Steve Jurvetson, the first investor in Elon Musk and an early backer of SpaceX and Tesla Motors (ASX: TSLA), with a 29-year friendship with Musk, recently shared his views on the future of AI and his current investment focus.

Jurvetson believes the exponential growth of AI-driven computing will disrupt three major industries within the next three years: energy, agriculture, and construction. These sectors represent the largest share of global GDP yet are the least digitized. His current investment portfolio is concentrated on nuclear fusion/fission energy, alternative proteins (cultivated meat/mycelium), epigenetic editing, new materials and critical minerals, and analog AI chips.

Initial Investments in SpaceX and Tesla

In the early 2000s, "private space" was not a recognized venture capital category. Jurvetson recalls that almost no investors were considering space; it wasn't listed on any investment website. The same logic applied to Tesla Motors—the automotive industry was not considered a typical VC domain.

His fundamental investment thesis was applying software and systems engineering thinking to transform traditional industries that had seen little change for decades. Aerospace and automotive were merely the first two test cases. He believes this logic will replay across almost every industry.

AI's Next Frontier: Three Undigitized Industries

When asked where AI will create the most significant change, Jurvetson didn't mention software but named three sectors: energy, agriculture, and construction. He noted these are massive contributors to GDP and among the least digitized industries on Earth, with healthcare following closely.

He referenced an exponential chart of computing power growth over 130 years, originally plotted by futurist Ray Kurzweil, to explain why this shift is happening. The chart shows a ten trillion-fold increase in computing power per dollar. He calls this the most important chart ever, illustrating how sustained exponential growth in computing is transforming low-margin "bad businesses" from the industrial era into information-centric operations, a path already proven by aerospace and automotive, with energy, agriculture, and construction next in line.

Regarding the technological driver, he admitted uncertainty but has an intuition about an architectural variant that might encompass current models. He specifically mentioned a new generation of labs in reinforcement learning, which he sees as a return to DeepMind's original mission before the large language model wave.

Current Investment Portfolio

When asked about his current investment focus, Jurvetson listed several areas, targeting old industries that have seen few new entrants.

Energy: Investments in nuclear fusion and a specific fission technology not triggering NRC regulation. His logic is that energy is the third major bottleneck for AI, after talent and computing power.

Alternative Protein: He believes that in 500 years, humans will not slaughter animals for meat. Cultivated meat, mycelium, and plant-based proteins are getting closer, and mycelium is the fastest-growing segment.

Epigenetic Editing: Described as "the software layer of biology, not the firmware layer of the genome," with applications in crop health, pesticide alternatives, and human health. This is one of his most active investment areas recently.

Critical Minerals & New Materials: From deep-sea mining to copper refining, he sees this as the foundation of the AI hardware supply chain, noting the US has lost significant capacity in this area over the years.

Analog AI Chips: He has three investments from different angles aiming for "100x and then another 100x" reductions in energy per computation. One company, Mythic, can perform 8-bit multiply-accumulate operations within a single transistor.

Healthcare & Life Sciences: Including organ cultivation, male contraceptives, and innovations falling through the cracks of traditional biotech VC. His overall portfolio is roughly 40% life sciences and 60% information technology.

He is also looking at construction but admits to having tried and failed a few times, though the search continues.

Superintelligence: A 30% Chance Next Year?

Jurvetson cited a specific figure from Anthropic co-founder Jack Clark, who gave a 30% probability for superintelligence emerging next year. Jurvetson finds this interesting as it's a definitive statement. His own stance is more cautious, assigning a vague future probability partly as an intellectual shortcut rather than a deeply considered prediction.

Lessons from 29 Years Observing Musk

When asked what he learned from Musk, Jurvetson highlighted three points.

First, extreme focus. Musk has an incredible ability to reject distractions. Jurvetson gave an example of trying to connect Musk with geneticist Craig Venter to discuss terraforming Mars; Musk refused, saying discussing what to do on Mars was meaningless until Starship could fly.

Second, compressing the innovation cycle. He considers this more important than focus. The core question is: How fast is your learning loop? He cited data showing that Tesla Motors vehicle cameras collect more data for AI training sets every four days than Waymo has collected in its entire history, with every car serving as a data collection device.

Third, attracting top talent with a grand vision. It's not just about building rockets or cars but about accelerating the transition to sustainable energy and making life multi-planetary. Such a vision attracts the brightest minds, creating a compounding effect as top talent attracts more top talent.

The Meaning of Life in an Automated World

Finally, when asked about the philosophical question of life's meaning when machines surpass humans at everything, Jurvetson responded that humans have a fundamental need for "symbolic immortality"—the belief that they leave something behind that outlives them, whether through children, writing, charity, or founding companies.

He believes human evolution is about the accumulation of knowledge, not biology. What we pass to the next generation are rules, laws, and understanding, not genes.

However, he acknowledged that the transition from full employment to zero employment will not be smooth, noting that no politician is seriously considering a transition period with 30%, 40%, or 50% unemployment. He doesn't want to end pessimistically but notes the path to a post-scarcity world could be difficult.