Oracle teams with Cerebras and expands AI cloud infrastructure
Oracle has unveiled a slate of major AI partnerships at Oracle AI World, including a collaboration with Cerebras Systems to offer high-speed inference on Oracle Cloud Infrastructure in 2027.
Oracle announced a major expansion of its cloud computing hardware at Oracle AI World in Las Vegas, teaming up with Cerebras Systems to bring its high-speed AI inference systems to Oracle Cloud Infrastructure in 2027. According to Bloomberg, Oracle will deliver cloud services running on Cerebras computers directly inside its data centers, securing another major client for the chipmaker as it competes against Nvidia's dominance in the AI hardware market.[1][4][6][7]
The planned Cerebras Inference service will be powered by the company's Wafer-Scale Engine. Emphasizing the growing demand for rapid computation, Cerebras Chief Executive Officer Andrew Feldman noted that every step an AI agent takes requires inference, making speed crucial for complex workloads. Oracle highlighted that AI inference demands hardware choice, and Crypto Briefing reported that the integration could significantly increase processing speed and bolster enterprise capabilities. Amid the announcement, shares of Cerebras Systems traded up 7.87% to $116.90.[1][3][5]
The Cerebras agreement forms part of a broader infrastructure push designed to widen customer hardware options. Oracle is expanding its Nvidia Vera Rubin Superclusters to multi-gigawatt scale, with Vera Rubin NVL72 systems and Vera CPUs now available. Starting in 2027, Oracle will also host Google's custom TPU 8t training and TPU 8i inference chips, becoming the first outside hyperscaler to provide them. Additional announcements include the addition of SambaNova inference processors, an expanded OpenAI alliance across four enterprise areas, and native availability of Snowflake's AI Data Cloud.[2][7]
Key facts
- Oracle will offer cloud services based on Cerebras computer systems powered by the Wafer-Scale Engine in its data centers starting in 2027.
- Cerebras Systems stock traded up 7.87% at $116.90 following the announcement.
- Oracle is expanding its Nvidia Vera Rubin Superclusters to multi-gigawatt scale, with Vera Rubin NVL72 systems and Vera CPUs currently available.
- Oracle will be the first hyperscaler outside Google to offer Google's custom TPU 8t and TPU 8i chips, slated for 2027.
- Oracle added SambaNova AI inference processors to its cloud platform.
- Oracle expanded its OpenAI partnership to four enterprise offerings and made Snowflake's AI Data Cloud available natively as a first-party service.
Sources · 7 sources
- AG
AGTP@AGTPinsightsPost on X ·
BREAKING: Oracle and Cerebras will bring Cerebras fast AI inference to Oracle Cloud Infrastructure, powered by the Wafer-Scale Engine. Cerebras Inference on OCI is planned for availability in 2027. https://t.co/qEfqn0QFeh
Open source - AI
Analytics Insight@analyticsinmePost on X ·
𝗢𝗿𝗮𝗰𝗹𝗲 𝗘𝘅𝗽𝗮𝗻𝗱𝘀 𝗔𝗜 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗪𝗶𝘁𝗵 𝗡𝗩𝗜𝗗𝗜𝗔, 𝗚𝗼𝗼𝗴𝗹𝗲 𝗮𝗻𝗱 𝗖𝗲𝗿𝗲𝗯𝗿𝗮𝘀 Oracle is expanding its AI chip strategy with a larger NVIDIA deal and plans to add Google TPUs, Cerebras and SambaNova processors to its cloud platform. The move gives customers more hardware choices as demand for AI computing grows. #Oracle #NVIDIA #GoogleTPU #Cerebras #ArtificialIntelligence #AIChips #CloudComputing #TechNews #AnalyticsInsight #AnalyticsInsightMagazine Read More 👇 https://t.co/vW57QZNM1w
Open source - CB
Crypto BriefingArticle ·
Oracle brings Cerebras inference chips into its AI cloud Oracle's integration of Cerebras chips into its AI cloud could significantly enhance AI processing speed, impacting enterprise AI capabilities. The post Oracle brings Cerebras inference chips into its AI cloud appeared first on Crypto Briefing .
Open source - EV
Evan@StockMKTNewzPost on X ·
Oracle $ORCL will offer cloud services based on Cerebras Systems’s $CBRS computers in its data centers - Bloomberg https://t.co/HserM9Vzsn
Open source - SB
Shay Boloor@StockSavvyShayPost on X ·
$CBRS and $ORCL are teaming up to bring Cerebras AI inference to Oracle Cloud in 2027. As Cerebras CEO Andrew Feldman says “every step an agent makes requires inference” which means speed matters even more as agents take on longer and more complex tasks. And that's what makes Oracle saying “AI inference demands choice” so interesting because Cerebras doesn't need to replace $NVDA or $AMD to win a meaningful share of AI workloads where speed matters most.
Open source - BL
Bloomberg@businessPost on X ·
Oracle will offer cloud services based on Cerebras’s computers in its data centers, adding another large customer for the fledgling challenger to Nvidia’s dominance of AI hardware https://t.co/K44Y8IePCV
Open source - WS
Wall St Engine@wallstenginePost on X ·
Oracle $ORCL announced several AI partnerships and infrastructure updates at Oracle AI World in Las Vegas today. Here's a breakdown of the major announcements: - $NVDA: Expanding NVIDIA Vera Rubin Superclusters to multi-gigawatt scale. Vera Rubin NVL72 systems and Vera CPUs are now available on Oracle Cloud. - $GOOGL: Bringing Google's custom TPU chips to Oracle Cloud starting in 2027, making Oracle the first hyperscaler outside Google to offer them. This includes TPU 8t for training and TPU 8i for inference. - $CBRS: Partnering with Cerebras to bring high-speed AI inference systems to Oracle Cloud in 2027. - SambaNova: Adding another specialized AI inference chip option to Oracle's cloud infrastructure. - OpenAI: Expanding their partnership with four enterprise offerings that integrate ChatGPT and Codex into business workflows, cybersecurity, application development, and small-business operations. - $SNOW: Bringing Snowflake's AI Data Cloud to Oracle Cloud as a native, first-party service, with customers able to purchase and manage it directly through Oracle.
Open source