Zhipu reportedly reached $1 billion in annual recurring revenue in July, roughly four times a March estimate, but the unconfirmed run rate is not annual sales and still sits far ahead of recognized cloud revenue while margins remain thin.
OpenAI issued a nominal $1 million in ChatGPT credits to 10,000 eligible participants who posted publicly about GPT-5.6, gaining reusable marketing content while keeping the incentive restricted to metered product usage.
OpenAI’s “Switch to Codex” campaign filled its 10,000-person allocation by exchanging restricted product credits for public GPT-5.6 posts and permission to reuse them in marketing. The response establishes the promotion’s reach; it does not disclose the credits’ redemption rate, OpenAI’s cost or whether the campaign changed users’ choice of coding agent.
OpenAI says on the campaign page that more than 10,000 people posted and that the first 10,000 eligible participants received $100 in ChatGPT credits. At face value, the grants therefore totaled $1 million.
The qualification rules were designed to produce attributable public material. ChatGPT Go, Plus and Pro subscribers were invited to post on X about what they liked about GPT-5.6 or what they had built, then submit the post URL and the email address associated with their account, as a July 15 account of the launch described.
OpenAI’s own terms were somewhat broader than a request for praise: a participant could publish an honest experience with GPT-5.6 or an honest comparison with other models. Reposts did not qualify, and each subscriber could receive only one grant. A paid subscription was necessary to receive and use the credits, although a person could subscribe after making the post.
In return, participants gave OpenAI permission to repost, share and feature their post and X handle in marketing, including paid advertising, without additional compensation. The campaign page now presents a gallery of posts that OpenAI identifies as part of the final reviewed eligible set.
That makes the collection useful as promotional material, but not as an independent performance comparison. The participants were selected for satisfying campaign rules, not through controlled testing, and the offer rewarded public participation regardless of whether an eligible post was an experience report or a comparison.
A contemporaneous report described each $100 grant as approximately 2,500 credits. It said the balance could not pay for a ChatGPT subscription and identified Codex and ChatGPT for Excel as principal uses.
OpenAI’s service-credit terms put tighter boundaries around that value. Promotional credits are not currency, cannot be redeemed or exchanged for money and are non-transferable. Unless an offer states otherwise, unused credits expire one year after issuance.
The amount of work those credits buy depends on the model and workload. OpenAI’s current Codex rate card prices most customers by token type rather than by message. GPT-5.6 Sol is listed at 125 credits per 1 million input tokens, 12.5 credits per 1 million cached input tokens and 750 credits per 1 million output tokens. Output-heavy work and fast mode consume credits faster; repeated context can be cheaper when input is cached.
OpenAI estimates average Codex spending at roughly $100 to $200 per developer per month, while warning of wide variation by model, the number of instances, automations and fast-mode use. That benchmark gives the grant some scale, but it does not promise a month of service or a fixed number of tasks.
The distinction matters to the campaign’s economics. OpenAI issued $1 million at the credits’ labeled value, but the retained sources do not disclose how much of the pool was redeemed or the marginal cost of serving that usage. Non-transferability, expiry and the paid-account requirement also make the grant less liquid than cash.
The campaign extends a pattern already visible in OpenAI’s team pricing. In an April announcement, the company offered eligible ChatGPT Business workspaces $100 in credits for each new Codex-only member who joined and began using the product, capped at $500 per team. OpenAI later stopped offering new pay-as-you-go Codex seats to Business plans on June 24 while leaving existing seats in place.
The offers are not directly comparable. The Business incentive rewarded adding a Codex-only team member; the July consumer campaign rewarded a public post and required access through a paid individual plan. Both, however, directed the subsidy back into OpenAI usage rather than paying cash.
Competition with Anthropic supplies the immediate context. In May, OpenAI CEO Sam Altman offered companies two months of free Codex usage after Anthropic put third-party coding-tool use behind a new credit meter, according to independent reporting. The July launch account said OpenAI executive Thibault Sottiaux promoted the consumer campaign while quoting an X user’s proposal to give a free month of Codex in exchange for a screenshot of a Claude cancellation screen. That screenshot was not one of the campaign’s published eligibility requirements.
This history supports a narrower strategic conclusion than the campaign’s “Switch” name suggests: OpenAI has repeatedly used temporary credits or free usage to lower the immediate cost of trying Codex. It does not show that recipients abandoned Claude Code or any other alternative after the subsidy ended.
OpenAI has disclosed a completed allocation and a gallery of eligible posts, but not the measures needed to judge conversion. The central questions are how many people subscribed only to qualify, how much of the credit pool they redeemed, whether they continued paying for agentic usage after exhausting it and whether Codex replaced another tool rather than supplementing one.
Those figures would separate an effective acquisition campaign from a short burst of subsidized activity. Independent, like-for-like performance evidence would answer a different question: whether GPT-5.6 earned the favorable claims featured in the gallery when the posting incentive is removed. Until either set of evidence appears, the demonstrated result is limited but concrete—10,000 eligible public posts, $1 million in nominal OpenAI credits and a body of content the company can reuse in its advertising.
Anthropic’s 20-year lease gives TeraWulf a customer for 401 megawatts of future AI infrastructure, but rent starts only after delivery and the proposed utility deal passes power and grid costs to TeraWulf, leaving financing and project margin unresolved.
OpenAI has put conversations and Projects back in its redesigned ChatGPT desktop app and enabled cloud Work threads to move across devices, correcting the launch's biggest usability failures without merging local Work or Codex histories.
Twenty-nine countries signed an agreement creating WAICO as an independent intergovernmental organization, while China paired the launch with capacity-building offers that are not yet confirmed as WAICO programs.
Meta reportedly plans to put departing AWS compute executive Dave Brown to work on its data-center buildout, adding hyperscale operating experience while leaving any customer-facing cloud business conditional and undefined.
Moonshot AI has made Kimi K3 available through its apps and API, pairing a 2.8-trillion-parameter architecture with early frontier-level results, but the model's open-weight claim cannot be tested until its weights and technical report arrive.
CIA Director John Ratcliffe said US intelligence is consistent with an estimate that Russian recruits last 20 to 30 minutes on Ukraine’s battlefield, but the public trail leads to an unsourced claim about assault troops and does not establish a representative average.
China has paired a five-year AI training offer for developing countries with cooperation centers, a weather-warning rollout and a new 29-country organization. The package gives Beijing a platform for influence, but no budget, selection rules or delivery timetable has been published.
Nebius has arranged its first senior secured facility against an operating GPU deployment and one customer’s cash flows. The deal adds project-level debt to its expansion toolkit, but an unnamed customer and a larger rival financing limit what it proves about the rest of Nebius’s backlog.
Huawei publicly displayed a 16-cabinet Atlas 950 configuration rated at 1 EFLOPS in FP8, providing tangible evidence of its system-scale AI strategy while leaving price, power use and sustained workload performance undisclosed ahead of the full system's planned fourth-quarter release.
UK testing places leading open-weight models four to seven months behind selected closed-model cyber results, yet longer attack chains, U.S. benchmarks and mixed cost comparisons show why that interval is a warning signal rather than a universal capability clock.
Moonshot AI's Kimi K3 added to a global technology selloff with near-frontier performance, but unreleased weights, mixed cost comparisons and a recommended 64-accelerator deployment leave its effect on chip demand unresolved.
Two binding EU decisions require Google to give rival AI services comparable access to 11 Android features and offer eligible search competitors a restricted, anonymized dataset, but phased deadlines, certification, pricing and privacy safeguards leave the competitive effect unproven.
Databricks has signed a term sheet for a Coatue-led financing at a $188 billion valuation, while unidentified sources put the round at $3 billion. The proposed capital would deepen its push into AI governance, data agents and operational databases, but the transaction remains open and the company supplied no new operating figures.
OpenAI says a handful of GPT-5.6 Sol file-deletion reports most commonly involved Full Access without sandboxing or Auto-review. Its own evaluations show a more complicated risk picture, and the company has not yet published an incident rate or evidence that its promised safeguards stop the failure.
Netflix says generative AI workflows were used on roughly 300 titles in 2026, but its only quantified example tied to that disclosure covers 17 minutes and does not establish how much finished material, spending or labor changed across the slate.
A rogue browser extension can trigger Claude for Chrome’s fixed Gmail, Docs, Calendar and business workflows; the attack is constrained by default approvals but can run silently for users who enabled unattended action.
Gemini 3.5 Pro missed its expected June rollout. An anonymous-source account says a late-June data change fell short of Google's coding goals, but Google has confirmed only partner testing—not the reported cause, a new date, or public results and pricing.
Gold Eagle has begun collecting and prioritizing AI-discovered software vulnerabilities, but the voluntary federal clearinghouse has not disclosed results, operating rules or the resources that would turn findings into deployed fixes.
Google is bringing Gemini Omni editing and a reusable face-and-voice avatar to Vids, but the sharper distinction is account-level identity across Vids and Gemini rather than a new category of AI video; Vids already offered Veo generation and customizable avatars, while specialist rivals already sell digital presenters.