Issue 113 - Simscale's Engineering AI Report, Tesla's Terafab, and MachineMetrics' Factory Experiments
Breaking the Bottleneck | 3/30/2026
Breaking the Bottleneck is a weekly newsletter and interview series covering manufacturing technology and physical AI. Want to chat? Reach out at aditya@machinafactory.org or connect with me on LinkedIn.
This newsletter is brought to you with support from our partners: Jiga, Industry 4.0 Club, Ironloop, and T51.
Simscale’s State of Engineering AI 2026
Simscale released The State of Engineering AI 2026 surveying 350 engineering leaders across the United States, the United Kingdom, and Germany. Below were some of the key takeaways:
Agentic AI adoption is concentrated in a few key areas: simulation and CAE (19%), followed by design and CAD (11%), and requirements engineering (10%).
92% of respondents report deploying surrogate models to accelerate simulation or design exploration workflows in 2025
AI simulation has had a profound effect.
On average, simulation requests take 17 hours with conventional workflows, compared with 6 hours with AI-enabled workflows. This is close to a 3× faster turnaround.
Furthermore, AI-enabled workflows are associated with one daily design iteration being tested in simulation, compared with one design every 2.5 days using conventional workflows.
Companies are now assessing 56 design variants per program, compared with 17 variants with conventional workflows.
Secure data governance and access controls (70%), clear cross-functional ownership for moving pilots into production (65%), and an executive mandate with a defined budget (56%) are key drivers for scaled AI adoption.
The survey identifies three primary blockers: data preparation/availability for AI (74%), governance and compliance concerns (48%), and software interoperability challenges (42%).
Check out the full report here!
Content I Enjoyed Last Week 🗞️🔬 📚
Some Interesting Reads:
McKinsey also dug deep into Engineering AI, interviewing 24 senior simulation leaders. Early adopters are already reporting up to 50% reductions in time to market, 15-60% improvements in product performance, and 30-50% gains in workforce productivity. One of the more interesting tensions in the report is around tooling, where 60% of interviewees expect a shift toward best-of-breed modular tools over monolithic software suites, though a vocal minority argues that workflow simplicity will keep companies anchored to comprehensive single-vendor platforms. [McKinsey]
A breakdown on the importance of REE. [Deloitte]
GM has built software that takes hand-drawn sketches from multiple angles and assembles a complete 3D model of the vehicle [Business Insider]
Fast Company revaled it’s most innovative manufacturing companies for 2026, with a few friends included. [Fast Company]
Apple is expanding its American Manufacturing Program, or AMP, with new programs with Bosch, Cirrus Logic, TDK, and Qnity Electronics to manufacture essential materials and components in the U.S [CNBC]
Another $1B in funding for Physical Intelligence [TechCrunch]
Elon is building a new Terafab facility near Tesla’s Austin headquarters, which is an absurdly hard problem, but excited to see how he’s able to reinvent the process. [TechCrunch]
An interesting transition from SES AI from advanced lithium metal batteries for major industries like electric vehicles to an AI materials discovery platform [MIT Tech Review]
Some examples of ways automotive manufacturers are accelerating smart factory investments [Automotive World]
The Helium crunch looming over the global economy [NYTimes]
Products & Research:
MachineMetrics ran a pretty neat experiment with five discrete manufacturers to build solutions that close the operational gaps that standard MES software never quite addresses. Harvey Performance built a lights-out scheduling dashboard that calculates dynamic setup times. Flexco built a forklift dispatch system from scratch. Johnstech’s build surfaced a decade of unprocessed tool-change data already sitting in the MachineMetrics platform, retroactively filling 90 days of tool-life history with no new hardware, while also closing a 10-to-15-year gap in planned-versus-actual production-time visibility via a live ERP integration. [Machine Metrics]
A great robotics market map and breakdown on the efficient frontier, the intelligence stack, and the robotics data problem from Jonathan Healy. [Cathay Innovation]
Augury and MaintainX partner to drive closed loop maintainence. When a machine anomaly is detected via Augury, a work order is automatically created in MaintainX with recommended actions, diagnostic context, and supporting data, so technicians can act immediately. [MaintainX]
Cognite adopts NVIDIA’s NV-Tesseract family of models to operationalize foundational forecasting models with early success at Celanese.[Automation.com]
WORC launches next week, check it out here! [emm0sh on X]
Blogs / Threads
An awesome breakdown of the ramifications of the Iran War on Supply Chains and the Industrial Economy fromDefy the Odds.
A really interesting essay from Burhop arguing that manufacturing is following the same AI adoption arc as software development, but will face more complex conditions that will make the transition slower.
Finance & Transactions 💵
Doss raised $55M from Madrona and Premji Invest
Manufacturing AI / ERP • Series B• USA
Normal Computing raised $50M from Samsung Catalyst Fund
Semiconductors • Seed • USA
Entrix raised €43M from Junction Growth Investors and Korys
Battery • Seed • Germany
Sift raised $42M from StepStone
Data / AI • Series B • USA
Planned Downtime 🧑🔧
Fred Again in London
Breaking the Bottleneck is brought to you with support from our partners.
Jiga - Is a platform that connects hardware teams with vetted manufacturers for custom parts. Instead of black-box instant quotes, you see exactly who’s making your parts and can talk to them directly.
Industry 4.0 Club - Leading cross-functional group accelerating the adoption of Industry 4.0, delivering better experiences to consumers, better profits to manufacturers, and better jobs to factory workers.
Ironloop - Ironloop is a modern OT/ICS software platform that delivers configuration-level intelligence, giving manufacturers ground-truth security visibility and compliance confidence across air-gapped networks.
T51 - T51 is an AI-native planning, scheduling, and optimization platform that integrates securely with your existing ERP, MES, and data stack.