Following his departure as the Chief Executive Officer of Intel Corporation in late 2024, Pat Gelsinger embarked on an intensive transition period characterized by what he described as 100 meetings in 100 days. This structured investigative process was designed to refine his professional trajectory after decades spent at the helm of some of the world’s most influential technology firms. By March of the following year, Gelsinger finalized his next move, announcing his appointment as a general partner at Playground Capital. The Silicon Valley-based venture capital firm is noted for its specialized focus on "deep tech"—a sector involving capital-intensive startups built on fundamental scientific breakthroughs and engineering innovations that often require long-term development cycles.
Gelsinger’s move into venture capital represents a strategic shift from managing a legacy silicon giant to fostering the next generation of semiconductor and hardware infrastructure. At Playground Capital, his primary objective is the revitalization of Moore’s Law, the industry-defining principle established by Intel co-founder Gordon Moore in 1965. Moore’s Law posits that the number of transistors on a microchip doubles approximately every two years, leading to exponential gains in computing power and efficiency. However, as transistor dimensions approach the atomic scale, the industry has faced significant physical and economic headwinds, leading many analysts to declare the law’s stagnation or eventual demise.
The Strategy to Reawaken Moore’s Law Through Advanced Lithography
The core of Gelsinger’s investment thesis at Playground Capital revolves around the belief that the current impasse in semiconductor scaling can be resolved through radical advancements in lithography. Lithography is the process of using light to etch microscopic patterns onto silicon wafers. Currently, the industry leader in this space is the Dutch firm ASML, which utilizes Extreme Ultraviolet (EUV) lithography. ASML’s state-of-the-art machines use light with a wavelength of 13.5 nanometers to print features on chips.
Gelsinger contends that to continue the trajectory of Moore’s Law, the industry must transition to even shorter wavelengths. Upon joining Playground, he assumed a board seat at xLight, a portfolio company focused on developing novel lithography techniques. xLight is working on free-electron laser technology that could potentially enable wavelengths as short as 2 to 5 nanometers. This leap would allow for significantly higher transistor density and more powerful processors than current EUV technology permits.
The significance of this technology has not gone unnoticed by federal regulators. xLight recently secured investment and support from the United States government via the Department of Commerce, a move aligned with the broader goals of the CHIPS and Science Act to secure domestic semiconductor supply chains and maintain a technological edge over global competitors. Gelsinger has emphasized that xLight’s mission is complementary to existing industry leaders, stating that the goal is to enhance the capabilities of ASML’s machines by providing superior light sources rather than replacing the established infrastructure entirely.
The Deep Tech Pivot in the Venture Capital Landscape
Gelsinger’s entry into venture capital coincides with a broader industry shift. For much of the last decade, venture capital was dominated by software-as-a-service (SaaS) and consumer internet applications. However, the rapid rise of generative artificial intelligence (AI) has begun to commoditize certain aspects of software development, prompting investors to look toward the underlying hardware and physical sciences—the "deep tech" layer—to find sustainable competitive advantages.
Data from the semiconductor industry indicates a massive acceleration in market value. In early 2024, the industry set a goal to reach $1 trillion in global revenue by 2030. However, the unprecedented demand for AI training and inference hardware has moved that timeline forward significantly, with some analysts and industry veterans like Gelsinger projecting the $1 trillion milestone could be reached as early as 2025 or 2026.
Gelsinger argues that while many venture firms are now pivoting toward deep tech, the sector requires a level of technical due diligence that many generalist firms lack. He maintains that his experience as Intel’s first Chief Technology Officer and later its CEO provides a unique vantage point for vetting founders who are operating at the absolute limits of known physics.
Challenging the Status Quo in AI Hardware and Inference
A major focus of Gelsinger’s work at Playground involves identifying alternatives to the current dominance of Graphics Processing Units (GPUs) in the AI sector. While Nvidia has established a near-monopoly on the chips used to train large language models (LLMs), the industry is shifting toward "inference"—the process of actually running the models for end-users.
Gelsinger believes that the hardware requirements for inference differ substantially from those for training. He suggests that today’s GPUs are not the most efficient tools for these tasks and predicts a future of heterogeneous hardware structures. This vision includes a multitude of specialized vendors whose processors harmonize within a single system, optimized for specific workloads. He has highlighted companies like d-Matrix, Fractile, and Cerebras as entities that are currently breaking traditional boundaries in memory and chip architecture.
One of the critical technical bottlenecks identified by Gelsinger is high-bandwidth memory (HBM). While HBM is currently the industry standard for AI accelerators, it is fraught with manufacturing complexities and thermal limitations. Gelsinger anticipates that by the end of the decade, stacked memory architectures will become the dominant solution, potentially offering a 10 to 100-fold improvement in efficiency.
The Energy Constraint: A Bottleneck for Economic Capacity
Beyond the silicon itself, Gelsinger has identified energy infrastructure as the single greatest threat to the continued expansion of AI and high-performance computing. He has criticized the stagnation of energy capacity expansion in the United States, noting that low single-digit growth over the past decade is insufficient for a digital economy driven by AI.
"In a digital AI age, energy capacity is economic capacity," Gelsinger stated during the RAISE Summit in Paris. He argues that the focus on climate goals, while necessary, has led to a neglect of total capacity requirements. The supply chain for new energy sources remains a hurdle: gas turbines have eight-year lead times, nuclear plants take a decade to commission, and solar energy remains heavily dependent on international supply chains that are subject to geopolitical tensions.
To address this, Playground Capital has invested in companies like Alva Energy, which focuses on nuclear upgrading—a process of harvesting more value and efficiency from existing nuclear footprints while advocating for a broader reignition of nuclear power construction. Gelsinger views the winners of the AI era as those who can secure the energy capacity necessary to power massive data centers and compute clusters.
Geopolitics, Regulation, and the Race for AI Leadership
Gelsinger’s transition also places him at the center of the ongoing technological "cold war" between the United States and China. The US administration has increasingly utilized chip export controls and interventions in the release of AI models as a means of maintaining national security and economic leadership.
Gelsinger supports the strategic necessity of American leadership in foundational AI models, arguing that it is essential for these technologies to be developed in alignment with Western values. However, he acknowledges the complexity of regulating an industry where major foundational models are released on a monthly basis.
On the subject of regulation, Gelsinger advocates for a rigorous benchmarking process. He suggests that if the industry cannot implement its own process for integrity and security testing, government intervention will be inevitable. He emphasizes the need for visibility into what proprietary models are trained on and vigorous testing to ensure security requirements are met before widespread release.
Broader Impact and Industry Implications
The move of a high-profile executive like Pat Gelsinger into the venture capital space signals a maturation of the deep tech ecosystem. His focus on the "whole stack"—from lithography and materials science to voltage regulation and power distribution—reflects a holistic approach to solving the performance-per-watt challenges that define modern computing.
If Gelsinger and the startups at Playground Capital succeed in "waking Moore’s Law from its nap," the implications for global productivity could be profound. A 10,000-fold improvement in AI efficiency, as Gelsinger envisions, would drastically lower the cost of intelligence, making advanced AI applications accessible to a wider array of industries and socio-economic groups.
As the semiconductor industry hurtles toward its trillion-dollar milestone, the focus has shifted from mere scaling to fundamental reinvention. Gelsinger’s departure from the corporate structure of Intel to the agile environment of venture capital highlights a belief that the next great breakthroughs in computing will likely emerge not from the incumbents, but from the specialized, science-led startups currently operating at the edges of the possible.
About the Author
