5YF Episode #23: Geminus CEO Greg Fallon

Future of Industrial AI: Simulating The Real World

subscribe and listen today

Today, we explore heavy industry’s embrace of AI to reimagine the physical economy.

ChatGPT was trained on 45 terabytes of data. A single oil field produces that in 9 days

The amount of data within an industrial conglomerate is equivalent to the amount of data on the internet. This realization is sparking a new wave of awareness among industrial leaders: in an era dominated by data-hungry AI models, they are sitting on vast, untapped digital oil fields. The convergence of proprietary operational data, AI models capable of becoming expert systems, and a renewed drive for modernization is fueling a revolutionary shift among industrial giants.

In this transformative landscape, we focus on one of the leading AI startups, Geminus, whose specialized models are poised to revolutionize industries ranging from energy to semiconductors. At the helm is CEO Greg Fallon, a veteran who has shaped the product strategies of major industrial software companies like Autodesk and ANSYS. Now, he aims to disrupt the very industry he helped build.

My 5 Year Outlook:

• Industrial giants will monetize their digital infrastructure: Following Amazon’s AWS playbook for immense profit.‍
• Generative Engineering will deliver heavy industry’s ChatGPT moment: Physics-constrained expert AI models trained on immense volumes of proprietary data.‍
• Our physical world will be precisely simulated: Digital twins grounded in the laws of physics.

Curious? Read on as I unpack each below 👇🏼

Industrial giants will monetize their digital infrastructure

Saudi Aramco announced a multiple tens of billions of dollars investment fund for digital technologies. They’re working to make a digital twin of their entire operation from the ground up.

Industrial giants are beginning to recognize the untapped value of the digital infrastructure they’ve developed in-house, much like Amazon did with AWS, the cloud infrastructure powering its eCommerce website. Initially built to support their own complex operations, these digital tools—ranging from advanced AI models to digital twin platforms and massive data systems— will be repurposed over the coming years and monetized as external offerings. By selling access to this infrastructure, these companies are creating entirely new revenue streams and marking a new chapter for the industrial economy.

Following Amazon’s playbook, where internal systems evolved into AWS and now contributes to $1T of the company’s market capitalization, industrial giants are poised to commercialize their proprietary tech and become pivotal players in industrial software. This shift transforms them from pure industrial manufacturers into platform providers, capable of selling cutting-edge digital solutions to other industrial companies seeking modernization.

Several factors uniquely position industrial giants for this digital renaissance: (1) complex and repetitive technical operations, (2) a workforce of highly trained and well-paid experts, (3) vast amounts of relevant data, (4) significant capital reserves, and (5) a pressing need to evolve beyond resource extraction and manual labor.

While it may seem obvious that a standalone industrial software company could emerge in this space, industrial companies are reluctant to share their data, viewing it as proprietary IP. Instead, this creates an opportunity for them to share a software product itself, safeguarding their data while still capitalizing on technological advancements.

Saudi Aramco, the world’s largest industrial company, offers a compelling first case study. Through Aramco Digital, they aim to transform their operations and modernize the broader Saudi industrial economy, before expanding globally. Backed by billions from their oil revenues, Aramco Digital is building a robust digital infrastructure focused on four key pillars: cloud, AI, connectivity, and security. Already, they have an ecosystem of 200 companies as both customers and solution providers. Saudi Aramco’s visionary strategy leverages its capital, talent, data, and deep expertise to create a digital future for its oil-dependent economy.

The next five years will see the rise of first movers looking to capitalize on exporting their digital expertise to peers across sectors. Mining giants like Rio Tinto and BHP, who already manage increasingly autonomous mines from hundreds of miles away, exemplify highly advanced organizations with the potential to modernize their entire industry. It’s not far-fetched to imagine these companies remotely operating mines in other countries. From mining and energy to agriculture and manufacturing, the wave of digitization is rapidly approaching in what could market a new industrial revolution.

Greg Fallon, CEO of Geminus

Geminus is a company at the forefront of simulating and automating industrial operations through cutting-edge AI technology. Geminus is revolutionizing how industries operate by bringing unprecedented speed and scalability to AI model deployment, with successful implementations in energy, oil and gas, manufacturing, and semiconductors. With heavyweight investors and partners like SLB and LAM Research backing them, Geminus is a key player in the digital industrial landscape. As a proud early investor in Geminus almost six years ago, alongside our friends at The Hive, I’ve had the privilege of watching this company innovate the industrial complex.

Greg Fallon is CEO of Geminus, he holds a Master’s in Science from the University of Virginia, has a wealth of experience, having held senior executive roles in product and commercialization at both Autodesk and ANSYS—two of the most influential companies in industrial software.

Generative Engineering will deliver heavy industry’s ChatGPT moment

While ChatGPT has captured the imagination of knowledge workers across various sectors, its value in the industrial world has been limited. ChatGPT is a general-purpose model, trained on a vast but generalized dataset—the entirety of internet data. However, what it lacks is the specialized expertise and precision required for deployment in industrial settings, where the stakes for accuracy are far higher.

The machine learning model for the physical world, i.e. for power plants, the requirements for accuracy and precision are very different from what we’d have for using ChatGPT.

Fortunately, industrial companies possess the specialized data and expert talent necessary to train highly accurate models. This is where companies like Geminus come into play, developing AI systems that can be rapidly scaled across organizations—often in weeks, not months or years.

The key to these models’ reliability lies in grounding them in the laws of physics, an approach Geminus calls AI-Augmented Computational Physics. By incorporating real-world constraints, these models can provide highly accurate simulations, such as predicting how a drill or pump will perform in specific geological conditions.

In addition to physical accuracy, these specialized models must excel in multi-step reasoning and problem-solving, a focus area for the AI industry. Recent advances, like ChatGPT-01, deploy AI agents that break down complex tasks into smaller, more manageable segments, collaborating to solve intricate problems with greater precision.

Geminus combines these two approaches—physics-constrained AI models and agent-based orchestration—to pioneer what it calls Generative Engineering, a nod to OpenAI’s Generative AI. The goal is to automate complex engineering tasks and provide an expert system that assists engineers and scientists. This ambitious vision could propel the industrial sector into a new era, but first, it must earn the trust of the industry’s key players.

Our physical world will be precisely simulated

Grounding AI models in the laws of physics will revolutionize not only the industrial economy but also how artificial intelligence interacts with and simulates the physical world. By embedding real-world constraints into AI, we create a more accurate digital representation of the environments in which these systems operate. This approach enables the creation of advanced simulations and digital twins—precise virtual replicas of physical systems—paving the way for AI to make decisions based on reliable, real-world context. With these foundations, AI systems can not only simulate physical processes but also predict outcomes with unprecedented accuracy.

This real-world context is critical in fields like autonomous driving and robotics, where AI must navigate and respond to complex physical environments in real time. Equally, it plays a crucial role in less obvious applications, such as space exploration and weather forecasting, where precise environmental modeling is essential. By ensuring AI understands how physical forces and laws impact operations—from friction and inertia to gravitational fields—we significantly improve the performance and safety of AI systems deployed in these demanding environments.

As AI continues to evolve, grounding it in first principles will become the key to unlocking more reliable and capable systems across industries. Whether orchestrating a constellation of satellites in deep space or forecasting increasingly extreme weather patterns, these physics-aware models will offer greater accuracy and trustworthiness.

The ability to simulate the real world with such precision is not just a technological leap—it’s a fundamental shift in how we apply AI to the physical challenges of the future.

Onwards!

‍