OpenAI Launches Jalapeño Chip and Secures Major HP Deal

OpenAI is scaling both its physical and digital footprint.

On the hardware front, OpenAI has unveiled Jalapeño, its very first custom AI chip, developed in record time alongside Broadcom and Celestica to push the boundaries of raw compute efficiency. 

Meanwhile, on the software front, the company is embedding itself into the bedrock of global business, securing tech giant HP as one of the first major enterprise customers for OpenAI Frontier, its new semantic operating system for autonomous AI agents. 

Unveiling Jalapeño: OpenAI’s first chip that was created in just nine months

Semiconductor manufacturing company Broadcom presented OpenAI’s first AI chip, named Jalapeño, last week. 

Jalapeño was delivered to OpenAI CEO Sam Altman and President Greg Brockman by Broadcom President and CEO Hock Tan and President Charlie Kawwas.

OpenAI designed the chip around its understanding of LLM fundamentals with partners Broadcom and Celestica contributing towards the chip implementation, board, rack system integration, networking and production systems.

It took just nine months to develop, thanks to support from Broadcom, Celestica and OpenAI’s own models to speed up the design process. 

“The same models served to users are helping improve the infrastructure used to run future models,” reads the OpenAI press release. “If AI can help engineers design better chips faster, it can lower the cost of compute across the industry and help democratize access to advanced AI.”

Under the hood of Jalapeño 

According to OpenAI, Jalapeño is designed with the flexibility to work across all LLMs, guided by their own insights into the inference needs of current and future AI models across the industry. 

“We optimised the architecture around the kernels, memory movement, networking, and serving patterns that matter most for frontier AI models,” says Richard Ho, who leads OpenAI’s hardware programme.

“Based on early testing, Jalapeño will efficiently execute our most important workloads close to the hardware’s theoretical limits.” 

Engineering samples of the Jalapeño chip are running ML workloads in the lab at production target frequency and power, including GPT‑5.3‑Codex‑Spark. 

“The world is moving to a compute-powered economy,” adds Greg Brockman, President and Co-Founder of OpenAI. 

“Jalapeño is part of our long-term full-stack infrastructure strategy to make compute more abundant, resulting in AI which is faster, more reliable, more affordable for people and businesses and can be used to solve more important problems. 

“By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access.” 

HP uses OpenAI Frontier to support across chat, voice and retail 

HP is using OpenAI Frontier to scale its customer support and digital operations, making it one of the first global enterprises to adopt the platform.

The goal is to use specialised AI agents to unify its fragmented commercial touchpoints into a single network. 

“With the use of Frontier platform, HP is planning to build a more consistent experience across store, partner, chat and voice experiences, giving customers and partners faster ways to get answers, complete routine workflows and move toward resolution,” says Prakash Arunkundrum, Chief Strategy and Transformation Officer at HP.

“With OpenAI there is an opportunity to fundamentally rethink how AI can deliver better outcomes.

What is OpenAI Frontier?

OpenAI Frontier was launched by OpenAI in February to function as a fleet-management platform designed to build, deploy, manage and monitor teams of AI agents across large organisations.

It was created to bridge what OpenAI calls the “AI opportunity gap”, which is the disconnect between highly capable AI models and the messy, fragmented data systems found inside massive corporations.

HP’s journey to using the Frontier tool

HP began using OpenAI Frontier when the tool was launched in February to explore how it could help assess its technical capabilities.

It also trialled several use cases through pilots of agentic capabilities, platform components, security and enterprise integrations. 

Ultimately, the firm’s goal is to deploy AI-driven solutions across customer- and partner-facing solutions and experiences, customer telemetry insights, employee productivity and software development. 

“HP is showing what enterprise transformation looks like when AI becomes an operating layer - connected to the systems and workflows where work already happens,” says Denise Dresser, Chief Revenue Officer at OpenAI. 

“HP has been an exceptional early partner, turning early value from OpenAI APIs and tools like ChatGPT and Codex into repeatable systems. 

“We’re thrilled to go deeper with them as they move beyond Frontier pilots to deliver measurable business impact at scale.”  

According to HP, AI is becoming a vital new layer for how work gets done across the company. 

It believes OpenAI Frontier provides this layer with the context, governance and execution capacity needed to move from early wins to true, enterprise-wide transformation. 

This, alongside its upcoming suite of agentic AI Devices that integrate into existing workflows, will establish OpenAI not just as a model developer, but as the foundational infrastructure for the autonomous enterprise.