In the gold rushes of the 19th century, a popular adage emerged: “When everyone is digging for gold, it’s good to be in the shovel business.” Just as miners needed picks and shovels, today’s AI-driven enterprises require certain foundational tools and infrastructure to power them. For an enterprise tech VC, there’s a compelling case to be made for investing in these essential tools of AI.
The Appeal of Infrastructure
- Use Case Agnostic: AI applications can deliver tremendous value in a variety of use cases and industries. Companies that create AI tools and infrastructure cater to a vast array of sectors. Whether it’s healthcare, finance, or e-commerce, these foundational tools have broad applicability.
- Consistent Demand: While AI applications may be subject to trends and market shifts, the tools that power them remain consistently in demand since they serve a range of applications.
- Recurring Revenue: Many AI infrastructure businesses are built on subscription or licensing models, offering more predictable, long-term revenue streams.
Infrastructure Categories That Run AI
As the industry matures, established incumbents such as Databricks, which is a SineWave portfolio company, have started to consolidate around capabilities that enable enterprises to efficiently and securely deploy AI. With their existing capabilities and the recent acquisition of Mosaic ML, they have made bets that allow them to capture a large part of the AI infrastructure value chain. However, several areas of opportunity exist for innovative startups.
- Compute Power: Training AI models, especially deep learning ones, requires tremendous computational power. Specialized hardware, such as GPUs and TPUs, have become the workhorses of AI, with companies like NVIDIA and Google leading the way. SineWave portfolio company, Rescale, has grown rapidly by helping customers deploy complex simulations on high-performance computing. While another portfolio company Paperspace was recently acquired by Digital Ocean because of its ability to abstract the complexity of cloud computing for machine learning applications.
- Data Storage and Management: The lifeblood of AI is data. Modern AI demands efficient data storage, retrieval, and management solutions. Major cloud service providers offer solutions tailored to high-volume, high-velocity data needs. However, there remain several opportunities in the domain of specialized data generation. Raw data is messy. Before it’s ready for AI, it often needs cleaning, labeling, and structuring. Platforms like DataRobot and Scale AI provide these essential data-prepping tools.
- Data Provenance Tools: The use of large amounts of data creates the need to have complete trust in the authenticity and security of data.
- Model Deployment and Monitoring: Once trained, AI models need to be deployed and monitored in real-time to ensure optimal performance and minimize model drift.
- Collaboration Tools: AI is a team sport. Tools that facilitate collaboration between data scientists, engineers, and domain experts are pivotal. Platforms like Hugging Face have stepped up to this challenge, offering communal spaces to share, tweak, and discuss models.
Navigating the Future of AI Investments
While the spotlight often shines brightest on the latest AI innovations and applications, it’s the underlying tools and infrastructure that make these innovations possible. From SineWave’s perspective, investing in these picks and shovels offers not just a stake in one company or application but a piece of the broader AI ecosystem. The AI gold rush is undoubtedly upon us, and sometimes, the most lucrative opportunities lie not in the gold itself but in the tools that unearth it.