PRODUCTS

OpenAI unveils agent builder capabilities

OpenAI is joining the agentic party. At its Dev Day event on Monday, the company debuted AgentKit, a toolkit designed for developers and enterprises to deploy AI agents, furthering the company’s push into the enterprise market. 

OpenAI CEO Sam Altman said in his keynote said the offering is “everything you need to build, deploy and optimize agentic workflows,” aiming to solve the issue of tool fragmentation that creates friction in the development of agents. 

“For all the excitement around agents and all the potential, very few are actually making it into production and into major use,” Altman said.

AgentKit comes with: 

• An agent builder, or a “visual canvas” for creating and iterating multi-agent workflows; 

• A connector registry, which lets admins manage the flow of data and tools between products; 

• And a ChatKit, which lets users embed chatbots into their products.

The package also comes with new capabilities to evaluate agent performance before launch, including prompt optimization, evaluations of third-party models and end-to-end assessments of agentic workflows. 

In its presentation of the AgentKit, the company successfully built an entire workflow, complete with two working agents, on stage in less than 8 minutes. 

With the debut of AgentKit, OpenAI is tapping into an industry trend that executives can’t get enough of: data from IDC, published last week, found that CEOs are broadly bullish on agents, despite the nascent nature of the tech.

Although OpenAI has long held the spotlight in the AI industry, it’s slightly late to the agentic party. Tech firms that have a longstanding footprint in the enterprise market, like Microsoft, ServiceNow, Google and Salesforce, already have their own agentic toolkits on the market. Startups in the space have also started to gain traction, such as n8n reportedly scoring a $2.3 billion valuation in August and Sierra Technologies raking in $350 million in funding at a $10 billion valuation in September.

Still, with OpenAI now boasting a roughly 4 million developer audience, the company’s sheer exposure and name recognition could shoot AgentKit to the top of enterprise competitors. Better late than never.

TOGETHER WITH STRIPE

Monetizing AI products: Nimble experimentation and scale

Rapid AI growth creates a hidden problem: your payments and billing infrastructure becomes the bottleneck to your growth. While some struggle with operational complexity, companies like ElevenLabs and Runway built systems that scale effortlessly with user growth. 

Stripe’s AI scaling playbook unpacks those companies’ blueprint for hyper-growth and eliminating operational drag. These leaders automated billing complexity and designed a pricing strategy that adapts with their business. As a result, teams can focus on building next-generation AI products for users, while their billing and revenue operation largely runs itself. 

Ready to eliminate the scaling bottleneck? Get the playbook right here

HARDWARE

AMD, OpenAI ink deal worth tens of billions

OpenAI is adding another AI infrastructure deal to its roster. On Monday, the AI giant announced it reached a deal to take a 10% stake in chipmaker AMD, a partnership worth tens of billions of dollars. As part of the deal, OpenAI will deploy 6 gigawatts of AMD’s Instinct GPUs, roughly enough to power 4.5 million homes, over several years and multiple generations of hardware focused on inference. The first gigawatt is expected to be operational in the second half of 2026. 

The chipmaker has issued a warrant to OpenAI for up to 160 million shares of its common stock, vested as each gigawatt of chips is deployed. 

You’re not having déjà vu: This deal marks OpenAI’s third major infrastructure partnership in as many weeks.

• As part of project Stargate, the company announced five new U.S. data center sites, bringing the project to nearly 7 gigawatts and $400 billion in investment over the next three years of 10 gigawatts and $500 billion planned. 

• It also announced a $100 billion partnership with Nvidia to develop upwards of 10 gigawatts of AI data centers. 

OpenAI is building up a massive arsenal of AI infrastructure. When all is said and done, between these deals alone, the company will have its hands in roughly 26 gigawatts of compute. Still, every last drop might be necessary: Greg Brockman, OpenAI’s president, told CNBC “If we really want to be able to scale to reach all of humanity, this is what we have to do.” Brockman told Bloomberg that this massive buildout aims to avoid a “compute desert.” 

By inking deals with several different infrastructure firms, OpenAI could be spreading its proverbial eggs among several baskets, something especially important as compute becomes more valuable and the firm toils away on its own custom chips. 

But from this deal, new competitors may emerge: Intel, the most direct competitor to AMD and Nvidia, and the U.S. government, with its 10% stake in the chipmaker.

TOGETHER WITH ACI.DEV

Gate22: Build Trust In Your AI Systems

Open-source MCP gateway and control plane to manage how your team uses AI with your existing systems. Safely automate what you couldn’t trust agents to do before.

• 100% open-source (Apache-2.0) and built for engineering teams.

• Ensure safe interaction between your AI tools and databases / other software.

• Govern which tools agents can use and what they can do—across agentic IDEs, or other agents and AI tools.

Star and use Gate22 now!

ENTERPRISE

Anthropic, Deloitte partner on enterprise AI

Anthropic scored its biggest enterprise win yet. On Monday, the company announced a partnership with consulting giant Deloitte to offer its Claude models to the firm’s 470,000 employees. 

The partnership expands a deal made between the two companies last year, in which Deloitte planned to train 15,000 professionals on Claude. 

Deloitte will establish a “Claude Center of Excellence” as part of the deal, which will include specialists to develop implementation frameworks to scale AI pilots within the company. 

Deloitte’s Chief Strategy and Technology Officer Ranjit Bawa stated in a press release that the partnership reflects the companies’ aligned approach to responsible AI. 

Terms for the partnership weren't disclosed, but Paul Smith, Anthropic’s chief commercial officer, told CNBC that “We are both investing a significant amount in this partnership, whether that’s financial or whether it is just simply the engineering resource that we’re going to put into this as well.” 

The deal signals a broader push into the enterprise market by major AI firms as they seek to monetize their massive models. 

However, enterprises are still largely struggling to derive value from these deployments. An MIT study published in August found that 95% of organizations surveyed have seen no returns on their investments in generative AI, despite $30 to $40 billion being poured into the technology.

LINKS

• OpenAI announces ChatGPT app development a Dev Day

• Google DeepMind unveils CodeMender, an agentic vulnerability detector 

• AstraZeneca signs $555 million deal with Algen for AI-developed therapies 

• Figure robots hit 5 months deployed on BMW production line

• Microsoft purchases 100 megawatts of solar power in Japan

• Sora gives users more control over AI doubles

• Elon Musk needs to spend $18 billion for 300,000 more Nvidia chips

• OpenAI Codex: The company’s Codex assistant is now available, with new admin tools, a software development kit and a Slack integration. 

• ElevenLabs Agent Workflows: A visual editor for designing complex agentic conversation flows 

• TimeFly: A progress activity monitor for developers to track projects, files, languages, interruptions and more

• PromptSignal: A tool for tracking how often AI models surface your content 

• Sorce: The “Tinder for Jobs,” an AI agent that helps candidates find roles. 

• Amazon: Deep Learning Architect, Generative AI Innovation Center

• Charles Schwab: AI Researcher

• Meta: Research Scientist Intern, AI Core Machine Learning

• Anthropic: Performance Engineer

GAMES

POLL RESULTS

What concerns you most about AI investments?

• Economic impact if the bubble bursts (43%)

• FOMO-driven investing in bad ideas (33%)

• Wealth inequality (21%)

• Not concerned (3%)