Agility Robotics Is Going Public at $2.5 Billion. Its 65,000 Factory Hours Reveal the Real Break-Even Math.

$38,462. Divide Agility Robotics' $2.5 billion pre-money equity value by the 65,000 cumulative deployment hours its Digit humanoid robot has logged across nine customer sites, and that is what you get per hour of factory work. It is probably the wrong way to evaluate this company, but it is a useful starting point for a question the glossy investor decks avoid entirely: at what utilization rate does a $250,000 bipedal robot actually beat a human manufacturing worker on cost, given what we know about loaded labor compensation, realistic uptime percentages, and the gap between what a robot can theoretically do and what it consistently delivers across an eight-hour production shift?

On June 24, Agility Robotics announced it would go public through a merger with Churchill Capital Corp XI, a blank-check company backed by Wall Street dealmaker Michael Klein, in a transaction that Reuters reported will generate more than $620 million in gross proceeds, including a $200 million PIPE investment led by Foxconn at $10 per share. Agility will trade under the ticker AGLT, becoming the only U.S.-listed pure-play humanoid robot company with live commercial deployments.

Existing customers include Schaeffler, GXO, Toyota Motor Manufacturing Canada, and Mercado Libre, and their combined deployments have racked up 65,000+ hours of operational runtime across logistics, manufacturing, and distribution workflows. More importantly, Agility has secured over $300 million in multi-year contracted orders for its updated Digit v5, with a pipeline of 30+ potential customers evaluating large-scale deployments, which at an estimated $250,000 per unit based on prior-generation pricing translates to roughly 1,200 robots.

Running the Break-Even Calculation

Start with what a human manufacturing worker actually costs an employer in the United States. Bureau of Labor Statistics data from May 2026 shows average hourly earnings for all manufacturing employees at $36.71, and when you add the standard benefits loading of 35% to cover health insurance, payroll taxes, workers' compensation, and paid leave, total loaded compensation runs to approximately $49.56 per hour, which over a standard 2,080-hour work year means one manufacturing worker costs an employer roughly $103,085 annually.

Now build the robot side: a $250,000 Digit running two shifts at 16 hours per day across 350 operating days per year logs 5,600 hours annually, and applying a conservative 90% uptime rate (AGIBot's G2 has demonstrated 96% uptime over 140+ hours of continuous operation at Longcheer Technology's tablet factory) yields 5,040 productive hours per year.

Spread the capital over five years:

At $22.14 per productive hour in years 2 through 5, compared against $49.56 for a human worker performing the same tasks, the robot looks like a bargain on paper. Except for one problem that nobody in the investor pitch seems eager to quantify: the robot does not do everything a human does, and the gap between "can perform task X in a controlled demo" and "performs task X at human speed, consistency, and judgment across a full shift" is the difference between a business and a science project.

Task Effectiveness Is the Whole Game

A Digit robot running 5,040 hours per year clocks 2.42 times the productive hours of a single human worker on a standard shift, which means that if the robot performs its assigned tasks at 100% human equivalence, it displaces 2.42 full-time workers, saving $249,466 in annual labor costs against $111,605 in robot operating costs for a $137,861 annual surplus, with payback arriving inside two years.

Drop task effectiveness to 50%, and the robot completes half as many useful operations per hour as a human would, displaces only 1.21 FTE, saves $124,733 against that same $111,605 in costs, producing a $13,128 annual surplus that stretches payback past 19 years, which is considerably longer than the robot will survive mechanically.

Solving for the crossover directly, at $22.14 per hour in robot operating costs and $49.56 per hour in human loaded compensation, running 5,040 robot-hours to replace 2,080 human-hours, the break-even task effectiveness lands at 54%. Below that, every robot loses money every year it operates. Above it, savings compound with each additional unit deployed, and the economics improve with every incremental percentage point of capability the AI learns.

AGIBot's G2 offers the best available benchmark for what that effectiveness looks like in practice, having demonstrated 310 units per hour at Longcheer Technology's tablet factory with a 99.5% success rate and cycle times of 19 to 20 seconds during a six-day global livestream running June 23 through June 28 that showed these robots embedded in a real production line, not on a demo stage. AGIBot claims each G2 replaces two human worker stations. Agility's Digit performs different tasks, primarily material handling and logistics rather than assembly inspection, but the throughput benchmark matters because it establishes a floor for what factory-deployed humanoids can achieve today in constrained, repetitive workflows. If Digit achieves even 65% effectiveness in logistics operations, the annual savings per robot hit approximately $50,000, and multiplying by 1,200 contracted units puts the fleet at $60 million per year in customer value above the total cost of the robots themselves.

What the Valuation Actually Requires

Industrial automation companies trade at 3 to 6 times revenue: Fanuc sits near 5x, ABB closer to 3x, Rockwell Automation stretches to 6x. A high-growth robotics startup naturally commands a premium, but SPACs carry an asterisk, because median DeSPAC stocks have historically declined 30 to 50 percent within two years of closing.

At an 8x revenue multiple, a reasonable growth-stage premium for a company with contracted orders and live deployments, the $2.5 billion valuation implies a target of roughly $312 million in annual revenue, which at $250,000 per Digit means shipping 1,250 units annually plus generating meaningful recurring software and support revenue. Agility's $300 million contracted backlog covers approximately one year at that run rate, but the backlog is spread over multiple years rather than concentrated in a single fiscal year, so reaching the implied revenue target demands tripling or quadrupling annual order intake within 24 to 36 months.

Foxconn's role as lead PIPE investor signals something beyond financial conviction. Foxconn operates the world's largest electronics manufacturing network, with 1.29 million employees across factories in China, India, Vietnam, and Mexico, and if a single Digit can replace even 1.5 FTE at 65% effectiveness while running two shifts, Foxconn's labor math improves across every facility in its network. At Chinese manufacturing wages of $6 to $8 per hour loaded, the break-even effectiveness drops below 30%, meaning a robot paying for itself inside a year against Chinese labor costs is entirely plausible. That explains both why Foxconn wrote the check and why AGIBot has already shipped 10,000 units domestically.

Against the Field

Agility is not alone in this race. Boston Dynamics priced its Atlas humanoid below $320,000, a figure that KED Global reported corresponds to the loaded cost of two U.S. manufacturing workers over two years, and Hyundai plans to deploy 25,000 Atlas robots across its own factories starting in 2028, with a 30,000-unit-per-year production facility under construction. But Hyundai's Korean union has declared that "not a single robot using new technology will be allowed to enter the workplace" without a prior labor agreement, injecting real deployment friction into what the spreadsheets treat as a smooth capital expenditure decision, a complication that Agility's North American customer base largely avoids because U.S. manufacturing unions are far weaker and Amazon warehouses are not unionized at all.

At scale, Hyundai Mobis plans to reduce Atlas actuator costs by 70%, pushing per-unit production costs to approximately $130,000 by 2030. If that happens, and if Atlas achieves the same operational reliability that AGIBot has demonstrated in Chinese factories, the per-hour economics shift dramatically: a $130,000 robot with 5,040 productive hours per year and reasonable maintenance burns less than $14 per productive hour, and at that number the break-even task effectiveness drops below 35%, which means automation becomes cheaper than labor for nearly every repetitive factory task that does not require fine judgment, improvisation, or the kind of social awareness that keeps humans from walking into each other.

Limitations

Several assumptions deserve scrutiny. Agility has not disclosed per-unit pricing for Digit v5, so the $250,000 figure is extrapolated from prior-generation estimates, and the actual price could differ substantially in either direction depending on how aggressively Agility prices for market share versus margin. Maintenance costs are modeled at 10% of capital annually, a standard assumption for industrial equipment that has no public validation from actual Agility deployments. Most critically, the 65,000 cumulative operational hours divide to roughly 7,200 hours per facility, and with an unknown number of robots per site, per-unit runtime could be as low as 700 hours, which is barely a month of two-shift operation. Industrial equipment typically requires 40,000+ hours of mean-time-between-failures data before customers commit to fleet-scale procurement, and Agility is one order of magnitude short.

SPAC valuations carry structural inflation: Churchill Capital's trust holds $420 million assuming no redemptions, but SPAC redemption rates have averaged 73 to 85 percent in recent years, and if substantial redemptions occur, Agility receives far less capital than the headline $620 million suggests, potentially constraining the production ramp that is prerequisite to justifying the $2.5 billion price tag in the first place.

What You Can Do

If you run a manufacturing or logistics operation: Request pilot data from Agility, AGIBot, or Apptronik before committing to volume purchases, and demand documented per-robot uptime logs, task completion rates, and maintenance costs from existing deployments rather than accepting marketing claims at face value. Run the break-even calculation above with your actual loaded labor costs and expected task coverage for your specific workflows. If your loaded worker cost exceeds $45 per hour and the robot achieves 60%+ task effectiveness on your applications, the three-year ROI is likely positive.

If you are evaluating AGLT as an investment: Watch quarterly unit shipments and cumulative fleet hours per robot with far more attention than you give to revenue growth, because a company shipping 500 robots that each run 4,000 hours annually is building a real business while a company shipping 2,000 robots that sit idle in customer warehouses awaiting integration is building a very expensive inventory problem.

If you are a manufacturing worker or union organizer: Hyundai's Korean union offers a tactical template worth studying closely, because their demand to tie robot deployment to guaranteed employment agreements and fixed-salary structures that decouple worker compensation from production hours is strategically sound: negotiating robot introduction terms now, while labor still has leverage and the technology still needs human oversight, is far more productive than attempting to oppose the technology after it crosses the 54% effectiveness threshold and the economics become irresistible to every CFO with a manufacturing line item on the balance sheet.

The Bottom Line

Agility's IPO is not a bet on a robot. It is a bet on a number: 54%. At that task effectiveness threshold, humanoid robots cross from expensive science project to cheaper-than-human labor, and the transition, once it begins, happens faster than most factory workers or investors or regulators expect. AGIBot has already crossed that line for narrow tasks in Chinese electronics manufacturing, processing over 3,000 tablets per shift with sub-4% downtime at a fraction of the local labor cost. Boston Dynamics is engineering its way there through a $130,000 cost target by 2030 backed by Hyundai's automotive supply chain. Agility is asking public-market investors to fund the bridge between 65,000 pilot hours and millions of production hours, pricing the bridge at $2.5 billion and betting that the chasm between a working demo and a profitable fleet is smaller than skeptics assume. Whether that is a visionary bet or SPAC-era optimism depends on how quickly Digit closes the gap between demonstration-grade and human-equivalent performance in the messy, variable, unforgiving environment of a real factory floor that doesn't pause for software updates. Right now, the entire thesis reduces to a single threshold: 54% task effectiveness. Cross it, and Agility is worth every penny. Miss it, and $2.5 billion buys a very expensive pilot program.