Agility Robotics Clears SPAC Vote as Factory Robots Move From Pilots to Contracts

Agility Robotics is close to becoming a public company. Shareholders of Churchill Capital Corp XI approved the proposed business combination at a special meeting on 2 November, according to The Robot Report, with more than 84% of votes cast in support. The SPAC is expected to close on 4 November, and the combined company should start trading on the Nasdaq on 5 November under the ticker AGLT. Churchill Capital stock closed up nearly 14% that day.

The money is smaller than originally advertised. The transaction is expected to give Agility about $465 million in gross capital before expenses, down from the more than $620 million anticipated when the merger was announced in June. The revised figure breaks down as roughly $164 million from Churchill's trust account, about $201 million from a PIPE led by Foxconn, and $100 million in equity Agility raised after the initial announcement. The original deal valued Agility at a $2.5 billion pre-money equity value.

Nine customer sites, and a contract figure that is not revenue

What separates Agility from most humanoid hopefuls is that its robots are already doing paid work. Digit units are operating in 9 customer environments, with logistics and manufacturing as the primary commercial markets, The Robot Report reported. The longest-running relationship is with GXO Logistics: the two companies began working together in 2023 and signed a multi-year Robots-as-a-Service agreement in 2024 at GXO's Flowery Branch, Ga. facility, where Digit moves totes onto conveyors.

In October the pair announced an expansion to two more GXO sites, in Statesville, N.C., and Lebanon, Tenn., with deployments expected to begin in early 2027 on Digit 4 and a later transition to Digit 5. The same month, agricultural equipment maker AGCO signed a commercial agreement to pilot Digit at its Hesston, Kan. plant from 2027, starting with sorting and stacking parts and placing totes for storage. Agility has claimed more than $300 million in committed multi-year orders for Digit 5, which was unveiled in September; early access is slated for the first half of 2027 and general availability for the end of that year. That order book is a signal of interest, not current revenue: the value is realised over time and depends on contractual milestones and product specifications.

Other robotics companies are raising money on the same pitch. Tutor Intelligence, a 2021 spinout from MIT CSAIL based in Watertown, Mass., said on 2 November that it had raised a $75 million Series B led by Andreessen Horowitz, bringing total capital raised to a figure The Robot Report truncated in its copy. Tutor's model is unusual: instead of robots-as-a-service, customers pay by the hour or by the pick, the way they pay for manual labour, with no upfront purchase and no long-term contract. The company said twice as many of its robots are working in almost four times as many facilities as a year ago, and that it is on track to have more than 1,000 robots under contract by the end of 2027, all built and deployed from the U.S.

"It's not only that Tutor's models let robots do jobs that robots have never been able to do before," said Josh Gruenstein, co-founder and CEO of Tutor Intelligence, in a statement carried by The Robot Report. "This type of intelligence also makes robots affordable on a whole new level." The company argues the physical world has no equivalent of the internet's text and image record to train on, which is why it runs fleets in production with remote human "tutors" feeding real-world data back into its models.

Power, chips and the cost of running all this

The constraint on factory and data centre automation is increasingly electrical. On 3 November, Tom's Hardware reported that Envision Energy has completed what it calls the world's first commercial data centre deployment of an integrated 800-volt DC power system, at its Galaxy AI data centre in Ulanqab, Inner Mongolia. The Dubhe system, tested since August with the Dubhe Controller and its Dubhe Energy Foundation Model, converts 10kV medium-voltage AC directly to an 800V DC bus, cutting conversion stages and shrinking the power apparatus. Envision says the campus will scale to 3GW and eventually "support up to one million AI accelerators".

The company lists three capabilities behind the claim: solid-state transformer conversion from 10kV to 800V across operating conditions, battery storage giving second-scale power support and minute-scale backup, and millisecond coordination between the two to absorb GPU load swings locally. Tom's Hardware notes the system is built for higher computing density in a smaller footprint. Whether it holds up at 3GW is a question the deployment has not yet answered.

The software layer is changing too, and not comfortably. SUSE CEO Dirk-Peter van Leeuwen told the company's 1,900 employees by internal email that SUSE will cut about 10% of its global workforce, citing "automation and AI", according to heise online, which credits The Register with breaking the story on Monday. "These decisions are painful for our employees, our corporate culture and our open source community," van Leeuwen wrote, per the same report. "However, they are necessary so that we can adapt and build a sustainable, resilient SUSE for the long term." Affected staff were to be informed by Wednesday, employee representatives named on Thursday, and individual consultations start early next week, with the rest of the company hearing the 2027 strategy on 16 November. The Register reported that SUSE had already asked long-serving staff in September whether they would take voluntary termination, and apparently did not get enough takers.

The security bill for physical AI

Automation also widens the attack surface. VicOne said it identified 28 zero-day findings in less than two days using AI-assisted research across software and tools supporting the physical AI ecosystem, including projects developed or maintained by Google, NVIDIA and Meta, The Robot Report wrote on 1 November. All 28 were classified as zero-day vulnerabilities on initial validation, though the count may change after vendor and authority review and some findings have no CVE numbers yet. VicOne said 75% of the findings involved development, training, simulation or validation tools, the stages before a robot is deployed. The consequences described are potential, not observed: a tampered model that behaves normally in testing, a compromised simulation that lets a robot pass a test it should have failed, or an operator console that goes dark mid-shift.

Hardware suppliers are making their own reliability arguments. In a sponsored piece published by The Robot Report on 1 November, 3M said its CoaXPress-compatible cable assemblies are flex-cycle tested to 45 million cycles and its USB3 Vision-compatible assemblies to 100 million, with flat-ribbon options spanning roughly 1 million to 100 million cycles. The company also claims its insulation displacement contact connectors, a technology it invented in 1961, can save up to 80% in wiring time compared with crimp-style connectors. Those numbers come from the vendor, not from independent testing.

Outside the robot vendors, the industrial logic is spreading. Volkswagen has opened an eight-stage industrial dismantling line at Zwickau, the plant it converted to EV production through a €1.2 billion programme begun in 2018, InsideEVs reported on 1 November. From 2027 the line will process 4,100 vehicles a year across two shifts with 100 retrained workers, recovering parts that VW has been selling through dealer and independent workshop channels since July, with doors, bumpers, motors and transmission components reportedly in high demand. VW plans to invest up to €90 million in the recycling business, with the Saxony state government pledging €10.7 million, and says it could eventually support 1,000 jobs. It has not disclosed prices, volumes, warranty terms or inspection procedures, and no vehicle production is planned at Zwickau beyond 2031.

The European startup scene is betting on faster setup rather than cheaper hardware. Paris-based Bleu Robotics raised a €5 million pre-seed round led by Contrarian Ventures and OVNI Capital, Sifted reported on 29 October, on the promise of cutting humanoid deployment time in factories from months to under an hour. Sifted's article is partly garbled in the copy available, so the specifics of the training method are best treated with caution; the funding figure and the investor names are clear.

Trade policy could yet reshape where this automation gets built. Several Chinese automakers have travelled to Beijing for talks with regulators on hybrid vehicle export quotas to the EU, CnEVPost reported on 2 November, citing Cailian. Bloomberg had reported on 11 October, citing a proposal document, that the first-year quota would be 350,000 vehicles, rising 6% annually and lasting four years, possibly from 1 December. The Chinese commerce ministry said on 12 October that the arrangement would allow growth from the 302,000 hybrids the EU imported from China in 2025. Sefcovic said on 9 October the understanding included halving Chinese HEV and PHEV exports to the EU but did not specify the baseline, so the two sides' public descriptions do not line up cleanly.

The through-line is less dramatic than the humanoid headlines. Robots are being sold by the hour, by the pick and by the multi-year order, and the companies that reach public markets will be judged on whether those contracts convert into revenue. Agility's listing on 5 November is the next data point.