Google ex-CEO Eric Schmidt jumps into the AI data center business with a failed, 150-year-old Texas railroad turned oil giant

Former Google CEO Eric Schmidt launched Bolt Data & Energy to build vertically integrated data-center campuses on Texas Pacific Land’s 882,000 acres, with TPL investing $50 million as part of Bolt’s $150 million initial raise. Bolt plans to start with natural-gas-fired generation ramping to 1 GW and ultimately target 10 GW of capacity (stated as enough to electrify ~7 million homes) while transitioning to wind, solar, batteries and future nuclear; the firm is marketing bespoke campuses to major hyperscalers and AI firms. The deal gives TPL (c. $20 billion market cap) a new revenue and asset-utilization vector (land, water and energy services) and positions Bolt to supply large-scale compute infrastructure in a high-demand AI market. Investors should view this as a strategic, medium-impact infrastructure/energy play with execution and permitting risks but clear upside if hyperscaler demand materializes.

Analysis

Market structure: Vertical integration of land+power+water (Bolt+TPL) tilts value to landholders and energy owners (TPL, Permian gas midstream, renewables/battery OEMs) and reduces long-term price power of grid-supplied data‑center tenants. Expect TPL to capture outsized rent/royalty-like economics; hyperscalers win only if they secure anchor pricing ~10-25% below market power costs and commit capital. Supply tightness for Permian gas and local grid capacity could raise local power prices 10-30% in stressed buildouts.

Risk assessment: Key tail risks are permitting/water-right litigation, capex overruns, and anchor-customer churn; a single failed anchor could wipe out early-stage returns. Immediate (days) impact is minimal; short-term (3–12 months) depends on LOIs and state approvals; long-term (2–5 years) realizes if Bolt hits 1–10 GW scale. Hidden dependencies: pipeline takeaway, transmission interconnect lead times (18–36 months), and federal/state permitting for turbines/nuclear.

Trade implications: Tactical: long TPL as primary direct play; complimentary exposure via selective hyperscaler optionality (GOOGL/MSFT) via 12–24 month call spreads to capture re‑rating if they sign. Hedge: long Permian gas futures/call options on a 3–6 month horizon if Henry Hub or regional basis tightens >15%. Avoid/short frothy pure-play AI energy IPOs and high-multiple data‑center REITs that priced in immediate 10+ GW rollouts.

Contrarian angles: Consensus underestimates execution drag and capital intensity — Fermi’s post‑IPO drawdown is a cautionary parallel. Market may overpay for 'AI land' narrative before anchor commitments; a realistic trigger is a signed multi‑year power purchase or lease (LOI) — wait for that before >3% position sizing.

Key Decisions for Investors

• Establish a 1.5–2.0% long position in Texas Pacific Land (TPL) within 30 days, target 30–50% upside over 6–18 months on anchor-customer news; set a tactical stop-loss at -20% and trim to 1% if no LOI within 9 months.
• Deploy 0.5–1.0% of capital into 12–24 month call spreads on GOOGL or MSFT (favor MSFT if cheaper volatility) to capture re-rating from potential anchor deals; risk small premium, take profits at +40% or on confirmed LOI within 12 months.
• Allocate 0.5% into short-dated (3–6 month) Permian-focused natural gas call options or calendar spreads to profit from regional takeaway constraints; add an equal notional if local basis widens or Henry Hub rises >15% within 90 days.
• Reduce exposure to high-multiple AI/energy infrastructure public names and data-center REITs by 25–50% vs. benchmark; consider shorting or buying puts on any new AI energy IPOs that priced >5x revenue without contracted offtake, and only revisit after signed multi-year contracts are disclosed.