Physical AI Absorbed $47.4 Billion in Six Months — Nearly 4x the Prior Half and More Than 2022–2024 Combined

Venture capital has found its next destination after chatbots and copilots: the physical world. According to data published by Crunchbase News on August 18, global venture funding for “physical AI” companies reached $47.4 billion across 521 deals in the first half of 2026 — a figure that reshapes how we understand the AI investment cycle.

The scale of the acceleration is difficult to overstate. In the second half of 2025, physical AI startups raised $12 billion across 470 deals. In the first half of 2025, they raised $26.4 billion across 436 deals. H1 2026’s total is nearly four times the immediately preceding six-month period and roughly 80% higher than the same window a year earlier. Perhaps the most striking comparison: across the entire three-year span from 2022 through 2024, venture investors put a combined $41.9 billion into physical AI companies — several billion less than what was raised in the first half of 2026 alone.

What counts as “physical AI”?

Crunchbase’s definition spans industries where intelligence is embedded into machines that act in the real world: robotics, autonomous vehicles, aerospace, drones, industrial automation, and sensors. This is the segment that many investors now treat as the next leg of the broader AI boom — and, as The Wall Street Journal recently reported, firms that built their reputations on early bets in software, internet services, and social media are increasingly writing checks for physical technologies and materials tied to the AI boom.

That marks a genuine philosophical shift for an asset class that spent a decade avoiding hardware’s capital intensity and long development cycles.

The megadeals behind the numbers

A handful of multibillion-dollar rounds drove the spike, and one deal alone accounted for nearly a third of all venture dollars in the category. In February, Mountain View-based Waymo raised a $16 billion Series D at a staggering $126 billion valuation, with Alphabet, Dragoneer Investment Group, DST Global, and Sequoia Capital co-leading.

Defense tech supplied the next tier of megadeals:

• Anduril Industries raised another $5 billion in May at a $61 billion valuation — double the $30.5 billion valuation it commanded less than a year earlier.
• San Diego-based Shield AI landed a $2 billion Series G in March, co-led by Advent International and JPMorgan Chase, lifting its valuation to $12.7 billion.
• Austin-based Saronic, which builds autonomous sea vessels, closed a $1.75 billion Series D in March led by Kleiner Perkins at a $9.25 billion valuation — more than double its 2025 Series C mark — bringing its total funding to around $2.6 billion.

The exit environment is real

Unlike many hyped categories, physical AI is already returning capital. SpaceX’s June IPO raised $75 billion at a $1.77 trillion valuation — the clear outlier of the year and one of the largest public debuts in market history. Space intelligence company HawkEye 360 raised $416 million in its IPO, and autonomous drone maker Aevex raised $320 million in its public debut.

M&A is stirring too. Mobileye’s roughly $900 million acquisition of Tel Aviv humanoid robotics startup Mentee Robotics stands out because the buyer explicitly tied the deal to its push into physical AI — a signal that incumbents in adjacent markets see embodied intelligence as a must-own capability.

Why now? Investors explain the shift

Ryan Ziegler, general partner at Edison Partners, argues that physical AI is far broader than the robotics, humanoids, defense, and foundation models where funding has historically concentrated. He frames it as the convergence of software, hardware, sensors, and IoT across real-world applications — with the change being AI’s ability to process data from those systems at a scale and speed that finally generates useful operational insight, while underlying hardware gets cheaper. “Even our mobile phones now have LIDAR scanners on them,” he noted, “democratizing the ability to map objects and spaces.”

Ziegler is drawn to high-value, traditionally analog industries — manufacturing, supply chain, utilities, agriculture, transportation, government, and physical/spatial intelligence — where physical AI becomes mission-critical infrastructure. Many of these companies, he says, resemble vertical software businesses with “attractive unit economics, large deal values and multi-year deployments,” and their combination of software, sensors, and hardware generates proprietary datasets that grow more valuable over time. In his formulation, hardware becomes “the distribution model for creating a data intelligence flywheel.”

Joe Fath, partner and head of growth at Eclipse, points to collapsing barriers: “Tech barriers are plummeting, experienced talent is pouring in, and market demand is rising.” The economics of building hardware companies have improved over the past two years in ways that echo what cloud infrastructure did for SaaS — accessible compute and foundation models, reusable models and physics-based simulation, more plentiful training data, and declining sensor and hardware costs. Business models are adapting as well, with companies bundling hardware into recurring or mixed-revenue models and moving toward outcome- or usage-based pricing. For Eclipse, physical AI is not a trend-chase but a core thesis dating back to the firm’s founding in 2015.

The bigger picture

The numbers slot into a broader arc. The Wall Street Journal counted $26.7 billion in global venture funding for physical AI through February 25 of this year — meaning roughly $20 billion arrived in just the four months that followed. Zoom out further and the trajectory is unmistakable: venture investment in robotics and physical AI grew from $4.2 billion in 2019 to $26 billion in 2025, before H1 2026 nearly doubled the annual record on its own.

Two currents are converging. First, frontier AI models now demonstrably handle multimodal perception, planning, and control — the missing ingredients that made past robotics investing a graveyard. Second, software investing itself has become a harder thesis as AI compresses the moats of traditional SaaS, pushing firms toward assets where defensible advantage comes from hardware integration, proprietary sensor data, and physical deployment footprints.

There are cautionary notes. Capital intensity hasn’t vanished — Fath acknowledges physical industries “remain capital intensive,” and the H1 total is heavily skewed by a handful of megadeals; without Waymo alone, the half-year figure drops below $31.4 billion. Concentration in aerospace and defense also means the exit activity hasn’t yet spread to areas like general-purpose robotics.

Still, the direction is clear. Funding is shifting, in Fath’s words, “away from experimentation and toward companies that can hit production milestones, land customers and scale efficiently.” After two years in which AI’s center of gravity was the text box, 2026 is the year capital decided the technology’s biggest market is the world outside the window — and it is paying robotics-era prices to get there early.