OpenAI $300B Oracle Deal, Nvidia GPU Pushback

OpenAI is reportedly set to secure a significant cloud computing deal with Oracle, potentially worth $300 billion over five years, starting in 2027. This partnership will involve building data centers with substantial power capacity. Meanwhile, Box is enhancing its content security platform with Box Shield Pro, introducing AI agents for data classification, threat analysis, and ransomware detection, aimed at industries like finance and healthcare. In the pharmaceutical sector, explainable AI (XAI) is becoming critical for building trust and transparency in AI-driven drug development, addressing concerns about bias and 'black box' algorithms. On the hardware front, a Chinese industry official is urging Asian countries to move away from Nvidia GPUs for AI development, advocating for custom ASICs and independent semiconductor ecosystems due to U.S. export controls. Oracle's CTO, Larry Ellison, predicts that AI inferencing will soon be more impactful than AI training, as the company reports strong financial results with significant remaining contract obligations. In the real estate sector, Kaz Nejatian, formerly of Shopify, is taking the helm at Opendoor, an AI-powered platform, with a focus on simplifying home buying and selling. Web development in 2025 is expected to be heavily influenced by AI and security, with tools automating coding and enhancing threat detection. Yext Scout is also helping brands navigate the evolving search landscape driven by AI. Separately, a survey indicates that while many Americans trust AI for medical advice, a notable percentage have followed incorrect guidance, highlighting the need for caution. Purdue University is adapting its curriculum to prepare students for an AI-integrated job market, viewing AI as a tool requiring human oversight.

Key Takeaways

• OpenAI and Oracle are reportedly in talks for a cloud computing deal estimated at $300 billion over five years, beginning in 2027.
• Box has launched Box Shield Pro, an AI-powered security upgrade featuring new AI agents for data classification, threat analysis, and ransomware detection.
• Oracle's CTO, Larry Ellison, anticipates AI inferencing will become more significant than AI training, as Oracle reports $14.9 billion in revenue and $455 billion in remaining contract obligations.
• A Chinese industry official is advising against the use of Nvidia GPUs for AI development in Asia, promoting the creation of custom ASICs and independent semiconductor ecosystems.
• Explainable AI (XAI) is deemed essential for fostering trust and transparency in pharmaceutical R&D, particularly for addressing bias in AI models.
• Kaz Nejatian, former COO of Shopify, is set to become the new CEO of Opendoor, an AI-powered real estate platform.
• Web development in 2025 will be significantly shaped by AI tools for coding assistance and enhanced security measures.
• Yext Scout is introduced as an AI agent to help brands optimize their visibility in the evolving AI-driven search environment.
• A survey reveals that a portion of Americans have received and acted upon incorrect medical advice from AI, underscoring potential risks.
• Purdue University is updating its curriculum to equip students for careers in an AI-integrated workforce, emphasizing AI as a supervised tool.

Box Shield Pro uses AI to boost content security

Box has launched Box Shield Pro, an AI-powered security upgrade to its content protection platform. It includes new AI agents for classifying sensitive data and analyzing threats. The system also detects ransomware activity by monitoring for mass file encryption. Box Shield Pro aims to help businesses like financial firms, healthcare organizations, and law firms protect sensitive information and comply with regulations. These new features will be available as an add-on for existing Box Shield customers.

Box Shield Pro adds AI security for industries

Box announced Box Shield Pro, an enhanced AI security suite built on its existing Shield platform. This new offering includes an AI Classification Agent for automatic content sorting, an AI Threat Analysis Agent for faster security alerts, and ransomware detection. These tools are designed to help industries like law firms, finance, and healthcare manage sensitive data and respond to threats more effectively. Box Shield Pro will be available as an add-on for current customers.

Box launches AI agents for smarter workflows

Box introduced new AI agent solutions to improve how work gets done, including Box Extract for data extraction from various content types and Box Automate for orchestrating work across teams and agents. Box Apps also gained new AI capabilities for building intelligent applications. These tools aim to help businesses automate tasks, gain insights from data, and manage complex business processes more efficiently. IBM is already using Box's data extraction powered by watsonx.

OpenAI and Oracle sign massive cloud deal

OpenAI and Oracle have reportedly agreed to a deal worth $300 billion for computing power over five years. This partnership, starting in 2027, involves building data centers with 4.5 gigawatts of power. Oracle CEO Safra Catz mentioned significant multi-billion dollar contracts in the first quarter, contributing to a 77 percent increase in cloud infrastructure revenue. OpenAI is also rumored to be working with Broadcom on custom AI chips.

Oracle CTO: AI inferencing will surpass AI training

Oracle's CTO Larry Ellison believes AI inferencing, the use of trained AI models for real-time business value, will become more important than AI training. He highlighted this during Oracle's Q1 2026 earnings call, where the company reported $14.9 billion in total revenue and a record $455 billion in remaining contract obligations. Ellison stated that AI inferencing will automate tasks in areas like factories, drug design, and finance. Oracle is investing heavily in AI infrastructure, including a new AI database and support for various AI models.

China Telecom Global uses AI for Belt and Road initiative

China Telecom Global (CTG) showcased its AI capabilities at the 10th Belt and Road Summit in Hong Kong, focusing on empowering the New Silk Road with 'AI for Good'. CTG presented its OneGrowth Global Partnership Initiative, AI global capabilities, and AI-driven applications. Key offerings included the OneTouch AI Engine and LLM Guardrail Platform. CTG highlighted its extensive digital infrastructure, including submarine cables and international Points-of-Presence, to support regional development and digital transformation.

Americans trust AI medical advice despite errors

A survey by Pearl.com found that 22 percent of Americans have followed incorrect medical advice from AI. Despite this, many people are turning to AI for health information, with some trusting AI more than doctors. While AI can offer quick information, experts caution that it lacks the nuanced judgment and personalized understanding of human medical professionals. The report also noted that 19 percent of respondents lost money due to bad AI advice.

Shopify COO Kaz Nejatian moves to lead Opendoor

Shopify's Chief Operating Officer, Kaz Nejatian, is leaving the company on September 12 to become the new CEO of Opendoor, an AI-powered real estate platform. Nejatian joined Shopify in 2019 and was instrumental in developing merchant services. Opendoor's co-founders, Keith Rabois and Eric Wu, will return to the board, with Rabois serving as Chairman. Nejatian expressed excitement about using AI to simplify the home buying and selling process.

Explainable AI needed for pharma R&D trust

Artificial intelligence (AI) is transforming pharmaceutical research and development, but challenges with bias, trust, and transparency remain due to complex 'black box' algorithms. Explainable AI (XAI) is crucial for making AI models understandable and trustworthy in this regulated industry. XAI can help identify bias, justify AI-driven decisions, and personalize medicine by explaining why certain patients might respond better to treatments. Implementing robust data governance, algorithmic auditing, and XAI integration is key for responsible AI use in drug development.

AI and security shape 2025 web development

Web development in 2025 will be shaped by AI, security, and core technologies like HTML, CSS, and JavaScript. Trends include responsive design using CSS container queries and smoother scroll-driven animations. AI tools will automate coding tasks and provide real-time assistance. Security remains a priority, with advancements in JavaScript validation and WebAssembly integration to counter threats. Voice search optimization and sustainable web practices are also key focuses for developers.

Yext Scout helps brands navigate AI search

Yext Scout is an AI search and competitive intelligence agent designed to help brands adapt to changing customer discovery methods, including AI search and agents. Launched earlier this year, Scout provides performance benchmarks against competitors and offers actionable recommendations. Yext will host a webinar on September 24th to discuss AI's impact on search and how Scout can improve brand visibility across local and AI platforms. This tool aims to help marketers optimize content for new AI-driven search channels.

Purdue adapts curriculum for AI in careers

Purdue University is adapting its curriculum and encouraging an optimistic mindset among students regarding AI's impact on future careers. Students and faculty view AI as a tool rather than a threat, emphasizing the need for human oversight in complex fields like engineering and business. The business school has integrated AI training, ethics, and bias education into its core courses. The consensus is that adapting to AI and treating it as a capable intern, requiring supervision, is key to success in the evolving job market.

China urged to ditch Nvidia GPUs for AI

Wei Shaojun, vice president of the China Semiconductor Industry Association, urges China and other Asian countries to stop using Nvidia GPUs for AI development. He warns that relying on U.S. hardware limits regional autonomy and could be detrimental. Wei suggests developing AI-specific ASICs instead of GPUs and highlights China's algorithmic advancements despite hardware restrictions. He believes China is committed to building its semiconductor ecosystem independently, despite U.S. export controls.

Sources

• Next-Gen AI Security: Box Shield Pro Brings Intelligent Agents to Enterprise Content Protection
• Box Unveils AI-Powered Security Suite Targeting Legal, Finance, Government and Other Industries
• Box Announces Next-Generation AI Agents to Drive Intelligent Workflows
• OpenAI reportedly signs $300 billion Project Stargate cloud deal with Oracle
• AI Inferencing Will Outpace AI Training
• China Telecom Global Shines at the 10th Belt and Road Summit, Empowering the New Silk Road via AI for Good
• Over 1 in 5 Americans following AI medical advice later proven wrong—survey
• Shopify COO Kaz Nejatian Exits For New Gig Leading AI-Powered Real Estate Platform Opendoor
• Making sense of AI: bias, trust and transparency in pharma R&D
• 2025 Web Development Trends: AI, Security, and Core Tech Evolutions
• Yext Scout Guides Brands Through AI Search Challenges
• Purdue community adapts curriculum and mindset for AI-integrated careers
• Top China silicon figure calls on country to stop using Nvidia GPUs for AI — says current AI development model could become 'lethal' if not addressed