Breaking the Bottleneck | The Next Year in Manufacturing
2025 Edition
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!
The Next Year in Manufacturing🏭🗞️🔬 📚
For AI Applications and Adoption…
Practical application of industrial AI will grow rapidly in 2025, moving beyond more basic asset health to process optimization, real-time operator guidance, and co-pilots to help create and configure automation and user interface applications.
- Rick Bullotta, Former Co-Founder of Thingworx, CTO of Wonderware
Forbes Magazine has identified 2025 as the "year of AI Agents." But it’s not enough for an AI Agent to answer Natural Language like LLMs can in the industrial space. In the industrial space, we need agents who are capable of causal reasoning, a critical skill that LLMs lack, according to a recent study by Apple. Instead, we'll need to look to technologies like Autonomous AI to provide AI Agents that can observe the plant floor, interpret complex situations, and then act in the same way humans do. Additionally, AI will play a crucial role in preserving expert knowledge, capturing decades of experience before it’s lost to retirement.
- Bryan DeBois, Director, Industrial AI RoviSys
I'm most excited about the second category: AI micro-agents that watch CAD and PLM data, providing insights, proactively simulating parts, and de-risking value chains. At Gradient Control Laboratories, we're contributing a reference architecture and implicit modeler for differentiable engineering that lets emerging hosts and AI/ML agents form enterprise-scale ecosystems.
- Blake Courter, Former CTO of NTop and MD of Gradient Control Labs
In 2025, manufacturers will ramp up the deployment of vision systems that can reliably understand the majority of tasks being done on the shop floor. Those who combine causal learning software, automation, and visual intelligence will benefit significantly.
- Cyrus Shaoul, CEO of Leela AI
Advances in physical AI and robotics will make it undeniably clear to hardware OEMs that they will be left behind (nationally and globally) unless they accelerate toward AI-embedded manufacturing. Hardware OEMs will start this journey using enterprise AI powered by language models. Still, they will quickly realize the importance of cleaning up their digital design and manufacturing data to prepare for an AI-first tech stack with new design, production, lifecycle management, and supply chain resilience capabilities.
- Sai Nelaturi, CEO of C-Infinity
One of the significant changes in 2025 will be from AI visual reasoning. We will see a transformation in efficiency and compliance on the factory floor through breakthroughs in visual reasoning models. The technology is highly accessible and will provide just-in-time insights via mobile devices that will be an exponential improvement over the more conventional static camera-based analytics already deployed (i.e., ergonomic support, hazard detection, etc).
- Devin Bhushan, CEO of Squint
The development of AI and robotics will continue to dominate 2025. The pressure on C-level manufacturing to drive investments in AI will lead to inflated expectations and to many new startups that pitch “AI co-pilots for manufacturing,” of which only a few will flourish. Manufacturing leaders will realize they must work on IT/OT infrastructure and data harmonization before benefiting more from AI.
- Robin Dechant, Angel Investor & Founder of the Manufacturing Newsletter
As we increasingly use large language models, we're beginning to recognize the value of incorporating explainability and subject matter expertise into their algorithms. For instance, these advancements allow us to revisit machine learning with SMEs, providing crucial situational awareness and context. The growing buzz and exponential development of talent, tools, and attention in this field will lead to genuine ROI and the adoption of more focused, explainable AI tools in 2025. This shift will occur as users realize the necessity of delivering measurable value, especially in hardware applications.
- Alex Reed, President and COO at Yokogawa Fluence
Manufacturing will rapidly enter Industry 5.0 not through disruption but by quietly embedding its principles using AI agents. Agentic workflows will automate time-consuming human-machine and human-human interactions, unlocking efficiency. Large manufacturers will leverage internal resources to automate processes and showcase case studies where agentic workflows will save millions of dollars with simple automation and unstructured data. This shift will become a mainstay of domestic manufacturing efforts – AI will be a part of the story for nations prioritizing their global industrial position.
- Karan Talati, CEO of First Resonance
Organizations that have started their digital transformation ahead of the curve will see significant benefits. These pioneers can expect 10-15% productivity gains for industrial workers, alleviating the technical labor shortage through AI copilots and comprehensive automation of repetitive tasks via AI agents. Such companies will set the standard, leveraging industrial AI to enhance asset uptime, improve quality, boost productivity, and drive sustainability.
- David Hahn, CEO of Remberg
For M&A and the Fundraising…
Mergers and acquisitions will pick up significantly in the industrial sector in 2025. Still, quite a few transactions will be disappointments for investors and founders who raised too much money at inflated valuations over the past couple of years. However, good companies will get fair value. Humanoid robotics will disappoint.
- Rick Bullotta, Former Co-Founder of Thingworx, CTO of Wonderware
With a modest reduction in interest rates expected in 2025, corporate spending should be increased on industrial technology and M&A. I’d expect further consolidation up and down the manufacturing software stack. However, some smaller-point solutions will fail.
- Brad Hafer, MD at Minuteman Advisory, VP of Corp Development at Plex
For Design & Simulation…
High-fidelity, first-principles-driven simulations will play an increasingly important role in industrial applications. They’re needed to provide training data for AI, which can be used to optimize designs, train staff, and more.
- Rick Bullotta, Former Co-Founder of Thingworx, CTO of Wonderware
Engineering analysis tools use AI to produce near-instantaneous results and farming data from previous analyses. The cloud and the use of GPUs have radically accelerated CAE performance. Now, AI will speed up understanding of how design variants affect product performance.
- Jon Stevenson, Former CTO of Stratasys, VP at GrabCad & PTC
2025 should be a year of clarity in engineering software. Due to the emergence of new interop libraries and standards, we will see more industrial applications of implicit modeling. The first crop of AI engineering software will be separated into two main categories. Those that accelerate workflows for existing markets, like ML-accelerated simulation, will become features in existing products.
- Blake Courter, Former CTO of NTop and MD of Gradient Control Labs
I'm most excited about non-LLM models improving design, simulation, etc., beyond the current text-centric LLMs, but those will still be betas in 2025. I expect to see LLMs commoditizing workflows across growth, operations, and data, but we won’t be at full auto, so successful apps will have human users and/or staff to fill the gaps. Adoption will remain the biggest bottleneck.
- Nick Pinkston, CEO of Volition
The current market for AI regarding electronics design is about to turn the corner. We’ve had enough time for companies to take a crack at building tools and gain early customer adoption. 2025 is the year the rubber meets the road for this first wave of electronic design AI companies that have been building over the last few years, and we will see if there truly is value in what has been built to accelerate design workflows and reduce risk.
- Ethan Pierce, Founder of Dodec Labs
Dassault is struggling, opening an unprecedented window of opportunity for other players to replace CATIA at scale. AI opens up new frontiers in simulations (but it’s early).
- István Csanády, CEO of Shapr3D
For Platform Adoption and Data…
In 2025, manufacturing companies will increasingly adopt low-code and no-code platforms to accelerate the customization of their software ecosystems and address unique challenges. Digital Thread solutions will become the backbone of manufacturing, enabling seamless data flow across the product lifecycle and improving collaboration between teams and systems. AI-driven automation will enhance decision-making in real-time, from supply chain optimization to predictive maintenance, driving efficiency and reducing costs. Additionally, the shift towards servitization—offering hardware as a service—will push manufacturers to prioritize software innovation and user-centric design to differentiate in competitive markets.
- Igal Kaptsan, SVP of Product at Aras, Former VP at Dassault and GE
Manufacturing will continue to digitize. Simplified digitized tools with lower entry points (Cost, Ease of Deployment, etc.) will gain more traction as awareness of next-generation technologies is disseminated. AI will be a piece of that puzzle but will evolve towards specific use cases and less GenAI.
- Bryan Bauw, COO at Pico MES
Adoption of cloud-based manufacturing software. Although almost all companies have adopted SaaS applications for other disciplines (payroll, HR, sales and marketing, etc.), adoption of engineering applications has been slow. This has been changing over the last few years but is now accelerating.
- Jon Stevenson, Former CTO of Stratasys, VP at GrabCad & PTC
Lastly, I see that IT and OT convergence is still not happening, perhaps getting worse, as the shop-floor data shop crowd continues pursuing their one-off science projects that don’t scale to transform at the enterprise level.
- Brad Hafer, MD at Minuteman Advisory, VP of Corp Development at Plex
Surprisingly, Apple Vision Pro is very popular and will expand its use this year.
- István Csanády, CEO of Shapr3D
By 2025, many organizations, particularly in the industrial sector, will face the harsh reality that they should have invested in digitalization and cloud adoption for their core processes earlier. To unlock the full potential of industrial AI and achieve a new level of productivity, it’s essential to first address these digital fundamentals. DataOps will be a top priority as manufacturers look to implement AI, with increased investments in best-of-breed MES, edge computing, cyber-physical protection, and the Unified Namespace.
- David Hahn, CEO of Remberg
For the Macro Environment…
The manufacturing industry will navigate a rapidly changing supply chain as the incoming administration wields import and export regulations as economic and diplomatic tools and introduces uncertainty into trade relationships with Canada and Mexico, complicating the nearshoring/onshoring conversation. Organizations will have to balance multi-sourcing more than ever while keeping an acute focus on the unit economics of their products. Meanwhile, tiered supplier relationships will be pushed to offer greater transparency with historically murky dependencies that may have been previously accepted but are now being viewed as threats. Companies that rely on long-term supply chain planning will look for tools to navigate that dynamic supply chain environment in near-real-time to meet their manufacturing requirements.
- Phillip Gulley & Matthew Haber, Co-Founders at Cofactr
Trump’s ‘America-First’ policies will increase input costs, complexity in the supply chain, and fewer subsidies as the Inflation Reduction Act is cut and tariffs are imposed. The American private sector will invest heavily to create a new sovereign manufacturing supply chain driven by booming space and defense technology industries. Given increased headwinds from low-cost EVs, a major automotive manufacturer will either go bust or be forced to restructure, sending ripples across the industry and supply chain.
- Sam Cash, Founder and Investor at Entropy Industrial
The significant X factor for 2025 is whether the Trump admin delivers on (1) their manufacturing promises, as some appointees have deep tech roots, which could make some startups big winners, and (2) how trade war escalations could affect US manufacturer inputs. The latter could create opportunities for marketplaces and manufacturing-as-a-service companies.
- Nick Pinkston, CEO of Volition
The incoming administration will prioritize public-private partnerships to enhance and restore critical manufacturing with an AI focus. This will reinvent modern manufacturing workflows similar to those in software engineering.
- Sai Nelaturi, CEO of C-Infinity
Tariffs and trade policies will be discussed more frequently than sustainability efforts.
- Robin Dechant, Angel Investor & Founder of the Manufacturing Newsletter
Tariffs and “reshoring” campaigns will drive manufacturers to think laterally about their supply chains.
- Bryan Bauw, COO at Pico MES
The changing macro environment (China, US tariffs, energy crisis, etc.) is putting unprecedented pressure on EU manufacturing, making companies consider unimaginable decisions. This accelerates the adoption of new tools. The EU automotive industry is literally fighting for its life.
- István Csanády, CEO of Shapr3D
For Robotics:
Robotics will have its "chat GPT moment" fueled by breakthroughs in open-source models, whereas humanoid robots will disappoint the market.
- Robin Dechant, Angel Investor & Founder of the Manufacturing Newsletter
 | A guest post by
|
AI is finally moving from buzzword to real impact in manufacturing. Love the point about micro-agents having AI that can quietly monitor CAD data and suggest optimizations could save so much time. But totally agree that none of this works without clean data and better IT/OT integration. Feels like 2025 will be less about flashy AI demos and more about the companies that actually fix the boring stuff under the hood.
Will lower-code platforms be enough to close the gap or will the upfront investment still be a major barrier?