Editor: GoodInfo Global News Team

Meta has officially completed its acquisition of humanoid robotics startup Assured Robot Intelligence, TechCrunch reports. This marks Meta’s first major acquisition in the embodied AI and robotics space, signaling the tech giant’s accelerating push to extend its AI capabilities from the digital world into the physical realm.

Deal Details and Strategic Intent

Meta confirmed the acquisition in a statement, noting that the Assured Robot Intelligence team will join Meta’s Fundamental AI Research (FAIR) division, focusing on developing foundational AI models for humanoid robots. The financial terms were not disclosed, though industry analysts estimate the deal is valued between $200 million and $500 million.

Founded in 2023, Assured Robot Intelligence’s core technology centers on real-time perception-decision-action loops for robotic manipulation, enabling complex task completion in unstructured environments. The founding team includes alumni from Carnegie Mellon University’s Robotics Institute and Boston Dynamics.

Meta’s Robotics Ambitions

The acquisition aligns with Meta’s broader AI strategy in recent years. From open-source large language models like the Llama series to AI-powered social assistants, Meta has been building a comprehensive multimodal AI technology stack. Humanoid robots are increasingly viewed as the ultimate physical embodiment of AI — and the next strategic frontier for tech giants.

Meta’s Chief AI Scientist wrote in an internal memo: “We are at a pivotal turning point where AI moves from screens into the physical world. Possessing powerful robotic AI models will be central to our next-generation product strategy.”

Intensifying Industry Competition

Meta’s move escalates its competition with Tesla, Google, and Amazon in the robotics arena. Tesla’s Optimus humanoid robot has already entered factory testing phases, while Google’s DeepMind is also advancing research into robotic manipulation models.

Analysts suggest the trend of tech giants acquiring robotics startups will accelerate throughout 2026. As foundational AI models rapidly improve their capabilities in robotic manipulation, embodied intelligence is approaching a critical threshold from academic research to industrial deployment.

Source: TechCrunch

Meta (formerly Facebook) has officially completed its acquisition of Assured Robot Intelligence, a company specializing in AI models for robots, Bloomberg and The Verge reported. The deal closed on Friday, marking Meta’s formal entry into the embodied intelligence and humanoid robotics space.

Acquisition Details

Assured Robot Intelligence develops AI models for robotic systems, with core technologies spanning robot motion control, environmental perception, and task planning — critical capabilities for autonomous robotic operations. The financial terms of the deal were not disclosed.

The Verge confirmed the acquisition and reported that Meta has already begun working on humanoid robot projects based on the acquired company’s technology.

Meta’s Robotics Strategy

This acquisition is not Meta’s first major investment in AI. In recent years, Meta has poured significant resources into open-source large language models (the Llama series), AI research, and computing infrastructure. The acquisition of Assured Robot Intelligence signals Meta’s ambition to extend its AI capabilities from the virtual world into the physical one.

Analysts note that Meta’s timing in entering the humanoid robot space is strategic. Tesla’s Optimus humanoid robot has already demonstrated the ability to perform complex tasks in manufacturing environments, while startups like Figure AI and 1X Technologies are advancing rapidly. Meta’s entry will further intensify competition in this emerging field.

Industry Context

Humanoid robots are widely regarded as the next major application frontier for AI. Unlike traditional industrial robots, humanoids must navigate and operate in unstructured environments designed for humans, placing extremely high demands on AI perception, reasoning, and control capabilities.

Recent advances in large language models and vision-language-action (VLA) models have opened new possibilities for humanoid robotics. By combining the semantic understanding of large models with the physical execution capabilities of robots, the industry hopes to achieve more general-purpose and adaptive intelligent machines.

Meta’s extensive AI research resources, massive computing infrastructure, and deep presence in the open-source community position it to develop unique advantages in the humanoid robotics field.

Looking Ahead

Meta has not disclosed specific timelines or product plans for its humanoid robot project. However, industry observers believe that given Meta’s R&D strength and resource commitments in AI, its humanoid robotics initiative warrants close attention.

Source: Bloomberg, The Verge

SoftBank Group announced on April 30, 2026, the formation of a new robotics company dedicated to building and operating specialized data centers designed for artificial intelligence and automation workloads. The company has set an ambitious goal: achieving an initial public offering (IPO) valued at up to $100 billion.

This strategic move marks SoftBank’s deepening commitment to AI infrastructure. As global demand for AI model training and inference grows exponentially, data centers are evolving from traditional computing facilities into integrated AI factories combining robotic automation, liquid cooling systems, and custom silicon.

SoftBank Chairman Masayoshi Son, who has repeatedly expressed his conviction in the AI revolution, sees the new company as a convergence point for the conglomerate’s diverse AI and robotics capabilities. The venture is expected to integrate technologies from portfolio companies including Boston Dynamics’ robotics expertise and Arm’s chip design capabilities, creating a full-stack AI infrastructure solution spanning hardware to software.

According to sources familiar with the matter, the company plans to deploy its first data centers in Japan and Southeast Asia, with subsequent expansion into North American and European markets. These facilities will operate with high levels of automation, extensively using robots for equipment maintenance and daily operations, significantly reducing labor costs and operational risks.

Industry analysts suggest that if the company achieves its IPO target, it would rank among the largest technology listings in history. For comparison, Equinix, the world’s largest data center operator, currently has a market capitalization of approximately $80 billion. A $100 billion valuation for SoftBank’s new venture would redefine industry standards for AI infrastructure companies.

Against the backdrop of an intensifying global AI arms race, data centers are being reassessed as the “power infrastructure” of the AI industry. SoftBank’s move is certain to further intensify competition in the global AI infrastructure landscape.

Source: TechCrunch

Robots Beat Human Half-Marathon Record

In April 2026, more than 100 humanoid robots competed in a half-marathon event held in an innovation zone on the outskirts of Beijing. A robot from Chinese smartphone maker Honor won the race with a time of 50 minutes and 26 seconds — a result that not only outpaced all other robotic competitors but also surpassed the human half-marathon world record of 57 minutes and 31 seconds.

The event marks a significant milestone in the athletic capabilities of humanoid robots. Just a few years ago, robots still struggled to navigate uneven terrain without stumbling. Today, they can complete a 21-kilometre course at a steady, competitive pace.

100m Sprint Record Within Reach

Even more remarkably, humanoid robots are making rapid progress in the 100-metre sprint. Chinese robotics company Unitree Robotics released video showing its H1 humanoid robot reaching a peak sprint speed of 10 metres per second — closing in on Usain Bolt’s human world record of 9.58 seconds (approximately 10.4 m/s).

New Scientist reports that robots are not only outperforming humans in endurance events but are rapidly narrowing the gap in speed events as well. Analysts suggest that, at the current rate of development, humanoid robots could break the human 100m world record within the next few years.

Technology Drivers

This progress is driven by breakthroughs across multiple technology domains. Advanced motor drivers, lightweight carbon-fibre materials, and AI-powered motion control algorithms enable modern robots to achieve far more agile and efficient movement.

Organisers of the Beijing half-marathon noted that competing robots must complete the race in fully autonomous mode, without any external control. This requires real-time environmental perception, obstacle avoidance, and dynamic balance adjustment — capabilities that were unimaginable in robotics just a few years ago.

Industry Significance

The leap in humanoid robot athletic ability is more than a technological showcase; it carries profound industrial implications. Robots capable of running stably on complex terrain are better equipped for real-world applications in disaster relief, exploration, logistics, and more.

However, experts caution that performance on a standardised track still differs from real-world deployment. Maintaining stability and adaptability in unstructured environments remains a core challenge that the robotics industry must address before widespread adoption becomes feasible.

Source: New Scientist, NPR