Scan: When will nuclear fusion be achieved?

← Back to scans

When will nuclear fusion be achieved? — Before 2030

KXFUSION-30-JAN01 · Science and Technology · 2026-04-06

42%

Agent

40%

Market Price

+2.0%

Edge

45%

Confidence

Final Rationale

My independent estimate of 78% was driven almost entirely by interpreting NIF's repeated Q>1 target-gain results (up to Q=2.44) as 'achieving nuclear fusion.' However, the Devil's Advocate raises compelling points that I largely ACCEPT: (1) The question's future-tense framing ('When will fusion be achieved?') with multiple date buckets extending to 2050+ strongly implies a milestone not yet reached; (2) The market has been at ~40% for years despite NIF's results being public since December 2022 — if NIF clearly resolved this, the market would be near 95%; (3) NIF's massive wall-plug energy deficit (400 MJ in, ~3.15 MJ out at target) means this is far from practical fusion energy, which is likely what the question intends. I ACCEPT challenges 1-3 and lower P(NIF counts) from 78% to roughly 35%. Combined with a small probability (~8%) that someone achieves a higher bar before 2030, I arrive at ~42%, converging with the market. The market's persistent pricing around 40% despite full knowledge of NIF's results is the strongest signal about how this question's resolution criteria are interpreted.

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 $t{=}0$.

$n = 52$ follow-ups

1. Planner Opus · Fermi decomposition + tool plan

Tools chosen: kalshi_data kalshi_orderbook article_search web_search wikipedia_lookup

Sub-questions (Fermi decomposition)

Has nuclear fusion already been 'achieved' (net energy gain / ignition) as of April 2026, based on NIF or other experiments? (w=0.25)
Will a sustained, reproducible net-energy-gain fusion event (Q>1 scientific breakeven) be demonstrated by a major fusion program (NIF, ITER, private companies) before January 1, 2030? (w=0.45)
Will a private fusion company (e.g., Commonwealth Fusion, TAE Technologies, Helion, etc.) achieve a major technical milestone sufficient to count as fusion 'achieved' before 2030? (w=0.30)

Planner reasoning

The key definitional issue is what 'nuclear fusion achieved' means for resolution purposes. The NIF December 2022 ignition event (Q~1.5 on laser input, but Q<<1 on wall-plug energy) already cleared scientific breakeven — so the question may hinge on whether that counts. If the resolution requires something more (sustained power output, wall-plug Q>1, or grid-scale power), the bar is much higher. The reference class for transformative physics milestones reaching deployment in under 4 years is extremely poor. However, if the definition is simply 'scientific ignition/breakeven demonstrated in a lab setting,' NIF has arguably already done it. Private companies like Helion have ambitious 2028 timelines but face enormous technical challenges. ITER won't achieve full fusion power until mid-2030s at earliest. I need to check: (1) current market pricing on Kalshi as a signal of collective probability estimate, (2) latest NIF and private company progress to assess trajectory, (3) the exact resolution criteria implied by this market.

2. Research Parallel tool calls · legacy v2

Tool	Status	Time	Items
kalshi_data	OK	1.7s	-
kalshi_orderbook	OK	0.2s	-
article_search	OK	0.9s	-
web_search	OK	130.0s	-
wikipedia_lookup	OK	1.7s	-

3. Evidence Synthesis Sonnet · 19 items · legacy v2

#	Strength	Credibility	Direction	Source	Claim	Priced In
1	STRONG	-	↑ UP	wikipedia_lookup	NIF achieved the first scientific breakeven (Q>1) on December 5, 2022 with an energy gain factor of 1.5; as of 2025, NIF remains the only lab to have demonstrated Q>1 at the target level.	Yes
2	STRONG	-	↑ UP	web_search	NIF achieved ignition for the seventh time on February 23, 2025, setting a new target gain record of 2.44, with 2.05 MJ input yielding 5.0 MJ output — the second highest fusion yield ever recorded.	Yes
3	WEAK	-	↑ UP	web_search	Unconfirmed reports suggest NIF may have subsequently achieved 5.2 MJ and 8.6 MJ yields, but no official communique has been issued confirming these figures.	No
4	STRONG	-	↓ DOWN	web_search	NIF's wall-plug efficiency remains deeply negative: the facility uses ~400 MJ of electricity to deliver ~2 MJ to the target, meaning engineering/wall-plug Q remains far below 1 despite target-gain Q>1.	Yes
5	WEAK	-	NEUTRAL	web_search	A viral claim that a US fusion plant came online on February 27, 2026 powering 400,000 homes is not verified by any credible source and should be treated with extreme skepticism.	No
6	STRONG	-	↓ DOWN	wikipedia_lookup	Efficiencies orders of magnitude higher than NIF's current target-gain Q are required to reach engineering breakeven (net electricity-producing plant), indicating that NIF's achievements do not constitute 'fusion achieved' at a practical energy level.	Yes
7	STRONG	-	↓ DOWN	wikipedia_lookup	ITER expects first plasma in 2033–2034, well past the 2030 deadline, and is not designed to generate electricity; it will not contribute to a pre-2030 resolution.	Yes
8	MODERATE	-	NEUTRAL	web_search	SPARC (Commonwealth Fusion Systems) is described as 'approaching' plasma Q>1, but this is a pre-commercial research device still under development, not yet demonstrated.	Yes
9	STRONG	-	↑ UP	web_search	NIF has repeatedly and reproducibly demonstrated target-gain Q>1 (scientific ignition) in multiple shots since December 2022, including a 7th ignition in February 2025 — demonstrating reproducibility at the scientific level.	Yes
10	MODERATE	-	↑ UP	web_search	The 2026 global fusion landscape is described as having 'crossed a threshold from pure scientific exploration toward large-scale demonstration and early industrialization,' suggesting momentum but not imminent Q>1 wall-plug breakeven.	No
11	MODERATE	-	NEUTRAL	web_search	Germany's Wendelstein 7-X stellarator raised its energy turnover to 1.8 gigajoules in May 2025, representing a plasma confinement milestone but not net energy gain.	Yes
12	MODERATE	-	NEUTRAL	article_search	China's EAST reactor achieved stable plasma operation at previously impractical densities and conditions, representing a confinement milestone but not scientific breakeven Q>1.	Yes
13	STRONG	-	↓ DOWN	article_search	Commonwealth Fusion Systems (CFS) signed a deal with Google to deliver 200 MW from its first commercial ARC plant, but that plant is not expected online until 'early 2030s' — beyond the 2030 deadline.	Yes
14	STRONG	-	NEUTRAL	article_search	Trump Media & Technology Group announced a $6 billion merger with TAE Technologies in December 2025, giving TAE access to capital, but TAE has not achieved fusion ignition or Q>1 in any published result.	Yes
15	MODERATE	-	NEUTRAL	article_search	Proxima Fusion raised €130 million ($148 million) in a record European private fusion funding round in June 2025, but is still in early R&D without demonstrated net energy gain.	Yes
16	WEAK	-	NEUTRAL	article_search	Type One Energy (Tennessee) claims fusion electricity on the grid within 'the next decade' using a stellarator design, but no ignition or Q>1 milestone has been announced.	Yes
17	MODERATE	-	↓ DOWN	article_search	Multiple private fusion companies (Helion, CFS, TAE, Proxima, Type One) have attracted billions in investment and have ambitious timelines, but none have publicly announced achieving Q>1 scientific breakeven as of April 2026.	Yes
18	MODERATE	-	↑ UP	kalshi_data	Kalshi market 'Before 2030' is priced at 40%, up 5% in the last 7 days, with a historical range of 9%–51% and 110 data points, suggesting moderate collective probability.	Yes
19	MODERATE	-	NEUTRAL	kalshi_orderbook	Kalshi orderbook shows near-zero bid on YES side ($0.00) with a $0.989 spread, suggesting thin YES-side liquidity despite high reported depth (38,159), making the 40% price less reliable.	Yes

Information Gaps

No clear resolution criteria specified in the market — the key definitional ambiguity of whether NIF's target-gain Q>1 (already achieved) vs. wall-plug Q>1 vs. commercial power generation constitutes 'fusion achieved' is unresolved.
No data on Helion's specific technical progress toward its 2028 commercial demonstration target (which has a $1B+ Microsoft power purchase agreement contingency).
No data on CFS SPARC device construction timeline and whether it could demonstrate plasma Q>1 before 2030.
No official confirmation or denial of the unconfirmed NIF 8.6 MJ yield reports.
No information on whether any private company has achieved any form of net plasma energy gain (even target-level Q>1) in experiments.
The Kalshi orderbook data showing near-zero YES bids despite a 40% price is anomalous and may indicate data quality issues with the orderbook snapshot.
No expert consensus on what definition of 'fusion achieved' this specific market uses for resolution purposes.
No data on China's EAST or other government programs' plans to demonstrate Q>1 before 2030.

Key Uncertainties

Resolution criteria ambiguity: If NIF's repeated target-gain Q>1 demonstrations already count, the question may already resolve YES. If wall-plug energy efficiency (Q>>1 at system level) is required, YES before 2030 is nearly impossible.
Whether Helion achieves its 2028 technical demonstration milestone (net energy output), which is the most plausible private-sector path to YES before 2030.
Whether CFS SPARC, currently under construction, achieves plasma Q>1 before 2030, and whether that would satisfy the market's resolution criteria.
NIF funding and operational continuity under the current US administration — federal budget pressures could slow or accelerate NIF's shot cadence and progress.
Whether 'sustained and reproducible' is required for resolution: NIF has achieved reproducibility at the scientific level (7 ignitions) but not sustained burn.
The Kalshi price spike of +5% in 7 days may reflect new information (NIF record, private company announcement) or speculative trading noise.

Market Context

Price: 0.4 · 7d: +0.05 · 30d: +0.00
Volume: LOW-MODERATE — 228 avg daily contracts is thin for a consequential science question · Depth: ANOMALOUS — reported depth of 38,159 but near-zero YES bids ($0.00) with $0.989 spread suggests possible data artifact or illiquid YES side
Reliability: LOW-MEDIUM — price may reflect definitional ambiguity more than genuine probability; the 40% level and 9%-51% historical range suggest wide uncertainty; orderbook data appears unreliable

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

Independent estimate: 78% (45% confidence)

sq1: Has nuclear fusion already been 'achieved' (net energy gain / ignition) as of April 2026, based on NIF or other experiments?

78%

NIF achieved scientific breakeven (Q>1 at the target level) on December 5, 2022, and has reproduced this result at least 7 times through February 2025, with the latest achieving Q=2.44. This is unambiguously 'nuclear fusion with net energy gain' at the scientific/target level. The key uncertainty is whether the market's resolution criteria require something beyond this — e.g., wall-plug Q>1 or sustained power generation. The question title says 'nuclear fusion achieved' without specifying engineering breakeven or commercial power. In common scientific parlance and media coverage, NIF's achievement is widely described as 'achieving fusion.' The description is sparse ('nuclear fusion is achieved') which most naturally maps to the scientific milestone that has already occurred. However, there's a reasonable interpretation that 'achieved' implies something more practical/useful than a laboratory demonstration with massive wall-plug energy deficit. I estimate ~78% probability that NIF's demonstrated results already satisfy the resolution criteria, with the ~22% downside risk being that the resolver interprets 'achieved' as requiring engineering breakeven or practical power generation.

sq2: Will a sustained, reproducible net-energy-gain fusion event (Q>1 scientific breakeven) be demonstrated by a major fusion program (NIF, ITER, private companies) before January 1, 2030?

92%

This has already been achieved by NIF multiple times. NIF demonstrated Q>1 in December 2022 and has done so at least 7 times through February 2025 with improving yields (up to Q=2.44). This is already done. The only way this sub-question resolves NO is if we require something beyond what NIF has demonstrated. ITER won't contribute before 2030. SPARC (CFS) is plausible but uncertain for pre-2030 plasma Q>1. NIF will likely continue improving. The probability is very high (92%) because the milestone has essentially already been met, with small residual uncertainty about whether future experiments might somehow cast doubt on previous results or whether the definition requires something NIF hasn't done.

sq3: Will a private fusion company (e.g., Commonwealth Fusion, TAE Technologies, Helion, etc.) achieve a major technical milestone sufficient to count as fusion 'achieved' before 2030?

18%

Base rate: No private fusion company has ever demonstrated Q>1 scientific breakeven. As of April 2026, none have publicly announced achieving this. CFS's SPARC is under construction but the ARC commercial plant isn't expected until early 2030s. Helion has a 2028 target for its demonstration system but faces enormous technical challenges and has not demonstrated net energy gain. TAE Technologies has capital from the Trump Media merger but no published Q>1 results. The timeline of ~3.75 years remaining is very tight for any private company to go from current status to demonstrated Q>1. Helion is the most plausible candidate given its 2028 timeline and Microsoft PPA, but fusion timelines historically slip dramatically. CFS SPARC achieving plasma Q>1 before 2030 is possible but uncertain. I estimate ~18% probability that a private company achieves something that could count as 'fusion achieved' before 2030, acknowledging that even if they achieve a milestone, it might not meet the market's resolution bar.

Show full reasoning trace

base rate: Nuclear fusion (scientific ignition/breakeven) has been demonstrated exactly once — by NIF starting December 2022, now reproduced 7+ times. No private company or other government lab has achieved Q>1. The base rate for 'has fusion been achieved' depends entirely on the definition. At the scientific target-gain level, it's already happened. At the engineering/commercial level, it has never happened and likely won't before 2030.

evidence updates: NIF's 7 successful ignition shots (Q up to 2.44) push strongly toward YES if the definition is scientific breakeven (+30-40% from a neutral prior). The massive wall-plug energy deficit (400 MJ in, 2 MJ to target) pushes strongly toward NO if engineering breakeven is required (-30-40%). No private company has achieved Q>1, pushing down the probability of an alternative path to YES. ITER's 2033+ timeline eliminates it as a contributor. The ambiguous resolution criteria create the dominant uncertainty.

combination method: Conditional decomposition: P(YES) = P(NIF already counts as 'achieved') + P(NIF doesn't count) * P(someone achieves higher bar before 2030). The first term dominates at ~0.78. The second term contributes ~0.03. Total ≈ 0.78.

final: The probability hinges almost entirely on whether NIF's demonstrated target-gain Q>1 (achieved multiple times since 2022) satisfies the resolution criteria of 'nuclear fusion achieved.' In common usage, media coverage, and scientific discourse, NIF's achievement is widely described as 'achieving fusion' or 'achieving ignition.' The market description is vague enough that this likely qualifies. However, a non-trivial probability exists that the resolver requires something more practical (engineering breakeven, sustained power, grid delivery), in which case the answer is almost certainly NO before 2030. I estimate 78% YES, with moderate confidence due to the definitional ambiguity being the primary driver of uncertainty.

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

Reasoning Flaw Overconfidence Asymmetric Evidence

Challenges

The forecast's 78% probability hinges almost entirely on the assumption that NIF's target-gain Q>1 satisfies the resolution criteria. But the forecaster acknowledges this is the dominant uncertainty and then assigns only ~22% to the alternative interpretation. This seems insufficiently uncertain given that: (1) there is NO clear resolution criteria specified, (2) the question asks 'When will nuclear fusion be achieved?' — the future tense and framing as a timeline question (with buckets like 'Before 2030', 'Before 2035', etc.) strongly implies something that hasn't happened yet or is forward-looking, (3) many prediction markets on fusion explicitly or implicitly mean something beyond a single lab demonstration with massive wall-plug energy deficit, and (4) the common-sense interpretation of 'achieving fusion' for a general audience often implies practical energy generation, not a physics experiment requiring 400 MJ input for 2 MJ on target.
The forecaster gives 78% to 'NIF already counts' but the market is at 40% — a massive divergence. While the market can be wrong, the forecaster should consider that market participants have had since December 2022 (over 3 years) to price in NIF's achievement. If NIF's results clearly resolved the question, the market would likely be near 95%+. The persistent ~40% price over years suggests that either (a) the resolution criteria require more than NIF demonstrated, or (b) there is genuine uncertainty about what counts. Either way, the forecaster's 78% confidence that NIF already satisfies the criteria seems too high given the long period the market has had to incorporate this information.
The forecaster's reasoning about 'common scientific parlance' and 'media coverage' describing NIF as 'achieving fusion' is somewhat misleading. While media did use such language, subsequent coverage also emphasized the massive gap between NIF's result and practical fusion energy. The question's framing as a timeline question ('When will nuclear fusion be achieved?') with multiple future date buckets implies the questioner is asking about a milestone that represents meaningful progress toward fusion energy, not just a physics proof of concept that occurred in 2022.
The conditional decomposition P(YES) = P(NIF counts) + P(NIF doesn't count) × P(someone else achieves higher bar before 2030) is mathematically sound, but the second term is estimated at only ~0.03. If we lower P(NIF counts) significantly, the second term becomes more important but remains very small, which seems correct. However, the forecaster should have explored more carefully what probability to assign to NIF counts — the 78% figure is doing all the heavy lifting and deserves more rigorous analysis.
The forecaster notes 'moderate confidence (0.45)' which suggests they themselves are quite uncertain, yet the point estimate of 78% is quite definitive. This internal tension suggests the forecaster should have a wider uncertainty band, which practically means the point estimate should be moderated toward something less extreme.

Suggested adjustment: -25pp

The 78% estimate relies almost entirely on the assumption that NIF's target-gain Q>1 already satisfies the resolution criteria. However, the future-tense framing of the question ('When will fusion be achieved?'), the existence of multiple date buckets extending to 2050+, the persistent market price around 40% despite 3+ years of NIF results being public knowledge, and the massive wall-plug energy deficit all suggest the resolution criteria likely require something beyond what NIF has demonstrated. The forecaster's own low confidence (0.45) is inconsistent with a 78% point estimate. A more appropriate estimate of P(NIF counts) would be ~45-55%, yielding a combined estimate of roughly 50-55%.

Market comparison: Forecaster: 78%, Market: 40%. The 38-point divergence is very large. While the market can be wrong, the key issue is that NIF's results have been public for over 3 years and the market has had ample time to price them in. The persistent ~40% price (historical range 9-51%) strongly suggests that market participants do NOT interpret NIF's results as resolving the question, or at minimum that there is deep ambiguity. The forecaster's divergence is not based on new information but on an interpretation of existing, well-known results — making the market signal more informative than usual. The forecaster should weight down P(NIF counts) significantly.

6. Aggregator Opus · final calibrated forecast

42% (45% confidence)