Altman calls OpenAI's own chip and its descendants the most important single bet for far cheaper, more efficient compute, alongside robots helping build data center racks.
OpenAI's in-house AI accelerator chip, developed to lower the cost of training and running its models; not publicly available.
Updated EU status checked Last shown
OpenAI custom AI chip at a glance
- What is OpenAI custom AI chip?
- OpenAI custom AI chip is a hardware from OpenAI. OpenAI's in-house AI accelerator chip, developed to lower the cost of training and running its models; not publicly available.
- Who makes OpenAI custom AI chip?
- OpenAI custom AI chip is made by OpenAI.
- Does OpenAI custom AI chip work in the EU?
- Unverified. No explicit evidence found. The chip is an internal, non-public hardware project; the EU pages found concern OpenAI's services (GDPR, EU Terms of Use), not this chip. Last verified 14 Nov 2026.
- How much does OpenAI custom AI chip cost?
- Pricing unknown. Not sold as a product; no pricing published.
- Where was OpenAI custom AI chip shown?
- In 1 creator video by Matthew Berman. Most recently by Matthew Berman on 21 Oct 2026, at 30:03 in “Sam Altman: RSI, AI Consciousness, OpenAI's New Hardware, Compute, and more!”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
The OpenAI custom AI chip is a hardware effort by OpenAI to design its own accelerator silicon for running and training its AI models. In a recent interview covered by Matthew Berman, Sam Altman described the chip and its successors as the most important single bet for much cheaper, more efficient compute, alongside robots helping assemble data center racks.
The provided material gives no technical specifications, manufacturing partner details, performance figures, or a public release date. It is an internal infrastructure project rather than a product sold to customers, so there is no pricing, public access, or developer documentation.
Any impact on users would come indirectly, through OpenAI services such as the API and ChatGPT running on this hardware, rather than through direct purchase or deployment of the chip.
- Custom accelerator designed in-house by OpenAI
- Aimed at cheaper, more efficient compute for training and inference
- Described by Altman as a key bet alongside robotics in data center buildout
- Internal use; not sold or offered to external customers
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Similar AI tools
Closest neighbours by meaning, not just by category.
OpenAI inference chip
Hardware · OpenAI
Custom AI inference chip announced by OpenAI and Broadcom, built to serve ChatGPT-style workloads faster and cheaper.
OpenAI Jalapeño Chip
Hardware · OpenAI
OpenAI's custom in-house AI accelerator chip, codenamed Jalapeño, discussed in industry analysis.
OpenAI consumer AI hardware device
Hardware · OpenAI
Unreleased OpenAI consumer AI device; no specifications, price or launch date have been published.
Google TPU
Hardware · Google
Google's custom AI accelerators, offered through Google Cloud for training and inference of machine learning models.
Private Intelligence
App · OpenAI
OpenAI confidential-computing option for running OpenAI models without OpenAI accessing or training on user content.
GPT Red
Model · OpenAI
OpenAI's internal red-teaming model trained via self-play to discover prompt-injection attacks; not publicly released.
