The AI boom is ushering in a new era for venture capital, moving away from the safe, predictable Software-as-a-Service (SaaS) investments that dominated the post-dotcom bust years. Investors are now embracing "moonshot capitalism," reminiscent of early Silicon Valley, by backing more ambitious, capital-intensive, and often out-of-this-world projects. This shift is driven by the concern that AI model companies like Anthropic and OpenAI will disrupt traditional enterprise software businesses, making hardware and deep tech more attractive.
Since the beginning of 2024, over $150 billion has been invested in deep-tech deals (excluding AI), significantly surpassing the roughly $130 billion invested in the entire decade leading up to the end of 2019. These new investments focus on areas that are either inputs or outputs of AI, such as technologies that can radically lower the cost of powering AI chips, put them in novel locations like space or nuclear power stations, and even brain-computer interface technologies. AI also acts as an enabler, making it easier to model complex systems for fields like nuclear fusion, thus reducing upfront capital costs.
While this trend offers high potential rewards, it also comes with multiplied risks. These deep-tech ventures require significantly more capital due to their physical nature and are technically more challenging to execute, often demanding a solution to a scientific problem before even reaching commercialization. The commercialization risk for these companies is as high as for any other venture-backed startup, meaning great scientific founders may not be great at selling their products. The risk is that if the AI boom slows down, the overflow of capital into these ambitious, "crazier" ideas could cease, limiting their potential.
