Breaking the Bottleneck | Issue 69
[1/13/2024] NVIDIA & Siemens at CES, Lemonade at Chick-Fil-A, AI in ECAD & More!
Breaking the Bottleneck is a weekly manufacturing technology newsletter with perspectives, interviews, news, funding announcements, manufacturing market maps, and a startup database!
💥 If you are building, operating, or investing in manufacturing, hardware, or robots, please reach out. My email is aditya@schematicventures.com – I’d love to chat!
🏭 If you were forwarded this and found it interesting, please sign up!
Content I Enjoyed Last Week 🏭🗞️🔬 📚
News:
Chick-fil-A’s Lemon-Squeezing Robots [Bloomberg]
Chick-fil-A has introduced cutting-edge automation to streamline lemonade production, using a highly advanced facility in Santa Clarita, California. The facility, dubbed Bay Center Foods, employs driverless forklifts, robotic arms, and state-of-the-art machinery to handle nearly every step of lemon processing, from washing and oil extraction to juicing and packaging. This automation has maximized lemon utilization from 40% to almost 100%, with extracted oils sold to the cosmetics and fragrance industries, creating a new revenue stream. The production process, which takes about 45 minutes, is monitored by 120 workers focused on maintenance and quality control. Robotic systems ensure precision, from sealing juice bags to stacking pallets. Innovations such as high-pressure pasteurization preserve the juice’s quality without using heat. The results are up to 1.6 million pounds of lemons daily with minimal human involvement and a reduction in labor-intensive tasks across Chick-fil-A’s 3,200 locations, where squeezing lemons previously consumed 10,000 worker hours daily and resulted in frequent finger injuries. By eliminating this tedious chore, Chick-fil-A aims to improve employee satisfaction, boost efficiency, and address labor shortages as it expands globally. Chick-fil-A’s investment in automation reflects a broader trend in the fast-food industry to address labor shortages and enhance efficiency. Other chains, including Chipotle and Wendy’s, also used automation for tasks like avocado preparation and AI-powered drive-thru orders. Chick-fil-A is considering additional measures, such as pre-chopped salads and premixed ingredients for breading chicken, to reduce in-store preparation further and improve speed.
Forklifts Hurt Workers Each Year, and Factories Seek Alternatives [WSJ]
American manufacturers are increasingly moving away from forklifts in factories and warehouses, driven by a desire to improve safety and productivity. Each year, around 7,500 forklift-related injuries and nearly 100 fatalities occur due to collisions, tip-overs, and other mishaps. This shift aligns with a broader effort to enhance workplace conditions and address safety risks. Retail orders for forklifts dropped by 28% in 2023, the steepest decline in 14 years, signaling a transition away from reliance on these vehicles. Companies such as Ipex, Mercedes-Benz, Tesla, and Whirlpool are leading the forklift-free movement. Ipex’s new North Carolina plant uses overhead cranes and electric pallet jacks, creating a safer and less stressful environment. Mercedes-Benz has been replacing forklifts with autonomous vehicles since 2018, while Tesla has adopted push carts and robotic tuggers to reduce traffic and injuries. Whirlpool has eliminated forklifts in production areas of its Ohio plant, using robotic tuggers to deliver parts efficiently and reduce workplace injuries. The future of material handling may lean toward automation. Companies like Toyota Material Handling North America and Hyster-Yale Materials Handling envision autonomous tuggers and robotic systems replacing forklifts, especially as storage methods evolve from pallets to boxes. However, not all manufacturers are ready for such changes. Forklifts remain indispensable in older facilities or for transporting large, bulky items.
The US Clean Energy Manufacturing Revolution is Real [Canary Media]
Over the past four years, the U.S. has made notable strides in domestic solar panel and battery production. The U.S. has rapidly transitioned from being heavily import-dependent to producing nearly 40 gigawatts of solar panels annually, meeting its domestic installation demand. However, upstream production, such as solar cell fabrication and silicon wafer processing, lags significantly behind, with only about 10 gigawatts of solar cell capacity under construction. The battery industry mirrors this trend, with investments in U.S. battery cell production surging to $95 billion between 2022 and 2023. Lithium-ion battery manufacturing capacity grew from 74 gigawatt-hours in 2023 to a projected 1,133 gigawatt-hours by 2030, aligning closely with anticipated domestic demand driven largely by electric vehicles (92% of demand). However, challenges remain in scaling battery production to meet this demand. Newly constructed gigafactories face hurdles in achieving full operational efficiency due to complex calibration processes, and political resistance has hindered some projects, such as Gotion’s $2 billion Michigan factory. This phased approach to onshoring, prioritizing cell production first, aims to create demand that will eventually spur upstream supply chain investments. While uncertainties remain, including factory delays, political opposition, and consumer hesitancy toward EVs, the U.S. is making significant progress in building a domestic clean energy manufacturing base. The ultimate test will be whether this foundational growth can draw sufficient investments in upstream material production to achieve greater self-sufficiency.
‘Mega’ Omniverse for Building Industrial Robot Fleet Digital Twins [NVIDIA]
NVIDIA announced the Mega Omniverse Blueprint at CES, a digital twin framework designed to optimize and manage large-scale autonomous systems in advanced factories and warehouses. At the core of Mega is a world simulator, which coordinates robot activities and sensor data, enabling real-time decision-making and operational optimization. Using Omniverse Cloud Sensor RTX APIs, developers can render sensor data across all intelligent machines in a facility, testing an infinite range of scenarios in a software-in-the-loop pipeline integrated with NVIDIA Isaac ROS. The first adopter of Mega, KION Group, is collaborating with NVIDIA and Accenture to enhance supply chain automation for industries such as retail, consumer goods, and parcel services. KION uses the Mega platform to create digital twins of warehouses by integrating data from CAD files, lidar, images, and AI-generated insights. These digital twins serve as virtual training environments for robotic systems, simulating tasks like load movement and route optimization. Robots in the simulation perceive, plan, and act within the digital environment, enabling precise tracking and iterative learning for enhanced productivity. Meanwhile, Accenture is leveraging Mega within its AI Refinery for Simulation and Robotics, delivering services such as Custom Robotics Training, AI-Powered Manufacturing Optimization, and Intelligent Humanoid Robotics. Thus enabling organizations to simulate hundreds of operational scenarios before implementing optimal solutions and to adapt to fluctuating market demands, workforce changes, and seasonal requirements.
Why World Foundation Models Will Be Key to Advancing Physical AI [NVIDIA]
NVIDIA’s Ming-Yu Liu, vice president of research, discussed the transformative potential of World Foundation Models (WFMs) in advancing physical AI systems. WFMs are powerful neural networks capable of simulating physical environments and predicting scene evolution through actions like prompts or control signals. A key advantage of WFMs is their ability to generate synthetic data, which reduces the cost and challenges of collecting real-world datasets. By creating rich and varied virtual environments, WFMs provide developers with controlled settings to train and test physical AI systems, minimizing risks and resource demands associated with real-world trials. At CES, NVIDIA introduced NVIDIA Cosmos, an open platform featuring pre-trained WFMs designed to accelerate physical AI development. Built on diffusion and auto-regressive architectures, Cosmos includes tools like tokenizers that compress videos into formats compatible with transformer models. This open-access framework enables enterprises and developers to fine-tune existing models or build new ones tailored to specific applications, enhancing flexibility and scalability.NVIDIA is collaborating with companies like 1X, Huobi, and XPENG to address challenges in physical AI and drive system advancements. While WFMs are still in their infancy, Liu emphasized their growing utility and the need for further integration into physical AI systems. By refining these models, NVIDIA aims to enable safer, more efficient real-world AI applications across industries.
Siemens Announcements at CES 2025
Siemens unveiled numerous new tools for its vision for the convergence of data, AI, and software-defined automation at CES 2025. Central to Siemens’ showcase was the integration of Industrial AI into factory environments, enabling real-time decision-making through tools like the Siemens Industrial Copilot for Operations. This ecosystem enhances human-machine collaboration, optimizing productivity across industries such as manufacturing, infrastructure, and mobility by deploying AI models directly at the edge of the factory floor. Siemens also announced a collaboration with JetZero, an aviation startup developing a sustainable blended-wing aircraft with 50% improved fuel efficiency and a vision for zero carbon emissions by 2035. JetZero will leverage Siemens’ Xcelerator platform for designing, manufacturing, and virtually simulating its aircraft and production processes using comprehensive digital twins. The Xcelerator platform leverages NVIDIA Omniverse for photorealistic visualization within product lifecycle management. This integration facilitates secure, real-time collaboration in a digital twin environment, improving workflows and reducing errors. Next, Siemens announced a deeper partnership with Sony to launch an immersive engineering solution combining Siemens NX software with Sony’s cutting-edge mixed-reality headset. This toolset enables high-fidelity 3D collaboration and product design for the industrial metaverse. Finally, Siemens introduced Designcenter, a unified software suite combining tools like Solid Edge and NX software, allowing businesses of all sizes to collaborate seamlessly using the industry-leading Parasolid modeling kernel. This solution ensures scalability and continuity across design and engineering tasks, making advanced industrial technology accessible to a wider range of enterprises.
Research / Blogs:
Podcasts/Video:
The Blueprint for an American Manufacturing Renaissance
Finance & Transactions 🏭💵
Venture Capital:
Biosphere - A company replacing the 1950's steam bioreactor with a modern, UV-enabled design that reduces the capital intensity of biomanufacturing.
$8.8 million [Seed] - Co-led by Lowercarbon Capital and VXI Capital and joined by Founders Fund, GS Futures, Caffeinated Capital, and B37 Ventures.
Tangible - A company that provides a carbon tracking platform to help builders collect, analyze and use your construction data.
$3 million [Seed] - Co-led by Prologis Ventures and Pi Labs and joined by Foundamental, Silence VC, and RE Angels.
Planned Downtime 🏭🧑🔧
Squid Game Season 3