Is this the most expensive business to start? (w/ Renée James) | Masters of Scale Summit 2025

In the rapidly evolving landscape of artificial intelligence, a single foundational element underpins every breakthrough: semiconductors. While much of the recent buzz centers on software and AI models, the physical chips that power these innovations remain the silent, indispensable architects of progress. Renée James, a seasoned veteran of the semiconductor industry and former President of Intel, offers a unique, candid perspective on the monumental challenges and profound opportunities within this critical sector, drawing from her decades of experience and her journey in founding a successful chip startup, Ampere Computing.

Key Takeaways

• Semiconductors are the bedrock of AI: Every significant leap in computing, from mainframes to modern AI, is fundamentally enabled by advanced chip technology.
• Entrepreneurship in semiconductors demands extraordinary capital and conviction: Starting a chip company requires billions in funding and an unwavering belief in one's mission, defying conventional startup wisdom.
• Innovation in chips can drive transformative real-world impact: Beyond mere processing power, chip advancements have direct applications in critical fields like medicine, significantly improving human lives.
• The industry is grappling with complex challenges: From the physical limits of power consumption ("power wall") to the geopolitical intricacies of global supply chains and export controls, semiconductors face multifaceted pressures.
• While AI's technology is definitive, its current market spending shows "frothiness": There's optimism for AI's long-term impact, but a need for more strategic and less "willy-nilly" investment in infrastructure.

The Indispensable Core: Why Chips Drive Every Technological Leap

The conversation around AI often centers on sophisticated algorithms and vast datasets, yet Renée James reminds us of the immutable truth: none of it is possible without the underlying hardware. "There isn't a lot of compute and all the software and AI and everything that we talked about all day without chips," she asserts, highlighting that this has been consistent throughout computing history. From the early days of mainframes to personal computers, mobile devices, and now the AI era, each generational leap has been powered by a corresponding advancement in semiconductor technology.

James, who has been immersed in the chip industry for over three decades, expresses a seasoned surprise at the sudden "heightened interest" in semiconductors. For her, the critical role of chips has always been a fundamental given, not a novel revelation. The current AI boom merely amplifies this long-standing reality, making the once-unseen engine of technology undeniable to a broader audience. As the demand for processing power continues its seemingly insatiable climb, so too does the reliance on cutting-edge silicon.

The Audacity of a Semiconductor Startup: Renée James's Journey with Ampere Computing

From Intel Veteran to Disruptive Founder

After a distinguished 28-year career at Intel, culminating in her role as President and working closely with the legendary Andy Grove, Renée James embarked on an entrepreneurial path that many deemed "insane." Before the current AI boom, she chose to venture into the incredibly capital-intensive world of semiconductor startups, aiming to compete in the processing arena dominated by CPUs and GPUs.

Her motivation stemmed from a deep-seated passion for invention and a belief in the profound impact chips can have. James recounts formative experiences in Intel's R&D, where innovations like video streaming on laptops—the precursor to applications like Zoom—felt like changing the world. More profoundly, she shared how Intel's work on synthesizing human DNA helped diagnose and cure cancer for a colleague. This personal connection cemented her conviction:

"When you do stuff like that, you're like, 'I'm going to build chips.' Keep going, right? This is like, this is, this is so, so that part of it, right? So like that, that to me always made me realize that this is like a noble mission."

This sense of mission was further fueled by Andy Grove himself, who, as her mentor, encouraged her to "do better" than the established players. Grove, one of Intel's founders, shared the historical imperative of innovation, harkening back to their departure from Fairchild Semiconductor to create Intel. This legacy of disruption resonated deeply with James.

Addressing the "Power Wall" with Ampere Computing

James identified a critical limitation in existing chip designs: the "power wall." For years, the industry understood that simply increasing chip performance without addressing power consumption was unsustainable, especially for burgeoning data centers. This wasn't merely a theoretical problem but a mathematical certainty that threatened to cap future growth. Ampere Computing was founded with a clear mission to overcome this challenge.

"Our mission was to pioneer the highest performance processors... AI acceleration at the lowest possible power."

Ampere's commitment to efficiency meant designing processors that were not only high-performing but also air-cooled, avoiding the need for complex, esoteric liquid cooling solutions. Against a chorus of skeptics who doubted her ability to transition from a large corporation to a startup, James successfully led Ampere for eight years, culminating in SoftBank's $6.5 billion acquisition.

The Astronomical Cost of Innovation: Capital and Scale in Semiconductors

Billions to Play the Game

The semiconductor industry stands apart in its capital requirements, a reality James underscores with stark clarity. Her Series A funding round for Ampere Computing was "a little bit north of $240 million," a figure that would be considered astronomical for most other startups but is merely the entry fee in semis.

"If you can't raise a billion or two billion or maybe a little more than two and a half billion... you can't really be in the game."

This immense financial barrier is a primary reason why many semiconductor ventures fail, either by running out of cash or struggling to gain market traction. The sheer cost of R&D, manufacturing infrastructure, and scaling production necessitates a level of investment that few industries demand. James notes that companies like Nvidia's Jensen Huang have spent 35 years building their empires, illustrating that organic growth in this sector is a multi-decade endeavor, not an overnight success story.

The Strategic Move to SoftBank

The decision to sell Ampere to SoftBank, even after achieving remarkable success against the odds, was a strategic one driven by the need for greater scale and sustained access to capital. James viewed the acquisition as the "best outcome" for the company's mission, especially given prevailing economic conditions. With SoftBank, Ampere gains the resources to accelerate its technology and expand its reach, ensuring its innovations can be deployed at the immense scale required by modern computing demands.

James expresses immense pride in Ampere's achievements, not only on the technical front but also in fostering a diverse and innovative work environment. "40% of our engineers are women," she highlights, positioning Ampere as a leader in diversity within the semiconductor space. The company also pioneered super-efficient processes, leveraging advanced tooling for emulation and simulation—methods that represent "the future of semis." This forward-thinking approach, coupled with the backing of a larger entity, positions Ampere for continued impact.

Navigating the AI Boom: Frothiness, Future, and Fragility

AI: Definitive Technology, Frothy Spending?

The current generative AI boom has undeniably captivated the tech world, yet James offers a nuanced perspective on its trajectory. She firmly believes that AI, as a technology, is "the definitive for the next several decades of computing," akin to how multimedia defined computing for previous generations. It represents the very "environment of compute," setting the stage for countless applications and refinements to come.

However, when it comes to the associated investment, James identifies an element of "frothiness."

"Is the spending a little bit willy-nilly right now? I'm going to say yes."

This stems from uncertainty about the precise infrastructure needed, particularly the balance between cloud-based solutions and enterprise adoption. While wealthy cloud providers can build massive infrastructure, the broader enterprise market, which accounts for a huge portion of semiconductor consumption, is still figuring out its AI strategy. This creates a scenario where a lot of capital is being deployed in a less-than-optimized manner, driven by a chasing mentality where "not everything's going to work."

Despite this caution, James remains an optimist, believing that humanity is only at the beginning of technological possibilities, not the end. She sees moments of "chasing" as inevitable in periods of intense innovation, emphasizing the importance of staying "directionally correct" even if individual bets don't always pay off.

Geopolitics and the Future of US Semiconductors

A National Treasure Facing Existential Questions

The strategic importance of the semiconductor industry extends far beyond economics, touching upon national security and global leadership. Renée James unequivocally labels Intel a "national treasure," acknowledging its foundational research in transistor technology and its critical role in the United States' technological dominance. While acknowledging Intel's recent challenges, she stresses that the US government's significant investment, though substantial in absolute terms, is a modest sum in the context of semiconductor capital requirements.

The US semiconductor industry is a complex ecosystem, comprising not just giants like Intel, Nvidia, and AMD, but also critical players like Broadcom, Marvell, and Micron (for memory). Each company contributes uniquely, and their collective strength forms the "raw horsepower" that fuels American innovation. James warns that this ecosystem, despite its robust history, remains "fragile" due to its inherent complexity and the global nature of its supply chains.

Navigating Export Controls and Global Competition

The AI race, particularly with China, introduces a layer of geopolitical complexity. The question of export controls on advanced chips to China is a contentious one, and James offers a nuanced perspective informed by decades of experience. She highlights that export controls on leading-edge technology are "not new," having been in place for over 30 years. The US government historically dictates who can and cannot receive certain advanced chip technologies.

However, she also points out the economic dilemma: China accounts for 50% of global semiconductor consumption, making it a vital market. An isolationist approach could harm the US semiconductor industry's health. Therefore, James advocates for a balanced approach, similar to proposals from industry leaders like Nvidia and AMD, which would allow for the sale of lower-performance chips while restricting the most advanced technologies. This strategy aims to maintain economic engagement without compromising national security.

"We have to continue being balanced between economics... versus not selling anything, right? Because isolationism isn't great for US semiconductor health."

James has absolute confidence in the United States' innovative capability, believing "we out innovate anybody." However, this capability relies entirely on a strong, well-supported domestic semiconductor industry. Protecting and fostering this industry is paramount to ensuring America's continued leadership in the AI era.

Conclusion

Renée James's insights paint a vivid picture of the semiconductor industry as simultaneously the most critical, challenging, and rewarding sector in technology. From the billions required to launch a startup to the profound impact chips have on human health and national security, the stakes are undeniably high. Her journey with Ampere Computing exemplifies the entrepreneurial spirit necessary to tackle these grand challenges, driven by a deep-seated passion for invention and a commitment to solving fundamental problems like the "power wall." While acknowledging the "frothiness" in current AI spending, James remains steadfastly optimistic about technology's future, viewing the current era not as an endpoint, but as a vibrant beginning. The continued success of the US semiconductor ecosystem, carefully navigated through economic and geopolitical complexities, will undoubtedly be the bedrock upon which the next decades of innovation are built.