Texas Freezes New Funding for Flock Cameras After Spending $30 Million

01Texas Governor Freezes New Funding for Flock Cameras After State Spent More Than $30 Million

Texas Governor Greg Abbott has frozen further state spending on Flock’s AI surveillance cameras, closing the funding tap for additional purchases after mounting scrutiny of how the systems are used. The action does not amount to a statewide ban, and the source does not say that existing cameras or contracts must be withdrawn.

The freeze came just before the publication of a Texas Tribune investigation that found Texas had already spent more than $30 million on Flock cameras. The money was raised primarily through a $1 fee added to insurance policies, purportedly to support efforts against catalytic converter theft.

Flock’s camera system has attracted opposition across the political spectrum amid concerns about privacy, access and transparency. At least six incidents in Texas have involved officers being placed on leave or criminally charged for misusing Flock systems. The source does not detail each case, but the number shows that the backlash extends beyond hypothetical fears about surveillance.

Other concerns include Flock’s data-sharing practices, the placement of some cameras near playgrounds, access by the company’s employees to video feeds, and video feeds that were left unsecured. The company has also faced scrutiny over its attempts to resist transparency, adding to questions about who can see footage and how oversight works.

The state-level funding freeze follows action by local governments. Cities around the country have canceled Flock contracts in recent months, including multiple cities in Texas. Abbott’s move therefore advances the response from scattered municipal contract cancellations to a restriction on further state purchases.

The political pressure has come from both parties. Gina Hinojosa, Abbott’s Democratic opponent in the governor’s race, has criticized him for expanding the cameras’ use in Texas. Republican Representative Keith Self has separately denounced Flock cameras as part of what he described as a growing surveillance state and invoked Fourth Amendment protections.

The immediate effect is to block more state spending on Flock cameras, limiting new deployments that would depend on that funding. Important boundaries remain unknown: the source does not specify the freeze’s legal form, duration, covered programs, or treatment of existing contracts. It also does not say whether local governments can continue buying cameras with their own funds.

02SpaceX Builds Gas-Turbine Blade Foundry in Texas; Musk Claims It Could Cut Deployment Times by Up to 18 Months

SpaceX is building a foundry in Bastrop, Texas, to cast blades and vanes for natural-gas turbines, Elon Musk confirmed on Saturday. The project targets a manufacturing bottleneck that is slowing new power generation as AI companies seek enough electricity to run expanding data centers. Musk said bringing the process in-house could get turbines online up to 18 months sooner, but that claim has not been independently verified and the foundry is not known to be operating yet.

The project emerged through job listings mentioning a “blades and vanes foundry” and research showing that SpaceX bought roughly 830 acres near its existing Bastrop Starlink factory between March and June. The source does not establish the foundry’s planned capacity, production date, yield or where its components would ultimately be deployed.

Turbine blades are unusually difficult to manufacture. Components in the hottest section operate at roughly 3,000 to 3,600 degrees Fahrenheit—about 800 degrees above the melting point of their metal alloy—using internal cooling channels and thermal-barrier coatings. Each blade must be grown slowly in a vacuum furnace as a single crystal, without microscopic seams that could crack under stress. Only four companies reportedly produce such blades at industrial scale, and their capacity is fully booked.

That constraint matters because data-center electricity use is projected by the International Energy Agency to roughly double by 2030. GE Vernova says its turbine production capacity is essentially sold out through that year, while hyperscalers including Amazon, Google, Meta, OpenAI and Microsoft are building private gas plants rather than waiting for grid connections. If SpaceX succeeds, it would control a capability on which other AI infrastructure developers depend on a small supplier group.

Faster turbine deployment would also intensify pollution concerns. In Memphis, the NAACP has repeatedly accused SpaceXAI of operating turbines without federally required permits or pollution controls; a limited University of Memphis analysis found local air pollution became “slightly worse.” Separately, a Virginia study using an Environmental Protection Agency model estimated that eight continuously operating turbines at one facility could cause 3.4 to 6.5 additional premature deaths annually and $53 million to $99 million in yearly health damages.

03Caterpillar Brings Mining Automation Experience to Jobsite AI and Plans $100 Million in Worker Training Over Five Years

Caterpillar is taking automation lessons learned in mines into more dynamic jobsites, quarries and construction sites, while preparing its workforce for the resulting changes. The equipment maker began its autonomy push in mining, where labor shortages and hazardous conditions gave automated machinery a clear purpose. It now sells autonomous haul trucks, drills, underground loaders and dozers, alongside remote-control equipment, command centers and fleet-management software.

The expansion is not simply about moving machines from one setting to another. Caterpillar CTO Jaime Mineart said the difficult part is integrating autonomous systems and “physical AI”—AI embodied in machinery—into customer workflows. That means redesigning how sites operate and how people work alongside equipment.

One example already entering frontline work is Cat AI Assistant. Customers, operators and technicians can use voice commands while standing beside a machine to retrieve repair procedures, troubleshoot potential faults and identify parts they may need before starting a repair. The assistant draws on Caterpillar’s proprietary information, including data from about 1.6 million connected assets worldwide and more than 16 petabytes of structured data, according to Mineart.

Caterpillar is applying AI elsewhere in industrial operations. It uses the technology to scan sites and create manufacturing digital twins, or virtual representations used to analyze physical operations. The company also deploys AI agents to modernize legacy code, generate and test software, and detect defects earlier.

Experienced operators are part of the deployment process rather than merely its end users. Caterpillar relies on their knowledge, accumulated over decades, to help train AI systems. As machinery becomes more autonomous, some operators could move from driving one machine to supervising several from a remote command center.

That shift is behind a planned $100 million investment over five years to train Caterpillar’s 118,000 employees in AI, automation and robotics. The company has not disclosed the courses, participation targets or assessment measures, nor has it provided adoption, efficiency or safety data for Cat AI Assistant.