Kevin O'Leary says his huge data center will create 10,000 construction jobs. We looked at the numbers.

Wonder Valley’s expected construction employment has been revised down to about 4,000 jobs over 10-15 years, far below Kevin O'Leary’s prior claim of 10,000 construction jobs; projected permanent jobs remain around 2,000, with an USC-based estimate nearer 1,350. The project could still reach 9 gigawatts at full buildout, but it has no tenant yet and final terms are still being negotiated. Local opposition remains strong over water use, air quality, and transparency, while the draft agreement includes up to 30 years of major tax rebates and no minimum job requirement.

Analysis

The market implication is less about one developer’s credibility and more about how aggressively state and local authorities are subsidizing capital-intensive infrastructure with weak labor intensity. A project like this can still be economically large while being politically small on employment, which means the true beneficiaries are likely to be power, electrical equipment, cooling, and grid-interconnect vendors rather than local labor or county tax bases. That also raises the odds that public opposition broadens from a land-use fight into a fiscal one if residents realize the job-creation story is the main political cover for a long-duration tax abatement.

Second-order, the biggest constraint is probably not construction labor but permitting and utility sequencing. If the project is genuinely multi-year and lacks a committed tenant, the valuation bridge from headline megawatts to cash flow is fragile: the capital can be announced quickly, but monetization depends on power availability, interconnect approvals, and AI/cloud demand persisting long enough to absorb the capacity. In that setup, the most exposed parties are adjacent infrastructure contractors and local service providers who may price in a buildout that gets stretched, while the less visible winners are utilities and specialized electrical gear firms that get paid on milestones even if the project is delayed.

Contrarian angle: the downside for the sponsor may actually be limited despite the job-estimate mismatch, because mega-project economics often survive narrative damage as long as there is political support and a path to scarce power. What’s more fragile is the community bargain; once residents anchor on lower job counts and high resource usage, every future delay or revision becomes a credibility event. That makes the next 1-3 months of permitting and agreement finalization more important than the long-run megawatt target, and it increases the probability of headline volatility rather than immediate financial impairment.

For investors, the cleanest expression is to fade the most obvious ‘AI data center boom’ beneficiaries in the local-infrastructure bucket until tenant commitments are public, while staying constructive on pure-play power and grid suppliers with diversified order books. The setup favors a barbell: avoid names that need a smooth Utah-style buildout story to justify multiples, and own companies that get paid whether the project is 4,000 jobs or 10,000. The trade is less about the specific site and more about separating tangible electrical capex from speculative real-estate-and-jobs narratives.

Key Decisions for Investors

• Short the most narrative-sensitive data-center construction / local-infrastructure names on strength over the next 1-4 weeks; the risk/reward improves if permitting headlines turn noisy and the market reprices schedule slippage.
• Long ETN/PHOTONICS-style electrical grid beneficiaries with existing backlogs (e.g., ETN, HUBB, VRT) on a 3-6 month horizon; they capture capex regardless of whether the project hits the advertised employment numbers, with cleaner downside protection from diversified demand.
• Pair trade: long VRT / short a local-regional industrial or contractor proxy most exposed to Utah buildout expectations; thesis is that earnings are more tied to delivered megawatt capacity than to public job promises.
• Avoid initiating longs in utility-adjacent local service names until a tenant is disclosed; if a hyperscaler signs, re-underwrite for a 6-12 month lag before meaningful revenue recognition.
• Optionality trade: buy put spreads on any publicly traded sponsor/proxy that has benefited from the headline narrative if the next agreement update removes job guarantees or extends timelines; the catalyst window is 1-2 months.