Apple’s OpenAI Lawsuit and AI’s Inflation Impact

Compact Conversations for 2026-07-13: 6 AI stories, ai news worth knowing in just 5 minutes.

[Audio embed placeholder]

The Lead: The 6 wildest claims in Apple’s lawsuit against OpenAI

Apple has filed a lawsuit accusing OpenAI of a coordinated scheme to steal trade secrets related to AI hardware. The allegations include OpenAI’s hardware head asking job candidates to bring unreleased Apple components to interviews, a former employee exploiting an authentication flaw to access and download confidential files weeks after leaving, and OpenAI coaching departing Apple employees on how to evade security checks.

Why it matters: This lawsuit highlights the intense competition and potential legal risks in the race to build AI hardware, with major implications for corporate security, intellectual property, and talent recruitment practices in the tech industry.

Source: The Verge

The Feed

Massive AI buildout poses latest inflation threat as consumers pay more for laptops and electricity

The massive investment in AI data centers—projected to top $700 billion this year—is driving up costs for memory chips, processors, and electricity. Economists estimate some memory chip costs have risen by up to 400% since 2024, contributing to price increases for consumer electronics like laptops and game consoles.

Why it matters: AI infrastructure demand is creating new inflationary pressures, affecting hardware costs for businesses and consumers, and could influence broader economic policy and interest rates.

Source: AP News

London-based Valarian, which allows companies to use US cloud providers for AI workloads but retain control of their data, raised a $50M Series A led by NEA

Valarian, a London startup, raised a $50 million Series A for its software layer that lets governments and enterprises use major U.S. cloud providers for AI while maintaining control over data access and flow, addressing concerns related to the U.S. CLOUD Act.

Why it matters: As geopolitical tensions rise, tools for data sovereignty become critical for organizations wanting to leverage U.S. cloud infrastructure without ceding control, especially in defense and sensitive sectors.

Source: Fortune

Sources: Nous Research, the startup behind Hermes, an open-source agent and OpenClaw rival, is raising $75M+ at a $1.5B valuation led by Robot Ventures

Nous Research, maker of the popular open-source Hermes AI agent, is finalizing a funding round of at least $75 million at a $1.5 billion valuation. Hermes, which rivals OpenClaw, allows users to automate tasks and has gained significant traction on GitHub.

Why it matters: The funding underscores the high investor interest and competitive landscape for AI agents that can automate workflows, a key area for developer productivity and enterprise automation.

Source: TechCrunch

Companies turn to Chinese AI models to cut costs

Major companies like DoorDash, Siemens, and Airbnb are exploring Chinese AI models to reduce costs and decrease reliance on U.S. technology, according to a Financial Times report.

Why it matters: This shift reflects both the intense cost pressure of using leading AI models and the growing geopolitical fragmentation of the AI supply chain, which could reshape vendor strategies and procurement.

Source: Financial Times

Anthropic extends free Fable 5 access for subscribers as OpenAI’s GPT-5.6 Sol heats up the pricing war

Anthropic is extending free access to its Claude Fable 5 model for subscribers through July 19th, delaying a planned shift to pay-per-use. The move is seen as a response to competitive pricing pressure, notably from OpenAI’s GPT-5.6 Sol.

Why it matters: The extension signals an intensifying price war among leading model providers, which could benefit developers and enterprises by delaying or reducing cost increases for advanced AI capabilities.

Source: The Decoder

One Thing to Try

Consider testing decommissioned NVIDIA enterprise GPUs like the P100 or V100 for local AI experimentation. A detailed community benchmark found these cards, available for under $200, offer substantial VRAM (e.g., 16GB) for running models, though they require workarounds for end-of-life software and have high power consumption.

Sources

Transcript

Host A: Welcome to Compact Conversations, the show that compresses the day’s AI news into 5 minutes.

Host A: [curious] Today’s lead is a blockbuster lawsuit from Apple accusing OpenAI of stealing trade secrets. According to The Verge, Apple alleges that OpenAI’s hardware head and a couple of former Apple employees ran a coordinated scheme—including asking job candidates to bring unreleased Apple components to interviews for show-and-tell sessions.

Host B: [thoughtful] The lawsuit gets specific. Apple alleges that one former employee accessed Apple’s cloud storage weeks after leaving, using an authentication vulnerability he found. He downloaded dozens of confidential files, and Apple claims he shared them with another employee still at Apple. The company also alleges OpenAI coached departing employees on how to avoid Apple’s security checks, advising them not to sign anything in exit interviews. OpenAI told The Verge they have no interest in other companies’ trade secrets.

Host B: Here’s one number to know: 700 billion dollars in AI data center investment this year alone. The AP reports that buildout is already pushing up prices for memory chips and electricity, with economists at JPMorgan Chase estimating memory chip costs have soared by as much as 400 percent since 2024.

Host A: [with a small lift] Starting with a funding story that touches on geopolitics: Fortune reports London-based Valarian has raised a 50 million dollar Series A. The company builds a software layer that lets governments and enterprises use U.S. cloud providers like Amazon or Microsoft for AI workloads, while Valarian’s system controls what data leaves and who can access it.

Host B: The idea is to address concerns around the U.S. CLOUD Act, which can compel American companies to hand over data. Valarian’s founders argue sovereignty needs to be built into the infrastructure itself, not just a settings toggle.

Host A: On the agent side, TechCrunch reports that Nous Research, the startup behind the popular open-source Hermes agent, is finalizing a new funding round. Sources say the company is raising at least 75 million dollars at a 1.5 billion dollar valuation, led by Robot Ventures. Hermes lets users run AI agents that can automate tasks, and it’s gained a massive following on GitHub with over 200,000 stars.

Host B: [conversational] The Financial Times reports that companies including DoorDash, Siemens, and Airbnb are turning to Chinese AI models to cut costs and reduce reliance on U.S. technology—a shift that reflects both pricing pressure and geopolitical concerns around AI infrastructure.

Host A: And on pricing, Anthropic is extending free access to Claude Fable 5 for subscribers through July 19th. It was supposed to switch to pay-per-use today. The Decoder sees this as a direct response to pricing pressure from OpenAI’s GPT-5.6 Sol and cheaper competing models.

Host B: [thoughtful] One thing to try if you’re building a homelab for AI work is looking at decommissioned enterprise GPUs. A detailed post on the LocalLLaMA subreddit benchmarks cards like the NVIDIA P100 and V100, which you can find for under 200 dollars.

Host A: The author spent a year testing these cards and found that while they have end-of-life software and poor power efficiency, you can work around the limits by compiling older AI software from source. For a lab you power off between tasks, they offer a massive amount of VRAM for the price—like 16 gigabytes on a V100—which is hard to ignore for local model testing.

Host A: That’s Compact Conversations for Monday. More AI news tomorrow. Until then, happy prompting.