A flagship data platform intended to transform the UK’s National Health Service (NHS) is facing scrutiny after a rights group revealed that many hospitals said they hadn’t adopted it ([1]). Medact, a health campaign group, obtained Freedom of Information responses from NHS hospital trusts indicating that 21% of those listed as users of Palantir’s Federated Data Platform reported they were not using it at all ([2]). The £330 million Palantir system is supposed to unify patient data and improve care, but this uptake gap casts doubt on the platform’s real impact. Palantir has touted impressive outcomes from the system, such as helping to deliver 100,000 extra operations and a 12% reduction in discharge delays ([3]), but the UK’s statistics regulator has already warned NHS England to ensure it does not overstate these benefits ([4]). For data leaders, the episode underscores that without genuine adoption and trust in data platforms on the ground, even well-funded AI initiatives will struggle to deliver promised value.
As the need for AI-ready data infrastructure grows, so do public concerns about its environmental footprint. On 4 November 2026, voters across multiple US states approved nearly all of 41 local ballot measures to restrict new data centre projects ([1]). These referendums were driven by community alarm over the massive electricity and water demands of AI supercomputing. One study warned that AI’s water consumption could equal the basic annual needs of 1.3 billion people by 2030 ([2]), highlighting the sustainability challenge. Even tech leaders have taken note of the backlash, Elon Musk quipped that “data centres in space [are] looking pretty appealing right now” in response to the election results ([3]). For organisations planning AI expansions, the message is clear: future data infrastructure projects must address environmental impacts and community concerns or risk delays, higher costs and regulatory hurdles.
On a more positive note, Thomson Reuters is showing how quality data can translate into AI success. In its 3 November earnings report, the company said that licensing its vast news archives to generative AI developers helped drive a 56% jump in Reuters News revenue for the third quarter ([1]). The Reuters news division saw sales climb 12% year-on-year, boosted by a surge in AI training data deals that lifted transaction revenues to $44 million ([2]). One market analyst noted “strong demand for Thomson Reuters’ products, including Westlaw and CoCounsel, with no evidence to date of a slowdown from AI-native entrants” ([3]), a sign that the firm’s 175-year trove of legal, tax and news data is providing a durable competitive edge. The takeaway for other leaders is that investing in unique, well-governed data assets can open up new high-margin AI opportunities, while those without such data may find themselves having to pay for access to someone else’s.
Even as enthusiasm for AI grows, studies show that few organisations have truly AI-ready data. In a global benchmark of 1,900 enterprises, only 3.1% were found to have reached the highest “optimised” level of AI maturity ([1]). Similarly, a survey of senior data and analytics leaders by Precisely and Drexel University revealed that 43% cite data readiness as their most significant barrier to aligning AI projects with business objectives ([2]). Finance chiefs share these concerns: in an EY poll, 61% of CFOs said poor data quality and bias are their top obstacle when seeking investment for new AI initiatives ([3]). The clear message is that while corporate confidence in AI may be high, the foundational work of improving data quality, governance and skills is lagging, and closing that gap is now an executive priority.