-2027, (b) achieving first plasma by late 2027, (c) demonstrating net energy gain by 2028-2029. Each step has historically slipped. The probability is non-trivial but low — perhaps 5-15% depending on definition strictness.

2. Research Parallel tool calls · legacy v2

Tool	Status	Time	Items
article_search	OK	0.8s	-
web_search	OK	64.8s	-
kalshi_data	OK	7.3s	-
wikipedia_lookup	OK	1.8s	-
code_execution	OK	0.0s	-

3. Evidence Synthesis Sonnet · 18 items · legacy v2

#	Strength	Credibility	Direction	Source	Claim	Priced In
1	STRONG	88	NEUTRAL	wikipedia_lookup	As of 2025, NIF is the only laboratory to have demonstrated a fusion energy gain factor above one (scientific Q>1), but efficiencies orders of magnitude higher are required for engineering or economic breakeven.	Yes
2	STRONG	72	NEUTRAL	web_search	NIF's December 2022 ignition shot produced 3.15 MJ from 2.05 MJ laser input (scientific Q~1.5); subsequent 2025 tests yielded up to 8.6 MJ output, but wall-plug efficiency remains orders of magnitude below 1.	Yes
3	MODERATE	75	↑ UP	web_search	China's EAST reactor broke plasma density limits in January 2026 using AI-controlled stabilization, removing a key obstacle to ignition — but this is not a net energy gain achievement.	Yes
4	WEAK	60	NEUTRAL	web_search	UK startup Pulsar Fusion achieved plasma ignition inside a fusion rocket engine in March 2026, a novel milestone but not a net energy gain demonstration.	No
5	MODERATE	70	↓ DOWN	article_search	As of January 2026, expert commentary frames fusion as still far from commercial viability despite recent milestones, with no program having achieved wall-plug Q>1.	Yes
6	STRONG	92	↓ DOWN	wikipedia_lookup	ITER's first plasma is now expected in 2033–2034, well outside the 2030 window, making it irrelevant for this question.	Yes
7	MODERATE	78	↑ UP	wikipedia_lookup	Commonwealth Fusion Systems (CFS) is constructing the SPARC tokamak in Devens, MA, targeting net energy gain; as of mid-2025 construction is ongoing with magnet assembly underway.	Yes
8	MODERATE	75	NEUTRAL	article_search	CFS's SPARC reactor is under construction with powerful superconducting magnets being assembled; no announcement of first plasma or net energy gain has been made as of mid-2026.	Yes
9	MODERATE	74	↓ DOWN	article_search	TAE Technologies is targeting a fusion power plant by 2031 per its CEO, placing net energy demonstration outside the 2030 window.	Yes
10	MODERATE	80	↓ DOWN	article_search	Trump Media merged with TAE Technologies in a $6B deal (Dec 2025), providing capital but not accelerating the technical timeline; TAE's first plant target is 2031.	Yes
11	STRONG	68	NEUTRAL	web_search	NIF's 2025 shots yielded up to 8.6 MJ fusion output — a 4x increase over the 2022 ignition shot — but the lasers still consume ~300x more grid energy than fusion output, making wall-plug Q>1 unachievable with current NIF laser technology.	No
12	WEAK	35	↑ UP	code_execution	A programmatic analysis estimated ~46% probability that at least one program achieves net energy gain by 2030, driven primarily by CFS/SPARC (~20%) and NIF (~15%) individual confidences — but these individual confidences reflect optimistic analyst priors, not calibrated forecasting.	No
13	MODERATE	65	↓ DOWN	kalshi_data	Kalshi prediction market prices the 'Before 2030' fusion achievement market at 27.6%, down 3.2% in 7 days and down 8.4% in 30 days, suggesting declining market confidence.	Yes
14	STRONG	93	↓ DOWN	wikipedia_lookup	ITER targets Q=10 momentarily but first plasma is 2033–2034, making any ignition milestone from ITER before end of 2029 essentially impossible.	Yes
15	MODERATE	74	NEUTRAL	article_search	As of June 2025, CFS SPARC is in construction with magnet assembly ongoing; the company is one of 43+ private fusion ventures racing toward commercialization, but no first plasma date has been confirmed.	Yes
16	MODERATE	72	↑ UP	web_search	China's EAST plasma density breakthrough (Jan 2026) removes a key barrier to achieving ignition conditions in tokamaks, potentially accelerating progress toward Q≥1 milestones globally.	Yes
17	STRONG	78	↓ DOWN	article_search	No fusion program has publicly confirmed achievement of ignition or Q≥1 on a sustained basis as of early 2026; industry commentary describes fusion as still years away from commercial viability.	Yes
18	MODERATE	60	NEUTRAL	kalshi_data	The Kalshi fusion market has traded in a wide range (9%–51%) over 140 days, suggesting substantial historical uncertainty and volatility in market estimates of this outcome.	Yes

Information Gaps

No confirmed CFS/SPARC first plasma date or updated construction timeline as of mid-2026 — critical for assessing whether Q≥1 by 2029 is feasible
Helion Energy's current technical status and timeline update (Microsoft contract target was 2028 — no 2025/2026 progress report found)
NIF's roadmap for improving laser wall-plug efficiency or plans for a next-generation facility that could achieve wall-plug Q>1
Updated status of Commonwealth Fusion's SPARC magnet testing and projected operational date
No data on TAE Technologies' technical milestones post-merger with Trump Media
No information on Chinese fusion programs (CFETR) timelines and their potential to demonstrate Q≥1 before 2030
No clarity on how this specific Kalshi market defines 'achieved' — the resolution criteria ambiguity is a major gap
No expert physicist consensus estimate (e.g., from Nature or Science editorials) on probability of pre-2030 net energy gain

Key Uncertainties

Resolution criteria ambiguity: does the market resolve on scientific Q>1 (laser energy only, as NIF already achieved), or wall-plug Q>1 (which no one has achieved and is far harder)?
CFS/SPARC timeline: construction is ongoing but no first plasma date confirmed; any slip past mid-2027 makes net energy demonstration by 2029 extremely difficult
Helion's 2028 commercial target is highly aggressive and has not been independently verified as on-track
Whether NIF's increasing yields (up to 8.6 MJ in 2025) represent a path to new laser technology capable of dramatically better wall-plug efficiency
Whether China's EAST density breakthrough translates into meaningful acceleration of tokamak ignition timelines globally
Historical pattern of fusion milestones slipping 5-15 years — strong base-rate evidence against pre-2030 achievement
Political/regulatory environment: Trump-TAE merger raises both funding and conflict-of-interest uncertainties that could help or hinder
Declining Kalshi market price (27.6%, -8.4% in 30 days) may reflect new negative information not captured in articles found

Market Context

Price: 0.276 · 7d: -0.032 · 30d: -0.084
Volume: LOW — 202 contracts/day average is thin · Depth: UNKNOWN — not provided in data
Reliability: LOW-MEDIUM — low volume and wide historical price range (9%-51%) suggest limited liquidity and high uncertainty; declining price trend over 30 days may reflect genuine updating but could also reflect thin market dynamics

4. Independent Forecaster Opus · No market price · legacy v2 Did NOT see market price

Independent estimate: 28% (35% confidence)

base rate

22%

40% conf

evidence driven

28%

35% conf

contrarian

34%

30% conf

Spread

12pp

Moderate

sq1: Has nuclear fusion already been 'achieved' as of May 2026?

35%

Base rate: ambiguous. Under a loose scientific Q>1 interpretation, NIF achieved this in Dec 2022 and repeatedly since (8.6 MJ in 2025). Under strict wall-plug or commercial interpretation, no. The market's framing 'When will nuclear fusion be achieved' suggests a future event, implying the resolver does NOT consider NIF's achievement sufficient — otherwise the market would have resolved YES already. The fact that this market is open and trading at non-trivial prices strongly suggests strict interpretation. I assign ~35% to reflect residual probability a resolver could retroactively count NIF or another recent milestone.

sq2: Will sustained net energy gain (Q>1 wall-plug or scientific) be demonstrated before Jan 1, 2030?

22%

Base rate: fusion milestones historically slip 5-15 years. ITER first plasma now 2033+. CFS/SPARC is the most credible candidate — first plasma targeted ~2027, Q>1 ~2028. Helion targets 2028 commercially (very aggressive). TAE explicitly targets 2031. NIF cannot achieve wall-plug Q>1 with current lasers. If interpretation is scientific Q>1, NIF already counts — boosts to ~50%+. If interpretation is sustained/wall-plug, ~10-15%. Weighting these interpretations: ~30% loose × 0.85 + ~70% strict × 0.12 ≈ 0.34. But base-rate slippage and SPARC schedule risk drag this down. Estimate 22%.

sq3: Will a leading program reach a meaningful fusion milestone (ignition or Q≥1) by end of 2029?

45%

This is a softer bar — 'meaningful milestone' including ignition. NIF has already done this and continues improving (8.6 MJ in 2025). SPARC achieving first plasma by 2027-2028 is plausible (~50%). China's EAST progress and 43+ private ventures racing. Under loose interpretation, this is already near 100% (NIF). Under strict 'Q≥1 by a new program' interpretation, maybe 30-40%. Blended: ~45%.

Show full reasoning trace

base rate: Historical fusion timeline base rate: major milestones slip 5-15 years from original estimates. No program has achieved sustained wall-plug Q>1 in 70+ years of fusion research. NIF achieved scientific ignition Dec 2022 but this is widely seen as a step, not 'fusion achieved.' Private fusion timelines (Helion 2028, TAE 2031, CFS 2027-28) historically miss by years.

evidence updates: Strong DOWN: ITER pushed to 2033+ (irrelevant). NIF cannot reach wall-plug Q>1 with current lasers. TAE explicitly targets 2031. The market remaining open at moderate prices strongly suggests the resolver does NOT count NIF's existing ignition shots — this is the most important signal. Moderate UP: SPARC magnet assembly progressing, EAST density breakthrough, well-funded private sector with 43+ ventures. Net: stay well below 50%.

combination method: Weighted average per spec, then adjusted for positive correlation through resolution criteria. The resolution ambiguity is the dominant variable — if strict (likely given market is open and uses future tense), probability ~12-18%; if loose, ~60-80%. Probability-weighting strict at ~75% gives blended ~22-25%.

final: Final estimate 22%. The market being actively traded and using future-tense language strongly implies a strict resolution standard (sustained net energy or commercial demonstration), under which pre-2030 achievement requires SPARC or Helion to hit aggressive timelines that historically slip. Some residual probability from loose-interpretation scenarios and breakthrough optionality from multiple well-funded private programs.

ensemble: {'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.22, 'confidence': 0.4}, 'evidence_driven': {'p_yes': 0.28, 'confidence': 0.35}, 'contrarian': {'p_yes': 0.34, 'confidence': 0.3}}, 'spread': 0.12, 'individual_p_yes': {'base_rate': 0.22, 'evidence_driven': 0.28, 'contrarian': 0.34}}

5. Devil's Advocate Sonnet · legacy v2 Market price revealed

Reasoning Flaw Missing Info

Challenges

The resolution ambiguity is correctly identified as the dominant variable, but the forecaster's framing that 'the market is open at moderate prices, therefore the resolver uses strict interpretation' is circular reasoning. The market could be open precisely BECAUSE the resolution is ambiguous, not because traders have inside info on a strict interpretation. Don't double-count this signal by both using it to infer strict interpretation AND then anchoring to the market price.
The weighting (~75% strict, ~25% loose) is asserted without clear justification. Given that NIF has demonstrably achieved scientific Q>1 multiple times with 8.6 MJ outputs, a Kalshi-style resolver might well count this — the loose interpretation deserves more than 25% weight, perhaps 35-45%, especially given the question's plain-language reading of 'achieved'.
Sub-question combination is opaque. sq1=0.35, sq2=0.22, sq3=0.45 — but the final 0.28 doesn't clearly follow from any stated combination rule. If sq1 OR sq2 satisfies resolution, the disjunction should be higher than either alone (e.g., 1 - (1-0.35)(1-0.22) ≈ 0.49 if independent). The drop to 0.28 needs explicit justification.
Potential 'never happened' fallacy lite: citing '70+ years no sustained wall-plug Q>1' as base rate, but current conditions ARE structurally different — 43+ private ventures, $10B+ private capital, HTS magnets enabling new tokamak designs. The base rate weighting may be too pessimistic given regime change.
Information gap on Helion (Microsoft 2028 commitment) is significant — Helion's pulsed approach could plausibly hit a 'meaningful milestone' by 2029 and is not adequately weighted in the analysis.

Suggested adjustment: +3pp

Small upward adjustment (~+3pp) is warranted because (a) the disjunctive combination of sq1/sq2/sq3 appears to undershoot what the sub-forecasts imply, and (b) the strict-interpretation weighting is partly justified by circular market-based reasoning. This is NOT about market conformity — the forecast already matches the market — but about internal consistency of the sub-forecast combination.

Market comparison: Forecaster: 28.0%, Market: 27.6%. Near-perfect alignment. Divergence is negligible and the forecast appears well-calibrated to the market consensus. The main risk is not directional bias but that both the forecaster and market may be under-weighting the loose-interpretation scenario where NIF already counts.

6. Aggregator Opus · final calibrated forecast

30% (35% confidence)