Turning $1.78 million into a $2.5 billion valuation—how do you price that? On November 5, Nasdaq's pricing document for Agility Robotics put the pre-money equity valuation at $2.5 billion. Divided by the company's full-year 2025 net sales of $1.782 million, that's roughly 1,403 times. Alongside it came another set of numbers: the company's robots have accumulated 65,000 running hours and moved 100,000 totes in warehouses.
One ledger, two columns. The left column records work the machines have done, measured in hours and totes; the right column records money coming in and going out, measured in dollars. What makes the IPO weighty is that this two-column ledger is now in a glass case—someone will open it every quarter. Until now, humanoid robot valuations rested on private-market narrative; now there is a publicly checkable counterparty, updated daily.
The Left Column: Hours Accumulated
Agility was founded in 2015, spun out of Oregon State University's Dynamic Robotics Laboratory. The founders originally studied efficient robot locomotion in human environments—not building a humanoid for its own sake. Cassie in 2017 had only two legs; in 2019 a torso and arms were added, and Digit entered warehousing and manufacturing. Former Microsoft executive Peggy Johnson became CEO in 2024, pushing commercialization. The company has roughly 400 employees, headquartered in Oregon.
Per Tencent News and Tonghuashun (both November 5), Digit has accumulated over 65,000 hours of real operation across 9 customer sites; at GXO's Flowery Branch, Georgia warehouse it has moved more than 100,000 totes with roughly 98% accuracy. Customers include GXO, Schaeffler, Toyota Motor Manufacturing Canada, Amazon, and Mercado Libre; AGCO plans a pilot in 2027 at its Hesston, Kansas farm-equipment plant. Yahoo Finance adds that Amazon began trialing Digit in 2023, initially moving empty totes. The 98% figure comes without disclosed test methodology—it should be read as a company/media-reported operating figure, not an audited result.
Digit 5, released September 2026, secured over $300 million in multi-year orders, corresponding to roughly 1,000 units—a three-year Robot-as-a-Service (RaaS) contract at $8,500 per month plus a $25,000 deployment fee, with a list price of $200,000 per unit. Under RaaS, robots remain Agility's property; customers pay monthly, with software, maintenance, and battery swaps included. Deployment to existing partners is planned for H1 2027, general availability by end of 2027, with two new GXO sites starting in early 2027.
TechRepublic's comparison of major players' 2027 deployment progress (via Tonghuashun) concludes that "based on public operating records, Agility has the strongest evidence of sustained commercial deployment." Figure AI's BMW pilot follows, while Tesla Optimus's external commercial deployment "has not yet been established." This is a third-party assessment, not an exchange classification. Its point is precisely the value of the left-column numbers: while most peers remain at the demo-video stage, someone has converted credibility into run-time records at customer sites.
The Right Column: Revenue, Losses, Cash
The prospectus is equally blunt. Full-year 2025 net sales were $1.782 million (2024: $310,000); about 64% came from related parties—$1.132 million per Humanoid Analytics citing the filing. Yahoo Finance also notes that reportedly over half of revenue came from 5 robots purchased by one of the company's investors (the source used "reportedly"). Net loss for the period was $138.1 million; gross margin was -151%, with cost of sales at 2.51x revenue—1 minus 2.51 gives -1.51, so the numbers are internally consistent.
Further down: 2025 R&D expense ~$91.64 million; SG&A ~$45.84 million; operating expenses rose from $71 million (2024) to $111 million; operating cash outflow ~$99.84 million; year-end cash and equivalents ~$103 million; accumulated deficit $313.1 million; lifetime operating cash used ~$261.9 million; total liabilities $56.5758 million, 3.17x the prior-year level. The year-end cash increase came from ~$162 million net Series C financing and a $100 million SAFE in July 2026—not from operations. Management explicitly warns: without the SPAC merger financing, there is substantial doubt about going concern.
Unaudited H1 2026 is rougher: net sales of $174,000 (vs. $625,000 a year earlier); operating cash burn of $86.2 million (vs. $46.2 million). The filing attributes the sales decline mainly to one-off robot sales in the prior-year period. Humanoid Analytics notes accounting losses include substantial non-cash compensation—cash burn must be read separately.
| Frequently cited big number | Value | Which ledger | Date/nature | |---|---|---|---| | Pre-money equity valuation | $2.5B | Deal terms | Equity value to existing shareholders in the merger | | Multi-year orders | >$300M | Conditional forward contract | Single customer, with delivery/product milestones | | Cumulative operation | >65,000 hours | Operating record | 9 customer sites, over multiple years | | 2025 net sales | $1.782M | Confirmed revenue | 64% from related parties | | 2025 net loss | $138.1M | Accounting | Includes non-cash compensation |
The point of this table: the three most eye-catching numbers are not in the same ledger. The valuation comes from deal terms; the orders come from a contract that has not begun delivery; the hours come from production records. Writing "a company valued at $2.5 billion has run 65,000 hours and holds $300 million in orders" reads smoothly, but the accounts don't actually reconcile.
$2.5B and $300M: Different Denominators
The $2.5 billion is not post-listing market cap. Humanoid Analytics is explicit: the deal assigns $2.5 billion in equity value to existing shareholders and does not establish a live public market valuation. November 2 filings estimate ~$465.1 million in total funds from the deal and related financing, of which $100 million was already received, before transaction fees. "IPO-day pricing" headlines strictly require post-listing trading.
The $300 million is not receivables. The prospectus attributes the commitment to a single customer; revenue depends on contract milestones and promised product capabilities, and the buyer received warrants vesting with robot deployments. Digit 5 is expected to be commercially available only by end of 2026. As Yahoo Finance puts it: committed order value and confirmed revenue are not the same number, and in early industrial tech they can diverge for years.
The ~$465 million raise has structure: SPAC trust ~$164 million, Foxconn-led PIPE ~$201 million (at $10/share), plus $100 million of previously received equity financing. Churchill Capital Corp XI shareholders approved the merger with an 84% yes vote; closing was November 4, listing under AGLT on November 5 (warrants: AGLTW). The investor roster also includes Nvidia, Amazon, and SoftBank Vision Fund 2.
The >1,400x sales-multiple math: $2.5B ÷ $1.782M = 1,402.9. Caveat: the numerator is pre-money equity valuation, the denominator is full-year 2025 net sales—different dates, different natures; the multiple is only order-of-magnitude illustrative. A standard P/S uses market cap against current or trailing revenue, and H1 2026 revenue is still declining.
A Contrast in Capital Paths
Chinese media paired this listing with domestic moves the same week. One editor's observation (Toutiao AI Daily, November 5): SPAC anchors on "deployed paying customers," while IPO anchors on "revenue structure and going concern"—noting Zhiyuan (Agibot) delaying its HK listing, Weiyi halting pre-listing, and Unitree's Shanghai listing in August. The same day, JD announced purchasing 3 million robots over five years, and Unitree reported nine-month revenue of 1.817 billion yuan (+55.71% YoY) and net profit of 387 million yuan (+270.54% YoY)—secondary compilations, not checked against original filings; context for demand-side only.
Such contrasts are useful but have limits. One side is the US market giving the first pure-play humanoid a public ledger; the other is Chinese buyers making multi-million-unit purchase commitments. They measure different things: one measures a single company's unit economics, the other the sector's order pipeline. Turning that into "who leads" overreaches.
The Next Checkable Numbers
Back to the two-column ledger: the left column has 65,000 hours and 100,000 totes; the right column still shows $1.782 million in revenue and a $138.1 million loss. The IPO doesn't improve either column immediately—it only fixes in place the requirement to lay them side by side every quarter.
Humanoid Analytics lists the next metrics worth watching: customer-accepted deliveries, customer-confirmed running hours, repeat-purchase revenue, and whether gross margin improves after deducting maintenance and human support. Until those four numbers arrive, 1,400x and -151% will keep being quoted—and their respective denominators must be filled in by the reader.
Sources
- Tencent News, 2026-11-04 and 2026-11-05 coverage of the listing
- Tonghuashun advisory platform, 2026-11-05 (including TechRepublic deployment comparison)
- Yahoo Finance, 2026-11-05: "A Humanoid Robot Company Just Hit Wall Street With $1.8 Million in Sales and $138 Million in Losses"
- Humanoid Analytics, 2026-11-05: "Agility's Nasdaq Debut Puts Humanoid Economics to the Test" (citing Nasdaq Trader announcements and the 10-13 prospectus)
- NetEase / IT Home, 2026-11-05
- Toutiao AI Daily 2026-11-05 and WeChat morning briefings (capital path comparisons and domestic procurement data, secondary compilations)