Energy Transition: My Predictions for 2026

Staying deliberately away from geopolitics — this is about technology, markets, and physical reality.

2026 is shaping up to be the year the energy transition finally stops being a culture‑war talking point and becomes what it actually is: an industrial execution problem.

Grid queues, transformer shortages, mine supply curves, factory automation, AI power demand — everything converges. Not hypotheticals. Schedules.

This isn’t a hype piece. It’s a base‑case map of where the world is already heading, with realistic ranges rather than fantasy precision.

When I say execution, I mean the unglamorous work on the ground: faster permitting and interconnection, scalable financing models, standardised hardware, and supply chains that actually deliver transformers, cells, and cables on time.

1. Commodities: tight stuff and tiny stuff

The big backbone metals

Copper sits at the center of everything in 2026: grids, chargers, transformers, motors, windings, data centers. There is no fast substitute and no fast supply response. Even modest delays in new mines ripple straight through prices.

Lithium remains volatile but structurally demanded. EVs are no longer the whole story — grid storage and AI‑era electricity smoothing are now permanent demand pillars. Oversupply narratives fade once deployment outruns project timelines.

Monetary metals still matter

Gold enters 2026 as a macro stress gauge rather than an inflation hedge alone. Rates, currencies, geopolitics and capital rotation keep it relevant — but expect volatility, not a straight line.

Silver remains the most misunderstood metal of the cycle. It behaves like a monetary asset during fear spikes and like an industrial metal during build‑outs. Solar, electrification and grids keep the floor firm, while price swings stay violent.

Strategic gremlins

These are small‑volume, high‑leverage materials: they rarely move markets in tonnage terms, but when supply tightens or performance margins matter, they can quietly dictate outcomes across semiconductors, batteries, and next‑generation energy systems.

Gallium is geopolitical, not geological. Tiny market, concentrated supply, huge strategic value for semiconductors and RF systems. Price action in 2026 is driven by policy, not demand curves.

Cesium never makes headlines, but it punches far above its weight in niche high‑performance applications. Supply is extremely concentrated and volumes are tiny, which means price behaviour is discontinuous rather than cyclical.

In solar, cesium shows up in advanced perovskite and tandem PV research, where cesium‑based compounds improve thermal stability, lifetime, and efficiency. It’s not about tonnes — it’s about enabling next‑generation modules to survive real‑world heat, UV, and moisture.

Beyond PV, cesium is critical in vacuum electronics, atomic clocks, sensors, and specialty optics, all of which quietly benefit from electrification, grid digitisation, and AI‑era infrastructure. Demand growth is incremental, but supply fragility means even small increases matter.

Cesium won’t trend with the energy transition — it will surface as a constraint metal when performance margins matter most.

Magnesium is the quiet efficiency metal — but in batteries it also has a chemistry-level role, not just a structural one.

In LMFP (Lithium Manganese Iron Phosphate) — a higher-voltage evolution of LFP — magnesium is increasingly used as a dopant and stabiliser. Small amounts of Mg improve crystal stability, reduce lattice distortion during cycling, and help suppress manganese dissolution — one of LMFP’s key degradation pathways. The payoff is better cycle life, improved thermal behaviour, and higher usable voltage without sacrificing LFP’s safety and cost advantages.

In LMR (Lithium‑rich manganese‑based cathodes) — next‑generation, high‑capacity cathode materials — magnesium plays a similar but even more critical role. Mg doping helps stabilise the oxygen lattice, mitigate voltage fade, and improve structural integrity during high‑capacity cycling. LMR promises very high energy density at lower cost, but only if its stability issues are controlled — and magnesium is one of the quiet tools engineers are using to make that viable.

This puts magnesium in a rare category: a performance enabler at the atomic scale. It doesn’t drive demand in tonnes, but it unlocks next‑generation cathode chemistries that matter for EVs and grid storage alike. Like cesium, it won’t trend — it will matter precisely where margins are tight and failure isn’t tolerated.

Theme: 2026 rewards understanding where supply is brittle, not just where demand is big.

2. EV adoption: normalisation, not novelty

(All figures below refer to global passenger and light‑duty plug‑in vehicles.)

By 2026, EVs are no longer proving themselves — they’re competing on price, experience and availability.

Global plug‑in sales: ~23–28 million
Global plug‑in share of new car sales: ~27–33%

China continues to lead on volume and affordability. EVs behave like appliances — chosen by default, not ideology.

Europe lives or dies by pricing and fleet economics. Affordable BEVs (battery electric vehicles) matter more than premium specs.

The US moves slower, but directionally stays positive as charging, grid upgrades and used‑EV pricing mature.

The real shift in 2026 is psychological: EV ownership stops being explained. It’s assumed.

A key signal of this shift already arrived in 2025, when two countries crossed the 90% EV sales threshold: Norway (which effectively reached this level in 2024) and Denmark. Once markets enter this zone, reversal becomes structurally unlikely — the ecosystem has flipped.

My base case for 2026 is that more than one additional country joins the 90% club. Candidates are smaller, wealthy, grid‑stable markets with strong urban density and policy continuity, where EVs are already the default choice rather than an alternative. This is how S‑curves end: not with drama, but with inevitability.

3. Robotics: ROI beats spectacle

Robotics accelerates in 2026 not because the technology is flashy, but because labour scarcity, wage pressure, and reliability demands leave fewer viable alternatives.

Autonomy and FSD quietly cross a line

Alongside robotics, vehicle autonomy makes a less visible but important step-change by 2026.

China is already trialling Level 3 (L3) autonomy on public roads in controlled domains, and this matters more than flashy demos. L3 shifts responsibility from driver to system under defined conditions — a legal, regulatory, and technical threshold that most regions have not yet crossed.

By 2026, expect expanded L3 deployment in geofenced, rule‑dense environments (urban highways, ring roads, logistics corridors), particularly in China. Progress elsewhere will be slower and more fragmented, but the direction is set.

This does not mean full autonomy everywhere. What it means is that hands‑off driving becomes normal in specific contexts, and the conversation quietly moves from “can it work?” to “where is it allowed?”

Autonomy, like robotics, advances through constrained, economically useful domains first — not through universal freedom, but through execution.

2026 is not about robots doing backflips. It’s about robots doing boring work reliably.

Warehouses, logistics hubs, factories and ports keep automating because labour volatility and efficiency demands leave no alternative.

Humanoids continue to advance, but deployments stay controlled and domain‑specific. The winners are the teams that ship uptime, not stage demos.

Robotics in 2026 becomes infrastructure — invisible, expected, and economically justified.

4. AI: power becomes the bottleneck

By 2026, AI capability is no longer the constraint. Electricity is.

Data centers scale faster than grids can connect. Cooling joins power as a secondary constraint — air, water, and immersion systems increasingly shape where and how capacity can be deployed. Transformers, interconnection permits, and local generation become the limiting factors.

This triggers a feedback loop:

More grid investment
More storage deployment
More copper, lithium and silver demand

AI doesn’t slow — it drags the physical world forward with it.

5. Renewables and grid storage: the decade reveals itself

Renewables keep adding capacity at record pace, but 2026 makes one thing clear:

Storage is no longer optional.

Solar and wind increasingly ship paired with batteries by default. Curtailment becomes a design failure, not a market surprise.

LFP goes mainstream — and stays there

LFP (Lithium Iron Phosphate) firmly establishes itself as the workhorse chemistry of the energy system. In 2026 it dominates grid storage, entry‑level and mid‑range EVs, commercial fleets, and energy‑intensive applications where safety, longevity, and cost matter more than maximum energy density.

Cycle life, thermal stability, low degradation, and predictable behaviour make LFP the default choice wherever uptime and total cost of ownership dominate decision‑making. LMFP builds directly on this foundation, extending voltage and performance without abandoning LFP’s core advantages.

In short: LFP doesn’t get displaced — it becomes infrastructure.

Sodium-ion quietly enters at the bottom

Sodium-ion doesn’t replace lithium in 2026, but it earns a permanent foothold.

Its role is clear and limited: low-cost, lower-energy-density storage where weight and range don’t matter — stationary storage, short‑range mobility, two‑ and three‑wheelers, backup systems, and regions where lithium supply chains or cost volatility are problematic.

Sodium’s advantage is not performance; it’s abundance, cost stability, and supply-chain simplicity. In a world building storage at scale, that matters.

By 2026, sodium-ion shifts from “interesting alternative” to accepted secondary chemistry — not everywhere, but exactly where it makes sense.

Global grid‑scale storage deployment accelerates sharply, moving from niche projects to system‑critical infrastructure. Markets evolve, revenue stacking improves, and storage graduates into a recognised asset class.

6. Transport electrifies beyond cars

A crucial but often overlooked pillar of this shift is mining and construction equipment electrification. Haul trucks, loaders, excavators, and site machinery electrify first in controlled environments where fuel logistics are costly, utilisation is high, and downtime is expensive. These deployments are execution-led: fixed routes, on-site charging or trolley systems, immediate operating cost savings, and measurable emissions reductions — not ideology, just economics.

By 2026, the electrification story clearly expands beyond passenger vehicles.

BEV trucking continues its steady transition from pilot programs to scaled fleet deployment. Short‑haul, regional, port, and depot‑based trucking electrifies fastest, driven by predictable routes, charging at base, lower operating costs, and tightening urban emissions constraints. Megawatt‑scale charging moves from trials into early commercial reality.

Buses are already deep into the S‑curve in many regions and continue to electrify almost by default. Transit agencies increasingly treat diesel as a legacy procurement choice rather than a neutral option.

Rail benefits indirectly as electrified freight corridors, hybrid systems, and battery‑assisted locomotives expand where full overhead electrification is impractical.

Maritime sees continued growth in electric ferries, port equipment, and short‑range coastal vessels, where energy density requirements are manageable and charging infrastructure is localised.

Aviation remains the most constrained — but not static. Electric and hybrid‑electric aircraft expand in training, regional hops, cargo drones, EVTOLs, and airport‑adjacent operations. These are not gimmicks; they are the first economically viable niches where batteries outperform liquid fuels on efficiency, maintenance, and noise.

Across all modes, the pattern is the same: electrification enters where routes are known, utilisation is high, and infrastructure can be planned. Scale follows execution, not headlines.

2026 isn’t about proving the transition works — it’s about executing it at scale.

The 2026 takeaway

2026 isn’t a constraint story — it’s a build‑out story.

EVs keep eating the car market, AI keeps eating electricity, storage keeps eating curtailment — and commodities get dragged along behind the tow bar.

The debate phase is over.
Execution has arrived.

So, what’s your 2026 vector?

In a year defined by execution, there are only three strategic postures:

Build the bottleneck — own the transformers, the interconnection queue, the charging power, the mines.
Simplify the stack — back LFP, sodium‑ion, standardised hardware, and systems that scale cleanly.
Automate the inevitable — deploy robotics and autonomy where labour is scarce, costly, or unreliable.

The worst position is to be a spectator to the spectacle — mistaking humanoid demos for deployment, or commodity volatility for a lack of direction.

One final note: expect humanoid robotics demos to reach a new level of frenzy in 2026. As AI models, sensors, and actuators converge, demonstrations will look dramatically more capable and will dominate headlines. Most of this will still be pre‑scale and tightly scripted — but it matters psychologically. These demos will accelerate capital flow, talent migration, and public belief, even if widespread deployment lags the spectacle.

Bring on 2026. ⚡🔋🐂