ChatGPT Plus users got a taste of fewer restrictions, but OpenAI has now put the clock back on Codex and ChatGPT Work.

OpenAI restored a usage allowance that resets every five hours for ChatGPT Plus subscribers using Codex and ChatGPT Work on Aug. 25, ending a temporary period when only the weekly quota applied.

Thibault “Tibo” Sottiaux, an OpenAI engineering lead working on Codex and ChatGPT, announced the change on X after the company temporarily removed the five-hour restriction in July. The move gave Plus users more freedom to use the tools during individual sessions, but that period has now ended.

Sottiaux said the five-hour window helps OpenAI spread computing demand more evenly while preserving a relatively generous weekly allowance.

He also said some newer and more casual Plus subscribers were unintentionally consuming their entire weekly allowance in a single stretch, leaving them confused when they could no longer use the tools.

What happens when users hit the limit

The five-hour restriction works alongside the weekly quota. Once a Plus subscriber exhausts either allowance, the user must wait for the relevant reset or purchase additional credits to continue using Codex.

OpenAI had temporarily removed the five-hour window in July, while also resetting weekly allowances early on some occasions as Codex and ChatGPT Work reached usage milestones. That gave developers and other heavy users a short period of greater flexibility before the restriction returned.

For now, Sottiaux said the five-hour restriction will remain disabled “for the upcoming months” for users on the plans he identified as the $100 and $200 tiers. Enterprise and Edu accounts use a separate credit-based system and are not covered by the Plus-plan change.

More must-read AI coverage

A less predictable experience for developers

For frequent Codex users, the restored five-hour window reduces flexibility. A developer working through a demanding project could exhaust that window’s allowance even when weekly usage remains available.

But there is a practical reason for the restriction. AI coding tools can consume significant computing resources, and allowing users to concentrate large amounts of usage into short periods can make demand harder to manage. Spreading that usage across five-hour windows gives OpenAI more control over its infrastructure while preserving a larger weekly pool.

Plus subscribers working on demanding projects now need to monitor both their five-hour and weekly allowances. Before beginning a long coding session, users should check their remaining capacity and plan for a reset or additional credit purchase if either allowance is running low. The change gives OpenAI more control over computing demand, but it also makes usage less predictable for developers who rely on Codex throughout the workday.

Read more: OpenAI’s Codex Windows app brings its AI coding workspace to more developers, with tools for managing multiple coding tasks from a dedicated desktop interface.

Amazon and Nvidia just got a lot closer. The two companies announced Wednesday an expanded partnership that includes a deal to add another 2 million Nvidia GPU chips to Amazon’s data centers.

These GPUs, which are designed to handle the heavy compute demands of training and running AI models, include Nvidia Blackwell Ultra, Rubin, and Rubin Ultra GPUs. The chips will head to Amazon Web Services’ data centers in 2027 and 2028. 

The announcement, made during Nvidia’s quarterly earnings call, comes just five months after Amazon agreed to deploy more than 1 million Nvidia GPUs across AWS infrastructure starting this year. Nvidia said in a statement that since then, “demand has exceeded those expectations.”

Neither company shared financial terms. It’s unclear what the exact return will be for Nvidia. But considering GPU units costs, the deal is worth tens of billions of dollars.

The announcement is notable not just for its size and the speed in which it grew, but also because it extends beyond Amazon buying more Nvidia chips. And it’s happening even as Amazon invests in its own potentially competing AI chips. 

Nvidia said Wednesday that its technology, including the networking hardware that connects thousands of GPUs into one system, as well as its open models, CPUs, data processing software, and robotics platform, will also be integrated across AWS. 

The companies said “surging demand” from startups, enterprises, AI labs, and even governments influenced the decision to work more closely. 

The expanded partnership comes as Amazon ramps up its own AI chip efforts — particularly with CPUs, which are the general purpose processors at the heart of servers. 

Amazon has been building its own chips to lessen its dependence on Nvidia and even compete with the chip giant. Amazon’s AI chief Peter DeSantis has said that AWS is in talks to sell its Trainium chips — which are a direct alternative to Nvidia’s H100 or Blackwell chips for deep learning workloads — to other companies for use in data centers. Amazon’s Arm-built Graviton CPU is also seen as a challenger to traditional server chips from Intel and AMD. 

Amazon has said its custom chip business is growing, noting on its last earnings call that it crossed a $25 billion annualized revenue run rate, driven by $225 billion in total commitments from AI labs like Anthropic and OpenAI. 

But, it seems Nvidia is still the GOAT in the world of AI chips. 

With the 2 million GPU chips Amazon is adding to AWS starting in the third quarter, Nvidia also plans to send an unspecified number of Vera CPUs, “some integrated with Rubin, others standalone,” according to Nvidia CFO Colette Kress. 

Nvidia CEO Jensen Huang has big plans for the company’s Vera CPUs, boasting back in May that he had found a “brand new $200 billion TAM” for the company.

Aside from AWS, Kress said Wednesday that Nvidia expects Vera to be deployed by “every major hyperscaler, neocloud, AI lab, and system OEM, with shipments already underway to our lead partners,” which include Oracle and SpaceX AI.

The partnership is also extending to Amazon’s warehouse robots and enterprise offerings. 

Kress said Amazon plans to adopt Nvidia’s full physical AI stack to power its fleet of robots. The stack includes Omniverse (its simulation and digital twin platform); Cosmos (its world model platform); Isaac (its robotics development platform); and Jetson (computing hardware for robots and edge AI). This week, Nvidia also introduced a new version of Jetson designed as a more accessible robotics computer for “entry-level edge AI.”

On the enterprise side, AWS will serve Nvidia’s Nemotron family of open models on Amazon Bedrock, its managed foundation model platform, and SageMaker, its managed cloud service. 

Nvidia also reported Wednesday that it recorded sales of $96.2 billion for the second quarter, beating analyst estimates. Data center revenue made up the majority of Nvidia’s sales for the quarter at $89 billion, up 117% from a year ago. 

Nvidia said it expects revenue to reach $108 billion in the third quarter, some of which will come from its next-gen Rubin GPUs. Nvidia said it began production shipments this quarter. Investors have been looking out for Rubin’s initial Q3 sales for signs that demand will continue into Nvidia’s next generation of hardware. 

Nvidia has committed $279 billion to secure supply and manufacturing capacity for current and future data-center projects, up substantially from $119 billion last quarter, as the chipmaker looks to secure memory and manufacturing capacity to meet AI demand over the next few years. That commitment includes $92 billion in projected spending for the rest of the fiscal year and another $87 billion in fiscal year 2028. 

“The thing that matters for the industry is that AI is now doing productive and useful work,” Huang said during Wednesday’s call. “AI is generating profitable tokens…If we had more compute, we could generate more profitable tokens, which results in more profit for all of the services. This is the exact phase where we’re at, which is the reason why everybody’s leaning in.”

Investors will be watching to see if additional compute indeed translates so neatly into additional profits as AI companies pour hundreds of billions of dollars into infrastructure. 

Google is putting Gemini to work in the legal trenches, giving law firms AI agents designed to review contracts, track regulations and tackle other time-consuming legal tasks.

Google Cloud on Tuesday launched Gemini Enterprise for Legal, a purpose-built version of its enterprise AI platform aimed at law firms and corporate legal departments.

Available in preview, the platform combines specialized AI skills, connections to legal software, third-party agents and centralized governance controls. Google said the goal is to move beyond chatbots that simply answer questions and toward agents that can carry out legal workflows.

“General-purpose AI, however capable, does not meet that standard on its own. Foundational model intelligence is necessary. For legal work, it is nowhere near sufficient,” Google Cloud CEO Thomas Kurian wrote in the company’s announcement.

The platform can assist with contract review and redlining, legal research, regulatory monitoring, data subject access requests, document redaction and NDA drafting. It can also turn older agreements into reusable contracting playbooks.

Google plugs into existing legal systems

Rather than asking firms to replace their existing technology, Google is connecting Gemini Enterprise for Legal to systems already used by legal teams.

The company lists integrations with iManage, NetDocuments, Docusign, Everlaw, RelativityOne, Thomson Reuters HighQ, CourtListener, Harvey, Legora and other platforms. Google said its secure Model Context Protocol connectors inherit existing permissions and access controls.

That approach could be important for firms that have already invested heavily in specialist legal software. The platform is also supported by technology and consulting partners including Accenture, Deloitte and KPMG.

Google said client data, firm-specific playbooks, intellectual property, custom agents and model outputs remain private to the organization and are never used to train or fine-tune its foundation models.

More Google coverage

The bottom line

Google’s most interesting move may be its decision to sit above existing legal technology rather than trying to replace it.

That could make Gemini Enterprise for Legal easier for large firms to adopt, particularly where replacing document management, research or litigation systems would be costly and disruptive. But it also leaves Google competing for a place in a market where specialist vendors already understand legal workflows deeply.

There is another hurdle: AI-agent governance remains an unresolved enterprise challenge when mistakes can affect clients, court filings or regulatory obligations. Google’s emphasis on verifiable grounding, traceable citations, data isolation and permission controls shows that trust is central to its product pitch, not an optional extra.

Google’s announcement does not include pricing, making it difficult to assess the platform’s cost or accessibility beyond the firms participating in its early rollout. For legal IT teams, the immediate task is to evaluate whether its permission controls, citations and audit trails meet their confidentiality and compliance requirements before allowing agents to handle client work.

Read more: Google Cloud Next’s five biggest AI takeaways show how Gemini Enterprise fits into the company’s broader plan for governed, enterprise-scale agents.