In partnership with
DAILY TECH BRIEFING // THURSDAY 06.25.2026
Tech Daily
Your daily briefing on the stories that actually matter.
TODAY'S HEADLINE: The company behind ChatGPT just designed its own computer chip, and it has a spicy name.
For years, OpenAI has been one of the biggest buyers of Nvidia's expensive chips, the engines that run ChatGPT. This week it took a big step toward making its own. Alongside chip giant Broadcom, OpenAI unveiled its first custom processor, codenamed Jalapeño, designed specifically to run its AI models faster and cheaper. It is a major move that could reshape who really controls the hardware behind AI. Here is what happened and why it matters.
Your prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.
SECTION 01 // What actually happened
Meet Jalapeño
On Wednesday, OpenAI and Broadcom revealed Jalapeño, OpenAI's first custom-built chip. It is what the industry calls an inference chip, meaning it is built to run already-trained AI models when you type a prompt, rather than to train them in the first place. OpenAI designed the chip from scratch around the specific needs of its own models, while Broadcom handled the silicon engineering and manufacturing.
The companies say they went from initial design to finished blueprint in just nine months, which they claim is one of the fastest chip-development cycles ever for advanced semiconductors. Notably, OpenAI even used its own AI models to help speed up parts of the design. First deployment is targeted for late 2026.
The announcement: https://www.cnbc.com/2026/06/24/openai-and-broadcom-reveal-jalapeno-first-ai-chip-in-partnership.html
SECTION 02 // Why OpenAI built it
Breaking Up With Nvidia
The core reason is cost and control. Running ChatGPT for hundreds of millions of people is enormously expensive, and almost all of that has run on Nvidia's pricey chips. By designing its own processor tuned for exactly its workloads, OpenAI aims to cut those running costs, with one report suggesting roughly 50 percent savings on inference versus standard GPUs.
It is also about not depending on a single supplier. Demand for AI computing has exploded, and Nvidia chips are in short supply and high demand. Building its own silicon, with Broadcom, gives OpenAI another source and more leverage. It mirrors what Google, Amazon, and Meta have already done with their own custom AI chips.
Why it matters: https://techcrunch.com/2026/06/24/openai-unveils-its-first-custom-chip-built-by-broadcom/
SECTION 03 // The reality check
Read the Fine Print
A few caveats keep this grounded. OpenAI has not released hard performance numbers yet, only early claims that Jalapeño's efficiency is "substantially better" than current top chips, with a full technical report promised later. Claims without benchmarks deserve patience. The chip is also purpose-built for inference, so the heavy work of training new models will likely still lean on Nvidia for now.
And this is a starting point, not a finished product at scale. Broadcom's CEO described "small prototype development" late in 2026 before it ramps up. So while the direction is clear, the real test is whether Jalapeño performs in the wild and can be produced in the volumes OpenAI needs.
Tempered view: https://www.tomshardware.com/tech-industry/artificial-intelligence/broadcom-and-openai-unveil-custom-built-jalapeno-inference-processor-openais-first-chip-is-a-massive-reticle-sized-asic-built-in-an-ultra-fast-nine-month-development-cycle
SECTION 04 // Why it matters now
The Race to Own the Stack
This is part of a bigger shift: the most powerful AI companies want to own every layer, from the chip to the chatbot. OpenAI already makes the models, the apps, and increasingly the data centers. Adding custom chips lets it tune the whole system together for speed and cost, the same playbook that made Apple's in-house chips so effective.
For Nvidia, it is another sign that its biggest customers are also becoming competitors. For everyone else, cheaper inference could eventually mean cheaper, faster AI products. The AI arms race is moving down into the silicon, and that is where a lot of the next few years will be decided.
Strategic picture: https://venturebeat.com/infrastructure/openai-unveils-first-custom-ai-inference-chip-jalapeno-with-broadcom-and-its-development-was-sped-up-with-openais-own-models
THE TAKEAWAY
What This Means For You
First, cheaper AI to run can mean cheaper AI to use. If OpenAI cuts its computing costs, those savings can eventually reach you through lower prices or more generous free tiers. The economics of running AI are a big reason these tools cost what they do.
Second, the giants want to own everything. From chips to apps, the biggest AI firms are building the whole stack themselves. That can make their products faster and more polished, but it also concentrates a lot of power in a few hands.
Third, do not over-read a launch. A flashy unveiling with no benchmarks is a promise, not a result. Jalapeño is real and significant, but watch for actual performance data and real-world deployment before assuming it changes everything.
FAQ // Quick answers
Frequently Asked Questions
What is OpenAI's Jalapeño chip?
Jalapeño is OpenAI's first custom-designed computer chip, built in partnership with Broadcom. It is an inference processor, meaning it is optimized to run already-trained AI models, like the ones behind ChatGPT, quickly and efficiently rather than to train new models from scratch.
Why is OpenAI making its own chip?
The main goals are lowering cost and reducing reliance on Nvidia. Running AI models at OpenAI's scale is hugely expensive, and a chip tuned to its exact workloads can cut inference costs, with one estimate around 50 percent. It also gives OpenAI a second source of advanced silicon amid tight supply.
Does this mean OpenAI is dropping Nvidia?
Not in the near term. Jalapeño is built for inference, while the demanding work of training new models is expected to keep relying on Nvidia hardware. OpenAI is adding its own option to the mix, not fully replacing Nvidia.
When will the chip actually be used?
OpenAI and Broadcom are targeting initial deployment by the end of 2026, starting with small-scale prototype work before ramping up. Full performance details are expected in a technical report in the coming months.
How does this compare to Google and Amazon?
Google, Amazon, and Meta have already built their own custom AI accelerators to cut costs and reduce dependence on outside suppliers. OpenAI's move follows that same playbook of designing in-house silicon tuned to its own software.
We will keep tracking this and bring you the next chapter as it lands. Stay sharp out there.
This newsletter is for general information only and is not investment advice. Always do your own research before making financial decisions.
TECH DAILY // www.techdailynews.org
Free Weekly AI Sessions for Experienced Software Engineers.
Every Wednesday at 5 PM CT, Gauntlet AI professors teach a live, hands-on AI engineering session — completely free. If you're nontechnical, this isn't for you. New topic every week, built for engineers who want to build, not just watch. See upcoming sessions.
Free Weekly AI Sessions for Experienced Software Engineers.
Every Wednesday at 5 PM CT, Gauntlet AI professors teach a live, hands-on AI engineering session — completely free. If you're nontechnical, this isn't for you. New topic every week, built for engineers who want to build, not just watch. See upcoming sessions.