SambaNova Brand Guidelines: The Stanford “New Dance” in AI Inference Chips

A Stanford project named for a new dance.

SambaNova was founded in 2017. English Wikipedia dates its founding to November 2017 and names three co-founders: Kunle Olukotun, Christopher Ré and Rodrigo Liang. The company tells the story the same way. Its about page says the idea came from two Stanford professors working on the same problem from different sides — Olukotun on chip multiprocessor designs, Ré on the machine-learning systems that ran on them — and that Liang, a semiconductor executive, completed the founding team. It says it started operations in Palo Alto; its press releases now give its headquarters as San Jose, California. Liang is chief executive and Olukotun chief technologist.

The name is part of the story. The company says “SambaNova” is Portuguese for “new dance,” chosen to signal a new way of designing AI chips. Wikipedia adds that the technical foundation came from research on processor arrays for machine learning that had received funding from DARPA, the US defence research agency.

Early backers came from the chip and internet industries. Wikipedia records an investment from Intel and Alphabet in 2019, a further round in 2020, and a 2021 round led by SoftBank Vision Fund 2, with Temasek and GIC among the other investors.

The Reconfigurable Dataflow Unit.

SambaNova’s chips are called Reconfigurable Dataflow Units, or RDUs. The company says it designs the RDU accelerators, builds the racks and systems they run in and writes the compiler and orchestration software that drives them, then sells the whole stack as a platform. Its product names follow the brand: SambaCloud for its cloud service, SambaStack and SambaRack for its systems.

The SN40L was introduced in 2023, according to Wikipedia. SiliconANGLE reported in February 2026 that it is made on Taiwan Semiconductor Manufacturing Co.’s five-nanometer process, with its cores split across two silicon dies mounted on an interposer built with TSMC’s CoWoS packaging, alongside high-bandwidth memory, and that the company ships it in an air-cooled system holding 16 accelerators. Its successor, the SN50, was unveiled in February 2026; SambaNova said SoftBank Corp. would be the first customer to deploy it, in its AI data centres in Japan.

The software side has its own milestones. Wikipedia says Samba-1, a bundle of 56 generative AI models, was announced in February 2024, that Time named the SambaNova Suite to its Best Inventions list in 2024, and that the company was included in the Forbes AI 50 in 2026. Research centres have used its systems: Wikipedia names the AI Testbed at the US Department of Energy’s Argonne National Laboratory and Japan’s RIKEN Center for Computational Science, where SambaNova systems run alongside the Fugaku supercomputer.

Recent partners listed in the company’s newsroom include TEPCO Systems in Japan in April 2026, Oracle Cloud Infrastructure in October 2026 and Supermicro, whose inference racks the company said in November 2026 would use SN50 chips. In July 2026 SambaNova said JPMorganChase had chosen it as an inference infrastructure partner for on-premises AI. SiliconANGLE reported that a managed offering, SambaManaged, introduced in July 2025, was followed within six months by hardware deals with four data-centre operators, one of them supplying chips for a planned sovereign cloud campus in Scotland.

How AI chip companies present themselves.

AI chip companies sell something most of their customers will never see. The processor sits inside a server, the server sits in a rack, and the rack sits in a data hall far from the people who use the models running on it. So the industry’s visual language has grown up around making that hidden hardware visible: a whole silicon wafer, as wide as a dinner plate, held up on a keynote stage; a single chip pinched between finger and thumb for the cameras; a die shot recoloured in neon false colour so that its blocks of memory and compute read like a city map at night.

The second set of images is the data centre. Press photographs tend to show long rows of racks in dark halls, lit by the glow of status lights and cold aisles, with cabling arranged as neatly as a diagram. These pictures stand in for scale, which is what buyers of AI compute are being asked to imagine, and they appear again and again in launch announcements, partner news and investor material.

Keynote slides do much of the remaining work. Model names, token counts and benchmark charts appear beside the chip’s own product name, and the companies’ identities are usually kept spare to let those numbers and images lead: black or white grounds, a single accent colour, a sans-serif wordmark and very little ornament. Several firms have adopted coined product categories, names for their processors that set them apart from the graphics processor.

The shop window is the technical conference. Hot Chips, NeurIPS and the AI Infra Summit are among the events where architectures are presented, demonstrations run and booths compete for engineers’ attention. This guide looks at how each company in the group presents itself, and keeps to what can be checked about who founded it, who owns it and what it makes.

From takeover talks to an Intel partnership.

The most closely watched chapter in SambaNova’s recent history is its relationship with Intel. In December 2025, SiliconANGLE recalled, reports emerged that Intel had offered to acquire the company. Those talks stalled. Instead, on 24 February 2026, SambaNova announced a new funding round led by the private-equity firm Vista Equity Partners and Cambium Capital, with Intel’s investment arm, Intel Capital, taking part, together with a multi-year agreement with Intel to deliver AI inference products, Reuters reported. Intel’s chief executive, Lip-Bu Tan, also serves as SambaNova’s executive chairman, according to the same report. In April 2026 the two companies published a joint blueprint pairing GPUs, SambaNova RDUs and Intel Xeon 6 processors for different stages of AI inference.

The company has continued to raise money since. On 8 July 2026 it announced the first close of a Series F round led by General Atlantic, with Seligman Ventures and T. Rowe Price among the investors and Vista, Cambium and Intel Capital returning. It said the money would expand capacity and scale deployments for enterprises, “neo-clouds” and sovereign AI customers.

Its description of itself has widened. Wikipedia notes that in 2021 the company remained focused on AI chip design while beginning to offer “dataflow-as-a-service”; its 2026 press releases call it a provider of “next-generation AI infrastructure” with a full-stack platform for AI inference, and Liang’s statements stress inference as the centre of enterprise AI. Olukotun, in a quotation on the company’s about page, frames the project as a rethink of the foundation “from the silicon up.”

For a design guide, SambaNova shows how a research spin-out can carry an academic origin story, a memorable name and a family of product names into a full infrastructure brand. Compare Intel, its investor and partner, and TSMC, which makes the SN40L. For this guide we use a deep slate, ember orange and warm sand with a sans, an unofficial interpretation; no product or logo is reproduced.

Keep SambaNova’s identity on hand — save it to a board in the Made Good inspiration library.