Mira Murati’s AI startup Thinking Machines has unveiled its first general-purpose artificial intelligence model, marking a significant attempt by a Western company to narrow China’s growing lead in the open-weight AI ecosystem.

The release comes at a time when Chinese laboratories have emerged as the dominant force in open-source AI, while several leading U.S. developers have increasingly shifted toward proprietary models.

The San Francisco-based startup, founded last year by former OpenAI Chief Technology Officer Mira Murati, introduced Inkling, an open-weight large language model that developers can download, run and customize. Unlike closed-source systems from companies such as OpenAI and Anthropic, open-weight models give enterprises, researchers and governments greater control over deployment, security, data privacy and fine-tuning.

Inkling represents Thinking Machines’ first foundation model after the company debuted its AI customization platform, Tinker, in October last year. The new model is available through Tinker as well as other developer platforms.

With 975 billion parameters, Inkling ranks among the largest open-weight AI models released to date. Parameters are the mathematical variables that determine how an AI system processes information, learns patterns, and generates responses. Although parameter count alone no longer determines model quality, models of this scale typically require enormous computing resources and sophisticated training infrastructure.

The launch comes as competition in open-weight AI has shifted dramatically toward China.

Over the past year, Chinese developers, including Alibaba, DeepSeek, Baidu, and Tencent, have steadily gained market share by releasing capable open models that offer performance approaching proprietary U.S. systems at significantly lower operating costs. Their rapid progress has been aided by aggressive optimization techniques and a strategy centered on making advanced AI widely accessible rather than keeping it behind commercial APIs.

That momentum has accelerated after Meta largely retreated from its previous open-model strategy. Following the disappointing reception of Llama 4, Meta moved toward a more proprietary development approach, creating a gap among Western companies offering high-performance open models.

Thinking Machines is positioning Inkling as one of the few credible Western alternatives in that space.

The company published benchmark results comparing Inkling against proprietary models from Anthropic, Google and OpenAI, as well as leading open-weight models, many of which originate from Chinese AI laboratories. While Thinking Machines acknowledged that top proprietary systems continue to outperform Inkling overall, it said the model delivered competitive results, particularly in agent-related tasks that require AI systems to plan, reason and execute multi-step workflows.

The emphasis on AI agents reflects one of the industry’s fastest-growing areas. Enterprises are increasingly deploying autonomous AI systems capable of completing complex business processes with minimal human intervention, making agent performance an important competitive differentiator.

The launch also underscores the changing economics of enterprise AI adoption.

Many businesses have begun favoring open-weight models because they eliminate recurring API costs, allow deployment inside private data centers, and reduce dependence on a single vendor. Organizations operating in regulated industries often prefer open models because sensitive information never has to leave their own infrastructure.

Thinking Machines cited Bridgewater Associates as an example of that trend. The hedge fund used the company’s Tinker platform to build a customized version of Alibaba’s Qwen model, which it said outperformed leading proprietary systems while operating at lower cost.

Across the AI industry, many enterprises are increasingly adapting open models to specialized use cases, improving performance while reducing inference expenses rather than relying exclusively on frontier closed models from OpenAI, Anthropic or Google.

The release also comes amid intensifying geopolitical competition over AI.

Washington has tightened export restrictions on advanced AI chips and semiconductor equipment in an effort to slow China’s development of frontier AI systems. At the same time, Beijing has doubled down on open-source AI as a strategic tool to expand its global influence, particularly across emerging markets.

Chinese firms have recently accelerated the rollout of capable models and computing infrastructure, including Huawei-powered AI clusters designed to reduce reliance on Nvidia’s most advanced processors. Chinese officials have also promoted open AI as a public good capable of expanding access to artificial intelligence worldwide.

Against that backdrop, Inkling represents more than just another foundation model. It signals an effort by a new generation of U.S. startups to rebuild Western competitiveness in open AI after China’s rapid advances reshaped the landscape.

Whether Inkling can significantly slow China’s momentum remains uncertain. However, its release broadens the pool of high-end open-weight models available to enterprises and developers at a time when demand for customizable AI systems continues to grow rapidly.

Nvidia is accelerating its expansion into Japan’s industrial technology sector through new partnerships with some of the country’s biggest robotics manufacturers, as Chief Executive Jensen Huang broadens the company’s strategy beyond AI chips and cloud computing into what he describes as the next major growth frontier: physical AI.

Speaking at a media event in Tokyo on Thursday, Huang announced collaborations with industrial automation leaders Fanuc and Yaskawa Electric to advance robotics and artificial intelligence, underscoring Nvidia’s ambition to become the foundational technology provider for intelligent machines as factories worldwide become increasingly automated.

“With AI, robots will become smart, easily adaptable and accessible,” Huang said.

The partnerships represent an important step for Nvidia as it seeks to extend its dominance from training large language models to powering autonomous robots capable of operating in factories, warehouses, logistics centers and eventually homes. Rather than focusing solely on supplying graphics processors, Nvidia is positioning its hardware, software platforms, and AI models as the operating system for a new generation of intelligent industrial machines.

The announcement also aligns with Japan’s long-standing strengths in robotics. Fanuc and Yaskawa are among the world’s largest industrial robot manufacturers, supplying automation equipment to automotive, electronics, and manufacturing companies across the globe. Integrating Nvidia’s AI computing platforms into those systems could enable robots to move beyond repetitive programmed tasks toward machines capable of perception, reasoning, and real-time decision-making.

The collaborations come as governments and companies increasingly see “physical AI” as the next phase of artificial intelligence, where AI systems move from generating text and images to interacting directly with the physical world.

Separately, government-backed AI infrastructure company Noetra announced plans to purchase 27,500 Nvidia Rubin AI chips as it develops physical AI capabilities. The company counts Sony among its investors, highlighting growing public and private sector efforts to build domestic AI infrastructure in Japan.

The Rubin processors represent Nvidia’s next-generation AI architecture, succeeding its Blackwell platform, and are expected to power some of the world’s largest AI data centers and robotics applications. A deployment of 27,500 chips would constitute one of Japan’s largest announced AI hardware investments and would further strengthen Nvidia’s presence in one of Asia’s most technologically advanced markets.

Japan has emerged as a priority for Nvidia as countries race to establish sovereign AI capabilities. Rather than relying entirely on foreign cloud providers, governments are investing heavily in domestic computing infrastructure, advanced semiconductors and AI research to secure long-term technological competitiveness.

Huang’s visit also points to Nvidia’s broader effort to deepen relationships across Japan’s semiconductor ecosystem. Although Japan no longer dominates global chip manufacturing as it did during the 1980s, it remains indispensable to the semiconductor supply chain through its leadership in chipmaking equipment, specialty chemicals, precision materials and advanced manufacturing technologies.

During the trip, Huang met executives from several key suppliers, including the chief executives of memory chipmaker Kioxia and semiconductor equipment manufacturer Tokyo Electron, companies that play critical roles in producing advanced AI chips.

The visit comes as momentum across the global semiconductor industry continues to strengthen.

Dutch chip equipment maker ASML raised its sales forecast this week and announced further capacity expansion to meet growing demand from AI chip manufacturers. Meanwhile, Taiwan Semiconductor Manufacturing Co. reported record earnings on Thursday and increased its capital expenditure forecast, reflecting sustained investment by customers building AI infrastructure worldwide.

Together, those developments amplify expectations that the AI investment cycle remains robust despite concerns over supply constraints, valuations and geopolitical risks.

Huang also appeared alongside Japan’s Minister of Economy, Trade and Industry, Ryosei Akazawa, at a government AI event, marking what many believe to be a close coordination between Nvidia and Japanese policymakers.

The partnerships announced in Tokyo show that Nvidia’s strategy is evolving beyond selling AI accelerators to cloud providers. The company has been embedding its technology across the entire AI value chain, spanning data centers, robotics, industrial automation, autonomous systems and sovereign AI infrastructure.

Against the backdrop of the widening gap left by China’s robotics leadership, Japan sees collaborating with Nvidia as an opportunity to combine world-leading robotics expertise with cutting-edge AI computing, potentially strengthening the country’s competitiveness.

On the other hand, as manufacturers worldwide shift toward more autonomous, AI-driven production systems, Nvidia sees the alliances as an opportunity to deepen its presence in one of the world’s most sophisticated manufacturing economies. This will help the chipmaker to boost its ambition to become the foundational technology platform not only for generative AI but also for the emerging era of intelligent machines.

OpenAI is reportedly developing its first dedicated consumer hardware product, a screen-free artificial intelligence companion designed for the home that could mark the company’s biggest step yet beyond software and put it in position to compete in the emerging market for AI-native consumer devices.

According to a Bloomberg report, the device is being developed as a mobile smart speaker with deep ChatGPT integration. However, it is intended to be far more than a conventional voice assistant. Internally, OpenAI is said to describe it as a “humanlike AI companion that lives in the home,” capable of learning about its owner over time and becoming increasingly personalized through continuous interaction.

The project signals OpenAI’s ambition to build an integrated AI ecosystem where proprietary hardware, software and cloud-based models work together, echoing the strategy y that helped Apple dominate the smartphone era through tight integration between its devices and operating systems.

Unlike existing smart speakers from Amazon, Google and Apple, OpenAI’s reported device would be screen-free and designed around conversational AI rather than voice commands alone. Bloomberg said the product would have access to a user’s digital life, including emails and other personal information, enabling it to deliver highly contextual responses, anticipate user needs and automate everyday tasks.

The report also described the device as incorporating “mechanical elements that can move on their own,” suggesting OpenAI may be experimenting with a more expressive physical interface aimed at making interactions feel more natural. Rather than functioning as a passive speaker waiting for commands, the product is reportedly envisioned as a proactive assistant that develops a relationship with its owner over time.

That approach represents a significant shift in consumer AI. Today’s smart speakers generally operate as voice-controlled utilities that respond to explicit requests. OpenAI’s reported vision is closer to an always-present AI agent that continuously learns preferences, manages information, and acts autonomously on behalf of users.

Such capabilities could transform the device into a central hub for the connected home, integrating scheduling, communications, entertainment, shopping and smart-home controls while serving as a physical extension of ChatGPT.

The project indicates that OpenAI is pursuing a strategic goal of reducing its dependence on third-party hardware platforms.

For years, ChatGPT has primarily reached users through smartphones, computers, and web browsers controlled by companies such as Apple and Google. Building proprietary hardware would allow OpenAI to own the entire customer experience, gather richer contextual data and introduce new AI interaction models that are difficult to replicate through existing devices.

Industry analysts have described AI hardware as the next major battleground following the rapid adoption of generative AI software. While smartphones remain the dominant computing platform, technology companies are searching for new device categories that place artificial intelligence at the center of the user experience rather than treating it as another application.

The reported device is being developed with the assistance of several former Apple engineers who played key roles in creating products including the iPhone and Mac, according to Bloomberg. Their involvement suggests OpenAI is attempting to build a long-term consumer hardware business rather than launch a single experimental product.

The timing, however, is particularly sensitive.

Last week, Apple sued OpenAI in federal court, accusing the AI company of misappropriating confidential information and trade secrets related to consumer hardware development. Apple alleged the claims outlined in its complaint represented only “the tip of the iceberg” and said additional evidence would emerge during the discovery process.

OpenAI, in response, has denied the allegations.

Bloomberg reported that OpenAI believes its planned hardware product differs substantially from Apple’s existing products and does not infringe on Apple’s intellectual property. The legal dispute introduces an additional layer of uncertainty to OpenAI’s hardware ambitions, particularly if the company intends to launch multiple consumer devices over the coming years.

The reported home companion also enters an increasingly crowded race to define AI-native hardware. Technology companies have spent more than a decade developing smart speakers and voice assistants, yet products such as Amazon Echo, Google Nest, and Apple’s HomePod have generally remained secondary devices rather than replacing smartphones or computers.

Generative AI has renewed optimism that more capable conversational models could finally unlock broader consumer adoption by making digital assistants significantly more useful, proactive, and personalized.

That optimism has fueled substantial investor interest.

Hark, an AI startup founded by Brett Adcock, raised an oversubscribed $700 million Series A funding round in May at a $6 billion valuation to develop what it describes as “personal intelligence” through proprietary AI models paired with custom hardware. Although Hark has disclosed few details about its product, the financing illustrates the growing willingness of investors to fund AI hardware companies before commercial products reach the market.

The emergence of multiple well-funded startups suggests investors increasingly believe the next major computing platform could revolve around AI-first devices rather than incremental improvements to smartphones.

For OpenAI, launching dedicated hardware would create new revenue opportunities beyond subscriptions and enterprise software while strengthening customer engagement through continuous interaction. It would also give the company another avenue to distribute its AI models as competition intensifies among OpenAI, Google, Anthropic, Meta and other developers seeking to establish dominant consumer ecosystems.

The reported device remains under development, and OpenAI has not publicly confirmed its existence or provided a timeline for a commercial launch.

If successful, however, the product could represent an important milestone in the evolution of consumer artificial intelligence, shifting AI from an application users occasionally open to an always-available companion embedded in everyday life. Such a transition would also intensify competition among technology companies seeking to define the next generation of personal computing after the smartphone.