Hardware has long been venture capital’s ugly duckling. Why? Software has historically offered the cleaner economic model: lower upfront capital requirements, higher gross margins, and recurring revenue. But the market is starting to pay less for theoretical capital efficiency and more for defensibility due to the emergence of Artificial Intelligence.
Carta says AI startups captured 44% of all U.S. startup capital in 2025, while Peter Walker of Carta notes that hardware, driven in large part by robotics, rose from 9% of startup funding in 2023 to 15% in 2025 and 23% at the start of 2026. That is a meaningful change in capital allocation, not just a change in narrative.
AI helps explain the shift. In software, development cycles have compressed, product iteration is faster, and distribution advantages can erode quickly when AI-native challengers emerge. Public markets have responded accordingly as investors have rewarded AI “picks and shovels” and re-rated parts of software exposed to substitution risk. That does not mean all software is equally vulnerable, but it does mean investors need to separate durable moats from feature layers.
This is where Physical AI begins to look structurally different. As we argued in Is the Future of AI Physical?, the next frontier is no longer AI that only generates content, but AI that can perceive, decide, and act in the real world: robots, autonomous vehicles, drones, and intelligent machines deployed across logistics, manufacturing, healthcare, and infrastructure. For investors, that matters because Physical AI bundles AI models, embedded compute, specialized hardware, software, and real-world deployment data into a single product. That complexity creates a larger moat than software alone, while also enabling recurring software, analytics, and edge-compute revenue on top of hardware.
Venture dollars are beginning to validate that thesis. PitchBook says robotics and Physical AI attracted roughly $27.6 billion across 1,009 deals in 2025. What makes that especially notable is that hardware’s momentum is building in a market where so much capital is already being deployed into AI. PitchBook reported AI/ML accounted for 65.6% of all U.S. VC deal value in 2025. In other words, investors are not rotating out of AI; they are becoming more selective about where within AI durable value is likely to accrue while also making new bets on AI-adjacent hardware investments.
Space belongs in the same conversation. As we wrote in The Modern Space Race, the investable case is not just about launch economics and commercial travel. It is about satellite operators challenging incumbent telecom models, the rise of dual-use technologies that serve both commercial and government customers, and geopolitical competition that keeps capital flowing into the sector. Those are exactly the kinds of forces long-duration investors should pay attention to: large end markets, strategic relevance, and barriers to entry that cannot be replicated by AI in a few hours.
The market data now supports this narrative. The Space Foundation says the global space economy reached a record $613 billion in 2024, with commercial activity accounting for 78% of the total and government budgets contributing $132 billion. Looking further out, the World Economic Forum and McKinsey project the space economy could reach $1.8 trillion by 2035. Private investment is moving in the same direction: Reuters, reported that private investment in space technology rose 48% to a record $12.4 billion in 2025, with the U.S. accounting for $7.3 billion, or about 60% of global funding.
Recent developments have only strengthened the case. On April 14, 2026, Amazon agreed to acquire Globalstar for $11.57 billion to deepen its low-Earth-orbit and direct-to-device ambitions. On April 1, Reuters reported that SpaceX had confidentially filed for an IPO that could value the company at more than $1.75 trillion and become the largest public listing on record. On the private side, Starcloud recently raised $170 million at a $1.1 billion valuation to pursue orbital data centers.
None of this guarantees easy returns. Physical AI and space both require more capital, longer development cycles, and higher execution discipline than pure software. But that is precisely why they are attracting investor attention. In a market that is increasingly skeptical of undifferentiated software, capital is moving toward businesses with larger moats, scarcer capabilities, and clearer strategic relevance.
Venture spent the last two decades optimizing for bits. The next phase may belong to companies that build in atoms.
DISCLOSURE: This material has been prepared or is distributed solely for informational purposes only and is not a solicitation or an offer to buy any security or instrument or to participate in any trading strategy. Any opinions, recommendations, and assumptions included in this presentation are based upon current market conditions, reflect our judgment as of the date of this presentation, and are subject to change. Past performance is no guarantee of future results. All investments involve risk including the loss of principal. All material presented is compiled from sources believed to be reliable, but accuracy cannot be guaranteed and Evergreen makes no representation as to its accuracy or completeness. Securities highlighted or discussed in this communication are mentioned for illustrative purposes only and are not a recommendation for these securities. Evergreen actively manages client portfolios and securities discussed in this communication may or may not be held in such portfolios at any given time.