AI Infrastructure
AI Infrastructure
Data centers, power generation, networking, compute capacity
Embedded Files
PUBLIC MARKET COVERAGE
Vistra Corp. (VST): The Hedge Book Holds and the Estimates Do Not
Vistra Corp. (VST): The Hedge Book Holds and the Estimates Do Not
Sector: Utilities, Independent Power Producers
Date: September 23, 2026
Verdict: CONDITIONAL GO · Fair Value $171 · Entry below $136.80
Abstract: First-half generation earnings rose 47% while the $472m hedge mark that took GAAP net income down 6.7% is the accounting shadow of a book sold forward two years, roughly 100% of 2026 and 94% of 2027. What moves the mark is the sell side's 8% cut to 2027 earnings in sixty days, not the price; at multiples of 14x and 19x fixed at initiation, that revision alone sets the fair value. The September 22 close of $140.43 sits in the Fair band, 2.7% above the accumulation band's top edge, with the 28% open 2028 book and a policy channel running through ten states as the tail. Sell baselines are $177.47, $186.94 and $193.23, with the scenario floor at $94.60.
IREN Limited (IREN): The Contracted Number Is Real and the Recognised Number Is Not It: Initiating at Avoid on a Funding Gap Wider Than the Company
IREN Limited (IREN): The Contracted Number Is Real and the Recognised Number Is Not It: Initiating at Avoid on a Funding Gap Wider Than the Company
Sector: AI Infrastructure, Data Centers
Date: August 28, 2026
Verdict: AVOID · Fair Value $28.25 · No entry at current levels
Abstract: 87.7% of reported operating cash flow is the increase in deferred revenue rather than operating earnings, so the cash statement describes prepayment timing and not contract economics. The $4bn contracted ARR contains two opposite blocks: a 200MW anchor at $9.50m per MW that loses $0.34m per MW pre-tax on contract-term depreciation, and a 100MW repriced block at $21.00m earning $1.39m. The $40.54 close embeds 57.8% weight on the build-out scenario, pricing 288.8MW of the 2027 increment as contracted when none of it is. FY27 capex guidance of $25bn to $30bn against a $14.5bn market capitalisation makes the funding gap, not demand, the deciding variable.
Vistra Corp. (VST): The Multiple De-Rated, the Earnings Did Not: A 38% Drawdown That Carried the Close Through the Entry Line
Vistra Corp. (VST): The Multiple De-Rated, the Earnings Did Not: A 38% Drawdown That Carried the Close Through the Entry Line
Sector: Utilities, Independent Power Producers
Date: August 24, 2026
Verdict: CONDITIONAL GO · Fair Value $178 · Entry below $142.40
Abstract: The stock has fallen 38% from its October 2025 high while second-quarter adjusted EBITDA of $1,767m beat both the street median and the covering house's estimate. The hedge book is what decides the rating: approximately 100% for 2026, 94% for 2027 and 72% for 2028, which places two years of earnings outside spot power risk while the equity still trades on the AI power narrative. Guidance reflects neither the Meta power purchase agreements nor the Cogentrix acquisition, and the $1.0bn Helix investment alongside NVIDIA and KKR carries a right of first choice to supply power with no obligation to participate. Sell baselines are $184.87, $194.58 and $200.96.
SK Networks Co., Ltd. (KRX: 001740): Two Limit-Up Closes in Ten Sessions Reprice a Conglomerate as an AI Proxy; the Market Now Pays ~KRW 1.2T for a Pre-IPO Stake Marked Near KRW 0.26T; Fair Value ~KRW 9,000: AVOID, Do Not Chase
SK Networks Co., Ltd. (KRX: 001740): Two Limit-Up Closes in Ten Sessions Reprice a Conglomerate as an AI Proxy; the Market Now Pays ~KRW 1.2T for a Pre-IPO Stake Marked Near KRW 0.26T; Fair Value ~KRW 9,000: AVOID, Do Not Chase
Sector: Conglomerate, AI Exposure
Date: June 10, 2026
Verdict: AVOID · Fair Value KRW 9,000 · Do not chase
Abstract: SK Networks closed at KRW 14,290 on June 9, capping a five-week 117% advance with two limit-up closes in ten trading sessions: May 26 on the Upstage Series C disclosure, and June 8 after opening down 12.8% and closing at the ceiling. The June 9 session described the regime, with an 8.4% gap open, a new fifty-two week high of 16,240, a fade to 12,900, and a 0.8% gain at the close on 54.4 million shares, 27% of every share outstanding. What is being repriced is not the operating company, expected to earn KRW 128B in 2026, but a 12.9% minority stake in Upstage, the AI-model developer marked near KRW 2T and expected to list in the second half. The KRW 2.87T market value exceeds the operating business by roughly KRW 1.2T, about 4.6 times the stake's current mark and nearly double its value even in the KRW 5T listing scenario. Blending an earnings lens near KRW 6,400 and an asset lens near KRW 8,400 and adding a probability-weighted stake value yields fair value near KRW 9,000, about 37% below a close that stands 30% above the highest published target. The earnings recovery is real and small relative to the price: first-quarter operating profit of KRW 33.4B rose 102.4% and beat consensus by 26%, yet the forward multiple of 36.2x sits against 15.65x for sector peers, and consensus 2026 earnings per share of KRW 395 sits below trailing KRW 436.
Arista Networks (ANET): AI Ethernet Misread as Growth Ceiling
Arista Networks (ANET): AI Ethernet Misread as Growth Ceiling
Sector: AI Networking, Communications Equipment
Date: May 6, 2026
Verdict: CONDITIONAL GO · Target $181 · Entry within $135 to $185
Abstract: The 15.4% post-earnings sell-off on May 6 reflects multiple compression on a misread of supply-constraint commentary rather than an earnings break. Three quarters of revenue acceleration to 35.1% year on year, deferred revenue at $6.2bn after growing 101%, and a raised FY26 guide of $11.5bn at 27.7% growth all point to demand strength rather than a ceiling. The probability-weighted twelve-month target is $181 against a $144 entry, with pre-committed thesis-break triggers at second-quarter revenue below $2.65bn, Spectrum-X share above 35%, or single-customer concentration above 30%.
IonQ (IONQ): Quantum Computing Pure-Play Entry Framework
IonQ (IONQ): Quantum Computing Pure-Play Entry Framework
Sector: Emerging Technology, Quantum Computing
Date: April 17, 2026
Verdict: WATCH · Trigger-conditional · Position capped at 3%, hard 30% stop
Abstract: This is a company where discounted cash flow and forward earnings multiples do not apply: 2025 revenue of $130m sits against a $510m GAAP loss, and the stock trades at 50 to 120 times sales. The note therefore builds a separate rule set for milestone-driven equities rather than forcing a valuation framework onto a pre-commercial business. Two entry gates are defined in advance: first-quarter revenue above the $51m guidance ceiling, and commercial customer mix holding at 60% or better. Probability-weighted 2030 fair value is approximately $114 against a $30 to $35 market, with a 30% downside scenario priced in through a hard stop-loss and a strict size limit rather than through the fair value itself.
Constellation Energy (CEG): Regulatory Reset Creates a Band, Not a Break
Constellation Energy (CEG): Regulatory Reset Creates a Band, Not a Break
Sector: Power & Utilities
Date: April 16, 2026
Verdict: CONDITIONAL GO · Target $280 · Entry within $260 to $305, no chase above $325
Abstract: Recommended $260–$305 entry after a 28% selloff from $402.95 driven by FERC and PJM regulatory repricing, not a fundamental break. Anchored the downside on 32.4 GW of operating nuclear capacity valued at ~$2,500/kW versus $12,000/kW new-build replacement cost, with the 20-year Microsoft Crane PPA and the post-Calpine 55 GW combined platform underwriting a 13%+ EPS growth path. Probability-weighted 12-month expected value of $354 (bull $465 / base $360 / bear $230) implies ~20% upside; thesis reassessment is triggered if the mid-2026 FERC ruling on PJM large-load interconnection is categorically restrictive.
Alphabet Inc. (GOOGL): Contrarian Long: Search Moat Intact, Peer Discount Unjustified
Alphabet Inc. (GOOGL): Contrarian Long: Search Moat Intact, Peer Discount Unjustified
Sector: AI Infrastructure, Megacap Technology
Date: April 16, 2026
Verdict: CONDITIONAL GO · Fair Value $385 · Entry within $275 to $340
Abstract: Three concerns drive the 18% peer discount: AI cannibalization of Search, $180B CapEx burden, and DOJ antitrust overhang. Each fails under examination. Search revenue growth accelerated from 9.8% to 17% YoY through 2025 while ChatGPT, Claude, and Perplexity scaled, contradicting the cannibalization thesis. CapEx intensity at 42.9% of revenue is in line with MSFT (43.5%) and below META (45.5%), so the discount is not relative. Chrome divestiture was rejected in September 2025, removing the structural tail risk. Probability-weighted fair value of $377 against $317, with a tighter bear case (-8%) than bull case (+47%).
Banking Sector Cloud Migration: Implications for Hyperscaler Equity
Banking Sector Cloud Migration: Implications for Hyperscaler Equity
Sector: AI Infrastructure
Date: February 1, 2026
Verdict: GOOGL GO · MSFT CONDITIONAL GO · AMZN WATCH
Abstract: Multi-cloud has shifted from best practice to regulatory mandate after the October 2025 AWS US-EAST-1 and Azure Front Door outages, combined with the Bank of England and FCA critical third party regime, DORA, and the EU Data Act. Microsoft's banking lock-in is eroding at the margin while Google Cloud captures disproportionate share through TPU and Gemini vertical integration, validated by a $155bn backlog after 82% growth and operating margin expansion to 23.7%. Sum-of-the-parts analysis identifies Google Cloud as the cleanest re-rating opportunity at 12 times revenue against the discount embedded in the current consolidated multiple.
Page updated
Google Sites
Report abuse