It’s widely understood among business leaders today that AI is becoming table stakes. For it to make a difference, there needs to be a strategy for how it will provide an advantage, and then those plans need to be put in place. Cisco Chief Strategy Officer Ammar Maraqa told me the strategy needed for an AI transformation isn’t all that different from adding any new technology or process to an enterprise—but the speed of AI advancement and the acceleration it can provide businesses make this transformation all the more critical. There’s more from my conversation with him later in this newsletter.

We’re taking a summer break, and Forbes CIO will not publish next week, July 16. We’ll be back in your inbox on Thursday, July 23.

This is the published version of Forbes’ CIO newsletter, which offers the latest news for chief innovation officers and other technology-focused leaders. Click here to get it delivered to your inbox every Thursday.

Artificial Intelligence

On Wednesday, SpaceXAI—the rebrand of Elon Musk’s AI company—released Grok 4.5, its most advanced model. This version of Grok was designed for coding, agentic tasks and knowledge work, and the company says it can also handle financial and legal items. It was trained on datasets dealing with code, science, engineering and math, a company blog states. SpaceXAI CEO Elon Musk said on social media that Grok 4.5 is “Opus class,” referring to Anthropic’s frontier model, but faster, more efficient on tokens and less expensive. Grok 4.5 was jointly trained and released with SpaceXAI-owned Cursor.

But Grok 4.5 isn’t the only new frontier model that’s being released this week. OpenAI’s newest models—GPT-5.6 Sol, and its less involved companion Terra and Luna models—will see wide public release today, OpenAI CEO Sam Altman posted on X. This model was first announced by OpenAI late last month, and the company complied with a request from the Trump administration to initially limit its release for security reasons. In a blog post about GPT-5.6 Sol, OpenAI calls it “our strongest model yet,” with advanced capabilities in coding, biology and cybersecurity. The model has two new reasoning levels: max, which allows more time for deep research and reasoning; and ultra, which uses subagents to complete complex work.

The June blog post about GPT-5.6 Sol acknowledges the Trump administration’s current form of AI regulation—an executive order requiring AI companies to voluntarily participate in a 30-day government review process before public release of new models to assess “advanced cyber capabilities”—but OpenAI makes it clear that isn’t ideal. This kind of regulatory stop keeps the “best tools” from being available to users who need them, the post........

© Forbes