AI adoption gains offset by rising security risks

That tension is becoming harder to ignore after researchers uncovered about 900,000 manipulated AI skills that can create fresh vulnerabilities for AI agents. The finding underscores a broader problem for investors and operators alike: as businesses embed AI into workflows, they are also expanding the number of ways systems can be tampered with, misled or breached.

The market backdrop reflects both enthusiasm and caution. Nvidia, the key supplier to much of the AI buildout, closed at $212.50 on July 15, up from $203.53 two days earlier and well above its 50-day moving average of $209.50, while its RSI reading of 62.4 points to renewed momentum. Microsoft, a major AI platform player, finished at $395.63, still below its 50-day moving average of $401.88, suggesting investors remain selective even as the AI trade stays central to broader equity positioning.

Adalytica’s proprietary sentiment gauge shows investors are still constructive on Nvidia, with AI-related sentiment at 67 and awareness at 44, both neutral but improving over the past week. By contrast, Microsoft’s earnings sentiment has cooled to 33, reflecting a more uneven read-through for software and cloud names tied to AI spending.

For investors, the key question is no longer whether AI adoption continues, but which companies and countries can convert it into durable earnings without exposing themselves to security, compliance and model-risk setbacks. That makes policy moves like Kemkomdigi’s more than a tech headline: they are part of the competitive infrastructure race shaping who captures AI’s productivity upside.

The next catalyst is execution, not rhetoric. Markets will be watching whether governments tighten AI standards fast enough to reduce misuse while still encouraging deployment, and whether the next wave of AI products can prove they are both commercially useful and secure.