AI kept moving across models, products, and money over the past day. DeepSeek shipped an official, agent-focused version of its V4 model, OpenAI and Anthropic pushed new speed and self-hosting options to developers, and the sector's capital story stayed outsized — a fresh $2 billion raise and a plan to marshal more than $500 billion for AI compute. Here's the rundown →

TOGETHER WITH OUR SPONSOR

Leadership is asking: are we getting value from AI? Which tools are worth the spend? Where are we exposed? Right now, most teams have no idea.

Harmonic Security Usage Explorer changes that.

You get a complete picture of how your organization uses AI, automatically categorized into custom tasks and use cases.

You’ll see the projects being worked on, who’s using what tools, where AI investments are driving value, and where employees are engaging in risky behavior.

CIOs can rationalize spending and cut wasted licenses. CISOs can pinpoint where risk exists and neutralize it. AI committees can show exactly how their efforts are paying off.

See it in action

DeepSeek releases its official V4-Pro model. On August 13, DeepSeek made its V4-Pro model, designated V4-Pro-0813, generally available through its API, app, and website, framing the launch as a push to keep pace with fast-moving domestic rivals as it expands hiring, computing capacity, and fundraising. The company said the model is built for agentic work — using tools, executing code, and completing multi-step tasks — and supports a context window of up to 1 million tokens with outputs as long as 384,000 tokens, in either a "thinking" or "non-thinking" mode. DeepSeek reported agent-focused benchmark scores including 87.9 on Terminal Bench 2.1, and said a price increase for the V4 family, along with new peak and off-peak billing, takes effect August 16. Source

OpenAI previews an "Ultrafast" mode for ChatGPT. OpenAI began previewing an Ultrafast mode built on its GPT-5.6 Sol model that the company says runs at up to 14 times the speed of the standard configuration, according to its release notes. The option is aimed at users who want lower-latency answers for interactive and high-volume tasks, and sits alongside recent updates that added a response-effort slider and broader free access. Source

Anthropic adds self-hosted runners to Claude Code. Anthropic expanded Claude Code with self-hosted environments, introducing a self-hosted runner that lets teams turn their own machines or containers into places where Claude Code web, mobile, and desktop sessions can run; the capability is available on Team and Enterprise plans. The company also began rolling out beta security scanning that checks third-party skills and plugins for malicious content on Enterprise plans, and made Claude Opus 5 the default model in Claude Code. Source

OpenAI-backed Thrive Holdings raises $2 billion. Thrive Holdings, the AI-focused holding company founded by Thrive Capital's Josh Kushner, raised $2 billion at a $12 billion valuation from investors including SoftBank, D1 Capital Partners, and Altimeter Capital, according to TechCrunch. The firm acquires traditional businesses — starting in accounting and IT — and works to embed AI into their operations, and said it will use part of the new capital to expand into physical-asset industries. OpenAI took an ownership stake in Thrive in December 2025 and has assigned staff to help its portfolio companies adopt AI. Source

Nvidia lines up more than $500 billion for AI infrastructure. Nvidia said it is partnering with six financial firms — Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR — to establish financing platforms intended to mobilize more than $500 billion in third-party capital for AI compute infrastructure. Under the arrangement, each firm would independently underwrite funding and direct it to Nvidia customers building data centers, which the company frames as a way to give those customers access to large-scale, long-term financing. Specific commitments and collateral terms were not disclosed. Source

TOGETHER WITH OUR SPONSOR

Creators lock in holiday calendars 90 days out. Structure commissions, recruit creators, and optimize your creator affiliate strategy before the rush with Levanta's 90-Day Holiday Sprint. Get the Free Guide.

That's today's AI Current. Thanks for reading — reply and tell us what you'd like us to cover next. See you tomorrow.

TOGETHER WITH OUR SPONSOR

Opening entities in every country can be slow, expensive, and hard to scale.

That's why more companies are using EOR to hire globally faster.

See how Oyster helps teams hire, pay, and support talent in 180+ countries while staying compliant along the way.

Get all the answers