HARDWARE

Apple lawsuit threatens OpenAI’s hardware plans

If OpenAI and Apple still appeared to be loyal partners, that illusion evaporated in a July 10 court filing. 

Apple brought a lawsuit against OpenAI on Friday in US District Court, presenting a vast trove of evidence that former Apple employees systematically absconded with trade secrets during and after being hired by OpenAI to work on its hardware devices designed by Jony Ive. 

Ive isn't mentioned in the lawsuit. But at the center of the complaint are Tang Tan, who worked at Apple for 24 years and is now OpenAI's chief hardware engineer, and Chang Liu, who worked at Apple for eight years and joined OpenAI's hardware team in January. 

The legal filing, seen by The Deep View, alleges that these two led an illegal campaign to take confidential files, secret product designs, and insider information about how Apple builds its devices, chooses suppliers, and manufactures products. The suit also alleges that these two coached former Apple employees to bring physical Apple parts to OpenAI interviews and showed them how to circumvent Apple's security protocols to exfiltrate valuable information before leaving the company.

"Our teams are constantly developing breakthrough technologies," an Apple spokesperson said in a statement shared with The Deep View. "Significant evidence has emerged suggesting individuals employed by OpenAI wrongfully took Apple's secret and confidential information regarding our unreleased technologies, processes, and products. We will always defend our teams' hard work and innovations, and we are taking all appropriate steps to do so."

The Deep View also reached out to OpenAI for comment and will update this story if and when there's a response.

In the suit, Apple requested that OpenAI:

• Stop using all Apple trade secrets or disclosing them to others

• Destroy illegally obtained information 

• Return stolen hardware

• Pay damages for the value OpenAI got by skipping years of R&D

• Guarantee future OpenAI products are not based on Apple plans 

It's unclear which devices OpenAI is building, but the reports sound very similar to the kinds of devices Apple is rumored to be considering: smart glasses, a wearable pendant, and earbuds with cameras. 

OpenAI CEO Sam Altman has explained that a different kind of device is needed for the AI era than anything Apple or other device makers have built so far. But he's also never been shy about his admiration for Apple's products.

"I think the iPhone is currently the greatest piece of consumer hardware ever made, by a lot. But it was not meant for a world where you need a piece of hardware that could absorb all of the context of your life," Altman said in a recent interview. "[This is] a very interesting conversation. I would love this to be referenceable by my personal AGI later. But my phone is in my pocket, and it's not going to understand. And I would like a device that, if I want it to, can participate and understand and know about this conversation."

Apple is requesting that the merits of its case be decided in a trial by jury. If Apple were to prevail in the case, it could significantly delay and/or impair OpenAI's hardware ambitions by limiting its options to product directions that Apple hasn't planned to pursue. A successful Apple outcome would put the onus on OpenAI to prove the originality of its hardware and device plans. It's also significant to note that the lawsuit reported that OpenAI has hired over 400 former Apple employees. That's a stunning admission, and perhaps a wake-up call for Apple and its incoming CEO John Ternus. Nevertheless, in the lawsuit document, Apple stated something that appears to cut to the heart of the matter: "That OpenAI now employs people who were once entrusted with Apple’s trade secrets does not entitle OpenAI to use that information to jumpstart its hardware efforts." If OpenAI viewed devices as a key piece of its path to a profitable future, those plans are now in doubt.

TOGETHER WITH IBM

Tokenmaxxing is dead, long live valuemaxxing

Enterprise AI’s first phase in software development was all about tokenmaxxing: driving AI adoption, increasing prompt volume and maximizing usage across coding workflows. But as AI becomes a core part of engineering, many teams are rethinking how they measure success. 

If AI is getting smarter, why focus on how many tokens developers consume? The shift to valuemaxxing puts the emphasis on outcomes instead—accelerating delivery, improving code quality, reducing rework and creating measurable business value.

In this next phase of AI-powered development, impact matters more than activity. The goal is no longer to generate more code, but to help developers create more value with every line they ship.

Read the full article

PRODUCTS

User backlash kills Meta's AI use of your photos

Meta launched Muse Image, its answer to viral AI tools like Google's Nana Banana and ChatGPT's Images 2.0 models, but the new app came with some controversy.

Last week, Meta Superintelligence Labs released its first image-generation model, which promises to produce high-quality images that understand complex prompts and multi-image references. Meta's competitive edge is that users can access these features directly in their favorite social apps, starting with Instagram and WhatsApp, and soon on Facebook and Messenger. However, at launch, this also came with a major caveat for all Instagram users.

A brief note in the release said that users could generate images from public Instagram accounts with a simple @ mention. The idea was that you could create images that were "rooted in your world." Of course, this was immediately met with online backlash, with users outraged that the toggle was automatically turned on or "opted in" without their knowledge, not being told when someone creates an AI image of you, and no way of deleting existing images.

Meta has since pulled the feature, saying in a statement: "Our intent was to provide a useful creative tool and to give people control over whether their public ​content could be referenced in this way."

The fact that the feature was ultimately removed is a compelling example of community backlash producing real, tangible impact—especially on issues as sensitive as personal identity. This underscores the need for greater transparency around what companies are doing, so the public can respond appropriately and, ideally, drive positive change. It also illustrates the need for keeping good habits when posting online. 

"You can reduce the risk. Avoid posting high-resolution close-ups of faces or hands. Strengthen privacy settings so only trusted people can view your content," Raj Ananthanpillai, founder and CEO of Trua, and author of the new book The Trust Crisis told The Deep View.

This issue goes far beyond Meta, which is not new to being under fire for violating user privacy settings. Ever since AI image generation models first popped onto the scene, the biggest question has been about the datasets used to train them. Oftentimes, these models just scrape whatever is available on the open web, which, of course, can include publicly available images of you and copyrighted materials. This opens a broader discussion about whether it is possible to maintain control over your image and likeness in the age of AI. This issue is likely to be exacerbated by the rise of AI wearables, which require cameras to see the world around them and therefore can capture passersby or non-consenting human content. In addition to new laws and regulations that will eventually be needed to govern this area, societal norms and expectations will also need to be established at the level of culture.

TOGETHER WITH CDATA

Secure MCP Architecture for AI Data Connectivity

AI agents need access to enterprise data, but unmanaged MCP servers introduce user and agent permission risks, credential sprawl, and zero audit visibility. CData's 2026 Security Best Practices Guide gives IT and security teams a concrete framework for deploying AI safely. 

Learn how identity-first access, role-and-attribute-based access controls, and comprehensive audit trails operate within Connect AI to keep you in control.

Download the free guide and learn more about security with CData Connect AI.

PODCAST

Why sovereign AI is so urgent in the agentic era

Is it possible for enterprises to build AI agents they can actually trust?

In this episode of The Deep View Conversations, Senior Reporter Sabrina Ortiz sat down with Priya Srinivasan of IBM to discuss Sovereign Core, IBM’s software designed to enable enterprises to deploy AI in a secure, compliant, and sovereign way.

As AI moves from chatbots to agents that can take real action inside organizations, companies need more than speed. They need to know where their data lives, where their models run, who has access, and whether their AI systems comply with internal policies and external regulations.

Srinivasan explains why digital sovereignty is becoming more urgent in the AI era, especially for governments, regulated industries, and enterprises trying to move AI projects from proof of concept to production. She also breaks down how Sovereign Core is designed to bring the control plane, security, access, compliance evidence, and deployment flexibility inside an organization’s own boundaries.

Topics covered include:

• What digital sovereignty means in the age of AI

• Why AI agents raise new questions around governance and trust

• How IBM Sovereign Core helps enterprises deploy AI workloads

• Why compliance can slow AI projects from reaching production

• What "sovereignty with receipts" means

• How companies can balance speed, cost, compliance, and innovation

• Why regulated industries need stronger AI governance from day one

If you’re interested in enterprise AI, agents, governance, compliance, or how major companies are trying to make AI production-ready, this conversation offers a look at what the next stage of trustworthy AI deployment will require.

📺 Watch on YouTube
🎧 Listen in your favorite podcast player

Subscribe to Deep View Conversations for interviews with the leaders shaping the future of AI, business, and technology: tdv.transitor.fm 

LINKS

• 69% of Americans now support AI firms creating a public stock fund

• Anthropic is extending access to Fable 5 for another week, through July 19

• OpenAI and Google sell models to blacklisted Chinese groups, FT report 

• MiniMax is seeking HK$9.5 billion from sale of shares and bonds 

• UK names Microsoft, Google, Amazon, Oracle critical finance suppliers

• Google will now disclose ads made with generative AI tools 

• Microsoft 365 Copilot: GPT-5.6 is now the preferred model in Microsoft 365 Copilot

• GPT-5.6 Sol: OpenAI's new model with reset limits across ChatGPT Work and Codex

• Reve 2.1: The 4K image model got upgrades, including greater prompt understanding

• ChatCut: Agentic video editor got a new plug-in, the Codex app

• Google AI Studio: Rolling out; developers can now get pretty URL

• Fixity Technologies: Artificial Intelligence Engineer

• Morgan Stanley: AI / ML Engineer

• JPMorganChase: AI ML Engineering Analyst

• Mirage: ML Engineer, Agentic Systems

GAMES

A QUICK POLL BEFORE YOU GO