Photo Credit: Gabriele Malaspina, Unsplash
Witnessing Thirty Years of Robotics Evolution
Genesis is approaching this robotics cycle as both investor and operator-side financier. Our portfolio company Botsync, a Singapore-based industrial robotics startup, has deployed more than 200 systems across 30 enterprise clients in six countries. Alongside that exposure, we are evaluating additional robotics platforms as the sector accelerates.
What follows is our house view on why robotics financing has become one of venture capital’s largest stories, how far the machines have progressed since our early university experiments, and what financing will turn technical capability into durable deployment across Southeast Asia. Capability and readiness are not the same: the billions flowing into robotics are outpacing answers to the harder question of whether homes, hospitals and hospitality businesses are ready to adopt these machines at scale.”
My own reference point stretches back thirty years – to when I was in a lab coat. In 1996, my research applied Professor Rodney Brooks’ behavior-based robotics paradigm to mobile robots whose neural-network controllers mapped sensors directly to motors. Simple behaviors – obstacle avoidance and following – produced coordinated group movement without a central planner. A year later, I helped build amphibious legged robots for naval tasks. Field tests in open seawater and dense forest made the constraints tangible: power, payload, battery life and precision actuators turned every engineering choice into an economic one.
Today, the contrast is extraordinary. Figure AI has raised more than a billion dollars in a round at a reported $39 billion valuation and unveiled Figure 03, a roughly $20,000 home humanoid designed to load a dishwasher, fold laundry and water plants. Its soft shell and sensitive fingertips are far removed from the rigid mechanisms I once watched researchers assemble to mimic a human neck. Yet the underlying lesson remains: enormous complexity hides inside actions people perform without thinking

Photo credit: Figure AI
The Robotics Capital Stack Has Never Been This Busy
Robotics venture funding has reached a record even though many funded robots still struggle with basic tasks without human support. F-Prime Capital’s tally across the Americas, Europe and Israel shows a V-shaped cycle: a 2021 peak, a two-year correction and a new high of $20.4 billion in 2025. Capital is moving faster than manufacturing maturity, safety certification and dexterity.

Source: F-Prime Capital / PitchBook.
Ten Robots, Ten Different Bets
A widely circulated comparison places ten humanoid platforms side by side, from Boston Dynamics’ Atlas models to 1X’s NEO, Figure 01, Agility’s Digit, Unitree’s H1 and Tesla’s Optimus Gen 2. Their similar physical profiles conceal radically different financing models (Figure 2).

Physical size says little about financial size. Sanctuary AI’s Phoenix is valued at roughly $230 million versus Figure’s $39 billion – a gap of more than 150 times between machines of comparable stature. Boston Dynamics has not raised a conventional venture round, yet Hyundai’s stake purchases imply a valuation near $20 billion, financed from a corporate balance sheet. These are not ten versions of one bet; they are ten different bets on technology, manufacturing and capital access.
The Brain Is Worth More Than the Body
Humanoids dominate headlines, but capital increasingly rewards the brain rather than the body. In the twelve months to July 2026, Robotic Foundation Models and General Purpose Robots captured 77.6% of disclosed physical-AI capital, versus 17.1% for humanoid hardware. Skild AI underscored the thesis by raising $1.4 billion in January 2026 for an ‘omni-bodied’ brain intended to operate many types of robots.
The shift is recent: F-Prime’s segment data shows humanoid and robotic-foundation-model funding across the Americas, Europe and Israel rising from $100 million in 2021 to $6.1 billion in 2025, with most of the increase in the last two years.

Source: F-Prime Capital / PitchBook

Regional patterns complicate the brain-premium thesis. Europe’s largest disclosed round belongs to NEURA Robotics, whose $1.4 billion Series C valued it at $7 billion. US rounds cluster around software and integration, while China’s smaller disclosed rounds understate its advantage: manufacturers are shipping humanoids at a fraction of Western unit cost. America is monetising the mind; China is building the muscle more cheaply. Singapore sits strategically between these capital and supply-chain pools.
The Shipment Numbers Say What the Funding Numbers Can’t
Figure AI’s valuation commands attention, but shipment volume tells a different story. Six Chinese manufacturers, led by Unitree and AgiBot, shipped a combined 12,868 humanoids in 2025. Three named American platforms shipped 450 – a ratio of roughly 29 to one, based on Omdia data reported by Rest of World.

Source: Unitree, Omdia via Rest of World / Visual Capitalist
Unitree’s entry-level R1 lists at $5,900, while competing platforms range from $20,000 to above $40,000. The race has split into two contests: valuation and deployment. Shipment volume creates field data and manufacturing learning that compound with every unit. If Chinese manufacturers sustain their cost and volume advantage, that learning curve may become harder to close than the funding gap.
The Last Mile Is a Forearm
Most conference-stage humanoids can walk; few can reliably pick up an egg. Actuation accounts for 40% to 60% of a humanoid’s bill of materials, and the dexterous hand alone for 15% to 25%. Advanced hands often last less than a year in continuous industrial use. Mass commercialisation may require 50% to 90% cost reductions in components such as cycloidal reducers and axial-flux motors.

Capital is following this bottleneck. Hangzhou-based Xynova raised nearly RMB 1 billion in six months and targets annual capacity of 10,000 dexterous hands by end-2026; 1X has unveiled a 25-degree-of-freedom hand for NEO with a similar unit ambition. Engineers call the hand an ‘impossible triangle’ of performance, cost and reliability. It may set the real timeline for trusted factory deployment.
From Warehouse Floor to Hotel Lobby
Industrial customers have moved from pilots to purchase orders faster than consumers. Warehousing and logistics reached 41,000 deployed units in 2025, up 28%, while food service reached 8,200, up 61%. Botsync reflects this shift: more than 200 systems across 30 enterprise clients in six countries, one million live production trips in 2025 and 230% revenue growth. Its customers include Ford, Caterpillar and Coca-Cola; SyncOS, its vendor-agnostic orchestration layer, manages mixed robot fleets through one control plane.

The Workforce Question — and Where Venture Debt Fits
Robotics enters markets that simultaneously fear job displacement and labour scarcity. The IMF estimates AI could significantly affect nearly 40% of jobs worldwide. Older projections of 400 million to 800 million jobs displaced by automation by 2030 should be treated as directional, but the transition is real.
Evidence so far points more to task reallocation than wholesale replacement. The International Trade Administration finds that a 1% increase in industrial robot density drives roughly a 0.8% productivity gain, while 2024 research on China’s robot adoption found gains in productivity, profit, employment and wages. Humans still supervise, troubleshoot and manage exceptions.
Hospitals and hotels make that division of labour visible. Diligent Robotics’ Moxi has completed more than a million deliveries across roughly 25 US hospitals, freeing nurses from routine fetching; Aethon’s TUG carts serve 37 VA hospitals and hundreds more worldwide. Hotel robots deliver towels and room service, but technicians and staff still intervene when a robot jams, loses charge or cannot communicate with an elevator. Automation changes what scarce workers do before it eliminates their roles.
Robotics companies finance two cash gaps at once: inventory and manufacturing before shipment, then receivables after deployment because customers often prefer leases or Robotics-as-a-Service contracts. Both gaps widen as sales accelerate. Funding them entirely with equity creates dilution precisely when commercial traction should strengthen founders’ position.
Venture debt is therefore not simply extra runway; it is a tool for matching capital to assets and cash flows. Facilities tied to inventory, purchase orders or receivables can fund repeatable deployments, preserve equity for R&D and hiring, and bridge companies to milestones that support a stronger next round. Botsync used venture debt to finance inventory and receivables with materially less dilution.
This matters especially in Southeast Asia. The region’s robotics companies sell across fragmented markets, import components in foreign currencies, face long enterprise procurement and payment cycles, and often lack the deep late-stage equity pools available in the US or China. Local banks usually seek profitability, hard collateral and operating histories that young hardware companies do not yet have. Venture debt can occupy that missing middle: more flexible than bank credit and less dilutive than equity, while imposing the discipline to finance contracted demand rather than speculative manufacturing.
For Genesis, the opportunity is not only the headline humanoid platform. It is financing the inventory, deployments and invoices that allow credible regional robotics companies to scale across borders.
Written by Jeremy Loh, with research and drafting support from Claude AI




































