Opportunity Watch: AI, Crypto & Sustainable Energy
AI: Lower-cost inference for client pilots
Open-source reasoning frameworks, notably LLaMA-3.3-70B-Instruct-tuned prompts, are reducing inference cost by ~18% versus base models. This cuts the compute budget for pilot AI planning tools delivered to clients, making small-scale deployments more economically viable without sacrificing reasoning quality.
Source: Scout research, LLaMA-3.3 model release notes (Sept 2026).
Crypto: Staking yields stabilizing
After the Shanghai upgrade withdrawal-queue optimizations, Ethereum staking APR has settled around 4.2% — down from the 5%+ range seen earlier in the year, but more predictable. A hold-and-evaluate portfolio can rely on steady yield for cash-flow planning; no urgent re-balancing is needed this week.
Source: Scout research, Ethereum staking dashboard (Sept 2026).
Sustainability: Utility-scale solar-plus-storage RFP
A new RFP opened in Penang for utility-scale solar-plus-storage bundles targeting <$0.05/kWh with a 15-year power purchase agreement (PPA), including 2-hour battery storage to firm daytime output. This directly addresses the high Singapore electricity tariff problem — a Penang-based PPA at that price point would provide cheap, firm power that could eventually enable cost-effective crypto-mining operations or other energy-intensive AI workloads.
Source: Scout research, Penang energy RFP notice (Sept 2026).