Apple Sues OpenAI for Trade Secret Theft as Meta Muscles Up on AI
Apple OpenAI trade secrets lawsuit: Apple alleges hardware technology theft
The Apple OpenAI trade secrets dispute just escalated into a sweeping new lawsuit. Apple accuses the company of a sustained effort to obtain confidential Apple hardware technology, according to Techstrong.ai. OpenAI is expanding into consumer hardware, and Apple says that push relied on stolen know-how. Apple filed the suit in the US District Court for the Northern District of California. It names OpenAI and its hardware unit io Products. Two individuals are also named: Tang Tan, OpenAI’s Chief Hardware Officer and a 24-year Apple veteran, and Chang Liu, a former OpenAI technical staff member.
According to the filing, Tang Tan emailed supplier information and internal industry materials to himself before leaving Apple. He also allegedly told Apple employees interviewing with OpenAI to bring physical hardware components and prototypes to their interviews. Separately, Apple accuses Chang Liu of accessing its internal systems after his employment ended. Chang exploited an authentication flaw, Apple says, to obtain confidential engineering files. Apple raised concerns with OpenAI in February but says it received no response. The company now claims more than 400 former Apple employees work at OpenAI. “This is the tip of the iceberg,” Apple alleges in the filing.
However, OpenAI spokesperson Drew Pusateri rejected the allegations. He said OpenAI has no interest in other companies’ trade secrets and remains focused on building technology that benefits users. Still, the lawsuit marks a sharp reversal from 2024, when Apple and OpenAI partnered to integrate ChatGPT into Apple’s operating systems. That relationship has since deteriorated: Apple’s latest Siri rollout relies on Google’s Gemini technology instead. Meanwhile, OpenAI acquired Jony Ive’s hardware startup io Products last year for roughly $6.5 billion. Ive himself is not named in the suit.
Meta flexes new AI muscle: Muse Spark 1.1 targets OpenAI and Anthropic’s developer market
Meta rolled out Muse Spark 1.1 this week. Techstrong.ai reports it’s the company’s most aggressive proprietary push yet into coding and autonomous agentic workflows. The model can write and debug code. It can also manage multi-step digital workflows and orchestrate tasks across third-party applications with minimal human intervention. Meta priced it at $1.25 per million input tokens and $4.25 per million output tokens. That positions it below premium models like Anthropic’s Claude Sonnet, but above lower-tier options like OpenAI’s GPT-5 mini. Overall, the launch fits Meta’s broader shift away from its open-source Llama family toward a proprietary, pay-as-you-go developer ecosystem.
Meta’s “Watermelon” model reportedly closes the gap with GPT-5.5, as OpenAI trade secrets fight rages
Meta’s Superintelligence Chief Alexandr Wang made a bold claim at a recent internal town hall. He told employees that the company’s forthcoming model, codenamed Watermelon, has reached performance parity with OpenAI’s GPT-5.5, according to Techstrong.ai. That claim rests on internal benchmarks only. Watermelon reportedly requires “an order of magnitude more compute” than Muse Spark, which is internally codenamed Avocado. Meta hasn’t disclosed the specific benchmarks behind the parity claim.
For context, Mark Zuckerberg has already acknowledged leadership was overly optimistic about the pace of agentic development. So even if Watermelon matches GPT-5.5, the bar keeps moving. OpenAI has already debuted GPT-5.6, though the company still restricts that model from public release at the request of the U.S. government.
Meta eyes an AI cloud business to monetize its infrastructure buildout
Meta is preparing to turn its AI infrastructure into a commercial cloud business, per Techstrong.ai. The plan could offer both API access to Meta-hosted models and direct rental of raw GPU capacity. An internal division called Meta Compute is developing the initiative. As a result, Meta would compete directly with AWS, Microsoft, Google, CoreWeave and Nebius. Meta has forecast capital expenditures of $125 billion to $145 billion in 2026 for AI data centers and next-gen GPUs. Analysts say monetizing that excess capacity could meaningfully boost revenue and cash flow.
Meta’s Muse Image tool draws privacy backlash
Meta is also facing backlash over Muse Image. The free AI photo tool is available on WhatsApp, Instagram Stories and the Meta AI app, and it lets users manipulate other people’s public Instagram photos without their explicit knowledge or consent, Techstrong.ai reports. Meta opts public accounts in by default and won’t notify affected users. Meta automatically protects private accounts, however. Foxglove’s Donald Campbell called it an “obvious recipe for disaster.” Similarly, Privacy International said the feature shows AI companies “see people’s images and data as raw material to be exploited.”
SAP avoids EU antitrust fine with a decade of support policy changes
Finally, SAP has agreed to overhaul its on-premises software maintenance and licensing policies. The move follows an agreement with the European Commission that ends a 2025 antitrust investigation, according to Techstrong.it. The commitments stay in force for 10 years. They let customers divide SAP environments into separate commercial installations, opt out of SAP support for specific systems, and use third-party maintenance providers. SAP will also expand single-metric licensing contracts. Plus, it’s adding flexibility for companies that need to reduce unused licenses during workforce reductions, divestitures or bankruptcy. SAP’s cloud offerings, including S/4HANA Cloud and RISE with SAP, remain unaffected.
Watch the full episode
Hosts: Mike Vizard, Alan Shimel
Guests: Stephen Foskett, Elizabeth Safran, Yvette Schmitter
Catch this and every episode on the Techstrong Gang show page, where the Apple OpenAI trade secrets lawsuit and the rest of today’s AI news get broken down live.
Read More:
- Apple Sues OpenAI, Alleging Theft of Hardware Trade Secrets
- Meta Challenges Rivals with New Proprietary Muse Spark 1.1 AI Model
- Meta’s Upcoming ‘Watermelon’ AI Model Draws Even with OpenAI’s GPT-5.5: Report
- Meta Is Developing AI Cloud Business to Monetize Infrastructure Investment
- Meta Faces Backlash Over AI Tool That Alters Users’ Photos Without Explicit Consent
- SAP Avoids EU Antitrust Fine with Global Changes to Customer Support Policies
