Spending on executive-protection within AI companies is up big. Palantir increased its spending by 150%, and Oracle's 85.5% per cent citing "specific threats and safety concerns" in their disclosures.

This follows incidents such as, a 20-year-old man throwing Molotov cocktails at the San Francisco home of Sam Altman, a man entering Anthropic's lobby with a warning that senior executives were going to be killed, and an Indianapolis councilman's home being struck by 13 gunshots, with a note reading "No Data Centers" left at the scene.

Recently, opponents of data-centre construction planned demonstrations approx. 125 locations across the United States. These protests do not bode well for data centre builds where it is estimated that opponents blocked or delayed roughly 75 projects worth approximately $130 billion in the first quarter of 2026 alone, a figure comparable to the total recorded for all of 2025.

A parallel movement through a viral social-media campaign target the data-collection layer directly. Flock Safety, whose automated license-plate readers operate in more than 12,000 communities, has seen its cameras destroyed across multiple US states in 2026.

The same theme surfaced at the top of the industry with Satya Nadella’s "The Reverse Information Paradox" essay on X warning against allowing AI companies access to your internal intellectual property.

Indeed, sentiment towards AI remains low in the US. An NBC News survey of 1,000 registered voters, conducted from 27 February to 3 March 2026, found that 26 per cent of respondents viewed AI positively against 46 per cent negatively, a net favourability of minus 20. That placed AI below Immigration and Customs Enforcement, and President Trump.

At Traffyk, we may be one of the best placed to provide an assessment of the current landscape because it’s, in large part, a reputation problem. We’re power users of all the frontier models and are constantly experimenting with what’s possible. We can say with confidence that the coding capabilities of models such as Fable, Sol, K3 are already a step change for humanity, yet the majority of the population has little to no exposure to these models let alone the benefits. There is still a question as to whether these capabilities will extend to other types of work and whether the large-scale development of data centres is feasible.

For now, the core message from AI leaders should be that they’re focused on human-centric, sustainable implementations and committed to responsible use of data. For the companies that operate otherwise, regulation should be applied. Hopefully Albanese’s recent announcement for a national approach to AI will address these points directly.