Slashdot: OpenAI Plans To Shift Compute Needs From Microsoft To SoftBank

Feb 21, 2025

—

Source URL: https://slashdot.org/story/25/02/21/2131244/openai-plans-to-shift-compute-needs-from-microsoft-to-softbank?utm_source=rss1.0mainlinkanon&utm_medium=feed
Source: Slashdot
Title: OpenAI Plans To Shift Compute Needs From Microsoft To SoftBank

Feedly Summary:

AI Summary and Description: Yes

Summary: OpenAI is planning a significant shift in its computing strategy, moving its primary resource needs from Microsoft to SoftBank-backed Stargate by 2030. This transition indicates a major transformation in the operational landscape for cloud computing, especially for AI service providers.

Detailed Description: OpenAI’s strategy change encompasses important aspects in the realms of cloud computing, AI infrastructure, and potentially impacts on security and compliance:

– **Transition of Computing Infrastructure**:
– OpenAI will be transferring most of its computing power from Microsoft, which currently supports its operational needs, to Stargate.
– This shift suggests a reevaluation of partnerships and dependency on cloud service providers, which can influence overall security postures and compliance with regulations.

– **Timeline and Financial Projection**:
– The transition is not immediate; OpenAI plans to continue increasing spending on Microsoft data centers in the upcoming years.
– A notable financial forecast indicates a substantial rise in operational costs, estimating a leap from $5 billion in 2024 to a projected $20 billion in 2027.
– This raises questions about the sustainability and financial health of AI infrastructure operations, particularly concerning cost management strategies in the cloud.

– **Implications for AI Model Operations**:
– OpenAI anticipates that costs associated with running AI models (inference) will surpass training expenses, indicating a shift in resource allocation and potential impacts on the efficiency of AI service delivery.

This strategic shift could have several implications for professionals in security and compliance, emphasizing the need for robust risk assessments associated with vendor changes and heightened operational costs in AI infrastructure.

1 2 2024 24 3 4 5 7 a Act AI ai model AI models and anti art as assessment by C CERN CIA Cloud cloud computing cloud service cloud service providers compliance compute Computing computing infrastructure computing power cost cost management Costs Current D data data center data centers de DoT e efficiency end evaluation exp financial for g health http HTTPS implications in Inference Influence infrastructure infrastructure operations J k l land led Link lm man management management strategies media Micro Microsoft model model operations models no non o of on open openai operation operational cost Operational Costs ory out over partnership partnerships planning post potential Power professionals project question R rate RCE real Regulation regulations resource resource allocation Risk Risk Assessment risk assessments Ro s sec security security and compliance security posture security postures service service delivery service providers Sig SoC SoftBank source SSE Strategy sustainability T the Time to Tor TP training transformation transition up US V val Valuation vendor vendor changes Wi x