Senior People Operations Specialist

Nscale has agreed to supply at least US$3.5 billion worth of AI cloud computing capacity to robotics startup Figure, whilst making an undisclosed investment in the company. The London-based cloud provider will use Nvidia chips at a Texas site to fulfil the contract from the second half of 2027, with potential to scale beyond US$6 billion. The deal mirrors Nvidia's strategy of investing in customers and partners. Figure, founded in 2022, develops humanoid robots and raised over US$1 billion last year at a US$39 billion valuation. Nvidia is an investor in both Nscale and Figure. Nscale is reportedly seeking to raise up to US$3 billion in a US IPO, potentially as soon as September.

Figure launches Index, paying people to film everyday chores. Figure's Index app has paid $15M to 264,000 users filming household and workplace tasks, building training data for its Helix robotics stack. August 26, 2026 by HowAIWorks Team Introduction. On August 25, 2026, humanoid robotics company Figure came out of stealth with Index, a smartphone app that pays ordinary people to film themselves doing everyday chores. The company says it has already paid $15 million to users after four months of quiet operation, and that it intends to spend more than $1 billion on data and compute over the next 12 months. The pitch is a straightforward bet on scale: the physical data a general-purpose robot needs is not on the internet, so Figure is buying it directly from the people who generate it. What Figure has collected. The numbers Figure published for its four-month stealth period: * 264,000 app downloads across 108 countries * 44,000+ weekly active users, whom Figure calls Creators * 16 million videos uploaded * 30 minutes of video ingested every second - what Figure describes as 4.9 years of human work uploaded per day * $15 million paid out to Creators to date Figure also reports a diversity measure rather than a raw volume one: per 1,000 hours collected, the dataset contains 373 unique tasks, 1,146 unique manipulated objects and 116 unique environments. That framing matters more than total hours, because a million clips of the same kitchen teaches a robot very little. The tasks range from household work - cooking, cleaning, laundry - to commercial settings including logistics centers, restaurants, factories and offices. Figure cites cleaning kitty litter, changing oil and busing restaurant tables among the more obscure submissions. Creators can record in their own home or workplace, and the app also lets users book a Creator to come to their home or business and do chores while filming. Why buy data this way. Figure says it tried the conventional route first: "Prior to this we tried buying data. Vendors couldn't hit the throughput, diversity, or quality bar Helix requires, so we built the pipeline ourselves." The underlying claim is that generalization in robotics is a data problem before it is an architecture problem - the same scaling laws story that played out in language models, applied to physical tasks. Crowdsourced human video is one answer to the embodied AI data shortage; synthetic data from simulation is the other, and most labs use both. The long tail is the argument for crowdsourcing specifically. As Figure puts it, every new Creator brings "an unseen environment, unfamiliar objects, and their own idiosyncratic way of completing a task" - variation that is nearly impossible to specify in advance in a lab or a simulator. Ingesting consumer video at this rate forced Figure to rebuild its data infrastructure around consumer-app constraints. The processing pipeline has five stages: automated filtering for technical, visual and semantic quality; human fraud review auditing users for deliberate evasion; embedding-based deduplication that discards clips too similar to accepted data; rebalancing against task quotas and embedding clusters; and finally hierarchical text captioning of every episode. What is not in the announcement. Figure says the generalization results from training Helix on this data are "already validating this thesis" and that it will share more soon. That is the load-bearing claim of the whole post, and it arrives with no benchmarks, no evaluation protocol and no comparison against models trained without Index data. Until those appear, the verifiable part of the announcement is the collection operation, not its effect on robot performance. Two other gaps are worth noting. First, human egocentric video is not robot demonstration data: it lacks the joint states, forces and action labels a policy ultimately needs, and bridging that embodiment gap is an open research problem rather than a solved preprocessing step. Second, a global network of paid contributors filming inside homes and workplaces raises consent and privacy questions - for bystanders and employers as much as for Creators - that the announcement does not address. Figure frames Index as "laying the groundwork for ordering robots as a service," with today's human Creators as a placeholder for tomorrow's machines. That is a long way from 16 million phone videos, but the funding commitment behind it is specific: 100x the current collection rate, and over $1 billion in the coming year. Conclusion. Index is the largest public attempt yet to solve robot training data by paying consumers for it, and the operational numbers - $15 million out the door, 30 minutes of video per second - are real and substantial. Whether the data translates into robots that generalize is the question Figure has promised to answer and has not yet answered. The next release, with actual Helix results attached, is the one that will settle it. Sources.

The 370-billion-dollar race: How integrated design can help humanoid manufacturers succeed in a rapidly growing market. Imagine a manufacturing floor where robots don't just repeat the same motion thousands of times but adapt to new tasks. They're able to navigate cluttered spaces, manipulate unfamiliar objects and switch between tasks as fluidly as human workers. They can climb stairs, squeeze into tight assembly areas and collaborate seamlessly alongside people in environments designed for human bodies. This is the future of humanoid robotics, and it's closer than most people realize. The implications are staggering. Humanoid robots could revolutionize manufacturing by bringing unprecedented flexibility to production lines. Unlike traditional industrial robots bolted to factory floors, humanoid robots can move between workstations, handle diverse products and adapt to changing production needs without expensive retooling. They could transform warehouses, logistics operations and hazardous environments where human workers face risk. The market agrees. In their report "The future of robotics: Intelligent, adaptable, and on your team," McKinsey reported that while the general-purpose robotics market is valued at under $1 billion today, if progress continues at the current rate, it could reach a value of $370 billion by 2040. Major players from Tesla to Figure AI are rushing to commercialize humanoid platforms. However, due to the sheer complexity of these robots, significant hurdles remain before they are broadly commercially adopted. Bringing to life the most complex systems ever designed for mass production. The technical challenges of building humanoid robots are immense. An average humanoid robot has dozens of articulated joints which must be carefully coordinated with precise mass, inertia and center of gravity control. Even the slightest change in structure can lead to cascading impacts on balance, gait, reach and energy efficiency. To effectively navigate these design hurdles, manufacturers need sophisticated software tools that seamlessly integrate design and simulation. Unfortunately, while most manufacturers have mature computer-aided design (CAD) processes, their overarching workflows were built for a completely different environment and era. One of the key challenges is that most humanoid programs still lean on disjointed methods to bring mechanical designs into robotic simulation and control environments. On many teams, kinematic models are still rebuilt by hand, with key variables like mass and inertia approximated rather than derived from real geometry and joint definitions. This causes drift between design and control representations, which compromise simulation accuracy and lead to early freezes of mechanical designs to avoid complicated downstream reworks. And while virtual prototyping can save millions for manufacturers, it's only a viable option if those simulations stay perfectly synchronized with evolving designs. For manufacturers racing against time and competitors, overcoming these inefficiencies will be key to staking out broader market share. To create viable, fully functioning humanoids for the factory, manufacturers need a solution that eliminates systemic design obstacles associated with such disconnected systems. Integrated design for integrated machines. Siemens now offers an exciting path forward for addressing this Gordian knot of design complexity. With Siemens Xcelerator, humanoid manufacturers can connect mechanical design, motion simulation, PLM, manufacturing and service so your team can move from concept to deployed robot faster, all on one platform. In Siemens Designcenter, humanoid robot manufacturers can leverage native URDF support, which allows for exporting URDF models from a CAD platform with a single click, while mapping links, joints and kinematic hierarchies. These URDF models are compatible with PhysX-based simulation engines. With this unparalleled design simulation interoperability, manufacturers can perform earlier validation of balance, gait and reachability, reducing reliance on costly physical prototypes. Altogether, Siemens Xcelerator creates a unified digital thread connecting every workflow from mechanical design, electrical systems, simulation and validation to software development and manufacturing planning. Design changes propagate automatically. Simulations stay synchronized. Teams collaborate in a shared digital environment globally and seamlessly. With these capabilities, teams can move beyond "design it right once" to "design it right, train it right and deploy it right - every time." The window is closing. The humanoid robotics market is entering its defining phase. The companies that reach production-ready, commercially viable robots first will capture enormous market share. Those that get bogged down in development inefficiencies risk becoming footnotes in someone else's success story. The technical challenges are formidable, but they are solvable. The real question is whether your development process can keep pace with your ambitions and your competition. In a race measured in months, not years, the teams that eliminate design bottlenecks will cross the finish line first. Download its e-book From design concept to robot intelligence to learn more about how Siemens Designcenter enables one-time-right humanoid design. Sponsored content by Siemens Industry Software