Forget Nvidia: This Utility Stock Could Actually Outperform It From Here

Oklo is presented as a high-risk, long-duration beneficiary of AI data-center power demand through its planned small nuclear reactors and long-term electricity sales model. The company is valued at roughly $7 billion—more than 5,500x current sales—and has no commercial reactor, making near-term execution highly uncertain. The article argues that a 10x gain to a $67 billion market capitalization by 2036 is plausible if Oklo commercializes and scales its reactors, while Nvidia reaching 10x would imply a $53 trillion valuation.

Analysis

The investable implication is not that AI load automatically accrues to pre-revenue reactor developers; it is that firm, deliverable power has become a scarcity asset. CEG is the cleaner listed beneficiary because it can monetize tightening capacity markets and data-center contracting on an operating fleet, whereas OKLO must clear licensing, construction, fuel, financing and execution hurdles before it has merchant or contracted cash flow. Hyperscalers may ultimately favor a portfolio of grid-connected nuclear, gas-backed generation, storage and transmission rather than rely on first-of-a-kind on-site SMRs.

For the next 1-3 months, promotional attention can sustain OKLO's retail-driven beta to AI and nuclear headlines, but its valuation leaves little room for delays in regulatory milestones, customer conversion, or capital raises. The key second-order risk is financing: a higher equity price lowers dilution cost, but a reversal can create a reflexive funding problem because reactor development requires substantial capital well before recurring revenue. A credible long-term power purchase agreement with creditworthy counterparties would matter more than additional nonbinding pipeline announcements.

Over 6-18 months, the likely bottleneck is not reactor demand but supply-chain and project-delivery capacity: HALEU fuel availability, qualified nuclear components, EPC labor and NRC timing. These constraints favor incumbent generators and fuel-cycle exposure before they benefit early-stage SMR equity holders. The consensus is likely underestimating how much data-center operators will pay for reliability, but overestimating the speed at which a novel reactor can translate that willingness into earnings.

Key Decisions for Investors

• Prefer long CEG over OKLO for AI-power exposure over 6-18 months; CEG has nearer-term earnings sensitivity to contracted clean-power demand while avoiding first-of-a-kind reactor risk. Reassess if power-price hedging, forward guidance, or new data-center contracts fail to support incremental EBITDA.
• Use OKLO only as a small, venture-style satellite position or a tactical momentum trade around independently verifiable licensing/customer milestones; avoid treating it as a core AI infrastructure holding until financing needs, fuel sourcing and commercial-operation timing are disclosed. Thesis is falsified by material schedule slippage or equity issuance at a meaningful discount.
• Consider a relative-value pair: long CEG / short OKLO in equal dollar volatility over 3-6 months if OKLO materially outperforms without a binding, creditworthy PPA or regulatory de-risking event. The trade captures likely multiple compression in speculative nuclear versus cash-flow nuclear; stop out on an approved, funded project with contracted economics.
• Maintain NVDA exposure separately from the power thesis. Power constraints are a medium-term cap on aggregate AI deployment, not an immediate determinant of NVDA revenue; watch hyperscaler capex guidance and data-center commissioning delays for evidence that electricity availability is becoming a 2027-28 GPU demand constraint.