Selling Dreams in the Age of Speculative Technology

In today’s financial markets, a troubling asymmetry has emerged between those who sell technological dreams and those who buy them. Founders, CEOs, and early investors often understand how distant true commercialization remains, while many young retail investors—driven by optimism and headlines—see only the promise, not the timeline. This imbalance creates a structural divide: the informed side monetizes expectations, while the uninformed side absorbs the losses.

Quantum computing offers a striking example. The 2025 Nobel Prize in Physics was awarded to John Clarke, Michel H. Devoret, and John M. Martinis “for the discovery of macroscopic quantum mechanical tunnelling and energy quantisation in an electric circuit.” Their work revealed that quantum effects—once confined to the microscopic world—can emerge in circuits large enough to see, laying the foundation for today’s superconducting qubits. Quantum systems promise radically new ways of computation through superposition, entanglement, and tunnelling—phenomena that classical chips cannot truly replicate. Yet the road from elegant physics to practical machines remains brutally hard. Qubits are fragile, error rates stubbornly high, and environmental noise constantly threatens to erase their delicate states. The Nobel Prize rightly honors a fundamental scientific breakthrough, but it doesn’t mean quantum computers are anywhere near replacing GPUs and HBM-powered systems. The discovery is profound—but the commercialization story is still very much a dream under construction.

Interestingly, quantum computing stocks have surged sharply over the past eight months (Feb 2, 2025-October 2, 2025), recovering from a major crash earlier this year. The rebound came after a steep decline triggered by Jensen Huang’s remarks suggesting that quantum computing is unlikely to become a reality in the near future. Although Huang later softened his position, I still agree with his initial assessment. The excitement that followed feels less like renewed scientific optimism and more like another speculative cycle fueled by narrative rather than progress. Most of the field’s current funding still comes from government research programs, not private demand, yet market narratives continue to portray quantum computing as an imminent revolution. When expectations race ahead of the underlying science, disappointment becomes inevitable. And when executives quietly sell shares while retail investors rush in, the market stops being a platform for innovation and instead becomes a mechanism for wealth transfer—from hope to disillusionment.

At the technical level, many current efforts in quantum computing for scientific computation neglect fundamental concepts from numerical analysis, including conditioning, stability, and error propagation. A common misconception—shaped by the success of AI—is the assumption that quantum computing will naturally follow a similar trajectory; this is a flawed analogy rather than a reasoned inference. The issue is further compounded by repeated appeals to “exponential speedups” without adequate consideration of constants, input/output costs, noise, and overall practical feasibility. In this context, claims that numerical analysis problems can be solved using quantum computing effectively amount to asserting that stable continuum computations can be carried out on a noisy, probabilistic device—a proposition that is internally inconsistent from a numerical analysis perspective.

Quantum computing may indeed prove transformative in theory, but in practice—when measured by economic efficiency, energy use, and scalability—it is likely to be outperformed, at least for now and for the foreseeable future, by the relentless progress of GPUs coupled with HBM memory.

Robotics will undoubtedly play a central role in future industries. The misconception lies not in this importance, but in assuming that humanoid, bipedal machines are the inevitable or optimal embodiment of that future. Human-like robots attract attention because they resemble us; however, their energy demands and control complexity make them inefficient for most practical applications. Airplanes do not flap their wings because biological flight relies on elastic deformation and continuous compliance that rigid structures cannot replicate. Likewise, robots lacking intrinsic elasticity face structural limits in reproducing human performance. Bipedal locomotion requires constant stabilization, precise foot transitions, and tightly coupled sensorimotor control, all of which consume substantial computational and electrical power. The core constraint is not merely engineering immaturity but energetic scaling. Battery energy density remains far below biological metabolic efficiency when normalized by mass and sustained output. Humans exploit passive dynamics and elastic energy recycling; most humanoid systems instead depend on stiff actuators and active feedback, replacing mechanical compliance with torque and computation. Stability must therefore be actively maintained rather than mechanically inherited. As environments become more unpredictable, control demands grow nonlinearly, increasing both power consumption and failure risk. Without advances in embodied elasticity, actuator efficiency, and energy storage, humanoid robots are unlikely to progress beyond demonstration toward economically viable autonomy. Engineering priorities should center on minimizing total system energy consumption rather than reproducing human form.

This critique does not diminish robotics’ broader necessity. In aging societies with shrinking labor forces, automation is increasingly indispensable. In hospitals, nursing facilities, and homes, assistive systems can reduce caregiver strain, manage routine logistics, and support independent living. Function-oriented robotics offers clear social and economic value where human labor is scarce. Recent demonstrations of fine hand manipulation—among the field’s most challenging problems—are technically impressive. Yet it remains uncertain whether extreme dexterity is required in most industrial, medical, or service contexts, where “good-enough” precision often suffices. Pursuing human-level manipulation risks over-engineering at the expense of robustness and reliability. In contrast, simpler platforms—wheeled systems, tracked robots, and articulated arms—may lack theatrical appeal but already deliver measurable gains in manufacturing, logistics, and healthcare. Nevertheless, investment continues to concentrate on humanoid systems, reflecting a familiar pattern in technology markets: spectacle attracts capital more readily than incremental but durable productivity gains.

Another example can be found in nuclear fusion research, a field that has promised a breakthrough for decades. Fusion is often described as the ultimate clean energy source—“the power of the sun on Earth.” The vision is inspiring, but the reality remains punishingly complex. Each new experimental reactor demands tens of billions of dollars and decades of construction, yet delivers only marginal progress toward the elusive goal of sustained net energy gain. Despite these persistent difficulties, the narrative of “near-future success” keeps the funding cycle alive, attracting new investment every few years. In practice, fusion has become less a race toward commercialization and more a long-running story that sells hope—a scientific ideal repackaged as a financial instrument.

While it is important to pursue ambitious dreams, it is equally crucial to acknowledge reality and advance them with the right balance of expertise and accountability—optimizing personnel and funding to safeguard public resources rather than draining a nation’s precious tax money. 

The contrast is instructive: technological progress often advances through practicality, not spectacle. Across fields, the message is the same. Imagination drives innovation, but expectation must remain grounded in physics, economics, and time. When markets price in dreams faster than science can deliver, hope turns volatile. Investors — especially younger ones — deserve not cynicism but realism: a recognition that even the most promising technologies evolve incrementally, not explosively. Innovation indeed thrives on belief, but belief is strongest when guided by understanding