newsfilter.io
Interview, Fireside Chat

Finding the Right Problems: How Sutter Hill Ventures’ Mike Speiser Creates Great Companies

  • Mike Spicer joined Sutter Hill Ventures in 2008 as Managing Director, transitioning from a background as an entrepreneur who sold his company, Bix, to Yahoo.
  • Spicer views his primary role not as a traditional investor but as a "builder," noting that 90% of his returns historically came from ventures where he actively helped construct the business rather than purely investing capital.
  • Sutter Hill has evolved into a product incubator and "product studio" rather than a traditional venture firm, now employing approximately 100 staff members.
  • The firm's workforce composition includes a majority of engineers, researchers, designers, and salespeople, with only three to four pure investors among them.
  • Sutter Hill operates as an evergreen fund with a highly concentrated limited partner (LP) base, requiring no external capital raises and allowing for multi-decade investment horizons.
  • Spicer and the firm personally invest a minimum of 25 cents of every dollar raised, meaning a significant portion of capital deployed is their own.
  • The firm prioritizes technical and scientific risk over demand risk, leveraging their ability to solve technical challenges through the selection of the right talent.
  • Current Sutter Hill portfolio companies include NVIDIA, Pure Storage, Snowflake, and the recently publicized Astera Labs.
  • Spicer is currently serving as the founding CEO for three active projects, one of which consumes 80% of his time.
  • Project "REV" (meaning "dream" in French) focuses on generative AI for creativity, specifically enabling users to output multimodal media like images, videos, and stories from internal "fuzzy ideas."
  • Project "Bench" targets enterprise AI for work, aiming to integrate diverse AI tools and proprietary enterprise data to enhance team productivity and workflow management.
  • Both REV and Bench operate on extremely tight iteration cycles, with REV achieving model-to-product updates every 48 hours.
  • Spicer emphasizes the importance of intellectual honesty and avoiding premature "victory declarations" to maintain product-market fit.
  • Snowflake originated from a realization that the critical database problem lay in analytics (OLAP) rather than transactional processing (OLTP), requiring cloud-based parallelization of compute resources.
  • Spicer stepped down as Snowflake's CEO after roughly 18 months to recruit Bob Muglia, believing a different skillset was required to scale the company to production.
  • Sutter Hill maintains a deep strategic involvement in portfolio companies, assisting with the development of second-generation products like Pure Storage's FlashBlade.
  • The firm remains physically concentrated in Palo Alto, California, based on Spicer's conclusion that Silicon Valley offers a unique cultural ecosystem where ambitious ideas are encouraged rather than dismissed.
  • Spicer argues that the "time value of time" compounds similarly to money, requiring continuous, lifelong learning to stay competitive in the rapidly shifting technology landscape.
  • Current AI shifts, particularly regarding post-training and data quality, are forcing the firm to reinvent its building processes and methodologies.
  • Data strategy is now viewed as a prerequisite for any successful AI strategy, with unique, differentiated data becoming the primary driver of model differentiation.
  • Spicer attributes his success to a philosophy of "lifelong learning," where the ability to learn and adapt rapidly is more critical than formal education or early career achievements.
  • Spicer writes publicly on social media primarily as "self-therapy" to process his own challenges, aiming to inspire the specific teams he works with to take calculated risks and embrace failure as part of innovation.
  • The firm advocates for small, protected teams to foster creativity, positing that large groups are ill-suited for inventing new products due to organizational conservatism.
  • Spicer cites the "10X rule" as a benchmark for new ventures, noting that Snowflake's ability to perform in two minutes tasks that previously took 30 days on $20M systems exceeded even 10,000X improvements in specific contexts.
  • Spicer identifies human health and longevity as the most exciting future application of AI, anticipating that breakthroughs in the field will be accelerated by AI-driven scientific discovery.

In discussing the firm's culture, Spicer highlights that Sutter Hill maintains high standards of integrity and "old-fashioned" business practices, such as transacting on handshakes. He notes that the firm's ability to make long-term bets is directly correlated with the trust established through its transparent, evergreen structure. While acknowledging the rise of technology hubs in India and China, Spicer maintains that Silicon Valley remains the singular location where the convergence of talent, capital, and a culture of ambition creates "magic" that cannot be replicated elsewhere.

Regarding the current AI landscape, Spicer observes that the speed of innovation is staggering, with product iteration cycles collapsing from months to days. He warns that this acceleration creates a reset for the industry where those who do not commit to relentless hard work and continuous learning will quickly fall behind. He asserts that while macroeconomic trends are often discussed, Sutter Hill's focus remains strictly on microeconomics, solving one specific technical problem at a time with the right people and processes.