The robots we saw at CES 2026: The lovable, the creepy and the utterly confusing

CES has always been a magnet for attention-grabbing robots, but Engadget framed 2026 as a real inflection point. With AI materially improving robot “brains” and autonomy, the show moved beyond novelty and leaned into systems that can actually do more on their own. Humanoids were front and center: Agibot’s lineup stood out, with the smaller X2 learning complex choreography and the larger A2 holding real conversations while actively operating booths. On the consumer side, robotics stretched past the standard vacuum demos. Dreame showed a multipurpose extendable arm capable of picking up objects and cleaning corners, along with a stair-climbing concept that handled full-size steps. Roborock introduced its own stair-climbing vacuum concept, while LG’s CLOiD ran an extended demo folding laundry, fetching drinks, and assisting in the kitchen, impressive, even if still conceptual. AI companions were everywhere as well, from Sharp’s Poketomo to Takway’s Sweekar pocket pet and Ludens AI’s Cocomo. At the other end of the spectrum, Realbotix leaned into realism, showcasing humanoids with facial tracking and expression-reading vision tech that many found unsettling.

Major Takeaway: CES 2026 made it clear that robotics is accelerating quickly, with more autonomy and a broader range of form factors, but also highlighted the tension between practical, value-driving machines and human-like robots that still feel difficult to trust. Read More

Robotics market projected to reach US$ 199.50 billion by 2035

A new market outlook from Astute Analytica projects the global robotics market will grow from $51.51 billion in 2025 to $199.50 billion by 2035, driven by sustained investment and clear industrial demand. The report argues robotics has moved beyond pilots and demos into core operational infrastructure, powered by embodied AI and use cases that are increasingly non-optional for companies. It highlights rapid industrial expansion, noting that the operational stock of industrial robots reached roughly 4.7 million units with 9% year-over-year growth, and that China accounted for 54% of global robot supply. Hardware represents the largest share of the market at 44.7%, while industrial robots remain the dominant category at 35.5%, led by manufacturing use cases such as welding, painting, and assembly. The outlook also points to growing deployments in logistics and healthcare, and ties future growth to the convergence of generative AI with humanoid platforms capable of more complex work.

Major Takeaway: Robotics is increasingly being treated as critical infrastructure, not experimentation, driven by industrial demand, hardware-heavy economics, and the next wave of embodied AI expanding what robots can realistically handle. Read More

Walmart expands drone delivery with Wing to 150 more stores

Wing is significantly expanding its Walmart drone delivery operations, adding service to 150 more U.S. stores and extending coverage from Los Angeles to Miami. Axios notes that what once sounded futuristic is already routine for some customers, with certain Dallas users ordering Wing drone deliveries multiple times per week. Walmart and Wing project that by the end of 2026, roughly 40 million Americans could have access to the service, with upcoming metro launches planned for Los Angeles, St. Louis, Cincinnati, and Miami. The operating model is designed for scale and repeatability: customers order through Wing’s app, items are packed into a small basket, loaded into a fenced parking-lot “nest,” and flown autonomously while human pilots oversee flights from a centralized hub. Deliveries are currently free as Wing prioritizes adoption, with plans to integrate the service directly into Walmart’s app as the network grows toward more than 270 locations by 2027.

Major Takeaway: Walmart and Wing are pushing drone delivery toward the mainstream by scaling aggressively, standardizing operations, and betting that speed and convenience for everyday essentials will drive consistent, repeat usage. Read More

About Lucid Bots

Founded in 2018, Lucid Bots is an AI robotics company that is committed to uplifting humanity by building the world's most productive and responsible robots that can do dangerous and demanding tasks.

Headquartered in Charlotte, the company engineers, manufactures, and supports its products domestically, which include the Sherpa, a cleaning drone, and the Lavo, a pressure-washing robot.

Lucid Bots' products are elevating safety and efficiency for a growing number of customers around the world. Lucid is a Y Combinator-backed company, with investments from Cubit Capital, Idea Fund Partners, Danu Ventures, and others. Lucid Bots was recently recognized as the fastest growing robotics manufacturer in the United States.

‍

Drone startups set sights on sustained growth in 2026

Drone startups are heading into 2026 expecting steadier growth as the industry moves beyond experimentation and into repeatable commercial deployments. Founders say the next phase will be defined less by pilots and more by long-term contracts, performance at scale, and reliable operations across real job sites. Clearer regulations and increased demand for domestically produced systems have helped reduce uncertainty and attract new investment. While defense remains a key driver, companies serving agriculture, infrastructure, logistics, and commercial cleaning and inspection are also positioning themselves for broader adoption by offering bundled hardware, software, and services designed for professional use.

Major Takeaway: Drone startups are shifting from early trials to commercial traction, with growth increasingly tied to reliability, long-term customer relationships, and real-world applications such as industrial facility operations and commercial services in 2026. Read More

FCC foreign drone ban could impact U.S. farmers

The Federal Communications Commission has placed new restrictions on foreign-made drones and components, effectively blocking many new models from being authorized for sale or import in the United States. The move, driven by national security concerns, means drones that dominate agriculture and commercial facility operations may no longer be available unless approved by U.S. security agencies. Farmers and industry groups warn the timing could be difficult, as many rely on drones for precision spraying, mapping, crop monitoring, and other automated workflows. While drones already in use remain legal, limited replacement options could increase costs and uncertainty for operators who depend on drones as core infrastructure.

Major Takeaway: A security-driven policy shift is colliding with industries that rely on drones as essential tools, including agriculture and commercial operations. The ruling highlights how urgently domestic manufacturing must scale to support precision farming and other professional drone use cases. Read More

NVIDIA releases new physical AI models as global partners unveil next-generation robots

At CES 2026 in Las Vegas, NVIDIA announced a major expansion of its physical_toggle AI stack, unveiling open models, frameworks, and tools designed to accelerate how robots perceive and act in the real world. The company released new open-source models like Cosmos and GR00T to support world understanding and task execution, along with simulation tools that shorten the path from research to deployment. Global partners including Boston Dynamics, Caterpillar, Franka Robotics, and others demonstrated next-generation machines spanning humanoids, industrial robots, and autonomous systems built for factories, warehouses, and industrial facility operations. NVIDIA also introduced new edge hardware designed to support high-performance robotics in environments where reliability and uptime matter.

Major Takeaway: NVIDIA is helping push robotics toward real-world deployment at scale. By combining physical AI, simulation, and edge computing, the company is accelerating adoption in industries that depend on automation for industrial cleaning, inspection, logistics, and facility operations. Read More

About Lucid Bots

Founded in 2018, Lucid Bots is an AI robotics company that is committed to uplifting humanity by building the world's most productive and responsible robots that can do dangerous and demanding tasks.

Headquartered in Charlotte, the company engineers, manufactures, and supports its products domestically, which include the Sherpa, a cleaning drone, and the Lavo, a pressure-washing robot.

Lucid Bots' products are elevating safety and efficiency for a growing number of customers around the world. Lucid is a Y Combinator-backed company, with investments from Cubit Capital, Idea Fund Partners, Danu Ventures, and others. Lucid Bots was recently recognized as the fastest growing robotics manufacturer in the United States.

Robots are not apps. You cannot A/B test your way to a machine that changes an industry. You have to ship, stand next to customers while it works, and take responsibility for the outcome.

We have deployed 500 robots into real businesses. It took five years to go from 0 to 100. We went from 400 to 500 in four months.

We are sharing a non-exhaustive list of what we learned in hopes that it helps others think clearly about robotics. It is an important and still nascent industry, and our vantage point is unique because we build, deploy, service, and scale robots in the field every day, from Charlotte, NC.

Our three fundamental beliefs about the future

1. Robots are the future of dull, dirty, dangerous, and demeaning jobs. Robots increase human dignity by removing work that ruins bodies and shortens careers.
2. Robots must be built in the United States. A domestic supply chain is essential for economic resilience and national security.
3. Robots plus automation unlock human abundance. Everything that moves will be automated over time.

1) Master one job, design for expansion

Reputation is currency in the trades. You earn it by mastering one task first and becoming the reliable tool crews ask for. That first win proves you can deliver a robot people trust and businesses rely on.

Think LEGO, not Swiss Army knife. A Swiss Army knife can technically do many things, but none as well as a purpose-built tool and it rarely holds up in industrial or professional use. Build a common brain and stable base with standardized connection points, then add specialized payloads only when the job requires them. Architect for expansion while you execute, so new capabilities click into the same system without a ground-up rebuild.

This is the NVIDIA arc: start with a concrete performance problem in graphics, build a repeatable architecture, then extend the same core into new workloads and markets. Robotics should follow the same pattern: master one job, keep the base constant, and add specialized payloads and software as new bricks.

Generalist robots that are mediocre at everything always lose to robots that are great at the task at hand in real commercial and industrial settings. I don’t care what the headlines say. Adoption in B2B robotics always comes down to ROI.

Marketing doesn’t win. Proof and clear ROI in the field wins at scale.

2) Focus on transformative ROI

Marketing gets you a meeting. Proven outcomes get you adoption. In serious commercial settings, pilots are auditions, not victories.

Our bar

• Up to 5x faster. A radical shift in how the work is done, not an incremental improvement. Customers need a compelling reason to switch from the status quo.
• We emphasize growth and efficiency over simple labor reduction: throughput per crew, revenue per day, job cycle time, consistency, and rework avoided.
• Proof comes from production jobs, not demos.

Sidebar: Why B2B adoption beats B2C

Household robotics has allure, but adoption lags even for products that work and are affordable. Quick show of hands: what percent of your neighbors run robot lawn mowers? They have been around for years, often cost under $1,000, and still see low penetration. In B2B, the buyer has an active P&L and adopts when the job gets done faster, safer, and more profitably. That is where the next wave of robotic adoption will compound.

3) Own the customer relationship

Integrate hardware, software, training, service, and parts so you own the outcome. That earns trust and creates a distribution moat through workflow integration, training, and the internal playbooks your customers build around your platform.

Owning the outcome also makes you faster. If you do not own repairs or training, every product update creates drag and cost. Control the loop, iterate quickly, and keep customers happy through change.

4) Control the full stack: software is not enough

Early advice we heard: focus on software and let someone else handle hardware. That is the quickest way to be three layers removed from your user and to lose control of reliability and iteration.

Hard take: If you want to sell to customers who use the product every day, outsourcing the physical stack is the dumbest thing you can do. You will be slow to fix issues, slow to improve, and you will not own the future of your product.

Why full stack control wins

• Reliability and safety. You control tolerances, power, thermals, EMI, sensing, and failure modes. That is where uptime is won.
• Iteration speed. Hardware, firmware, and software evolve together. You ship coordinated updates instead of waiting on vendor roadmaps.
• Serviceability and telemetry. Consistent interfaces and LTE connected diagnostics make field support fast and predictable.
• Independence. You are not gated by third party repair loops or parts bans. You can fix defects and improve the product on your schedule.

Owning the stack is not about perfection. It is about earning the right to improve fast in the real world.

5) Reliability and customer success are non-negotiable

You win only when your customer wins. Customer wins and reliability are your marketing. Measure it and organize the company around improving it.

How we measure progress

• Our north star is Robot Operating Minutes (ROM). More minutes doing productive work means customers are winning.
• We track the ratio of operating hours to customer issues and work to improve it continually.

Most customers do not expect perfection with frontier tech. They expect visible progress, quick action, and clear communication when something goes wrong.

6) Robots are learning and data platforms

A production robot should both do work and produce a stream of high quality operational data. This data improves autonomy, reduces downtime, informs preventive maintenance, and opens software revenue. Tesla is a useful parallel: large scale deployments created the driving data that now powers Full Self Driving. The same logic applies in industrial robotics.

The goal is not to simply sell data. Use it to deliver better outcomes and, where appropriate, monetize the insight stream.

7) Supply chain is strategy

You cannot service customers if you cannot build and repair machines. Component choices, vendor depth, domestic availability, and test coverage matter. Invest early in:

• Second sources for critical parts
• Clear QA and QC at incoming, in process, and final assembly
• Bills of materials that are stable and serviceable in the field
• Spare parts logistics that match your installed base

A great supply chain turns reliability into a habit, not an act of heroism.

8) Build in the USA: resilience, security, velocity

Building domestically is not about symbolism. It is a performance choice.

• Economic resilience. Domestic sourcing shortens lead times, reduces single points of failure, and avoids customs and port delays that stall revenue.
• National security. Critical infrastructure depends on trustworthy supply chains. A U.S. based stack reduces exposure to geopolitical shocks and policy whiplash.
• Data privacy and control. When hardware, firmware, and fleet data are under domestic control, compliance is simpler and customer risk is lower.

Service speed is the most visible difference. Before we built our own systems, we relied on international technology. Repairs routinely took months. With DJI, our average turnaround was about 4.5 months, and some units never returned. International shipping and queue times erased revenue opportunities. When a robot is a P&L asset, a quarter without it is unacceptable. Domestic design, parts, and repair centers mean fixes in days, not months.

This shift is happening across categories. For products that matter to operations, domestic build is becoming the default expectation.

9) Operations are as hard as the tech

Standing up a facility, production line, vendor management, technician training, and repair flows is real work. Scalable hardware powered by software and electronics is even harder. The best way to deliver a scalable robot is to focus relentlessly on quality. Treat service as a product. Design for maintainability: fast access to wear parts, diagnostic telemetry, and procedures that a field tech can run in minutes. Connectivity is critical. If you send robots into the wild without LTE, your ability to troubleshoot and support customers will struggle. Ubiquitous connectivity is how you manage a growing fleet.

Robots are like cars. Even with a century of learning, cars still need tires, oil, and brake pads regularly changed. Your robots will need planned maintenance and unplanned repairs. Design the machine and the org for that reality.

10) Training beats turnover

Many industrial customers have high crew turnover. If your training assumes a stable team, the deployment will stall. Make training easy, accessible, and scalable so a new operator can become productive quickly and safely.

11) Field time compounds

One hour on a job site is worth ten in the office. You will learn more standing with a customer than you will in any simulator. Ship, observe, fix, and repeat. The feedback loop is where moats form.

We require every employee to attend at least one job site per year, regardless of role. If you are in engineering, customer success, sales, and adjacent teams, it is many more. Field time sharpens judgment and keeps the whole company grounded in customer outcomes.

12) Move fast to learn, not just to move

Speed without learning is noise. Move fast when you can extract a new truth from the field. Slow down when changes stop teaching you.

A simple guardrail: every change ships with a hypothesis and a metric. If you cannot write both in a sentence, you are guessing.

13) Solve for problems, not features

Customers change behavior when you change their economics. Focus on big, frequent, painful work and aim for step changes, not marginal gains. Always anchor on the customer problem. Robots that do not solve real world problems are useless. Features that impress in demos but fail on job sites do not matter. Keep asking: does this change an outcome the P&L can feel?

14) Be inevitable, not just innovative

Pick markets where adoption is a matter of when, not if: labor shortages, safety exposure, and chronic inefficiency. Show up with a working product, a clear ROI, and a plan to support it for years. That is what leadership looks like.

15) Build the team you can trust, then listen to them

Robotics is a team sport. Hire people who take ownership, who will meet you at midnight at a customer site, and who tell you the truth

The next decade will be decided by deployed minutes, not demos or press cycles. Our north star is Robot Operating Minutes because minutes compound into learning, reliability, and customer trust. That is how Tesla turned fleet miles into Full Self Driving and how NVIDIA turned a graphics engine into a compute platform. Robotics follows the same pattern: specialize, deploy, learn, expand.

We build in the United States because resilience matters. We focus on B2B because outcomes are measurable. We design like LEGO so new payloads click into a proven base. We own the outcome so iteration is fast and support is real.

If you are building, chase customer outcomes over features. If you are buying, demand radical performance, not incrementalism. If you are investing, back teams that ship, service, and learn in the field.

Robots are the arms and legs of AI. Done right, they raise productivity, improve safety, and increase human prosperity. We are just getting started. See you at 1,000.

‍