The landscape of venture capital is undergoing a significant transformation, particularly within the capital-intensive realm of deep tech. As startups increasingly opt to remain private for extended periods, the secondary market has emerged as a crucial engine for liquidity, reshaping strategies for investors and founders alike. Sriram Viswanathan, founding managing partner of Celesta Capital, a San Francisco-based venture firm exclusively focused on deep tech, offers an insightful perspective on this accelerating trend and its profound implications.

Celesta Capital, established in 2013, has built its investment philosophy around the conviction that monumental technological shifts are rooted in innovation at the infrastructure layer. With an impressive $1.1 billion in assets under management, the operator-led firm boasts a portfolio of 110 investments and 43 exits, showcasing a remarkable exit-to-investment ratio that underscores its expertise in navigating complex deep tech landscapes. Notable successes include the semiconductor companies Credo and Innovium, AI processor developer Habana Labs, drone systems pioneer IdeaForge, and digital cell biology innovator Berkeley Lights. These exits highlight Celesta’s ability to identify and nurture foundational technologies that drive future growth.

Viswanathan, a seasoned veteran in deep tech, observes that the very core of Celesta’s investment focus – hardware, semiconductors, and systems – is now at the epicenter of market opportunities. This confluence of specialized expertise and burgeoning market demand positions Celesta at the forefront of the deep tech wave. He recently shared his perspectives on why the secondary market is gaining momentum, its specific impact on deep tech companies, and where the most compelling opportunities lie within this rapidly evolving sector.

The Surge in Secondary Market Activity: A Pendulum Swing

The heightened activity in the secondary market is a direct consequence of current economic dynamics: an abundance of capital pursuing a limited number of high-quality deals. "It’s a pretty obvious outcome that when you have more capital chasing fewer deals, more investors are looking for exposure," Viswanathan explained. He characterizes this as a pendulum swing, where periods of capital scarcity alternate with periods of abundance. Currently, fueled by an explosion of innovation in artificial intelligence (AI), healthcare, and data center infrastructure, an "incredible amount of capital" is flowing into the innovation space.

This influx means that while entrepreneurs might find it challenging to secure initial funding in a discerning market, the demand for exposure to promising ventures remains high. As a result, more "discriminating investors" are seeking avenues for greater exposure, often through secondary transactions that allow earlier investors to realize returns. Historically, early-stage venture capital firms primarily relied on mergers and acquisitions (M&A) or initial public offerings (IPOs) for liquidity. However, secondary transactions are now rivaling these traditional exit paths, offering a new, robust mechanism for capital recycling. This development brings a "newer mix of investors" – often deeper-pocketed institutions – who provide crucial liquidity for earlier-stage venture capitalists and angel investors.

Deep Tech and the Unique Dynamics of the Secondary Market

The secondary market’s engagement with deep tech presents a distinct departure from its traditional modus operandi. Conventionally, secondary market players, typically larger asset groups with different return expectations, gravitate towards mature companies with proven business models, predictable cash flows, and a shorter time horizon to liquidity (typically two to three years). Their focus has historically been on de-risked assets.

However, the current environment, particularly driven by the AI revolution, has altered this approach. "As it relates to deep tech, what’s happening is that there’s a feeding frenzy, because people see what happens in AI training. They think that a similar opportunity might exist in AI inference," Viswanathan noted. Investors are making strategic bets on the AI space, not necessarily prioritizing immediate revenue streams or cash flow. Instead, the focus is on identifying potential "winners" in a market often characterized by a "take-all" dynamic. This phenomenon is evident in the rise of a few dominant large language model (LLM) players, suggesting that the market will consolidate around a select few, rather than supporting a multitude of competitors.

While this speculative fervor currently fuels secondary market demand in AI, Viswanathan anticipates that it will eventually normalize. The current phase is less about traditional financial metrics and more about securing a stake in what are perceived as the foundational technologies of the next era.

Market Sustainability: Froth, Innovation, and Overbuild Concerns

Addressing the longevity of this hot market, Viswanathan invoked the principle of "reversal to the mean." He acknowledged the existence of "froth" in the market, where valuations can inflate beyond sustainable levels due to excessive capital chasing opportunities. "Valuations creep up. There’s more capital chasing it, and people are forgetting the back to basics, which are building a meaningful business, meaningful revenue, meaningful free cash, and all of that. And the hype cycle sets in," he cautioned, suggesting that certain segments of the AI infrastructure buildout might be in such a period.

Despite these concerns, he highlighted a significant counter-trend: the dramatic reduction in the "time to revenue" and the "cost of development" for deep tech companies. Advances in cloud computing, open-source AI frameworks, and accessible development tools have significantly lowered the barriers to innovation. This enables smaller teams, even a "one-person, billion-dollar company," to emerge, a phenomenon that will become increasingly common as the cost of innovation continues to plummet.

However, this rapid development also carries a risk. The race to build out infrastructure, particularly in AI, is provisioning a supply side that could easily exceed demand in many cases. If demand fails to catch up with this rapid overbuild, "capacity in excess of demand" will become a "pretty big problem." While deeper-pocketed investors, such as hyperscalers and neocloud providers, might weather such a storm, smaller players are likely to experience "whiplash" from market corrections.

Beyond AI: The Expansive Horizons of Deep Tech

While AI undoubtedly dominates current attention, Viswanathan emphasized that deep tech encompasses a much broader spectrum of innovation. He identified three expansive "sandboxes" where significant opportunities are emerging:

1. Hardware Systems, Intelligence, Infrastructure, and Data Center AI: This foundational sandbox encompasses the entire physical hardware infrastructure necessary for advanced computing. This includes next-generation semiconductors, specialized processors, high-performance computing architectures, advanced networking solutions, and the physical infrastructure required for massive data centers. These innovations are the bedrock upon which AI and other advanced technologies are built, extending beyond just AI-specific hardware to broader computational and data management needs.

2. Software Layer Above Frontier Models: Beyond the large language models and other frontier AI models, a vast software ecosystem is exploding. This includes "AI tools or AI software capability as an operating system," supporting a myriad of applications and services across diverse verticals. Examples include AI-assisted radiology in healthcare, intelligent marketing and go-to-market strategies in retail, advanced diagnostics, optimized supply chain management, and AI-enhanced financial services, including payment infrastructure. This layer focuses on applying AI’s analytical and predictive power to solve real-world problems and enhance existing industries.

3. Biology Influenced by Deep Tech: The convergence of deep tech with biology is revolutionizing healthcare and life sciences. This includes advancements in robotics for precision surgery, sophisticated diagnostic tools that leverage AI for faster and more accurate disease detection, and the miniaturization and advanced manufacturing of complex medical systems. Imagine AI-powered ultrasounds, connected CT scan and X-ray machines, and other radiology equipment that are not only smaller and more efficient but also enhanced by cloud connectivity and intelligent algorithms. This area promises transformative improvements in patient care, drug discovery, and biological research.

Prolonged Hold Periods Reshaping Behavior

The extended private lifecycle of startups has fundamentally altered the expectations and behaviors of Limited Partners (LPs), General Partners (GPs), and founders. LPs, who commit capital to venture funds, are increasingly prioritizing "DPI" (Distributed-to-Paid-in capital) over mere "IRR" (Internal Rate of Return). Viswanathan humorously cited a popular sentiment: "DPI is the new IRR." This signifies a shift from focusing solely on theoretical paper returns to demanding tangible cash distributions.

This push for earlier distributions means LPs have "greater impatience" for traditional venture timelines to shrink. Consequently, leveraging secondary markets becomes "a very important part of that market-clearing that has to occur." The robust performance of public markets over the past 7-10 years, offering attractive liquidity, has further intensified this demand for distributions from private investments. While this trend is healthy in driving accountability for capital returns, Viswanathan expects the attractiveness of public markets to fluctuate over time, potentially drawing investors back to the illiquid asset class of early-stage venture.

What Deep Tech Companies Uniquely Need

In this dynamic environment, deep tech companies must return to fundamental business principles. "You can’t continue to build on AI as a sector if you cannot show revenue growth, if you cannot show margin growth, if you cannot show scale," Viswanathan asserted. Core technology, while necessary, is "not sufficient." The true measure of success lies in technology adoption supported by demonstrable revenue growth.

He drew a parallel to the SaaS market, where once-inflated valuations (e.g., 20x annual recurring revenue, or ARR) have corrected to more modest multiples (5x-7x ARR in a good market). A similar shift is anticipated in the AI sector. Investors will increasingly demand "revenue growth" and "profitability growth" rather than solely focusing on technological breakthroughs or user acquisition at all costs. Deep tech companies, often characterized by long development cycles and significant upfront capital, must therefore integrate a clear path to commercialization and sustainable financial performance into their strategies from the outset.

In essence, the deep tech investment landscape is at an inflection point. The secondary market is providing a vital liquidity valve, while the promise of AI and other foundational technologies drives unprecedented capital inflows. However, success will ultimately hinge on a balanced approach that combines groundbreaking innovation with sound business fundamentals, ensuring that the "back to basics" philosophy prevails amidst the excitement of technological revolution.