Nuclear Fusion Adoption

Before scanning, we get the obvious out into the room. KANCILS is a seven-prompt warm-up adapted from Eddie Choo's open-source futures practice (see Double Hexagon guide). It surfaces what the group already half-believes so the rest of the work has something to push against.

For fusion adoption in 2045:

• K — Keep. Public investment in fusion science (NIF, ITER, JT-60SA). Independent regulatory oversight. International collaboration on tritium fuel cycles.
• A — Away with. The framing that fusion is "30 years away" or, conversely, "5 years away." Both close off thinking. The framing that fusion adoption is purely a physics problem.
• N — Never. A future where fusion is captured by a single firm or state and used as energy leverage. A future where fusion is built but its waste / tritium / activated structural materials are externalised the way fission's were.
• C — Challenging. Skilled-labour bottleneck for plant construction. Permitting fights with communities. Cost curves vs. ever-cheapening solar+storage. Tritium supply.
• I — Important. Whether fusion is a primary energy answer or a back-of-the-mix contributor changes the size of the bet societies should be making now.
• L — Learn. That mature renewables may be a harder competitor than fossil incumbents ever were. That the politics of energy infrastructure don't get easier with cleaner energy.
• S — Strengthen. Local grid operators' relationships with new generation types. Civil-society capacity to read fusion-specific safety arguments.

The discipline here is that none of the KANCILS answers are interesting on their own. The point is that everyone in the conversation can now see the same set of unspoken assumptions. We're starting from the same floor.

Try it yourself

Walk through the KANCILS prompts for the focal concern
"commercial fusion power adoption in 2045". Give 3–5 candidate answers
under each:
- Keep — what should be kept going
- Away — what should be done away with
- Never — futures we never want to reach
- Challenging — issues likely to be faced
- Important — what's personally at stake
- Learn — what we might learn
- Strengthen — present habits or relationships worth strengthening
Keep answers concrete, not aspirational. Avoid policy-paper register.

---

We now pull in signals from outside the usual feed. STEEP+++ is a sorting taxonomy (Social, Technological, Economic, Environmental, Political, plus anything else) — not a research method. We sort what we already know and notice gaps.

Technological

• December 5, 2022: the National Ignition Facility at LLNL achieved target gain — 3.15 MJ of fusion energy out from 2.05 MJ of laser energy delivered to the target. (1)
• Four further ignition shots followed through to February 2024; one reached an estimated 5.2 MJ from 2.2 MJ input. (2)
• Commonwealth Fusion Systems is assembling SPARC's high-temperature superconducting (REBCO) magnet array, targeting first plasma in 2027 and Q>10 net energy demonstration shortly after. (3)
• Helion's Polaris device demonstrated measurable deuterium-tritium fusion at 150M °C in February 2026 — the first private machine to do so. (4)
• ITER's first plasma has slipped to ~2035 with full-power D-T operation not expected until 2039. (5)

Economic

• Helion signed a 50 MW power purchase agreement with Microsoft in May 2023 — the first electricity-from-fusion PPA — for delivery starting 2028 from a plant near Malaga, Washington. (6)
• A subsequent Nucor offtake of 500 MW deepens commercial commitment. (7)
• The Fusion Industry Association reports private sector investment passing US$7 billion cumulative by 2024, dominated by ~five firms.
• Levelised cost projections for fusion remain unverified at commercial scale; comparison points like utility-scale solar continue falling (~US$30/MWh in 2024).

Political / Regulatory

• April 14, 2023: the US Nuclear Regulatory Commission voted unanimously to regulate fusion machines under 10 CFR Part 30 (byproduct materials), separating fusion's framework from fission's. (8)
• The 2024 ADVANCE Act / Fusion Energy Act codified this in US law, defining fusion machines as particle accelerators for regulatory purposes. (9)
• The UK (10) and Japan have moved toward similar separation; the EU is still consolidating.
• October 2025: the US DOE published a Fusion Science & Technology Roadmap aiming at a pilot plant in the 2030s. (11)

Environmental

• Tritium supply remains thin; only ~25 kg of usable tritium currently exists globally, much in CANDU reactor inventory.
• Activated structural materials are not high-level waste but require dedicated lifecycle planning.
• Fusion is firm-dispatchable like fission, which means its strongest argument is complement to intermittent renewables, not replacement.

Social

• Public polling on fusion is high but shallow — most respondents support it, few can articulate concerns.
• Local siting opposition to industrial infrastructure (data centres, transmission, hydrogen plants) is intensifying globally and would plausibly apply to fusion plants.
• Workforce: existing nuclear-trained operators are an aging cohort; fusion-specific apprenticeship pipelines are nascent.

Values / "+++"

• The "is this real?" question persists in mainstream press despite shipping science.
• An emerging argument frames fusion as the only technology that scales to civilisational energy needs without ecological compromise — this is a value-claim worth surfacing because it tends to delegitimise comparison.

Gap check. Our scan is heavy on Technological/Economic signals from US sources, lighter on Social/Values, and thin on non-Anglophone politics. That imbalance is itself a finding — and a reason to be modest about the eventual scenarios.

Try it yourself

For the focal concern "commercial fusion power adoption by 2045", list
5 signals or trends in each STEEP+++ category — Social, Technological,
Economic, Environmental, Political, plus one extra dimension that
matters here (Values, Infrastructural, Generational, etc).
For each signal, cite a source (or mark "unsourced — assumption").
Then state which buckets are over-represented in your scan and which
are likely under-represented because of where you read.

---

We now cluster signals and choose two driving uncertainties for the 2×2 in Step 4. A clean 2×2 requires its two axes to be:

1. Genuinely uncertain. Not "more X" vs. "less X" of a known direction — both ends must be plausible.
2. Independent. If one axis determines the other, the matrix collapses.
3. High-impact. Each axis must change the world meaningfully regardless of where the other lands.

Clustering the signals, three candidate axes emerge:

• Axis A — Engineering readiness: Does fusion deliver a reliable engineering-grade machine at commercial scale by ~2040? (Net energy is necessary but insufficient; tritium-self-sufficiency, plant availability, repair cycles, and unit cost all need to land.) Both directions are plausible: the gap between net-energy science and a reliable 24/7 plant is historically huge in energy tech.

• Axis B — Social/grid uptake: Does the surrounding system absorb fusion when it arrives? This covers regulatory clarity, transmission permitting, workforce, public legitimacy, and whether dispatchable firm power is still in demand in a grid that may be dominated by renewables+storage by 2040. Both directions are plausible.

• Axis C — Competitive landscape: Does fusion arrive into a world where solar+storage is so cheap that grid-scale fusion becomes a niche? This one is high-impact but partially dependent on Axis B (uptake depends on competitiveness), so we'd lose independence.

We pick A × B. They're independent (a perfect machine can still be blocked by permitting; uptake-ready systems can still wait on the physics) and they bracket different communities of practice (physicists/engineers vs. utility regulators/planners), which makes them productive to think with.

Why not just pick "fast/slow timeline" as one axis? Because timeline is the output we're trying to think about. Putting it on an axis would smuggle the answer into the question.

Try it yourself

Given these signals for "commercial fusion power 2045":
<paste your scan>

Propose 3 candidate axis pairs for a 2×2 Scenario Matrix.
For each pair, test:
- Are both ends of each axis genuinely plausible? (uncertainty test)
- Does one axis determine the other? (independence test)
- Does each axis change the world meaningfully? (impact test)
Pick the strongest pair and justify in 2–3 sentences.

---

We set up the matrix. The axes:

• Vertical axis: Fusion engineering readiness by ~2040 — Top = robust, multi-vendor, repeatable plants. Bottom = first-of-a-kind plants only, persistent reliability and tritium-supply issues.
• Horizontal axis: Social/grid uptake — Right = clear regulation, permitting, workforce, and dispatchable-firm-power demand. Left = regulatory friction, siting fights, and a grid that has already settled on renewables+storage+demand response.

This produces four quadrants. We give each a short name and a one-paragraph scenario.

Quadrant Q1 — Top-right · "Quiet Decade"

Fusion technology lands. SPARC-class machines achieve commercial Q>20 by the early 2030s; tritium breeding is solved. The regulatory split US-EU-Japan-UK aligns. Workforce pipelines (apprenticeships, university programs) ramp. By 2045, fusion is ~12% of OECD baseload, mostly co-sited with industrial heat users — data centres, hydrogen electrolysers, desalination. It does not displace solar; it complements it. The public narrative is boring. This is the scenario most fusion advocates secretly hope for and rarely state publicly because it requires admitting fusion is one tool, not the tool.

Quadrant Q2 — Top-left · "Built Without Buyers"

Engineering lands. Plants are built. But by the time they arrive, solar+storage+demand-response has eaten the dispatchable-baseload market in places that have grid flexibility (China, Northwest Europe, California). Regulatory clarity exists on paper but each new plant faces 5–8 year permitting fights from communities that have stopped trusting any "nuclear" word. Existing plants in the US and Korea run at ~60% capacity factor. Fusion finds its niche in industrial heat and desalination, not the wires. The story becomes "fusion works but doesn't matter" — much like advanced nuclear in 2025.

Quadrant Q3 — Bottom-right · "Permanent Demonstration"

The world is ready. Regulators have cleared the path, grids want firm power for AI and heavy industry, capital is lined up. But the physics turns out to be harder than the late-2020s milestones suggested — tritium breeding doesn't scale, magnet replacement cycles are too short, first-of-a-kind plants run at 40% availability. Each new "commercial" plant remains a demonstration. By 2045, fusion provides ~1% of global electricity; the field reorganises around materials science and rebuilds for a 2055 push. This is the "always 20 years away" quadrant — but with the specific texture of a frustrated, ready ecosystem rather than indifference.

Quadrant Q4 — Bottom-left · "Lost Decade"

Neither side delivers. Engineering struggles, public legitimacy frays, the energy transition runs ahead on other rails. Fusion shrinks back to large-state research budgets (ITER continues, NIF continues, a few private firms persist as deep-tech). The cultural story becomes one of disappointment — but quietly, because solar+storage+geothermal+small-modular-fission has filled the gap. Energy futures think tanks stop convening fusion sessions by 2042.

Wind-tunnelling: what survives across quadrants?

We stress-test plausible 2026 actions against the four worlds:

• Invest heavily in tritium fuel-cycle research. Holds up in Q1 and Q3 (it's a bottleneck); harmless in Q2 and Q4. Robust.
• Lobby for a special "fusion-fast-track" permitting regime. Wins big in Q1, useful in Q3 — but politically expensive in Q2 (where it would generate backlash) and wasted in Q4. Conditional.
• Pre-fund fusion workforce apprenticeships at scale. Pays off in Q1, gives the field continuity in Q3 even if plants stay demos. Robust if scoped.
• Build narratives that frame fusion as "the answer." Brittle in Q2 (where it sets up backlash when "the answer" is built but unused) and Q4 (where it discredits the field). Fragile.

The interesting move is not the winning strategy. It's the realisation that a brittle narrative may be more dangerous to the field than a slow timeline.

Try it yourself

Build a 2×2 Scenario Matrix for "commercial fusion power adoption by
2045" using these axes:
- Vertical: fusion engineering readiness by ~2040
- Horizontal: social/grid uptake

For each of the four quadrants:
1. Give the quadrant a short, evocative name (not "best case" / "worst
   case")
2. Write a 150-word scenario sketch — name the year, name a place,
   name a concrete artefact someone could touch
3. Note one strategy that wins in this quadrant and fails in another
Finish by wind-tunnelling 4 candidate present-day actions across all
four quadrants. Flag the most surprising failure mode.

---

The original brief proposes a backcasting variant: "fusion provides 20% of grid power by 2045 — what had to happen?"

This is a Q1-flavoured target. We work backward in five rough stages.

2045 — End state: Fusion delivers ~20% of OECD grid power. Roughly 80 GW commercial fusion online globally, ~60% in OECD, with rapid build-out in industrial-heat-paired co-generation.

2040 — Pre-conditions: ~30 GW operating. Three competing reactor designs at scale (a CFS-derivative, a Helion-derivative, one Asian state-led design). Tritium breeding at 1.05+ ratio routine. Regulatory regime: separate-from-fission, harmonised across US/UK/EU/Japan/Korea/UAE.

2035 — Inflection: First non-prototype commercial plant operating reliably at 80%+ capacity factor. Workforce pipeline outputting ~3,000 fusion-trained engineers/yr in OECD. Standardised tritium handling code adopted by IAEA. Public PPAs from anchor customers (data centres, electrolysers) demonstrate cash flows.

2030 — Crossroads: Multiple machines clearing Q>10 in repeated operation, not just shots. Manufacturing tooling for HTS magnets at GW-scale per year. First fusion-specific environmental impact statements completed and not derailed by litigation.

2026 — Today: What 2030 demands of us now. Tritium R&D funding doubled. Apprenticeship programs seeded in ~12 universities. Communities near plausible siting locations engaged in deliberative processes — not consultation theatre — to surface objections early. Regulator capacity built (people, not just rules).

Honest read: Each step's transition is physically possible. The total set of transitions is harder than the sum of its parts because they have to happen in a roughly coordinated way across jurisdictions. The backcast is useful not because it'll be right — it almost certainly won't — but because it makes visible which 2026 actions matter and which are theatre.

Try it yourself

Backcast from the target "fusion provides 20% of grid power by 2045".
Work in five stages: 2045, 2040, 2035, 2030, 2026.
For each stage, state:
- What is true (capacity, technology, regulation, workforce)
- What had to happen between this stage and the previous one
- What 2026 action this implies
End with an honest read: which transitions are physically possible
individually but hard to coordinate?

---

We now cross into Hexagon 2 — Design. The shift in posture is important: we stop arguing about what's likely and start asking what a specific quadrant feels like to inhabit. We deliberately pick Q2 ("Built Without Buyers") because it's the quadrant fusion advocates least want to think about, and where the most uncomfortable artifacts live.

Setting: Yorkshire, UK. October 2045. A 600 MW fusion plant, COMET-1, came online in 2041. It operates at 58% capacity factor — not because it can't run higher, but because the local grid increasingly clears with rooftop solar, regional storage, and demand-response. The plant's owners are restructuring its PPA.

Character: Priya, 41, plant operations engineer. Worked on SPARC's commissioning at MIT in 2027, joined the COMET project in 2034. Daughter, 9, is at the local primary school where one of the wings is named after the plant.

A day:

• 07:10. Walks the kid to school. Passes the plant fence; a hand-painted sign says "1 in 5 jobs here." A council-funded campaign — defensive.
• 08:30. Morning briefing. Yesterday's curtailment was higher than forecast; the grid operator dropped them to 32% output overnight because Scottish wind was overproducing. PPA penalty clause was triggered, again.
• 10:15. Calls with a hydrogen-electrolyser consortium about a co-located off-take agreement. The economics work only if the plant runs at 90%+; the consortium can't commit to that volume.
• 13:00. Quarterly review meeting. Headquarters wants to convert two of the plant's four reactors to industrial-heat mode for a desalination project — but desalination is in Spain, and the heat won't travel. The math is bad.
• 15:40. Walk-through with a school group. A 14-year-old asks, "If solar's cheaper, why did we build you?" Priya pauses. Says: "We didn't know solar would win the grid by the time we finished. We thought we were a backbone. We turned out to be a bridge." The teacher looks uncomfortable.
• 19:20. Home. Sees a notification: the plant's operating company has filed for "strategic restructuring" — language for an orderly wind-down by 2052. Her child asks if they have to move.

What this surfaces. Q2 is not a failure scenario in the technical sense — the machine works. It is a purpose failure. The discomfort the worldbuild generates is informational: it shows what kinds of policy and narrative the field is currently not preparing for. (A "strategic restructuring" of a working fusion plant is, in 2026, almost unsayable in the fusion advocacy register.)

Try it yourself

Pick the quadrant from your 2×2 that is hardest to think about.
Write a day-in-the-life for one character living inside it:
- Name the place, the date, the character
- 6–8 timestamps across a day
- At least one scene should be a small, plausible humiliation or
  ambiguity, not a dramatic crisis
- End with one notification or piece of news that lands at the end of
  the day and changes the read
Avoid stock dystopia furniture (drone deliveries, holograms). Use
mundane, specific 2045 textures.

---

We pivot from atmosphere to object. The Thing from the Future (Stuart Candy, Jeff Watson) is a generative tool — you fix a future-mood and a topic and force out a concrete artifact. We're not playing the full card game; we're using its discipline.

Prompt parameters (from Q2):

• Arc: Collapse-adjacent (a working thing being wound down)
• Terrain: Energy infrastructure
• Object: A document
• Mood: Resigned, with some pride

The forced juxtaposition produces a candidate artifact: a "Strategic Restructuring Notice" from a fusion plant operator to its workforce.

This is now a Design Fiction brief: produce the document as if you found it in 2045 Yorkshire. We write it in full so the texture lands. Where Design Fiction differs from sci-fi prop work is that the document is built to read as real — the bureaucratic register matters more than the speculative content.

---

COMET-1 Operating Company Ltd.

Notice of Strategic Restructuring — Issued under the Fusion Sector Transition Framework (UK FSTF 2043)

Reference: FSTF/COMET-1/2045-Q4-001 Date of issue: 14 October 2045 Affected personnel: All COMET-1 plant operations, maintenance, and tritium-handling staff (≈ 1,180 FTE)

1. Purpose of this notice

This notice initiates the formal Strategic Restructuring of COMET-1 in accordance with the Fusion Sector Transition Framework (FSTF) ratified by Parliament in March 2043 following the Sellafield-Lakes Inquiry. It does not in itself constitute termination of employment. Section 4 sets out the next-step entitlements.

2. Why this is happening

COMET-1 was commissioned in 2041 with a projected baseline capacity factor of 91%. The plant has operated safely and reliably since first power. However, the National Grid's 2044 Dispatch Reform has reduced firm-baseload procurement targets by 38% relative to the 2038 plan, primarily reflecting the maturation of regional inter-seasonal storage (Humber-Tees Caverns, Phase II) and continued cost decline of bifacial silicon-perovskite tandem PV. COMET-1's actual capacity factor in FY 2044 was 57.8%.

Under the PPA structure agreed with Anglian Power & Anglian Heavy Industries (effective 2040), the plant's revenue floor is no longer met. The Operating Company has, since Q2 2044, sought a redirected co-generation contract with the Humber Heat Cluster; that negotiation closed without agreement in September 2045.

3. Restructuring trajectory

The plant will move to Restructured Operations from 1 January 2046:

• 2 of 4 reactor units brought offline in stages through 2048
• Tritium inventory transferred to the Culham Strategic Reserve under IAEA-T protocol
• Industrial-heat configuration trialled with Anglian Steel (subject to consortium agreement)
• Wind-down to be completed by 30 June 2052

4. Personnel

All current staff have, under the FSTF, a guaranteed right to:

1. 18 months continuity of contract at present terms through 30 June 2047
2. Retraining placement in one of: HTS magnet manufacturing (Sheffield), tritium fuel-cycle services (Culham), or industrial-heat operations (Humber). Placement is matched to skill and family situation, not seniority.
3. Plant-tenure recognition when applying to next-generation fusion projects (currently: Cherwell-2, Manchester-A, Glasgow Heat-North)

5. Independent advice

The Fusion Workers' Council will hold open-floor briefings at the Visitor Centre on 21–24 October. Independent advocacy can be requested via the FSTF helpline.

6. The plant's record

In four years of operation, COMET-1 generated 17.4 TWh of zero-carbon electricity, supplied process heat to two industrial sites, and trained 64 fusion-qualified engineers. This is not a failure of the technology. It is a transition of the system that surrounds it.

---

Signed, Dr Helen Asare-Boatin · Chief Executive, COMET-1 Operating Company Ltd. Niamh O'Sullivan · Chair, Fusion Workers' Council (countersigning under FSTF §11.4)

---

Why this artifact pulls weight. The fact that it's bureaucratic is the move. A reader looking at it doesn't argue with the fusion in it — they argue with the Fusion Sector Transition Framework and what kind of country writes one. That secondary argument is exactly the conversation 2026's fusion advocacy doesn't yet know how to have. Design Fiction has done its job.

Try it yourself

Pick a scenario quadrant. Generate one Thing from the Future using
these parameters:
- Arc: <grow / collapse / discipline / transform>
- Terrain: <a sector or domain>
- Object: <document / device / interface / signage / ritual>
- Mood: <one feeling word>

Then write the artifact in full as if you found it in <year>. Use the
real bureaucratic register of that sector — payslips, receipts,
warnings, notices, manuals. Resist sci-fi vocabulary. The plausibility
of the form is the point.

---

You can stop at the artifact; many design-fiction projects do. But the next step is to put it in front of someone and watch what they do with it. An Experience Prototype is a short, contained encounter that tests whether the artifact provokes the conversation you hoped it would.

Prototype design:

• Participants: 6 — a mix of energy-sector professionals, climate advocates, and a couple of people with no particular fusion interest. (Mixed audiences read artifacts differently; that's the data.)
• Setting: A room with a printed copy of the Notice on each chair. No briefing. A camera and an audio recorder if consented.
• Time: 45 minutes.

Run:

1. 0–5 min. "Please read the document. Don't look anything up." Silence.
2. 5–15 min. "On a sticky note, write the first three questions that came to mind." Stick them on a wall. Cluster silently.
3. 15–30 min. Open discussion. Facilitator asks only: What kind of country produced this document? and Whose name is missing?
4. 30–40 min. Reveal: the document is from a 2045 scenario, and explain the 2×2 it sits in.
5. 40–45 min. "What did this surface that you hadn't been thinking about?"

What you're looking for:

• Which clauses provoked questions? (Likely §3 — the wind-down trajectory. Or §4.2 — placement matched to family situation.)
• Did anyone fight the technology, or did they argue about the framework? (If they argue the framework, the artifact is doing its job.)
• Did the climate advocates and the energy professionals react differently? (If yes, you've found a productive seam.)
• Did anyone ask "where are the workers who left in 2044?" — a missing-character question?

A note on harm. Design Fictions can land hard for people with skin in the game. If you're prototyping with people whose actual livelihoods are in fusion, brief them properly and offer an out.

Try it yourself

Design a 45-minute experience prototype for the artifact you produced
in Step 7.
- Who are 6 people whose differing readings would teach you something?
- What two facilitator questions are precise enough to provoke without
  steering?
- What's the reveal moment, and what's the last question that turns
  the encounter into useful data?
- What's the harm-aware brief for participants with skin in the game?

---

After the prototype runs, the closing move is to ask what the whole sequence has changed about your focal question. This is not a satisfaction survey on the method. It's the moment the futures work pays back.

A short debrief frame:

• What did the worked example let you see that you couldn't see before?
• What question do you now want to ask that you wouldn't have thought to ask in Step 1?
• Which of your 2026 actions look different in light of the Q2 worldbuild?
• What would you change about the focal question itself?

For this fusion example, an honest reflective read might be:

The original focal question — "What might commercial fusion adoption look like in 2045?" — was already too narrow. The real question is closer to: "What does the fusion field owe to the systems and people around it, in any of the four quadrants where it ends up?" That's a question about responsibility, not arrival. It's the question the artifact made unavoidable.

Reflection is the easiest step to skip and the most expensive to skip. It's where the methodology actually changes how the team operates next quarter.

Try it yourself

Debrief your worked example using these questions:
1. What did the example let you see that you couldn't see at Step 1?
2. What question do you now want to ask that you wouldn't have thought
   to ask at Step 1?
3. Which 2026 actions look different in light of the worldbuild?
4. What would you change about the focal question itself?
Keep answers under 60 words each. Resist the urge to summarise — name
what shifted.

---