The artificial intelligence boom is facing a new challenge as political resistance, infrastructure concerns and investor anxiety begin to reshape the industry’s future.
What once appeared to be an unstoppable race to build bigger data centres is increasingly becoming a debate about electricity, water, financial returns and the communities expected to accommodate these facilities.
The latest political shock came from Texas, where Democrat Gina Hinojosa defeated Republican incumbent Greg Abbott in the gubernatorial election.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Her victory represents a significant setback for the state’s rapidly expanding data centre industry, which has benefited from abundant land, relatively affordable energy and business-friendly policies.
Hinojosa campaigned on introducing a statewide moratorium on new data centre approvals until lawmakers establish stronger protections for residents.
She argued that communities should have a meaningful say in deciding whether these enormous facilities are built in their neighbourhoods. Her position reflects growing frustration among Texans worried about rising electricity bills, water consumption and pressure on the state’s power grid.
The scale of the challenge is substantial. Texas has attracted major infrastructure investments from technology giants including Microsoft, Google and Meta.
However, the state’s grid operator, ERCOT, has been tracking hundreds of gigawatts of proposed projects seeking electricity connections. The demand has raised questions about whether existing infrastructure can support the industry’s expansion without increasing costs for ordinary consumers.
Even Abbott, who previously welcomed technology investment, had introduced restrictions on new data centre approvals and environmental permits following mounting public opposition.
Hinojosa’s proposed moratorium could take that intervention further, creating uncertainty for developers with billions of dollars committed to projects across the state.
The election also delivered an uncomfortable message to Elon Musk, whose political spending helped make him the Republican Party’s biggest individual donor.
Musk reportedly spent about $445 million backing GOP candidates during the midterms, yet many of his preferred candidates lost. After Hinojosa’s victory, he posted that “data centers in space” were looking appealing, linking his political disappointment to an increasingly serious business idea.
Launch costs, radiation, thermal management and the difficulty of repairing equipment in orbit create formidable obstacles. Meanwhile, doubts about AI investment are spreading beyond Texas.
In China, a collapse in AI and semiconductor shares has erased hundreds of billions of dollars in market value as investors question demand and technology-export prospects globally.
On Wall Street, technology earnings and enormous capital expenditure plans have intensified scrutiny of whether spending will generate profits quickly enough. Investors are distinguishing between companies with durable revenue and those relying on borrowing or optimistic assumptions.
That shift does not mean artificial intelligence is finished. Demand for computing power remains significant, and major technology companies continue to invest.
But the market is moving from rewarding promises to demanding evidence: stronger cash flows, sustainable financing, reliable power supplies and measurable returns.
For investors, the lesson is clear. Safety may lie less in chasing every AI-linked stock and more in assessing balance sheets, diversification and businesses capable of earning money through the cycle.
Texas’s political revolt and Musk’s orbital quip expose the same reality: the AI revolution is not just a software race. It is a costly physical build-out, and communities, regulators and capital markets will determine how far it can go.