# Efficiency is the Climate Play — Doha Panel Brief My slots: Q1 (frame), Q6 (who crosses), Q7 (intelligence, 2 min), Q9 (trust layer), Q10 (who pays first), Q13 (close). 90 seconds is about 220 words: one claim, one number, one example, one instruction. --- ## Spine Efficiency is not a climate argument dressed up as a commercial one. It is the commercial argument, and the climate result follows. That is why it survives a change of government, oil price or fashion in the word ESG — and why it travels between SEA and the Gulf. Three moves: 1. SEA teaches you to price efficiency per unit. The Gulf lets you sell it per system. Same product, different denominator. 2. The binding constraint is no longer capital. It is power and time-to-energise. [[The Watt-Bit Spread and the Future of AI Power Markets]] 3. Intelligence is now cheap enough to attack problems we shelved for being too slow. That reopens a category of company that was unfundable. [[Qubits for Climate MOC]] --- ## Q1 — SEA discipline × Gulf scale SEA customers are fragmented and thin-margined, so nothing gets bought unless it lands on a bill. That forces you to price cost-per-unit-avoided — per kWh, per cubic metre, per tonne of heat. That number is portable. A sustainability narrative is not. The Gulf gives you the denominator: LNG trains, desalination, district cooling, ports, aviation. Cooling alone is roughly 70% of GCC peak summer demand *(verify)*. A 10% gain there is a national number. Build the proof in SEA where the customer is unforgiving. Sell the scale in the Gulf where the asset is enormous. Example: [Alternō](https://alterno.net), one of Alex's own — Vietnamese sand battery storing heat above 600°C, proved on tea and coffee processors, sold as heat-as-a-service so the customer never touches capex. Now in Mongolia, Japan, the Philippines. **Instruction:** write your value proposition as an equation before you write it as a sentence. If it doesn't reduce to a per-unit saving a finance team can audit, you don't have a Gulf business. [[Sustainabilty of Industrial Production]] · [[Cooling and Energy Efficiency Technologies]] · [[WebSummit Qatar x Regional Go To Market Reflections]] --- ## Q6 — Who makes the crossing work Not capital, network or product quality. Whether they treat localisation as redesign or as translation. - They rebuild the business model, not the deck. Alternō's SEA model is heat-as-a-service to smallholders; the Gulf model for the same physics is a capex sale or JV. Same molecule, different company. - They budget for the irrational first year — custom deployments before any platform. [[3 Hard Truths of Deep Tech Commercialization]] - They arrive with a local partner who has a P&L exposed to the outcome, not an advisor with equity. - They pick the boring buyer. The ones who fail chased the flagship customer. Negative signal: a founder who says the Gulf is the same market, just richer. --- ## Q7 — Abundance of intelligence (2 min) Open with the arc: > The industrial revolution took physical work off humans — machines, automation, factories, agriculture to transport. This century is the cognitive revolution, where our share of cognitive work goes to agents. What stays ours is the non-verifiable half: choices, trust, brand, relationships. Which is why rooms like this still matter. The big decisions still get made in rooms, not in models. What changed: 1. The cost of a hypothesis collapsed. DeepMind's GNoME produced 2.2m crystal structures, ~380,000 stable candidates — roughly 800 years of accumulated knowledge in one pass. 2. The bottleneck moved from theory to the wet lab and the pilot line. That is a capex and access problem, which the Gulf is well placed to solve. 3. Quantum extends this to what classical simulation can't reach. McKinsey: ~7 Gt CO2e of annual abatement by 2035. [[Quantum tech could positively impact the climate]] The example: Haber-Bosch is roughly 1.8% of global CO2, running on a catalyst optimised a century ago at 400–500°C and 200 atmospheres. Nature fixes nitrogen at ambient temperature. That gap stayed open for 100 years because the search space was too large. It isn't now. [[Ammonia Catalysts Innovation MOC]] **Instruction:** stop treating a decade-scale research programme as unfundable and start treating it as a 24-month de-risking sprint with a validation partner attached. Own the data and the validation loop, not the model — models commoditise in months, a proprietary dataset from a real asset does not. And note the corollary: your competitors' hypothesis costs collapsed too, so defensibility is now almost entirely physical and relational. --- ## Q9 — Trust layer or the asset? Yes, right now — where the asset already exists and the trust is asserted rather than proven. Ramos and Jamal are both that shape: assets deployed and under-monetised because nobody can verify what they did. Economically: the asset earns a utility return on capital. The verification layer earns a software return on trust established, with a fraction of the capex. With interconnection queues at 5–15 years, the layer that doesn't need a grid connection is structurally faster. The nuance, so it isn't glib: the trust layer is only investable where the measurement is hard to fake. MRV that reduces to a spreadsheet is a consultancy. MRV anchored in instrumented, tamper-evident data is infrastructure. [[AI Verification]] · [[Integrity Kernel MOC]] Say the caveat out loud, on a panel where Nicole has metal in the street: the trust layer is the better venture bet, not the more important thing. **Our own book — the same shape, with customers:** | Venture | The point | |---|---| | [[Azraq MOC\|Azraq]] | Risk intelligence for digital infrastructure — a $7tn+ asset class with no unified risk layer. The layer that makes an asset financeable is worth more than another asset. | | [[Renacore/Company Overview and Pivot\|Renacore]] | Physics-informed failure prediction over LNG assets, read-only off SCADA, live in under 30 days. Pivoted horizontal to vertical because Qatar access pulled it — the crossing done properly. | | [[Barrier Intelligence MOC\|Barrier Intelligence]] | Safety-barrier lifecycle for O&G. Qatar NFE alone is $30bn+, LNG scaling 77→142 mtpa by 2030. The buyer is regulated into caring. | | Dextrum | *(your one-liner — what layer, over which asset)* | None of them built the asset. They monetise one someone else paid for, by making it legible. --- ## Q10 — Who pays first Fair, and it's a sequencing problem rather than a capital one. Gulf: capital is available; the scarce thing is a large asset owner willing to go first. Nobody is rewarded for being the first buyer of an unproven system on a critical asset. SEA: the trial is easy, the capital to industrialise past it is the wall. Okapi (Radical Fund, Malaysia) is the proof — the financing *is* the product. What a founder does: - In the Gulf, sell the pilot and make somebody's career on it. Scope it small enough that failure is survivable for them, and price the first deployment as marketing rather than revenue. - Negotiate for the reference, not the revenue. Get the right to name them. - In SEA, instrument every pilot. Most founders run ten and can prove nothing from any of them. - Run both legs deliberately: SEA generates the proof, the Gulf generates the balance sheet. --- ## Q13 — Close (one sentence) > Get one customer in the new market to stake something real on you before you move — everything else you can rebuild, trust you can only earn in person. --- ## Fact bank **Grids and energy** - Global electricity demand grows 3.6%/yr 2026–2030, ~50% faster than the previous decade. Grid investment must rise ~50% from ~$400bn by 2030. ([IEA](https://www.iea.org/reports/electricity-2026/executive-summary)) - Over 2,500 GW stuck in global connection queues. Grids take 5–15 years; renewables 1–5; data centres 1–3. Component prices nearly doubled in five years. ([IEA](https://www.iea.org/reports/electricity-2026/grids)) **Data centres, if it comes up** - Just over 1% of global electricity, 0.5% of global CO2. To 2030 they add less demand growth than air conditioning (+651 TWh) or EVs (+838 TWh); they reach 945 TWh, about Japan's total. ([Carbon Brief](https://www.carbonbrief.org/ai-five-charts-that-put-data-centre-energy-use-and-emissions-into-context)) - The objection is local, not planetary: Ireland 21% of national electricity, Dublin 79%, Virginia 26%. - Affordability is the sharp version: PJM's capacity auction cleared at its $329/MW-day cap, $16.1bn, 7x two years earlier; its market monitor attributed a 174% price increase to data centre load. [[Data Centers could make or break electricity affordability]] - Water: US data centres ~163.7bn gallons direct (2021), ~211bn indirect via generation (2023). Golf is ~425bn/yr — but data centre use grows ~20%/yr against flat golf, so parity lands 2026–28. The honest line is "less than golf today, and that stops being true this decade." - Weak point, don't defend it: fossil fuels supply ~60% of data centre power. [[Jevons Paradox]] **Intelligence** - GNoME: 2.2m structures, ~380,000 stable candidates. ([DeepMind](https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/)) - Quantum: ~7 Gt CO2e abatement potential by 2035 (McKinsey). - COP26 commitments need ~$4tn/yr to 2030 and still land at 1.7–1.8°C. A technology gap, not a capital gap. **Verify before saying on stage:** GCC cooling ~70% of peak load; district cooling 40–50% more efficient; exact Haber-Bosch share. **Radical Fund bench** (Alex's book, for redirects): Alicia Bots — ship hull crawler, the cleanest Gulf-scale efficiency example. Sirsak and Arkadiah — trust layers on existing assets. Dash Electric mirrors Nicole; Waterhub mirrors Rhana. ([portfolio](https://www.theradicalfund.com/portfolio)) --- [[Data Center MoC]] · [[Physical Assets Revival]] · [[Time Price]] · [[Themes shaping 2025]] · [[Seamless web of well deserved trust]] · [[Technical Moat Assessment Framework]] · [[Virtual Power Plant (VPP) MOC]] · [[Context Layers MOC]] · [[Industrial AI MOC]] · [[Web Summit Panel Briefing - Infrastructure Intelligence and Sovereign Systems]]