Amazon Spends $5B Early on Nvidia AI Servers

Key Takeaways

• Amazon is accepting about $5 billion in Nvidia GB300 server racks early through a new Early Rack Receipt program
• The company plans to add roughly $3 billion in AI hardware monthly
• Some AWS data centers lack power capacity, so completed servers are being stored in warehouses
• Amazon projects $220 billion in capital expenditures this year and has entered negative free cash flow
• Google is bringing Gemini Personal Intelligence to cars via Android Auto, pulling from Gmail, Calendar, YouTube, Maps, and Search
• US and China agreed to set up a formal AI dialogue and hotline, with the next meeting planned for November
• Only 38 percent of Americans see more benefits than drawbacks in AI, compared with 85 percent in China
• Two Pennsylvania lawmakers introduced frontier AI safety bills (HB 2800 and HB 2829) following the OpenAI and Hugging Face attacks
• The bills require developers to test and validate AI models before deployment and disclose model capabilities and limits
• Cisco's president used personal AI agents to review $9,000 in yearly subscriptions, showing growing use of AI in personal finance

Amazon buys billions in Nvidia AI hardware before power is ready

Amazon is taking delivery of about $5 billion in Nvidia GB300 server racks early through a new Early Rack Receipt program. It expects to add roughly $3 billion in hardware each month. Finished equipment is being stored in warehouses because some AWS data centers lack the power capacity to run the servers. Amazon projects about $220 billion in capital expenditures this year, which has contributed to negative free cash flow. The company says early staging is a standard business practice to speed up future installation.

Amazon speeds up AI hardware buys despite data center power limits

Amazon is accelerating its purchase of AI hardware even though some AWS data centers face power delays. Internal plans show an Early Rack Receipt procedure to acquire around $5 billion in Nvidia GB300 server racks before deployment, with about $3 billion more expected monthly. Completed equipment is being temporarily stored because some sites cannot yet support the servers. Amazon has a Price-to-Sales ratio of 3.75 and a GF Score of 93 out of 100, reflecting strong growth expectations despite cash flow challenges.

Amazon stockpiles Nvidia AI hardware ahead of power availability

Amazon is changing how it buys hardware by agreeing to purchase billions of dollars in Nvidia graphics processors before its data centers have enough power to run them. Internal emails show a new Early Rack Receipt program to accept roughly $5 billion in Nvidia GB300 server racks early, with plans to take in about $3 billion more each month. The strategy ties up large amounts of capital in equipment that cannot yet earn revenue. Amazon projects around $220 billion in capital expenditures this year and has moved into negative free cash flow for the first time since 2023.

Gemini Personal Intelligence comes to cars

Google is adding Personal Intelligence to Gemini AI in cars, starting with Android Auto and later expanding to Google Built-in vehicles. The feature lets Gemini use data from services like Gmail, Google Calendar, YouTube, Maps, and Search to give more relevant answers. It was previously limited to paid users but is now available to all Gemini users. Drivers can choose to turn the feature off or select which apps to connect. Google says the goal is to help Gemini better understand context and provide more helpful responses.

Google adds personal Gemini AI features to cars

Google is upgrading Gemini AI in vehicles with Personal Intelligence, which can pull from Gmail, Google Calendar, YouTube, Maps, Search history, and other Google services. The feature will launch in Android Auto today and will expand to cars with Google Built-in in the coming months. Drivers must opt in for the feature to work. For example, a driver heading to the airport may be warned about a delayed flight, and Gemini can answer follow-up questions in natural language. Users can turn the feature off and choose which apps to connect.

US and China must deepen AI cooperation for safety

US-China AI talks opened a path for cooperation after a summit in Washington. Both presidents agreed to set up a formal AI dialogue and hotline. The next meeting is planned for November. American tech companies are spending heavily on chips and data centers while revenue models remain uncertain. Only 38 percent of Americans see more benefits than drawbacks in AI, compared with 85 percent in China.

Cisco president used AI agents to review subscriptions

Cisco's president said he used personal AI agents to screen $9,000 worth of yearly subscriptions. The AI tools helped him review and manage his spending on subscriptions. This shows how individuals are using AI agents for everyday financial decisions. The report was shared by Business Insider. It highlights growing use of AI assistants in personal finance.

FundGuard whitepaper explains AI architecture for investment accounting

FundGuard released a whitepaper on how investment firms can adopt AI effectively. It says legacy batch systems limit AI use in investment accounting. The paper covers real-time data, cloud-native design and governed AI for agentic workflows. It also discusses Model Context Protocol and Agent-to-Agent standards. The guide is aimed at asset managers, servicers and owners evaluating AI readiness.

IQVIA expands AI tools for clinical development

IQVIA Holdings expanded its use of Veeva systems across clinical development. It launched IQVIA Explorer and Life Science Models to support faster trial and commercial decisions. About 294 AI agents are already in production and 19 of the top 20 pharma customers use the tools. The company has a US$34.2 billion backlog. Analysts expect revenue of US$20.4 billion and earnings of US$2.1 billion by 2029.

Canadian DIY investors use AI and finfluencers for investing

A survey of Canadian DIY investors found 84 percent watch finfluencer content or use AI for investing. More than half use AI in their investment process. Among finfluencer viewers, 68 percent also use AI. Finfluencer content led 83 percent of viewers to take action like learning finance or starting to invest. The survey included 834 Canadian DIY investors with at least $10,000 in investable assets.

Artificial Intelligence

The article discusses artificial intelligence and website technologies like cookies. These tools enable site functionality, analytics, personalization, and targeted advertising. Users can learn more by viewing the provided link. The content also mentions permissions for editing collections.

Pennsylvania House introduces two AI safety bills

Two Pennsylvania lawmakers introduced frontier AI model safety bills in Harrisburg. The bills are HB 2800 and HB 2829. They came after the OpenAI and Hugging Face attacks. The bills require developers to test and validate AI models before deployment. They also require clear information about model capabilities and limits. These are the first such bills introduced after those attacks.

Sources

• Amazon stockpiles Nvidia AI hardware ahead of power availability -Business Insider
• Amazon (AMZN) Accelerates AI Hardware Acquisition Amid Data Center Power Constraints
• Amazon stockpiles Nvidia AI hardware ahead of power availability -Business Insider
• Gemini's Personal Intelligence is coming to cars
• Google is bringing a more personal version of its Gemini AI assistant to your car
• Opinion | US-China AI talks must ensure safety of both tech and business
• Cisco president used AI agents to filter $9,000 worth of subscriptions
• The AI Divide in Investment Accounting: Why Architecture Matters
• How New Clinical AI Tools At IQVIA Holdings (IQV) Have Changed Its Investment Story
• More than 80% of DIY investors watch finfluencers or use AI
• Artificial Intelligence
• Bill guide: Pennsylvania House considers two frontier AI safety bills, HB 2800 and HB 2829 — Transparency Coalition. Legislation for Transparency in AI Now.