Myuex

Long term energy storage solutions

The thesis that AI-driven electricity demand will drive adoption of long-duration energy storage is structurally sound and well-supported by a dense cluster of confirming news events. Multiple reports from BloombergNEF, PJM, and major tech firms project data center power consumption doubling or tripling by 2030-2035, with grid connection bottlenecks and emergency backup requests already visible. This creates a clear causal chain: surging AI load -> grid strain -> need for firming capacity -> demand for storage solutions. However, the regime is actively hostile to this thesis: inflation is reaccelerating at 3.3% CPI, the Fed is on hold with hike risk signaled by Chair Warsh, and bond yields are rising. Higher rates directly increase the cost of capital for capital-intensive storage projects, delay utility procurement cycles, and compress the valuation of pre-revenue growth stocks like GWH, EOSE, and IESVF. The basket's 30% drawdown since posting reflects this tension — the fundamental demand story is strengthening, but the macro-financial headwinds are overwhelming it. The strongest upside driver is the BloombergNEF forecast doubling US data center power to 194 GW by 2035, which directly supports the need for storage to manage intermittent renewables and grid peaks. The most probable risk is the Fed hiking rates, which would further compress the basket's valuations and delay project financing. The basket's composition is mixed: FLNC and NRGV have real revenue and some path to profitability, while GWH and IESVF are pre-revenue with high cash burn, making them acutely sensitive to rate-driven equity market rotation. BEP is a utility-scale renewable generator with only 30% storage exposure, diluting the thesis's direct impact. The author's claim that long-duration storage can stabilize the grid is coherent but remains speculative — no major utility has yet contracted iron or vanadium flow batteries at the scale needed to prove the thesis. The mechanical anchor is balanced because the upside demand evidence and downside rate risk are roughly equal in weight, and I hold at that level: the structural AI-demand story is real and intensifying, but the hostile rate regime and speculative nature of the specific technologies in this basket prevent conviction from rising above neutral.

GaN is the future of semiconductors

The thesis rests on two strong structural forces: AI data center power demand, which is confirmed by multiple high-relevance events (BloombergNEF projecting 20% of US electricity by 2035, OpenAI's $750B compute plan, Brookfield/NextEra's $100B campus), and China's gallium export controls, which are actively escalating and have a confirmed FID from Alcoa's Western Australia project. The regime supports this — the AI boom is accelerating, raw materials are very bullish, and the US-Iran war adds energy cost pressure that makes GaN efficiency more valuable. However, the basket's realized performance is poor (-9.8% vs S&P +3.3%), and the key risk is the enterprise procurement cycle: the Fed's hold with hike risk (confirmed by multiple news stories) keeps IT budgets frozen, and the author's own claim that volume production waits until 2027 is speculative with no confirming evidence. The transmission paths are credible: NVTS is a pure-play GaN designer with $450M in design wins, FPLSF refines high-purity gallium and is directly linked to the supply bottleneck, and AA extracts raw gallium as a byproduct of alumina refining — though AA's exposure is only 0.45 because aluminum is its dominant revenue driver. The strongest upside driver is the Alcoa gallium FID, which directly validates the non-Chinese supply chain thesis. The strongest downside is the Fed's persistent hawkish stance, which extends procurement delays and keeps the timeline risk alive. The mechanical anchor reflects balanced upside and downside on the force evidence, and I sit slightly below it because the enterprise procurement cycle pressure has high-probability confirming news (Fed hold stories) that directly delays the thesis's timeline, while the upside scenarios (GaN adoption, supply deals) remain unconfirmed beyond the Alcoa FID. The score is driven down by the gap between structural tailwinds and the near-term execution risk that the author themselves identified.