CHIPS

OpenAI takes AI chip supply into its own hands

It looks like OpenAI may have finally cracked the code on one of its biggest headaches: getting enough AI chips to build, train and roll out the next versions of ChatGPT.

According to the Financial Times, starting next year, OpenAI will be producing its own chips. The company is partnering with semiconductor developer and supplier Broadcom to develop custom AI hardware tailored to its needs.

OpenAI wasn’t explicitly named in Broadcom’s earnings call last Thursday, but the Financial Times says people familiar with the matter confirmed that OpenAI was the “mystery customer” Broadcom CEO Hock Tan mentioned, committing to a whopping $10 billion in orders.

It’s no surprise OpenAI is heading in this direction. The company has been dealing with serious supply issues with AI chips, which have created bottlenecks, making it harder to scale and keep up with growing demand. 

Earlier this year OpenAI’s Sam Altman posted on X that the company was “out of GPUs” as he announced that GPT-4.5 was ready. He went on to say that even though they wanted to launch it to both Plus and Pro at the same time, with all the company’s growth and GPT shortages, there were going to be delays.

That’s not how any company wants to run, which Altman acknowledged, “this isn't how we want to operate, but it's hard to perfectly predict growth surges that lead to GPU shortages,” he posted.

Having its own in-house chip puts OpenAI in a position where it doesn’t have to rely on others to secure enough hardware to run and scale its AI models, and gives it more stable access to resources. It also assures the situation it found itself in earlier this year doesn’t happen again. 

OpenAI isn’t the first tech company to find a personalized solution when it comes to chip shortages. Google, Apple and Amazon have also gone in that same direction to deal with the ups and downs of the semiconductor market. By developing proprietary chips, they can fine-tune performance for their specific needs, depend less on outside suppliers and maybe even save some money in the long run.

TOGETHER WITH CEREBRAS

Run OpenAI models 20x Faster than on GPUs

Cerebras Inference is the fastest at running the best models -  OpenAI, LLaMA, DeepSeek, Qwen3 and more.

Use cases include:

• Zero lag code generation

• Real time Agents

• Instant Search

• Conversational AI

Try Cerebras Inference today

AI ETHICS

Anthropic settlement raises stakes on ethical data use in AI

Anthropic’s $1.5 billion settlement with authors could change how AI companies handle copyrighted content, setting higher standards for ethical content sourcing.

Anthropic has agreed to potentially one of the biggest copyright payouts in history, a decision that could have far-reaching implications for the AI industry.

The AI company will pay $1.5 billion to resolve a class-action lawsuit from authors who accused it of using pirated copies of their books to train its chatbot Claude.

The proposed deal, announced Friday, is awaiting judicial approval.

If finalized, each of the authors behind the court case would receive $3,000 for the roughly 500,000 works involved, and could mark a turning point in AI companies’ responsibility when it comes to ensuring their systems source content ethically.

The Authors Guild, which backed the lawsuit, welcomed the result, with CEO Mary Rasenberger hailing the settlement calling it “historic.”

“This settlement sends a clear message that AI companies must pay for the books they use just as they pay for the other essential components of their LLMs,” she said. “We expect that the settlement will lead to more licensing that gives authors both compensation and control over the use of their work by AI companies.” 

The case began last year with authors Andrea Bartz, Charles Graeber and Kirk Wallace Johnson, before expanding into a broader fight over how AI models get trained. 

In June, U.S. District Judge William Alsup ruled that while training on copyrighted books isn’t inherently illegal, Anthropic had wrongfully downloaded millions of works from piracy sites like Books3 and Library Genesis.

The deal comes as AI companies face mounting scrutiny over the ethics and development of their chatbots. 

Apple and OpenAI are in the midst of an antitrust lawsuit from Elon Musk’s X over claims the companies are monopolizing the market, while Meta and Character.ai are under investigation in Texas over claims they misled children about their chatbots’ “therapeutic” capabilities.

TOGETHER WITH HEARTSCIENCES

A “bullish” diagnosis from Wall Street

Analyst thinks HeartSciences (Nasdaq: HSCS) could hit $15, but you can invest at just $3.50. Their patented AI tech enhances ECGs to better detect heart disease.

With FDA submission targeted for this year, invest now.

Get $3.50 units that include convertible preferred stock and a warrant.*

AI MODELS

Orson Welles’ lost film gets an AI resurrection

Amazon-backed startup Showrunner is using AI to bring lost Orson Welles footage back to life.

According to the Hollywood Reporter, the company will use its new FILM-1 model to stitch together the missing 43 minutes of “The Magnificent Ambersons,” a follow-up to “Citizen Kane,” which infamously had a third of its footage erased by the studio before its 1942 release. 

Showrunner’s platform will combine AI tools such as face and pose transfer with traditional filmmaking to recreate lost scenes, utilizing archived set photos as the foundation for these recreations. 

Voices will blend AI-generated content with live actor readings.

For the project, Showrunner will collaborate with researcher and filmmaker Brian Rose, who has spent the past several years digitally rebuilding the some 30,000 missing frames from the film using notes, set photographs and production records. 

The final results won’t be commercialized as Showrunner hasn’t received approval from Warner Bros. or Concord. However, CEO Edward Saatchi said it could act as a blueprint for how AI could reshape the film industry moving forward.

Saatchi confirmed the plans in an interview with CNBC last week, describing the film as a “ruined masterpiece,” and saying AI is a way to bring the film “back to life.”

Showrunner started out at Fable Studio as an experimental firm working with AI to generate prompt-based episodes, with the company attracting controversy by creating unauthorized “South Park” episodes. 

With this project, the company showcases its aim to move beyond short-form AI-content generation to feature-length footage. 

Describing itself as “the Netflix of AI,” Showrunner ultimately hopes to give its users the power to generate their own TV episodes from just a few words of a prompt, a vision that could be as disruptive to Hollywood as Welles was in his day.

LINKS

• AI-made animated feature film gets OpenAI backing

• Trump to be joined by OpenAI, Nvidia CEOs on U.K. trip

• Intel makes leadership moves, products chief exits

• Dell’s chief financial officer resigns after 30 years with company

• Microsoft inks $20 billion deal for AI cloud power from Nebius

• Anthropic backs California bill to regulate advanced AI

• Apple hit with AI copyright lawsuit 

• Hackers eye stolen art for AI training

• AI-made animated feature film gets OpenAI backing

• Google’s AI Mode adds five new languages

• Fewer AI Hallucinations Could Speed Business Adoption

• Dell Pro Max GB10: AI developer PC with NVIDIA stack, Blackwell chip, and 1 petaflop of FP4 compute*

• Guidde: AI-powered tool to auto-generate step-by-step video guides, docs, and voiceovers with a free browser extension*

• xpander.ai: A full-stack platform for AI engineers to design, test and deploy agents

• Incerto: An AI-native Copilot for databases that turns natural language into tasks using context-aware agents

• Meeting.ai: AI notetaker for visual learners automatically turns conversations into hand-drawn diagrams and visual summaries. 

• MetaSurvey: Card-style interactive surveys that boost engagement and completion. 

• Genspark: An AI Agent engine where specialized AI agents perform research and generate custom pages

• OpenAI: Account Director, Large Enterprise

• FanDuel: Data Scientist

• Prime Video: Applied Scientist, Personalization and Discovery Science 

• CapitalOne: Lead AI Engineer (GenAI Platform Services, Python)

GAMES

POLL RESULTS

Do you believe the investment numbers being thrown around?

• Yes (19%)

• No (70%

• Other (11%)