Nothing's $200M Bet: An AI OS

Here is what’s new in the AI world.

‘Nothing’ Secures $200M to Launch AI-Native OS for Future Devices

Based in London, consumer electronics company Nothing has secured a substantial round of funding and revealed plans to introduce a new lineup of devices equipped with native artificial intelligence, powered by an operating system that represents a clear departure from conventional platforms.

In a public statement, Nothing's founder and CEO, Carl Pei, announced that the company raised $200 million, bringing its valuation to around $1.3 billion.

This capital infusion is intended to support broader market distribution and faster product innovation.

A key element of Nothing’s future vision is an AI-centric operating system. According to Pei, the company is creating the “foundations of the future” through a purpose-built “AI OS” intended to offer a highly personalized user experience.

He clarified that it remains uncertain whether this new system will be based on Android, like the current ‘Nothing’ OS, promising further details at a later date.

Instead, ‘Nothing’ is developing what it describes as an “AI-native platform,” designed to be compatible with a wide range of devices, from smartphones, earphones, and wearables to more advanced products like smart glasses, electric vehicles, and robotics.

While the company is optimistic about the adaptability and relevance of its OS in a rapidly shifting tech landscape, it recognizes the growing competition.

This includes a rumored “screenless device” from former Apple design lead Jony Ive, in partnership with OpenAI.

‘Nothing’ believes it holds a unique advantage: direct control over the last mile of distribution and a nuanced understanding of user behavior.

These factors, it claims, will help the company develop an operating system that genuinely enhances everyday life.

Just as ‘Nothing’ leverages its control over hardware to build a better user experience, we believe the key to powerful AI isn't just the model, it's the data infrastructure beneath it.

Our platform is built on the principle that to develop AI agents that genuinely enhance everyday life, you need direct control and a nuanced understanding of your data's entire lifecycle.

We provide the intelligent data layer that allows companies to automate complex data workflows, ensure quality and governance, and ultimately, build AI that operates with context and purpose. It’s about moving from simply having data to truly understanding it.

New Open-Source AI Benchmarks for SOCs Unveiled by CrowdStrike and Meta

In a collaborative effort, CrowdStrike and Meta have launched CyberSOCEval—an open-source benchmark suite designed to assess how AI systems perform in security operations.

This initiative aims to provide organizations with clear, standardized metrics to choose the most effective AI tools for their Security Operations Centers (SOCs).

The new offering builds on Meta’s open-source CyberSecEval framework and incorporates CrowdStrike’s real-world threat intelligence. It addresses a critical industry need: defining what truly constitutes effective AI in cybersecurity.

CyberSOCEval tests large language models (LLMs) across essential SOC workflows, including incident response, malware analysis, and threat intelligence comprehension.

The benchmarks combine real attacker behavior with expert-crafted scenarios, helping organizations verify AI performance under realistic pressure and confirm operational readiness.

Cyber defenders today operate in a high-volume, fast-changing threat landscape. Adopting advanced AI and LLMs can help automate tasks and improve efficiency, yet many security teams are still early in their AI adoption journey.

A major hurdle has been the absence of reliable benchmarks, making it difficult to identify which AI systems and use cases provide a tangible advantage against actual attacks.

With CyberSOCEval, security teams can now pinpoint exactly where AI adds the most value, while model developers have a clear benchmark to guide improvements that boost both ROI and SOC effectiveness.

The suite is openly available to the AI and cybersecurity communities via Meta’s CyberSecEval framework.

By establishing a new standard for building and deploying AI in security operations, CyberSOCEval empowers defenders to proactively keep pace with and stay ahead of modern adversaries.

Google Open-Sources VaultGemma, an AI Model Built for Data Protection

Google has introduced a new open-access AI model named VaultGemma, developed using insights from research into differential privacy.

This compact 1-billion-parameter model builds on Google’s Gemma 2 architecture, which is part of the company’s family of small-scale language models.

Historically, applying differential privacy, a technique that adds controlled noise during model training, has been used to prevent AI systems from reproducing exact copies of their training data.

While effective for enhancing privacy, this method often comes with a trade-off: the introduced noise typically reduces the model’s accuracy.

This research explored how the ratio of artificial noise to dataset size affects model performance, testing various model scales and noise levels to determine how differential privacy can be scaled without significantly compromising output quality.

According to a Google blog post, VaultGemma is the largest openly available model that incorporates differential privacy while still aiming to deliver useful and accurate results.

It is intended to support the development of AI applications that prioritize both utility and data protection.

The model is now accessible for download on platforms like Hugging Face and Kaggle. Google has also released the model weights, enabling developers to fine-tune VaultGemma for customized use cases.

In a Strategic Swipe at Musk, OpenAI Hires xAI's Former CFO

OpenAI has hired Mike Liberatore, the former Chief Financial Officer of xAI, who stepped down from Elon Musk's company only a few months after taking the role, as reported by CNBC.

Liberatore will assume the position of business finance officer at the U.S. startup, where he will oversee billions of dollars in AI infrastructure spending. He is set to report to OpenAI’s CFO, Sarah Friar, starting this Tuesday.

Liberatore had joined xAI earlier this year, where he played a key role in fundraising initiatives and supported the expansion of a data center in Memphis.

However, his tenure was short-lived; he departed in July after just three months, according to The Wall Street Journal.

His exit creates a major gap at xAI, which has been actively raising capital to scale its AI infrastructure, including data centers and the development of computing chips.

This appointment is expected to intensify the ongoing friction between OpenAI CEO Sam Altman and Elon Musk, two figures at the center of one of tech’s most prominent rivalries.

Musk originally co-founded OpenAI with Altman a decade ago, before leaving to establish his own competing AI company, xAI.

Since then, Musk has made repeated attempts to thwart OpenAI’s shift to a for-profit structure. He has filed two lawsuits against the organization, accusing it of straying from its original mission, and has asked the court to prevent the restructuring.

In related developments, OpenAI and Microsoft are revamping their partnership. According to a report from The Information, OpenAI plans to reduce the share of revenue it allocates to commercial partners, mainly Microsoft, from 20% to around 8% by 2030.

This change would allow OpenAI to retain over $50 billion in additional revenue.

The two companies have also entered into a non-binding agreement that outlines new terms, facilitating OpenAI’s transition to a for-profit model.

Under the current setup, OpenAI’s nonprofit arm is slated to receive more than $100 billion, roughly 20% of the company’s targeted $500 billion valuation, which would make it one of the best-funded nonprofit entities globally.

Your opinion matters!

Hope you loved reading our piece of newsletter as much as we had fun writing it. 

Share your experience and feedback with us below ‘cause we take your critique very critically. 

Thank you for reading

-Shen & Towards AGI team