Analysis
Two robotics rounds this week point at the same underlying shift: Formic's $50 million round to expand its robot-as-a-service model, and Maxinsights' $30 million seed led by Peak XV Partners to build training data infrastructure for physical AI, both monetize robotics as a recurring service rather than a capital good.
Formic's pitch to manufacturers is that they pay for output -- parts produced, hours of uptime -- rather than buying a machine outright, moving industrial automation's economics closer to a SaaS subscription than a capital purchase. The round funds Formic's move beyond task-specific industrial automation and into general-purpose robotics, a broader addressable market than the single-task robot-leasing deals that built the company's early customer base.
Maxinsights raised a $30 million seed and takes a different layer of the same stack, selling access to a growing training-data asset as a subscription rather than a one-time deliverable -- Pulse's separate look at why robots need their own training-data supply chain covers that company's specifics in full. The short version: it is a sizable seed check for a company whose core product is data rather than hardware or a model, which is itself a sign of where investors think the scarce resource in physical AI actually sits.
“The pattern matters because it changes what investors are actually underwriting.”
Pricing a robot or a dataset as a subscription rather than a one-time sale is a deliberate choice both companies are making, not an accident of how their products happen to work. A subscription model gives investors a recurring-revenue multiple to underwrite against, instead of the lumpier, hardware-cycle-driven revenue that a one-time robot sale or data license would produce -- which is exactly why growth and venture investors are willing to write larger checks into this structure than they might for an equivalent one-time-sale business.
The pattern matters because it changes what investors are actually underwriting. A recurring-revenue robotics company still needs to buy or lease the physical hardware upfront, meaning the balance sheet still carries real capital costs even as the revenue model looks like software. That is a genuinely different risk profile than either a pure hardware company or a pure SaaS company, and it is one most venture funds have less practice pricing, since the comparable multiples investors reach for by habit come from software businesses with far lower fixed costs.
The bear case is straightforward: if manufacturing capital spending slows in a downturn, both the subscription revenue and the resale value of the underlying robots take a hit simultaneously, a correlation pure-software companies don't carry.