• Daily Zaps
  • Posts
  • OpenAI 'on track' to release hardware later in 2026

OpenAI 'on track' to release hardware later in 2026

First-ever gigawatt AI training cluster goes live, autonomous coding agents build a 3M-line web browser, OpenAI confirms its first hardware device is on track for 2026, and ads are officially coming to ChatGPT

In partnership with

Welcome back to Daily Zaps, your regularly-scheduled dose of AI news ⚡️ 

Here’s what we got for ya today:

  • First Gigawatt training cluster in the world

  • Coding AI agents build 3M line browser

  • OpenAI 'on track' to release hardware later in 2026

  • OpenAI to finally bring ads to ChatGPT

Let’s get right into it!

DATA CENTERS

First Gigawatt training cluster in the world

xAI has brought the first gigawatt-scale coherent AI training cluster online, consuming more electricity than San Francisco at peak demand and putting it well ahead of competitors that are still planning for 2027.

The pace of execution has been striking, with Colossus 1 going from bare ground to fully operational in just 122 days, and Colossus 2 already surpassing 1 GW with a target of 2 GW with 1.5 GW by April 2026. This reflects Elon Musk’s familiar approach of moving faster than rivals and scaling before others finish planning, making speed and large-scale execution the central strategy.

AI AGENTS

Coding AI agents build 3M line browser

Cursor, led by 25-year-old CEO Michael Truell, coordinated hundreds of GPT-5.2 agents to build a fully new web browser from scratch in one week, generating over 3 million lines of code across thousands of files, including a custom rendering engine, HTML and CSS systems, and a JavaScript virtual machine.

The effort initially failed due to agent indecision, but succeeded after introducing a strict hierarchy of planners, workers, and judges, allowing the agents to work autonomously, debug issues, and continue without human intervention. While the browser can render simple sites and demonstrates that AI agents can coordinate on extremely complex software, Truell admits it only “kind of” works and is far from production-ready, lacking the massive security, extension, accessibility, and maintenance layers found in mature browsers like Chromium.

The project is best seen as a proof of concept rather than a replacement for human engineering, but it highlights the growing potential of long-running, multi-agent AI systems, which Cursor plans to incorporate into its core product.

Better prompts. Better AI output.

AI gets smarter when your input is complete. Wispr Flow helps you think out loud and capture full context by voice, then turns that speech into a clean, structured prompt you can paste into ChatGPT, Claude, or any assistant. No more chopping up thoughts into typed paragraphs. Preserve constraints, examples, edge cases, and tone by speaking them once. The result is faster iteration, more precise outputs, and less time re-prompting. Try Wispr Flow for AI or see a 30-second demo.

GADGETS

OpenAI 'on track' to release hardware in 2026

At Axios House Davos, Chris Lehane said plans for OpenAI’s first artificial intelligence device are firmly “on track” for a 2026 release, emphasizing that the company is moving from experimentation into real-world deployment. Lehane framed the device as part of OpenAI’s broader push toward “practical adoption,” echoing comments from CFO Sarah Friar, who recently said 2026 will be about turning AI into durable, monetizable products rather than just advancing model capabilities.

He noted that OpenAI is focused on building technology people actually use daily, rather than flashy demos, and suggested hardware could provide a more natural, ambient way to interact with AI than phones or laptops alone. The remarks reinforce earlier hints from CEO Sam Altman, who has openly discussed OpenAI’s ambition to create a new category of AI-first devices following the acquisition of a startup founded by former Apple design chief Jony Ive.

BUSINESS

OpenAI to finally bring ads to ChatGPT

OpenAI has confirmed it will begin testing clearly labeled ads in ChatGPT for U.S. users on the Free and Go tiers, while keeping paid plans like Pro, Business, and Enterprise ad-free, framing the move as a necessary trade-off to expand access without charging everyone.

After months of user backlash over what looked like unsolicited product suggestions, the company says ads will be placed separately from responses, tied only to relevant sponsored products or services, dismissible, optional to personalize, excluded for users under 18, and barred from sensitive topics such as health and politics, while insisting answers will never be influenced by advertisers or shared with them.

The announcement follows internal ad experiments and prior comments from CEO Sam Altman signaling openness to ads, and reflects broader pressure on AI companies facing massive operating costs to find sustainable revenue, even as it raises questions about whether ChatGPT can avoid feeling like just another ad-driven platform.

In case you’re interested — we’ve got hundreds of cool AI tools listed over at the Daily Zaps Tool Hub. 

If you have any cool tools to share, feel free to submit them or get in touch with us by replying to this email.

🕸 Tech tidbits from around the web

How much did you enjoy this email?

Login or Subscribe to participate in polls.