One Subscription for Grok, Cursor, and X

X Bundles Grok, Cursor, and Premium for $30

Grok, Cursor, and X Premium+ are now being offered together for $30 a month, down from their previously combined $90 price.

Announced on October 12, The X Subscription combines AI chat, coding tools, AI agents, and X features under one plan. Prices start at $8 per month, with higher tiers reaching $200 and offering more AI usage across Grok and Cursor.

That's a big price difference. And it's worth watching.

I think the real advantage here is convenience, especially for people already paying separately for coding tools and AI assistants. A shared usage pool across Grok and Cursor could also make the subscription more attractive to developers who regularly switch between chat and coding tasks.

Existing subscribers can keep their plans, while new customers can sign up immediately, with broader availability expected over the next few days.

The bigger question is whether cheaper bundled access will convince users to move away from the AI tools they already prefer.

Microsoft Wants More AI Output Per Megawatt

Microsoft plans to make its own Maia chips the primary hardware for running AI across its products and services starting in 2027.

Announced on October 12, the company confirmed that Maia 200 is already powering production workloads, including Copilot, while Maia 300 is scheduled for deployment next year. Microsoft expects the new system to generate 40% more AI tokens per megawatt than leading GPU-based systems, with 96 accelerators packed into a single rack.

Power is becoming a constraint. And Microsoft knows it.

I think the important detail is Microsoft's growing ability to control the hardware costs behind products like Copilot, rather than relying entirely on external GPU suppliers. With AI usage expanding across Microsoft 365, GitHub, and Azure, better efficiency could make a meaningful difference to how much the company spends serving millions of users.

Microsoft also plans multi-gigawatt Maia deployments, suggesting this is becoming a major part of its infrastructure strategy.

The question is whether Maia 300 can deliver those efficiency gains at production scale.

NVIDIA and Cisco Want More Bandwidth From Every Switch

NVIDIA and Cisco are betting that a new optical connector can help AI data centers expand without adding more network switches.

On October 12, the companies announced plans for Ultra-Dense Small Form Factor Pluggable (USFP), a new optical module designed to support 400Gb/s per lane. They claim it can deliver 4x the bandwidth density and 1.3x the power efficiency of current 200Gb/s-per-lane modules, while fitting into a similarly sized slot on a switch.

The size matters. So does power.

AI data centers are connecting thousands of processors, and moving data between them is becoming increasingly expensive as networks expand. I think the interesting part is that NVIDIA and Cisco are targeting the physical connections between systems, where even small efficiency improvements could reduce infrastructure costs at scale.

They're also working with TE Connectivity and Amphenol on the hardware, with plans for an open industry specification.

The question is how quickly manufacturers can turn these designs into products that data center operators actually deploy.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.