According to @DowdEdward, companies leasing assets on a 3-year depreciation schedule are aiming to keep revenue momentum, which he says indicates the AI capex cycle is in its late innings, making revenue quality and financing structures key trading variables. Source: https://twitter.com/DowdEdward/status/1971983286554841314 Under US GAAP ASC 842, shifting from upfront equipment sales to 3-year leases recognizes income over the lease term and can smooth reported revenue while changing cash flow timing, a pattern traders monitor in late-cycle hardware markets. Source: FASB ASC 842 Major hyperscalers have recently increased server useful lives to 5–6 years, so a 3-year depreciation horizon is comparatively short and would front-load expense recognition versus peers, a useful benchmark for equity screening. Source: Amazon.com, Inc. 2023 Form 10-K; Microsoft Corporation 2023 Form 10-K; Alphabet Inc. 2023 Form 10-K For crypto markets, the state of AI capex can influence power and data center availability for Bitcoin miners, as US miners have signed AI/HPC hosting deals that tie mining infrastructure to AI compute demand. Source: Core Scientific company announcements 2024; Hut 8 Corp. investor materials 2024 Traders should track disclosures on lease mix versus capex in AI hardware vendors and data center operators, and watch any changes in hosting rates and capacity allocations that could affect BTC mining economics. Source: Edward Dowd on X; FASB ASC 842; public company filings cited above

According to @business, Nscale is a British data center company spun out of a crypto miner last year and is now at the forefront of a major AI-fueled data center boom, highlighting rapid AI infrastructure demand in the UK market (Source: Bloomberg @business). This positions Nscale as a notable AI infrastructure player in the UK and underscores the crossover between crypto mining operations and AI data center buildouts, which is relevant for capital allocation and sector rotation strategies (Source: Bloomberg @business). For traders, the report signals that entities with crypto mining roots can gain traction in AI infrastructure, warranting closer monitoring of data center operators and publicly listed miners pivoting to AI services (Source: Bloomberg @business). The development provides a tangible case study for evaluating dual-exposure narratives in equity and digital asset markets where mining competence, power access, and high-density compute capacity intersect with AI demand (Source: Bloomberg @business).

According to @StockMKTNewz, Mark Zuckerberg said Meta ($META) is planning to invest a “significantly higher number” than $600 billion in the United States through the end of the decade; source: @StockMKTNewz on X, Sep 6, 2025. That magnitude implies an annualized run-rate well above $100B (> $600B over ~5 years) versus Meta’s previously guided 2024 capex of $35–40B focused on AI infrastructure, a setup that historically boosts demand expectations for AI chips (NVDA, AMD) and U.S. data center buildouts; source: Meta Platforms 2024 capex guidance in Q1–Q2 2024 earnings materials; NVIDIA FY2025 earnings call commentary on hyperscaler demand. In crypto, AI-linked tokens such as RNDR, FET, and AKT have historically rallied around major AI equity catalysts, so traders often monitor these pairs when hyperscaler capex headlines hit; source: Kaiko Research 2024 analyses on AI token performance around NVIDIA earnings; Binance Research 2024 AI sector reports.

According to @LexSokolin, the AI grid is being built now, with compute centers as the new power plants, API calls as the new power lines, and prompts as the new switches, highlighting where infrastructure value may concentrate, source: @LexSokolin. According to @LexSokolin, this framing directs traders to focus on capacity, throughput, and reliability at the compute, API, and prompt layers when constructing exposure, source: @LexSokolin. According to @LexSokolin, the call to “bet accordingly” implies positioning in the infrastructure stack rather than purely application-layer bets as the intelligence “electrification” proceeds, source: @LexSokolin. According to @LexSokolin, crypto market participants can map this thesis to infrastructure-aligned themes that mirror power plants, grids, and switches, focusing on decentralized compute, data, and interface layers that align with the buildout he describes, source: @LexSokolin.

According to @nic__carter, the positioning can be read as either name-specific alpha versus the semiconductor basket or a barbell that is long AI infrastructure such as data centers and power while bearish chipmakers, with the book still net long despite a gigantic SMH put. Source: @nic__carter on X, Aug 15, 2025. SMH’s largest exposures are NVIDIA NVDA, Taiwan Semiconductor TSMC, and Broadcom AVGO, so a short via SMH put concentrates downside risk to these names. Source: VanEck Semiconductor ETF SMH fund documentation and holdings. A long AI infrastructure tilt is consistent with surging data center electricity needs, with the IEA estimating global data center power use could reach 620–1050 TWh in 2026 from about 460 TWh in 2022, which supports utilities and power-equipment beneficiaries over chipmakers in relative terms. Source: International Energy Agency, Electricity 2024 report. For crypto, AI infrastructure demand is already flowing to Bitcoin miners’ revenue lines as miners sign AI compute hosting deals, exemplified by Core Scientific’s multi-year agreements totaling 200 MW plus with CoreWeave, linking AI buildout to BTC miner cash flows. Source: Core Scientific press releases, June 2024. Trading takeaway: monitor the relative spread between SMH and listed power and data center beneficiaries, and watch BTC miner equities and hashrate-linked plays for flow-through when AI infrastructure outperforms chips. Source: @nic__carter on X; VanEck SMH documentation; IEA Electricity 2024; Core Scientific June 2024 press releases.