AI Bet, Energy Play, Factory Evolution

Betting on the Next Big Thing: Lessons from China's Hardware and Energy Sectors

As artificial intelligence and the clean energy transition reshape global industries, hardware manufacturers stand at a pivotal crossroads. The path from innovative concept to scalable, profitable product is fraught with technical, financial, and market-specific challenges. Two distinct narratives from China's technology landscape—one from a veteran consumer electronics builder venturing into AI hardware, another from a energy storage startup conquering overseas markets—illuminate the strategic calculations, persistent grit, and adaptive playbooks required to navigate this complex terrain. Their stories reveal a broader evolution in Chinese manufacturing: a shift from pure volume and efficiency towards deep technological integration, long-term strategic patience, and a nuanced understanding of divergent global market logics.

The Charging Bank Patriarch's AI Gambit: Embracing the "Hardware Winter"

Returning from the latest Consumer Electronics Show (CES), Zhao Guocheng, Chairman and CEO of Guangdong Pisen Electronics Co., carried a definitive conviction: "Without AI, there is no future." To the founder of a company often called the "father of China's charging bank," the current landscape of AI hardware feels strikingly familiar. He draws a direct parallel to the charging bank's early days in 2006, a product then dismissed as a solution in search of a problem. "We couldn't sell them. Phones had removable batteries back then," Zhao recalled in a recent dialogue on the podcast "Deep Practice." "But I was convinced that future electronics would be more power-hungry and compact. We kept iterating, and when Apple's iPhones arrived, Pisen was the only one with a full product line ready."

This history fuels his optimism about the current wave of AI-enabled devices. While many see a crowded field of uncertain gadgets, Zhao sees the dawn of "soul" for functional electronics. "The excitement is that for the first time, AI gives hardware a brain, a soul. It can communicate and act on our behalf," he stated. However, his optimism is tempered by two decades of hard lessons in manufacturing, a journey marked by costly missteps that inform his current strategy.

Zhao openly recounts a multi-million-dollar loss from the early 2010s. His team had developed a series of hardware solutions—like specialized card readers with built-in Wi-Fi—that allowed seamless file transfers between devices like cameras and phones, addressing a genuine pain point before cloud services were mainstream in China. "We thought it was a perfect solution," he said. "Then Apple's iCloud launched locally, and it was over. Our technical path was wrong, and we lost tens of millions." This experience underscored a critical lesson: a great hardware solution can be rendered obsolete overnight by a software or service-layer evolution, a "dimensional reduction strike" from platform giants.

A more intricate lesson came from collaboration with Wu Zhao (nicknamed "Wuzhao"), former head of Alibaba's DingTalk, on pioneering hardware like "glowing headphones" that pulsed with music. The project hit a classic innovation deadlock. No existing supply chain could produce the novel components. Suppliers demanded upfront development fees, while the startup client was unwilling to pay for unproven outcomes. As a manufacturer, Pisen stepped in to fund and manage the high-risk, iterative development. "You spend a fortune just to prove something won't work, then you change the approach," Zhao explained. The project also fell victim to the innovation time-cost. After two years battling an intractable antenna interference issue caused by the sleek, luminous design, the envisioned product never materialized. "The market appeal you have at the idea stage can completely evaporate by the time you finally, clumsily, bring it to life," he noted.

Yet, Zhao continues to work on the glowing headphone concept independently, believing it can be revitalized with AI features. This persistence stems from a core philosophy: true AI hardware must evolve from "functional intelligence" to "cognitive intelligence." This shift isn't merely about packing in more computing power and solving thermal dissipation, but about fundamentally re-architecting the "software-hardware integration" paradigm. He envisions concepts like "iDock," which would utilize idle smartphone processors as distributed computing brains for other devices.

Armed with these hard-won insights, Zhao is now positioning Pisen as a strategic partner for AI hardware startups, offering to be their "pitfall guide." His proposition is bold: for teams with a promising and technically sound roadmap, Pisen is willing to co-develop, co-invest, and even "lose money together" to navigate the manufacturing gauntlet. This is not charity but a calculated bet on learning, ecosystem positioning, and capturing the next "charging bank"-level necessity on its ground floor. His message is clear for aspiring hardware innovators: the brutal realities of supply chains, cross-disciplinary integration, and development timelines are non-negotiable hurdles that demand experienced, patient, and capital-ready manufacturing allies.

From Overseas Champion to Domestic Contender: Sigen's Split-Strategy in Energy Storage

While Pisen contemplates the AI hardware frontier, another Chinese company, Sigen New Energy, has already executed a lightning-fast capture of international markets, only to face a fundamentally different challenge at home. Founded less than three years ago, Sigen has achieved a remarkable feat: securing the number one market share position for distributed energy storage systems in Australia, Ireland, and South Africa, with cumulative revenue exceeding RMB 2.5 billion. Its weapon of choice is the SigenStor, a "five-in-one" integrated unit combining a solar inverter, battery, storage converter, charging module, and energy management system into a single, stackable module.

This product-market fit was precision-engineered for overseas, particularly residential, markets. It addressed key installer and homeowner demands for simplicity, safety, and smart functionality. The real differentiator, however, is its deeply integrated AI software. Sigen's system plugs into real-time dynamic electricity pricing data from over 60 utilities across 20+ countries, autonomously executing "buy low, sell high" strategies for stored energy. In Swedish trials, the system reportedly slashed users' electricity costs by an average of 70.3%. This deliverable, quantifiable economic benefit—rather than just hardware specs—propelled its premium product to rapid adoption, with overseas sales constituting over 99% of its revenue.

This staggering export dependence raises a pointed question: Is Sigen's model incompatible with China, or is it merely biding its time? The evidence points to a deliberate and distinct domestic strategy, necessitated by a starkly different market logic. In 2025, China moved away from mandatory energy storage allocations for new renewable projects, plunging the domestic sector from "policy-driven" to "market-driven" overnight. The market is now undergoing a painful consolidation, with awarded project scales down year-on-year. The new imperative is "real demand" driven strictly by levelized cost of storage and operational efficiency.

Sigen's domestic moves are tailored to this harsh new reality. It has pivoted from the residential focus of its overseas success to target the commercial and industrial (C&I) storage segment. At a recent technical exchange in Hainan, Sigen showcased a demonstration project at a fishery, emphasizing not its all-in-one appliance, but its "fully modular architecture" and "direct current (DC) coupling" technology. For C&I clients focused purely on return on investment, Sigen highlights technical advantages like support for higher DC input voltages (reducing cable costs) and a ~2% gain in system efficiency from DC coupling—all aimed at lowering the critical metrics of upfront capital expenditure and operational cost per cycle.

The contrast defines Sigen's bifurcated world. Overseas, it sells a "consumer-grade product" through installer channels, where brand, user experience, and future savings promise drive purchases. Domestically, it engages in an "investment-grade market," where hard-nosed calculations of efficiency, longevity, and cost-per-kilowatt-hour trump brand storytelling. The integrated "hardware-software" core remains constant, but its value proposition is translated into two different languages: one of convenience and savings for homeowners, another of IRR and payback periods for factory owners.

Sigen is not so much "cold" in China as it is playing a different, arguably harder, game. It is leveraging the advanced, AI-driven platform honed in competitive international markets to tackle the most economically sensitive segment of the Chinese storage sector, temporarily sidestepping the bloody price wars in more commoditized segments. Its domestic story is not a replication of overseas success, but a new experiment in applying globally competitive technology to a market governed by uniquely pragmatic and complex grid dynamics.

Convergence on Strategy: Patience, Integration, and Market Intelligence

The narratives of Pisen and Sigen, though operating in different spheres, converge on several strategic principles crucial for the next phase of tech hardware. Both emphasize long-term strategic patience. Zhao's faith in the charging bank and his continued tinkering with "failed" projects like glowing headphones mirror Sigen's willingness to forgo the easy volume of China's earlier policy-driven market to build a credible, economics-based case in the C&I sector.

Second, both are committed to deep "soft-hard" integration. For Zhao, the future is "cognitive intelligence," where hardware form is dictated by AI interaction models. For Sigen, AI is the non-negotiable brain that makes its hardware economically viable, whether in a Swedish home or a Hainan fishery. The hardware is the body, but the software is the central nervous system.

Finally, both demonstrate a sophisticated understanding of market dichotomy. Zhao warns against the universal application of a pure "foundry" model to innovative AI hardware, recognizing the need for strategic, risk-sharing partnerships. Sigen exemplifies this in practice, executing a virtually textbook-perfect case of product and go-to-market adaptation based on the mature, service-oriented demands of overseas markets versus the nascent, ROI-focused demands of the post-subsidy Chinese landscape.

Together, they illustrate a maturing dimension of China's tech hardware ecosystem. Beyond scale and speed, the new imperative involves navigating deep technological integration, tolerating the long gestation periods of foundational innovation, and possessing the analytical acuity to deploy different facets of a core competency across divergent global markets. As AI continues to seek its physical form and the global energy transition accelerates, these lessons in resilience, adaptation, and strategic partnership will become increasingly relevant far beyond China's borders. The journey from being a manufacturer of things to a builder of intelligent, adaptable systems is fraught with pitfalls, but as these companies show, it is a path defined as much by learning from failure and market nuance as by technological ambition.