SpaceX, AT&T, Apple steer into AI's currents—Grok 4.6, open models, DC lobbying

Key takeaway

• Three major US technology companies are strategically adapting to the accelerating AI wave: SpaceX released Grok 4.6, which benchmarks competitively against leading models from Anthropic and OpenAI, backed by hundreds of billions in infrastructure investment; AT&T is routing 45 billion daily AI tokens across a portfolio of open and closed models, with open-source models expected to power 70–80% of its workload over time and already cutting costs by 80–90% in some applications; and Apple hired a veteran government affairs leader to manage escalating issues around China, chips, and tariffs ahead of a September Xi-Trump meeting.

• The common theme is operators reading the direction of the AI current and steering into it rather than fighting it.

3 Key Points

FAQ

What is Grok 4.6 and how does it compare?

Grok 4.6 is SpaceX's new AI model that focuses on long-running agents and ambitious interactive and visual work. It benchmarks right up against Anthropic's latest Fable and Mythos models and OpenAI's GPT 5.6, with OpenAI's Astra next up for comparison.

How much is AT&T spending on AI tokens and where does the cost saving come from?

AT&T runs an average of 45 billion AI tokens every day across more than a thousand internal uses. Switching from closed to open models has already saved 80 to 90% in certain applications, and open models are expected to power 70 to 80% of its AI workload over time.

Why is Apple restructuring its government affairs team now?

The issues ahead—China, chip sourcing, supply chains, tariffs, and hundreds of billions in US-versus-overseas investment—are getting thornier by the day. The restructuring also aligns with John Ternus becoming CEO on September 1, Tim Cook moving to executive chairman while keeping government relationships, and a major Xi-Trump meeting at the White House in September.