Beyond Generative AI: The Physical Revolution in Energy and Computing

SINGAPORE – September 10, 2026 – While public fascination remains fixed on generative AI and its ability to create digital content, a quieter, more profound transformation is taking root in the physical world. This week, Singapore-based DeepCtrls announced a significant Series B+ financing round, a move that does more than just inject capital into a promising tech firm. It signals a powerful vote of confidence from global industrial titans in what many believe is AI's second act: the transition from the realm of bits to the realm of atoms.

The round, led by new energy giant CATL with strategic investment from Aramco Ventures, highlights a crucial pivot. The next phase of artificial intelligence won't just be about generating text or images; it will be about understanding, optimizing, and controlling the complex physical systems that power our world. It's a field being called 'Physical AI,' and it targets the foundational infrastructure of modern society—from semiconductor foundries to the sprawling data centers that house AI itself.

"The next phase of AI will be defined by the ability to control complex systems in the real world," said Li Hui, Founder and CEO of DeepCtrls, in a statement accompanying the announcement. "By connecting computing and energy through Physical AI, DeepCtrls aims to build the intelligent control layer for the infrastructure of the AI era." It’s a bold mission, and one that is attracting serious attention.

The Brains Behind the Brawn: What is Physical AI?

For years, industrial automation has relied on pre-programmed, rule-based systems. Physical AI represents a fundamental leap forward. It involves creating systems that can perceive their environment through sensors, use AI models to predict outcomes, and then autonomously act upon the physical world to achieve optimal results. It is the difference between a simple thermostat that follows a schedule and a system that learns a building's thermal dynamics, predicts occupancy patterns, and adjusts the climate in real-time to minimize energy use without sacrificing comfort.

DeepCtrls, founded in 2018 with research roots at Tsinghua University and Lawrence Berkeley National Laboratory, is a pioneer in this domain. The company’s core technology, the PhyAI™ Physical AI Engine, integrates advanced AI with the unyielding principles of physics. This allows it to create a closed-loop system of perception, prediction, and control for everything from advanced manufacturing lines to the intricate cooling systems of a data center. Industry analysts are taking note, with firms like Deloitte identifying “Physical AI / closed-loop systems” as a key part of the broadening technology mix in manufacturing, moving the industry from being data-rich to decision-rich.

This isn't just about making existing processes more efficient; it's about enabling a level of autonomous optimization that was previously impossible. The technology promises to handle the immense complexity and variability of real-world systems, learning and adapting on the fly in a way that rigid, pre-programmed logic never could.

A Strategic Bet on Real-World Infrastructure

The composition of this funding round is as telling as its size. The lead investor, CATL, is a global leader in new energy technologies and, notably, a long-standing strategic customer of DeepCtrls. This investment isn't a speculative punt; it’s a doubling-down on a proven partnership. For CATL, which has been strategically diversifying beyond its core electric vehicle battery business, the move aligns perfectly with its ambition to build a vertical energy chain for the power-hungry AI data center market. The battery giant’s recent investments in data center power suppliers and operators underscore a clear strategy: to power the AI revolution. DeepCtrls provides the intelligent software layer to optimize that power.

Equally significant is the strategic investment from Aramco Ventures, the venture arm of the Saudi Arabian energy behemoth. This investment provides more than just capital; it offers a stamp of international validation and a pathway for global expansion. Aramco Ventures has a clear mandate to invest in deep technologies that can drive industrial digitalization and sustainability. Its portfolio, which has recently grown to include other AI-driven energy management startups, shows a consistent thesis that the world's most critical industries are ripe for AI-powered optimization. For them, DeepCtrls represents a core enabling technology to enhance efficiency and reduce the carbon footprint of massive industrial operations.

From Validation to Value: The Tangible Impact of PhyAI™

For any technology to be truly transformative, it must move from the lab to large-scale commercial adoption. DeepCtrls appears to have crossed that chasm. The company reports serving hundreds of enterprise customers, a claim substantiated by a client roster that reads like a who's who of global technology and manufacturing: Tencent, ByteDance, NVIDIA, TSMC, and LG.

These are not companies that adopt new technology lightly. Their operations—from fabricating the world’s most advanced semiconductors to running massive cloud computing platforms—are defined by extreme complexity and a relentless drive for efficiency. While specific metrics are proprietary, the value proposition is clear. Industry-wide studies suggest the potential is enormous; a 2026 Deloitte survey found that AI could improve energy consumption in discrete manufacturing by an average of 22%. When applied to the colossal energy budgets of data centers and chip fabs, the savings are staggering.

Perhaps the most compelling application of DeepCtrls' technology is in solving a problem created by the AI industry itself. The computational power required for training and running large AI models has led to an explosion in energy consumption and heat generation within data centers. DeepCtrls is tackling this head-on with solutions for intelligent liquid cooling and the coordination of computing workloads with energy availability, effectively creating a smart grid inside the data center. It's a tangible, impactful application that addresses one of the most significant sustainability challenges of our time.

A New Frontier in a Crowded Field

DeepCtrls is not operating in a vacuum. The broader market for industrial automation is vast, projected to approach $600 billion by 2035. However, the specialized segment of Physical AI is where the most explosive growth is expected, with forecasts predicting the market could surge from around $30 billion in 2025 to over $640 billion in the next decade.

This potential has attracted a diverse field of players, from established industrial giants retrofitting their platforms with AI to a host of nimble startups targeting specific niches like grid balancing or building energy management. Yet, DeepCtrls has carved out a distinct and critical position. Rather than focusing on a single application, its PhyAI™ Engine is designed to be the foundational control layer that connects the two pillars of the digital age: computing and energy.

By enabling these two domains to communicate and coordinate intelligently, the company is doing more than just optimizing individual systems. It is building the autonomic nervous system for the physical infrastructure of the entire AI era, ensuring that as our digital ambitions grow, the real-world foundation supporting them can operate with unprecedented efficiency and intelligence.