Today on The Robot Beat, the hardware is finally catching up to the AI. From NVIDIA's RTX Spark superchip unifying memory to Apple's on-device AI silicon, the industry is rebuilding the stack from the metal up to make embodied intelligence a reality. Meanwhile, massive funding rounds and continued IPO momentum show investors are betting heavily on who can build the physical platforms to run it all.
Following the rapid surge past 3,000 deposits we covered earlier this week, Ubtech's U1 companion humanoid has now hit nearly 4,000 pre-orders, accumulating over 10 million yuan ($1.4 million) ahead of its June 30 reveal. The strong demand for the 88-DOF, emotion-driven robot underscores the market appetite for non-utilitarian personal robotics.
Why it matters
This is a significant early demand signal for the high-end consumer humanoid market. While industrial and logistics applications have dominated the humanoid conversation, the strong pre-order numbers for a non-utilitarian, companion-focused robot suggest a potentially large, untapped market for personal and social robotics. For entrepreneurs in the space, this validates the commercial viability of products centered on human-robot interaction and emotional connection, a market segment that has historically been viewed with skepticism. The U1's reception will be a key test case for whether consumers are ready to pay a premium for robotic companionship.
The high number of pre-orders suggests a market ready for sophisticated companion robots, potentially mirroring the early days of personal computers. However, some industry observers remain skeptical, pointing to videos showing the robot's somewhat stiff movements and questioning whether the final product can live up to the hype generated by its marketing. This situation highlights a recurring theme: the market's enthusiasm for humanoid robots may be outpacing the technology's current capabilities.
Following the recent wave of Chinese humanoid IPO filings we've been tracking—including Unitree, Deep Robotics, and EngineAI—public markets are signaling readiness to finance the massive capital costs of physical automation. However, a new arXiv paper warns that the AI agent benchmark scores driving some of these valuations may be inflated, even as market data shows Anthropic gaining enterprise ground on OpenAI.
Why it matters
The willingness of public markets to fund humanoid robotics marks a critical maturation point for the industry. It moves the sector beyond venture capital and into an era where large-scale manufacturing and deployment are seen as bankable investments. For a robotics entrepreneur, this is a double-edged sword: it validates the market and provides potential exit paths, but also signals the arrival of well-capitalized public competitors. The accompanying note on benchmark inflation is a crucial reminder to critically evaluate performance claims, especially as competition intensifies.
One financial perspective is that these IPOs are a bet on China's national strategy to automate its manufacturing and service sectors, creating a massive domestic market. A technical perspective, highlighted by the benchmark paper, warns that the metrics used to justify these valuations may not always reflect real-world performance, creating a potential 'hype bubble'.
Verified across 2 sources:
Asanify(Jun 14) · arXiv(Jun 14)
Click Copy for AI above, then paste the prompt
into your favorite AI chatbot — ChatGPT, Claude, Gemini, or
Perplexity all work well.
Boston Dynamics has pre-sold its entire 2026 production run of the electric Atlas humanoid, locking out external customers for the year. The units are split between Hyundai's Robotics Metaplant Application Center—accelerating the 25,000-unit Korean industrial deployment we covered last month—and Google DeepMind, which will use the fleet to advance the Gemini ER-1.6 foundation models we saw teased ahead of I/O.
Why it matters
This is a classic vertical integration play and a major strategic move that reshapes the competitive landscape. Instead of selling broadly, Boston Dynamics is focusing its entire initial commercial output on accelerating the capabilities of its own ecosystem. This creates a powerful feedback loop: Hyundai provides real-world industrial use cases and manufacturing scale, while DeepMind provides the AI brain. For competitors, this is a formidable moat. For the rest of the industry, it means the most advanced commercially-produced humanoid platform is off the market for at least a year, which could drive demand for alternatives from companies like Figure or Apptronik.
From a business strategy perspective, this prioritizes long-term competitive advantage over short-term revenue, betting that a vertically integrated, deeply optimized robot will ultimately dominate the market. An AI researcher's view is that giving DeepMind exclusive access to a fleet of Atlas robots could lead to a step-change in embodied AI, but it also centralizes progress within one corporate ecosystem. A supply chain analysis might point to the risks of this strategy, as it ties the fate of Atlas to the success of its two partners' internal projects.
Building on the dense cluster of sub-$25,000 domestic humanoids we've recently tracked—like the Astra-1 and Indra-X—India has officially integrated humanoid robotics into its National AI Mission. The new initiative targets a disruptive $10,000 price point for a locally produced humanoid within three years, aimed at SME automation and projecting over 50,000 new sector jobs.
Why it matters
This is a significant strategic move that could reposition India as a major player in the global robotics market, not just as a consumer but as a producer. The focus on a sub-$10k price point is critical; if achieved, it could dramatically accelerate adoption in manufacturing, logistics, and services, particularly in cost-sensitive emerging markets. For entrepreneurs, this opens up new supply chain opportunities and a potentially massive new market, while also signaling the rise of a new competitive force focused on affordability and scale.
From a national policy perspective, this is a clear move towards technological self-reliance, mirroring similar initiatives in the semiconductor and space industries. Economists will be watching to see if the program can truly create jobs on the scale projected, or if automation will lead to displacement. Robotics engineers may be skeptical of the three-year timeline for a $10,000 humanoid, given the current component costs and technical challenges, but the goal itself will spur intense innovation.
Despite recent milestones like Gatsby's first U.S. home cleaning deployment, a new analysis suggests that truly autonomous and generally useful household humanoids are still a decade away. While models like 1X's NEO are shipping in 2026, the report argues that widespread adoption will wait until the 2030s due to lingering hurdles in fine-motor dexterity, task coverage, safety certification, and cost reduction.
Why it matters
This article provides a much-needed dose of realism amidst the hype cycle. For an entrepreneur in the space, it's a crucial reminder of the long road ahead and the specific, difficult problems that need to be solved. The distinction between a robot that can perform some tasks under supervision (2026) and one that can be trusted to autonomously manage a household (2030s) is vast. Success in the near term will likely come from solving a narrow set of tasks reliably, rather than attempting to deliver a fully autonomous 'robotic butler' from day one.
The author breaks down the key hurdles, including the 'long tail' of household chores that are difficult to automate, the immense challenge of ensuring safety around children and pets, and the social and privacy issues that will arise. The article contrasts the current state of the art with the fictional portrayals that often drive consumer expectations, arguing for a more measured and incremental approach to development and marketing.
In a significant development from earlier this week, Hangzhou-based startup Spirit AI saw its Spirit v1.6 foundation model for embodied intelligence surpass NVIDIA's own Cosmos 3 model on the RoboArena global leaderboard. This benchmark measures the capabilities of robot control policies. The achievement is part of a broader trend, with other Chinese firms like Manifold AI and AgiBot also topping leaderboards for world models and perception, underscoring the country's rapid and focused push into physical AI.
Why it matters
This isn't just about one company winning one benchmark; it signifies a potential shift in the global center of gravity for embodied AI. While much of the focus has been on US-based labs, Chinese companies are demonstrating leadership in both AI model development and, crucially, the data infrastructure needed to train them. For the global robotics landscape, this intensifies competition, potentially accelerating innovation but also raising questions about supply chain dependencies and divergent technology ecosystems.
One viewpoint is that these benchmark victories are the direct result of China's coordinated national strategy and massive investment in AI and robotics. A more skeptical take suggests that benchmark performance doesn't always translate to robust real-world deployment, and the true test will be in commercial success and reliability. A geopolitical analysis notes that as physical AI becomes a key competitive frontier, we may see the emergence of distinct, and potentially incompatible, technology stacks.
Expanding on the OpenClaw initiative we've been tracking—which recently secured hardware backing from NVIDIA's Jensen Huang—the open-source humanoid project has released new architecture details. The 42-degree-of-freedom hardware design is integrated with an open software stack utilizing Llama 4 and Nova 1 for interaction, plus a novel 'SpatialGPT' model for visual processing and motor control via hierarchical reinforcement learning.
Why it matters
Projects like OpenClaw are crucial for democratizing research. By providing an accessible, fully open-source platform, it significantly lowers the barrier to entry for academics, independent researchers, and startups who can't afford the six-figure price tags of commercial humanoids. This fosters a more diverse and collaborative research environment, which can accelerate progress on fundamental challenges like sim-to-real transfer and the development of robust robot foundation models. It's the physical AI equivalent of open-source software libraries that have powered the last decade of digital AI.
This guide provides a comprehensive look at the project's architecture, highlighting its focus on creating a complete, integrated system rather than just open-sourcing individual components. It emphasizes the project's goal of enabling research into human-like interaction and learning through methods like imitation and novelty-driven exploration. The success of OpenClaw will depend heavily on its ability to build an active community of contributors to maintain and extend the platform.
Swiss startup Orbit Robotics has launched Helios, a specialized humanoid robot designed for work in microgravity. Revealed on May 20, Helios is legless and features four arms, optimized for tasks like maintenance and cargo handling on space stations. The company claims that the tasks Helios is designed for currently cost approximately $140,000 per hour when performed by human astronauts. The robot can operate for three hours on a single charge.
Why it matters
Helios represents a pragmatic and targeted approach to space robotics. Instead of a general-purpose humanoid, it's a purpose-built tool designed to automate high-cost, routine tasks, which is the most direct path to commercial viability in the space economy. By focusing on reducing the immense operational costs of human spaceflight, Orbit Robotics is addressing a clear market need. This development is a strong indicator of the growing maturity of the commercial space industry, where robotic automation is becoming essential for scalable and sustainable infrastructure.
The design choice to omit legs and add extra arms is a clever optimization for a zero-gravity environment where traditional locomotion is useless and multi-limb anchoring is key. While the three-hour battery life is a limitation, it is likely sufficient for specific, planned maintenance windows. This focus on a niche, high-value application could provide the revenue stream needed to fund the development of more advanced space robotics in the future.
German robotics company NEURA Robotics is raising up to $1.4 billion in a Series C round led by Tether, the issuer of the USDT stablecoin. The round, which includes co-investors like NVIDIA, Amazon, Qualcomm, and Bosch, values NEURA at approximately $7 billion. A key part of the deal is the integration of Tether's Wallet Development Kit and QVAC edge AI runtime into NEURA's humanoid and collaborative robots, which would enable the machines to make autonomous payments and financial decisions.
Why it matters
This is one of the largest private funding rounds in robotics history and signals a significant convergence of AI, robotics, and cryptocurrency. The vision is to create a 'machine economy' where robots can operate as independent economic agents—ordering their own spare parts, paying for charging, or bidding for jobs without human intervention. For an entrepreneur, this represents a paradigm shift from robots as capital equipment to robots as autonomous service providers. It opens up entirely new business models but also introduces complex challenges around security, regulation, and the governance of these autonomous economic actors.
Supporters frame this as the dawn of a truly autonomous economy, where friction in machine-to-machine transactions is eliminated. Critics raise significant security concerns, questioning the wisdom of giving machines with direct access to crypto wallets the autonomy to transact. A more pragmatic view from the industrial sector sees near-term value in simplifying internal logistics and payments, such as a factory robot autonomously paying for a new gripper from an internal supplier, long before any public-facing economic activity occurs.
NVIDIA unveiled the RTX Spark "superchip" on Tuesday, a new architecture that combines a Grace CPU and a Blackwell RTX GPU onto a single module with 128GB of unified, coherent memory. The connection is made via its high-speed NVLink-C2C interconnect, which effectively eliminates the traditional PCIe bus bottleneck that slows data transfer between the CPU and GPU. This design allows both processors to access the same memory pool simultaneously, drastically improving performance for on-device AI applications that are memory and bandwidth intensive.
Why it matters
This is a fundamental shift in hardware architecture aimed directly at the next generation of AI, particularly embodied AI. By removing the PCIe bottleneck, NVIDIA is enabling much larger and more complex models to run efficiently on edge devices, including robots. For robotics, this means faster inference, more sophisticated multi-modal processing, and the ability to run powerful planning and control models locally without relying on the cloud. This development could significantly accelerate the path to more autonomous and capable robots by solving a core hardware limitation.
One perspective highlights this as a direct response to Apple's success with its unified memory architecture, which has demonstrated significant performance gains in creative and AI workloads. Another view from the robotics community is that while promising, the cost and power consumption of such a high-performance chip will be critical factors for its adoption in mobile robots where battery life and thermal management are key constraints.
Apple is doubling down on a privacy-first, on-device AI strategy, systematically co-designing its silicon, foundation models, and developer APIs to run AI tasks locally on iPhones, iPads, and Macs. This approach minimizes latency, reduces reliance on cloud data centers, and, crucially, keeps user data on the device by default. This integrated hardware and software stack is positioned to enable a new class of powerful, local-first applications that can function without a constant internet connection.
Why it matters
Apple's approach represents a significant philosophical and architectural divergence from the cloud-centric AI models pursued by many competitors. For the robotics and edge computing ecosystem, this is a powerful validation of the local-first paradigm. It signals a massive market push towards optimizing models for performance and efficiency on-device, which will create spillover benefits for robotics developers. As Apple standardizes user expectations for private, responsive AI, it will increase pressure on all device makers—including those in robotics—to deliver similar capabilities without compromising user data.
Proponents see this as a necessary and 'quiet revolution' that correctly prioritizes user privacy and creates more resilient applications. Skeptics, however, question whether a purely on-device approach can keep pace with the rapid advancements of larger, cloud-based frontier models, potentially leaving Apple's ecosystem a generation behind in raw capability. A middle-ground view suggests Apple's hybrid model—using its 'Private Cloud Compute' for tasks that exceed device capabilities—offers a pragmatic balance of privacy and power.
Thanks to AMD's new Ryzen AI Max+ 395 chip and its unified memory architecture, compact mini PCs like the GMKtec EVO-X2 are now capable of running local inference for AI models with up to 235 billion parameters. This provides a powerful, cost-effective alternative to relying on expensive cloud-based AI subscriptions. For developers, the hardware investment could pay for itself in less than a year compared to the recurring costs of cloud services.
Why it matters
This is a significant milestone in democratizing access to large-scale AI. By enabling powerful models to run on affordable, local hardware, it drastically lowers the barrier to entry for developers, researchers, and startups. For robotics, this could be transformative, allowing for the deployment of highly sophisticated AI on edge devices or local control stations without the latency or privacy concerns of the cloud. It challenges the entire business model of cloud-exclusive AI and empowers a new wave of innovation at the edge.
Proponents argue this is the future of AI development, offering greater privacy, lower long-term costs, and offline functionality. Cloud providers would counter that their services offer scalability, access to the absolute largest models, and simplified management that local hardware can't match. A balanced view suggests a hybrid future where developers use the cloud for training and large-scale tasks, but rely on increasingly powerful local hardware for inference and real-time applications.
Two key announcements on Saturday are pushing towards a more standardized software layer for robotics. Numurus introduced its Edge Platform Interface (NEPI), a software abstraction layer designed to standardize AI operations across different processor brands like NVIDIA, AMD, and Qualcomm. Concurrently, NVIDIA unveiled a suite of platform components, including open-source models, the Isaac GR00T framework, and new Jetson T4000 hardware, with the explicit goal of becoming the default operating layer for robotics.
Why it matters
This is a battle for the 'Windows' or 'Android' of robotics. The lack of a standard operating layer is a massive source of friction in the industry, forcing developers to write custom, non-portable code for every hardware combination. NEPI's approach is to be a neutral, hardware-agnostic translator. NVIDIA's strategy is to make its own stack so compelling and comprehensive that it becomes the de facto standard. For a robotics entrepreneur, standardization is a massive win, as it reduces development costs and allows focus on value-add applications rather than low-level integration. The key question is whether the industry will coalesce around a single standard or remain fragmented.
Numurus CEO Jason Seawall pitches NEPI as a democratizing force, making it easier to use the best chip for the job without being locked into one vendor's ecosystem. NVIDIA's perspective is that a tightly integrated, full-stack solution from a single vendor provides the best performance and developer experience. The open-source community might argue that true standardization can only come from vendor-neutral projects like ROS, not from corporate platforms.
In response to U.S. export controls blocking access to advanced EUV lithography machines, Huawei's HiSilicon subsidiary has developed a new approach to chip scaling. Unveiled in late May, the 'Tau Scaling Law' and 'LogicFolding' 3D architecture prioritize reducing signal propagation time in vertically stacked chiplets over simply shrinking transistors. The company claims this approach can achieve performance equivalent to a 1.4-nanometer process by 2031 without relying on ASML's EUV technology.
Why it matters
This is a significant and creative attempt to engineer around geopolitical and supply chain constraints. If successful, it could provide a viable alternative path for semiconductor advancement, decoupling performance from the relentless, and increasingly expensive, pursuit of transistor density. For robotics and edge AI, this could lead to more power-efficient and performant custom silicon. It’s a powerful example of how constraints can drive innovation and could lead to a bifurcation in global chip design philosophy.
This is a major strategic pivot, shifting focus from a race Huawei couldn't win (EUV access) to one it could potentially lead (3D integration). Semiconductor experts are cautiously optimistic, noting that 3D stacking has its own immense challenges, including heat dissipation and manufacturing yield. From a geopolitical standpoint, this represents a major effort by China to achieve semiconductor self-sufficiency and insulate its tech industry from foreign sanctions.
Scientists at Nanyang Technological University (NTU) in Singapore have developed a tiny, 4.4 mm long medical robot capable of performing five distinct functions inside the body. Guided by external magnetic fields, the semi-rigid robot can navigate through complex and constrained areas, cut tissue, take biopsy samples, deliver medication, and use targeted heating (hyperthermia) to destroy cancer cells.
Why it matters
This is a major step towards making minimally invasive surgery even less invasive. By integrating multiple surgical tools into a single, tiny device, this microrobot could enable procedures in parts of the body that are currently difficult or impossible to reach, such as the brain or inner ear. Its multifunctional nature could also reduce procedure time and complexity by eliminating the need to swap out different instruments. For the field of medical robotics, this demonstrates the potential of microrobotics to move from the lab to practical clinical applications.
While a significant research achievement, the team acknowledges that a key challenge for clinical translation is real-time imaging to accurately track the device within the body. Medical device regulators will also require extensive safety and efficacy trials before such a technology could be approved for human use. The use of external magnetic fields for control, however, is a well-established technique, which could smooth the path to commercialization.
A research team from the University of Hong Kong (HKU) has developed the first vision platform that uses near-infrared II (NIR-II) fluorescence to navigate magnetic nanorobots. This technique provides clear, real-time visual feedback for guiding the tiny robots inside a living body, overcoming the limitations of previous imaging methods that struggled with penetration depth and resolution. The team has demonstrated potential applications in targeted drug delivery for inflammatory diseases.
Why it matters
One of the biggest hurdles for clinical use of microrobots has been the inability to see where they are and what they are doing in real time. This 'infrared GPS' system is a major breakthrough that directly addresses this challenge. By providing a clear, high-resolution view, it could enable the precision and control needed for complex in-vivo tasks like targeted drug delivery to tumors or breaking up blood clots. This could be the enabling technology that moves nanorobotics from a research curiosity to a viable therapeutic tool.
The use of the NIR-II window is key, as it offers deeper tissue penetration and lower background signal compared to visible light or NIR-I imaging. While the current research focuses on inflammatory diseases, the platform is broadly applicable to any therapy requiring precise delivery. The next step will be to integrate this navigation system with therapeutic nanorobots to demonstrate efficacy in pre-clinical models.
Researchers at the Italian Institute of Technology (IIT) have created a soft robotic arm that mimics the decentralized nervous system of an octopus. The arm features integrated sensors within its suckers that allow it to autonomously detect contact, assess force, and adapt its grip on objects. This all happens locally within the arm itself, without requiring centralized processing or commands from a main computer, enabling it to operate effectively in complex underwater environments.
Why it matters
This represents a significant leap forward for autonomous robotics in unstructured environments. Traditional robots struggle with the complexity and unpredictability of the real world, but this decentralized, bio-inspired approach offers a more robust and scalable solution. By embedding intelligence directly into the arm's structure, the system becomes more resilient to failure and can react faster than a centralized system. For applications like ocean exploration or disaster response, this paradigm of 'embodied intelligence' could be a game-changer, enabling robots to perform complex tasks that are currently impossible.
The researchers suggest this model could be applied to a wide range of robotic systems that need to interact with unpredictable environments. Biologists note that this is a prime example of how studying natural systems can lead to breakthroughs in engineering. A contrasting view from industrial robotics might question the precision and repeatability of such a system compared to traditional, rigid robots, highlighting the trade-offs between adaptability and structured performance.
Tesla's Robotaxi platform has been officially classified as an SAE Level 4 autonomous driving system in a 'Robotaxi First Responder Interaction Plan' document filed for Arizona. This designation means the vehicle is designed to operate without any human intervention within its specific operational domain. However, separate reports indicate the current service in Texas is struggling with operational bottlenecks, with a fleet of only 59 vehicles facing long waits and technical issues.
Why it matters
There's a growing chasm between Tesla's official classifications and its on-the-ground reality. The Level 4 designation is a legally and technically significant step, distinguishing the Robotaxi from the Level 2 consumer FSD product. Yet, the anemic fleet size and operational problems show that scaling an autonomous service is a monumental challenge that goes far beyond the AI model itself. This highlights the 'last mile' problem of AV deployment: even with a capable system, the logistics, maintenance, and customer service infrastructure required for a reliable service are immense hurdles.
One perspective is that the Level 4 classification and expansion into Arizona are bullish signals of Tesla's confidence and intent to compete directly with Waymo. Another, more critical view, is that the small fleet size and continued need for human safety monitors in some cases show the technology is far from mature. The first responder plan itself, which details procedures for remotely moving a disabled vehicle, underscores the many practical edge cases that must be solved before widespread, truly driverless deployment is feasible.
Penumbra has received FDA clearance for its THUNDERBOLT system, a computer-assisted vacuum thrombectomy device used for removing blood clots in acute ischemic stroke patients. The approval is a significant development in neurovascular treatment and a major win for Boston Scientific, which is in the process of acquiring Penumbra for approximately $14.5 billion.
Why it matters
This clearance highlights the accelerating trend of computer assistance and robotics in delicate surgical procedures. While not a fully autonomous robot, the THUNDERBOLT system uses AI to assist the surgeon, improving precision and potentially patient outcomes. For the medical robotics field, this is a prime example of the most viable path to market: augmenting, rather than replacing, the surgeon. The immediate acquisition by an industry giant like Boston Scientific underscores the high value placed on innovative, FDA-cleared robotic and assistive medical technologies.
From a clinical perspective, tools like THUNDERBOLT can make complex stroke interventions safer and more accessible to a wider range of hospitals. From a business perspective, the acquisition illustrates a common strategy in MedTech: startups focus on R&D and navigating the FDA approval process, while established players use acquisitions to expand their product portfolios and market reach. The high price tag reflects the significant clinical need and market opportunity for advanced stroke treatments.
Hardware Re-architected for Embodied AI A major trend is the fundamental redesign of silicon to support physical AI. NVIDIA's RTX Spark, Apple's on-device AI strategy, AMD's powerful mini PCs, and Huawei's LogicFolding architecture all show a move towards unified memory and specialized processors to eliminate data bottlenecks and enable efficient, powerful edge computing for robots.
Mega-Rounds Signal Confidence in European Robotics Two massive European funding rounds, NEURA Robotics' $1.4B Series C and THEKER's $85M Series A, demonstrate significant investor confidence in Europe's ability to compete in the global robotics market, particularly in creating cognitive and modular industrial systems.
The Duality of Consumer Robotics: Hype vs. Reality There's a clear tension in consumer robotics. While Ubtech's U1 companion robot is attracting thousands of pre-orders, signaling market appetite, critical analyses of robot vacuums and the long-term viability of home humanoids remind us that practical utility, cost, and maintenance remain significant hurdles to mass adoption.
Microrobotics Advances Toward Clinical Reality Multiple breakthroughs in microrobotics are pushing the field closer to medical applications. Innovations include an 'infrared GPS' for nanorobot navigation inside the body, a 4.4mm multifunctional surgical robot, and new programmable magnetic materials, all pointing toward less invasive and more precise future therapies.
The Robotaxi Rollout Hits Real-World Friction Despite official Level 4 classifications and expansion plans into new states like Arizona, Tesla's Robotaxi service is facing significant operational bottlenecks, with a tiny fleet and persistent technical issues. This illustrates the immense gap between autonomous driving promises and the complex reality of scaled commercial deployment.
What to Expect
2026-06-20—Dr. Sirko Straube of DFKI will lecture on AI agents and robotics decision-making.
2026-06-30—Ubtech is expected to fully reveal the U1 companion humanoid, including final pricing and specs.
2026-08-15—The International Joint Conference on Artificial Intelligence (IJCAI) 2026 kicks off in Bremen, Germany.
How We Built This Briefing
Every story, researched.
Every story verified across multiple sources before publication.
🔍
Scanned
Across multiple search engines and news databases
319
📖
Read in full
Every article opened, read, and evaluated
106
⭐
Published today
Ranked by importance and verified across sources
19
— The Robot Beat
🎙 Listen as a podcast
Subscribe in your favorite podcast app to get each new briefing delivered automatically as audio.
Apple Podcasts
Library tab → ••• menu → Follow a Show by URL → paste