Locus Robotics

A warehouse automation company with genuine commercial traction, a credible legacy product, and an autonomous manipulation claim that has yet to survive independent scrutiny.

---

How to Read This Report

This report applies a four-tier evidence framework throughout. Every material assertion is labelled according to the strength of its underlying source. Readers should weight conclusions accordingly.

Bracketed numerals [n] refer to the numbered source list in §14. Sources are drawn exclusively from the research dossier compiled for this report. No sources have been invented or extrapolated.

---

01Executive Overview

Locus Robotics occupies a specific and commercially validated niche in the warehouse automation market: it sells autonomous mobile robots (AMRs) and the software platform to orchestrate them, primarily to third-party logistics providers and large retailers operating high-throughput fulfilment centres. The company is headquartered in Wilmington, Massachusetts, and has raised over $330 million in venture funding across four rounds, reaching a stated valuation of approximately $2 billion as of its Series F close in November 2022 ¹⁰. It is privately held with no IPO on record ⁸.

The commercial foundation is solid by the standards of the AMR sector. The flagship deployment with GEODIS — 1,000 robots across 14 sites in the United States and Europe — is confirmed by a named customer press release ¹¹ and represents one of the larger documented AMR rollouts in the logistics industry. The company's pricing model, a robotics-as-a-service (RaaS) subscription at approximately $1,100 per robot per month on a 36-month contract ⁷, is consistent with independent analyst estimates ⁹ and provides a recurring revenue structure that investors in the sector have come to favour.

Where the picture becomes more complicated is at the frontier of the product line. Locus Robotics launched the Locus Array — a mobile manipulation system combining an omnidirectional base, a vision system, and a robot arm with a proprietary AI-powered gripper — with claims of fully autonomous picking, putaway, and replenishment without human involvement ¹⁴. The company claims productivity improvements of two to three times over baseline and a 90% reduction in picking and putaway labour ³. These figures are not independently verified for the Array specifically, and a community benchmark of AI picking models — the most directly comparable independent data point available — found the best-performing system achieving approximately 64 picks per hour against a human baseline of roughly 1,300 picks per hour ¹⁸. That is approximately 5% of human throughput, a gap that is not bridged by the vendor's marketing materials.

The editorial verdict is that Locus Robotics is a genuine, revenue-generating business with a proven human-assisted AMR product and an ambitious but unproven autonomous manipulation product. The legacy Origin and Vector robots are productivity tools that augment human pickers; they are not autonomous picking systems. The Locus Array may eventually deliver on its autonomous claims, but independent production-scale evidence does not yet exist. Investors, prospective customers, and competitors should treat the two product lines as categorically different propositions requiring different due-diligence standards.

Latest news

• How 8,000 robots are changing work inside logistics giant DHL Supply Chain

  Fortune·2026-05-20GENERAL
• Locus Robotics Acquires Nexera Robotics to Advance Mobile Manipulation

  Business Wire·2026-05-19FUNDING

---

02The Locus Robotics Story

Origins and Founding Context

Locus Robotics emerged from Quiet Logistics, a third-party logistics provider that had been an early customer of Kiva Systems — the warehouse robotics company acquired by Amazon in 2012 for $775 million. When Amazon absorbed Kiva and withdrew its robots from the open market, Quiet Logistics faced the loss of its core automation infrastructure. Rather than accept dependence on a competitor's technology roadmap, Quiet Logistics spun out a robotics engineering team to build a replacement. That team became Locus Robotics, incorporated and operating from Wilmington, Massachusetts ¹.

This origin story is commercially significant. Locus Robotics did not begin as a research project or a venture-backed concept looking for a problem to solve. It began as an operational necessity inside a functioning logistics business. The founding team understood warehouse workflows, picker behaviour, and the economics of fulfilment operations from the inside. That context shaped the product: rather than attempting to replicate the Kiva goods-to-person model immediately, Locus built a robot-to-picker system in which autonomous mobile robots navigate to pick locations and guide human associates, who perform the physical pick. The human remains in the loop for the manipulation task — the hardest part of the problem — while the robot handles navigation, tote transport, and task sequencing. This was a pragmatic engineering choice that allowed the company to deploy at scale before autonomous manipulation was commercially viable.

Funding History

The company's funding trajectory reflects the broader wave of investor enthusiasm for warehouse automation that peaked in 2021 and moderated thereafter.

The Series E raise of $151 million in February 2021 came at the height of pandemic-driven e-commerce growth, when warehouse automation was attracting capital at historically elevated multiples. The Series F, closed in November 2022, brought in $117 million at a valuation approaching $2 billion ¹⁰, but the timing is notable: by late 2022, the post-pandemic normalisation of e-commerce volumes was already underway, and the broader venture market was contracting. Whether the Series F valuation reflects durable enterprise value or peak-cycle pricing is a question that cannot be answered from public data alone. No subsequent funding round is documented in the dossier.

Key Milestones

The GEODIS partnership is the most significant independently verified commercial milestone. The agreement, confirmed by GEODIS's own press release, committed to deploying 1,000 LocusBots across 14 global warehouse sites over a 24-month period ¹¹. This is not a pilot or a proof-of-concept; it is a named, large-scale, multi-site deployment confirmed by the customer. It establishes that the Origin/Vector product line is operationally proven at meaningful scale.

The 2024 acquisition of Nexera Robotics — a company specialising in adaptive gripper technology — brought the NeuraGrasp AI-powered gripper into the Locus portfolio ¹⁴. This acquisition was the enabling technology for the Locus Array, which was subsequently launched and recognised with a 2026 RBR50 Robotics Innovation Award and a MODEX 2026 Best New Innovation finalist position ¹². Awards from trade publications and industry events are not evidence of commercial performance, but they do confirm that the product exists, has been publicly demonstrated, and has attracted attention from industry observers.

The SAP EWM integration, listed on the SAP marketplace ⁷, is a commercially meaningful signal. Enterprise warehouse management system integration is a prerequisite for deployment in large logistics operations, and SAP EWM is the dominant WMS in the enterprise segment. The listing confirms that the integration exists and has been through SAP's partner certification process.

---

03Product Portfolio: What Locus Robotics Actually Sells

Overview

Locus Robotics sells three hardware products and one software platform. The hardware products serve two distinct operational models: a human-assisted robot-to-picker model (Origin and Vector) and a claimed fully autonomous robots-to-goods model (Array). The software platform, LocusONE, orchestrates all three robot types and interfaces with customer warehouse management systems ¹².

Understanding which product does what is essential, because the company's marketing frequently blurs the boundary between the human-assisted and autonomous models when presenting productivity claims.

Locus Origin

VERIFIED: The Locus Origin is the company's entry-level AMR. It navigates autonomously to pick locations within a warehouse, guides human associates to the correct bin or shelf location, and carries totes between pick stations and induction or packing areas. The human associate performs the physical pick — reaching into the bin, selecting the correct item, and placing it in the tote. The robot handles navigation, sequencing, and transport ⁴⁹.

The Origin operates in shared human environments without requiring dedicated robot-only zones ⁹. This is a genuine operational advantage over goods-to-person systems that require significant infrastructure investment and zone separation.

Locus Vector

VERIFIED: The Locus Vector is a higher-capacity AMR in the same human-assisted model as the Origin, designed for higher-throughput environments or larger tote volumes. The fundamental operational model — human picker, robot navigator and carrier — is identical to the Origin ¹².

EDITORIAL INFERENCE: The Vector appears to be a capacity and throughput variant rather than a fundamentally different product. The dossier does not contain detailed independent specifications distinguishing the Vector from the Origin beyond the official site description.

Locus Array

VERIFIED (hardware description): The Locus Array combines an omnidirectional mobile base, a vision system, a robot arm, and the NeuraGrasp AI-powered gripper acquired from Nexera Robotics. It has a vertical reach of up to 10 feet and is described as capable of centimetre-level precision near double-deep shelving ¹⁴. The NeuraGrasp gripper is designed to adapt across variations in shape, surface texture, material, porosity, and weight ¹⁴.

COMPANY CLAIM: Locus Robotics states that the Array executes fully autonomous fulfilment in a robots-to-goods model, with no human performing picks, covering 100% of SKUs, and operating 24 hours a day, seven days a week ³¹⁴.

EDITORIAL INFERENCE: The Array is a recently launched product. It has received trade press coverage ¹⁴ and industry awards ¹², confirming it exists and has been publicly demonstrated. However, no independent source in the dossier confirms production-scale fully autonomous operation. The benchmark data available ¹⁸ — which tests AI picking models on a real warehouse task — found the best-performing system achieving 64 picks per hour against a human baseline of 1,300 picks per hour. The benchmark does not specifically test the Locus Array, and the dossier notes that the models tested were described as "mostly off-the-shelf and suboptimal." Nevertheless, the throughput gap is large enough to warrant scepticism about the vendor's productivity claims for the Array until independent production data is published.

UNKNOWN: Locus Robotics has not publicly disclosed Array-specific throughput data from production deployments, SKU coverage methodology, or failure mode rates.

LocusONE Platform

VERIFIED: LocusONE is the AI orchestration platform that manages fleet routing, task assignment, and integration with warehouse management systems including SAP EWM ⁷. System integrator partners include Hy-Tek, VARGO, and Zion Solutions Group ⁷.

COMPANY CLAIM: Locus Robotics describes LocusONE as providing AI-driven optimisation across the full fleet, adapting task sequencing in real time to warehouse conditions ¹².

EDITORIAL INFERENCE: The SAP EWM integration listing ⁷ is the strongest independent signal of the platform's enterprise readiness. WMS integration is technically demanding and commercially gating; the fact that it exists and is listed on the SAP marketplace suggests the platform has passed a meaningful certification threshold. The quality of the AI optimisation layer is not independently assessable from public sources.

Task Coverage

VERIFIED (with caveats): Locus Robotics claims coverage across picking (discrete, each, case, and batch), putaway, replenishment, induction, drop-off, slotting, returns processing, and inventory counting ²⁴. The MWPVL independent consultant review corroborates multi-task capability ⁹, though the review predates the Array launch and applies primarily to the Origin/Vector fleet.

Pricing and Deployment Economics

The Forrester TEI report ⁵ was published in 2019 and should be treated as indicative rather than current. The $1,100/month SAP marketplace figure is the most current available data point. The MWPVL all-in estimate of $35,000 per robot ⁹ may reflect a different contract structure or vintage; the dossier does not resolve this discrepancy definitively.

Products & versions

Locus Array

Fully autonomous robots-to-goods (R2G) mobile manipulation system featuring an omnidirectional base, vision system, robot arm, and NeuraGrasp AI-powered gripper; reaches up to 10 feet vertically with centimeter-level precision near double-deep shelving.

Locus Origin

Autonomous mobile robot that guides human associates to pick locations in the warehouse, carrying totes and navigating autonomously while workers perform the physical pick; part of the legacy human-assisted AMR fleet.

Locus Vector

Autonomous mobile robot in the Locus fleet that works alongside human associates for warehouse picking and putaway tasks, orchestrated by the LocusONE AI platform.

LocusONE

AI orchestration platform that manages and coordinates the entire Locus robot fleet (Array, Origin, Vector) across picking, putaway, replenishment, induction, slotting, returns, and inventory counting tasks.

---

04Technology Stack: Strengths and the Work That Remains

Navigation and Fleet Management

VERIFIED: The Origin and Vector robots navigate autonomously in shared human environments using onboard sensors and the LocusONE orchestration platform ¹⁹. The MWPVL independent review confirms that no human-exclusive zones are required and that the robots can be deployed on mezzanines ⁹. This is a meaningful operational advantage: it means the system can be installed in existing warehouse infrastructure without the capital expenditure required to build dedicated robot zones or install conveyor systems.

Fleet management — routing multiple robots through a shared space, avoiding collisions, managing charging cycles, and sequencing tasks — is a solved problem for the Origin/Vector fleet at the scale demonstrated by the GEODIS deployment ¹¹. Orchestrating 1,000 robots across 14 sites is a non-trivial systems engineering achievement, and the GEODIS deployment provides independent evidence that it works at that scale.

EDITORIAL INFERENCE: The navigation stack for the Origin/Vector fleet is the most technically mature component of the Locus Robotics system. It has been validated at scale by an independent customer. The LocusONE platform's AI optimisation layer is less independently assessable, but the existence of the SAP EWM integration ⁷ and the multi-site GEODIS deployment ¹¹ suggest it is operationally functional.

Manipulation and Grasping (Locus Array)

VERIFIED: The Locus Array uses the NeuraGrasp gripper, acquired from Nexera Robotics. The gripper is described as adapting across shape, surface, material, porosity, and weight ¹⁴. The arm has a vertical reach of up to 10 feet and centimetre-level positioning precision near double-deep shelving ¹⁴.

EDITORIAL INFERENCE: Adaptive grasping across arbitrary SKU populations is one of the hardest open problems in warehouse robotics. The claim of 100% SKU coverage ³ is the most aggressive in the vendor's portfolio and the least supported by independent evidence. The NeuraGrasp acquisition brings genuine IP — Nexera Robotics was a specialised gripper technology company — but acquisition of a technology does not automatically translate to production-scale performance. The benchmark data showing 64 picks per hour for the best AI picking model ¹⁸ is not specific to the Array, but it establishes a credible reference point for the difficulty of the task.

The gap between 64 picks per hour and the human baseline of 1,300 picks per hour ¹⁸ is not primarily a gripper problem; it reflects the combined challenge of perception, grasp planning, motion execution, and error recovery in a dynamic, unstructured environment. Even if the NeuraGrasp gripper is mechanically superior to the systems tested in the benchmark, the perception and planning stack must also perform at a level that closes this gap before the Array's autonomous claims are commercially meaningful.

Software and AI

VERIFIED: LocusONE integrates with SAP EWM on S/4HANA Cloud Private Edition ⁷. System integrator partners include Hy-Tek, VARGO, and Zion Solutions Group ⁷.

COMPANY CLAIM: Locus Robotics describes LocusONE as providing real-time AI-driven task optimisation, fleet routing, and adaptive sequencing ¹².

UNKNOWN: The architecture of the AI optimisation layer — whether it uses reinforcement learning, classical operations research methods, or a hybrid approach — is not publicly disclosed. The training data, model update frequency, and performance metrics for the LocusONE AI are not independently documented.

Safety Architecture

VERIFIED: The Origin and Vector robots are designed for human-shared environments and have been independently confirmed to operate without requiring human-exclusive zones ⁹. This implies the robots meet relevant safety standards for collaborative operation, though the specific certifications are not detailed in the dossier.

UNKNOWN: The safety architecture for the Locus Array — which introduces a robot arm capable of reaching 10 feet — is not independently documented. A mobile manipulation system operating in a shared human environment presents different safety challenges than a navigation-only AMR, and the dossier does not contain independent confirmation of the Array's safety certification status.

Strengths Summary

---

05Research, Papers, Authors and Labs

Academic and Research Footprint

The research dossier contains zero entries in the research category. Locus Robotics does not appear to have a significant academic publication record. This is not unusual for a commercially oriented robotics company — Kiva Systems, Boston Dynamics in its early commercial phase, and most AMR vendors have similarly thin publication records — but it is worth noting because it means the technical claims in the product literature cannot be cross-referenced against peer-reviewed methodology.

The NeuraGrasp gripper technology, acquired from Nexera Robotics, may have an associated research lineage, but no papers, authors, or laboratory affiliations are documented in the dossier. The dossier explicitly records zero research sources.

UNKNOWN: Whether Locus Robotics employs researchers who publish externally, collaborates with university laboratories, or has filed patents beyond the "patented single gripper" reference in the product documentation ¹⁴ is not publicly documented in the available sources.

Benchmark Data

The most directly relevant technical data point in the dossier is the community benchmark published on Reddit's robotics forum ¹⁸, in which a user tested four AI picking models on a real industrial warehouse task. The best-performing model achieved 64 picks per hour; the human baseline was 1,300 picks per hour. The benchmark author described the models as "mostly off-the-shelf and suboptimal."

This benchmark has significant limitations: it was conducted by an individual rather than a research institution, it was not peer-reviewed, it does not specifically test the Locus Array or NeuraGrasp, and the experimental conditions are not fully documented. It is included here because it is the only independent quantitative data point on AI picking performance in the dossier, and it establishes a credible order-of-magnitude reference for the difficulty of the autonomous picking problem ¹⁸.

EDITORIAL INFERENCE: The absence of peer-reviewed research from Locus Robotics is consistent with a company that has prioritised commercial deployment over academic publication. It is not evidence of technical weakness, but it does mean that independent technical scrutiny of the company's AI and manipulation claims is limited to trade press coverage and the community benchmark cited above.

Company-linked papers

This module is being compiled — no data to show yet.

Authors & labs

This module is being compiled — no data to show yet.

Code & simulation

This module is being compiled — no data to show yet.

Datasets & benchmarks

This module is being compiled — no data to show yet.

---

06Media Evidence Library: What the Videos Prove

Available Media Evidence

The research dossier records zero video sources. No video demonstrations, deployment footage, or media evidence from Locus Robotics are included in the compiled dossier. This section therefore cannot assess what specific demonstrations prove or fail to prove.

EDITORIAL NOTE: The absence of video sources in the dossier does not mean Locus Robotics has not produced demonstration videos — commercial robotics companies routinely publish promotional footage. It means that no video evidence was captured in the research compilation for this report, and this report will not speculate about the content or evidentiary value of videos not reviewed.

What Trade Press Coverage Establishes

The Robot Report's coverage of the Locus Array launch ¹⁴ provides the most detailed independent description of the product in the dossier. The Robot Report is a trade publication with editorial standards above marketing copy, and its coverage confirms: the hardware configuration (omnidirectional base, vision system, robot arm, NeuraGrasp gripper), the claimed vertical reach (10 feet), the claimed precision (centimetre-level near double-deep shelving), and the product's positioning as a fully autonomous fulfilment system. Trade press coverage of a product launch is not the same as independent performance validation, but it does confirm the product's existence and public specification.

Evidence Gaps

Media library

[COD20] Universal Locust Plague (Swarm Drone Infinite Stacking)

Bilibili56k views

---

07Commercial Reality

Revenue Model

VERIFIED: Locus Robotics operates a robotics-as-a-service (RaaS) model. The listed price on the SAP marketplace is $1,100 per robot per month on a 36-month contract, paid annually ⁷. This is a subscription model rather than a capital sale, which means revenue is recurring and predictable but requires the company to maintain and support the fleet throughout the contract term.

The RaaS model has structural advantages for customers — it converts capital expenditure to operating expenditure and transfers hardware obsolescence risk to the vendor — and structural advantages for Locus Robotics in terms of revenue visibility. It also creates a dependency: if a customer terminates or does not renew, the revenue disappears and the robots must be redeployed or written down.

Documented Deployments

VERIFIED: The GEODIS deployment is the only named, customer-confirmed large-scale deployment in the dossier. The agreement covers 1,000 LocusBots across 14 sites in the United States and Europe, deployed over a 24-month period ¹¹. This is a substantial deployment by any measure in the AMR sector.

UNKNOWN: The total number of customer deployments, the total number of robots deployed across all customers, and the renewal rate on existing contracts are not publicly disclosed. The dossier does not contain additional named customer confirmations beyond GEODIS.

EDITORIAL INFERENCE: A company that has raised $339 million and reached a $2 billion valuation almost certainly has more than one customer. The absence of additional named deployments in the public record may reflect customer confidentiality preferences rather than a thin customer base. However, the inability to independently verify the breadth of the customer base is a genuine due-diligence gap.

Productivity Claims: Claim vs. Evidence

The Forrester Total Economic Impact report ⁵ is the most detailed financial analysis in the dossier, but it was commissioned by Locus Robotics, published in 2019, and applies to the human-assisted Origin/Vector model rather than the Array. Commissioned analyst reports are not independent evidence; they are structured to present the vendor's case. The 2x–3x productivity figure from that report should be treated as a company claim with supporting analysis rather than independent verification.

Financial Position

VERIFIED: The company has raised over $330 million in total funding ¹⁰. The most recent round closed in November 2022 at a valuation approaching $2 billion ¹⁰. The company is privately held ⁸.

UNKNOWN: Revenue, EBITDA, cash burn rate, runway, and profitability are not publicly disclosed. No subsequent funding round after the November 2022 Series F is documented in the dossier. Whether the company is profitable, approaching profitability, or continuing to burn cash at a significant rate is not determinable from public sources.

EDITORIAL INFERENCE: The absence of a documented funding round since November 2022 — a period of more than three years — is ambiguous. It could indicate that the company has reached cash-flow sustainability and no longer requires external capital. It could also indicate that the company has been unable to raise at acceptable terms in a more challenging venture environment. The post-pandemic normalisation of e-commerce volumes, which reduced the urgency of warehouse automation investment for some operators, and the broader contraction in growth-stage venture funding since 2022 are relevant contextual factors. Neither interpretation can be confirmed from the available evidence.

SAP Partnership and Enterprise Integration

VERIFIED: Locus Robotics is listed on the SAP marketplace as a certified integration partner for SAP EWM on S/4HANA Cloud Private Edition ⁷. System integrator partners include Hy-Tek, VARGO, and Zion Solutions Group ⁷.

EDITORIAL INFERENCE: The SAP EWM integration is commercially significant beyond its technical function. SAP EWM is the WMS of choice for large enterprise logistics operators — the customer segment that can afford a $1,100/robot/month RaaS subscription at scale. Being listed on the SAP marketplace puts Locus Robotics in front of SAP's enterprise customer base and signals a level of technical and commercial maturity that smaller AMR vendors have not achieved. The system integrator partnerships are also meaningful: Hy-Tek, VARGO, and Zion Solutions Group are established logistics automation integrators with their own customer relationships, and their willingness to partner with Locus Robotics implies a degree of confidence in the product's deployability.

Customers & deployments

GEODISThird-Party Logistics (3PL) Provider

Signed an expanded agreement to deploy 1,000 LocusBots across 14 global warehouse sites in the U.S. and Europe over 24 months.

08Markets and Use Cases

Locus Robotics operates in a narrow but economically significant slice of the industrial automation market: the fulfilment and distribution centre segment of warehouse logistics. Understanding where the company's products actually fit — and where they do not — requires separating the addressable market from the served market, and the served market from the markets the company aspires to enter.

The Core Served Market

The company's documented deployments concentrate in two verticals: third-party logistics (3PL) operators and retail/e-commerce fulfilment. The GEODIS agreement — 1,000 LocusBots across 14 sites in the United States and Europe over 24 months — is the single most concrete public evidence of scale ¹¹. GEODIS is a 3PL, which is structurally important: 3PLs operate on thin margins, face intense labour cost pressure, and manage multiple client SKU catalogues simultaneously. These characteristics make them receptive to robotics-as-a-service (RaaS) pricing models, because capital expenditure avoidance matters more to them than to vertically integrated retailers who own their own fulfilment infrastructure.

The SAP EWM integration ⁷ signals a second target segment: mid-to-large enterprise shippers running SAP S/4HANA Cloud Private Edition. This is a meaningful qualification. SAP EWM is not a small-business warehouse management system; it is deployed by manufacturers, distributors, and retailers with complex multi-site inventory operations. Locus's certified integration with SAP EWM positions it for enterprise procurement cycles where WMS compatibility is a non-negotiable gate.

Use Case Taxonomy

The company's documented task coverage spans picking (discrete, each, case, and batch), putaway, replenishment, induction, drop-off, slotting, returns processing, and inventory counting ²³⁴. This breadth is a vendor claim rather than an independently verified operational reality across all task types, but the core picking and putaway functions are corroborated by the Forrester TEI report ⁵ and the MWPVL independent review ⁹.

The distinction between the legacy human-assisted model and the newer Locus Array autonomous model is commercially critical. For the Origin and Vector fleet, the robot is essentially a mobile tote carrier and navigation guide; the human worker remains the physical picking agent. This is a productivity amplification tool, not a labour replacement tool — a distinction the company's marketing language does not always make clear.

Facility Profile Requirements

The MWPVL independent review ⁹ and the Forrester TEI report ⁵ together sketch the facility profile where Locus deployments are viable. The system is designed for human-shared environments with no requirement for human-exclusion zones ⁹, which means it can be retrofitted into existing warehouse buildings without structural modification. The ability to deploy on mezzanines ⁹ extends the addressable footprint to multi-level facilities, which are common in urban fulfilment centres where land costs are high.

The deployment timeline of approximately three months ⁵ is relatively short by warehouse automation standards — fixed conveyor and sortation systems typically require 12 to 24 months of installation and commissioning. This speed-to-value proposition is a genuine differentiator for 3PLs managing seasonal volume spikes or short-term client contracts, where a multi-year capital project cannot be justified.

Seasonal and Peak Demand Dynamics

The RaaS model has a structural advantage in peak-demand environments. A 3PL operator running a retail fulfilment contract can, in principle, scale robot fleet size up for the fourth-quarter peak and down in the first quarter without stranding capital. Whether Locus's contracts actually permit this kind of elastic scaling is not publicly disclosed; the SAP marketplace listing references 36-month contracts ⁷, which implies a minimum commitment that partially offsets the flexibility narrative.

Markets Where Locus Is Not Present

The dossier contains no evidence of Locus deployments in manufacturing (assembly, quality inspection), healthcare (pharmacy, hospital logistics), food and beverage processing, or outdoor logistics. The company's sensor suite and gripper technology are calibrated for ambient-temperature, structured warehouse environments with predictable SKU geometries. Cold-chain fulfilment, fresh food picking, and unstructured outdoor environments are outside the documented operational envelope.

---

09Competitive Landscape

The warehouse AMR market is crowded, well-funded, and consolidating. Locus Robotics competes across two distinct product tiers: the human-assisted AMR tier (Origin, Vector) and the emerging autonomous manipulation tier (Locus Array). These two tiers have different competitive sets.

Human-Assisted AMR Tier

In the goods-to-person and robot-to-picker AMR segment, Locus's primary competitors are Fetch Robotics (acquired by Zebra Technologies in 2021), 6 River Systems (acquired by Shopify in 2019, subsequently sold), Geek+ (Chinese manufacturer with global deployments), and Körber's robotics division. Amazon Robotics (formerly Kiva Systems) dominates the captive market but does not sell to third parties, making it a market-shaping force rather than a direct competitor for external customers.

The competitive dynamics in this tier favour scale, integration breadth, and total cost of ownership. Locus's SAP EWM certification ⁷ is a meaningful moat for enterprise accounts already committed to the SAP ecosystem. Its RaaS pricing model at approximately $1,100 per robot per month ⁷ is broadly comparable to industry norms, though the MWPVL estimate of $35,000 all-in per robot ⁹ suggests that total deployment cost, including integration and deployment fees, is the more relevant comparison metric.

Autonomous Manipulation Tier

The Locus Array competes in a segment that is simultaneously more technically demanding and more commercially nascent. Direct competitors include Berkshire Grey (now part of SoftBank-backed portfolio), Covariant (AI-native robotic picking), Mujin, and RightHand Robotics. Larger system integrators such as Dematic and Vanderlande are also developing or integrating autonomous picking arms into their conveyor-based systems.

The critical competitive variable in this tier is picks-per-hour at acceptable error rates. The Reddit benchmark cited in the dossier ¹⁸ — 64 picks per hour for the best AI model tested, versus 1,300 for a human — is not specific to the Locus Array, but it is indicative of the state of the art in AI-driven robotic picking. At 64 picks per hour, a robotic picking system is economically viable only in narrow circumstances: extremely high labour costs, 24/7 operation requirements, or SKU profiles with very low variance. The vendor's claim of "2x–3x productivity improvement" ¹² almost certainly refers to the human-assisted Origin/Vector model, not the Array, but the marketing language does not consistently make this distinction.

Competitive Positioning Summary

The Consolidation Risk

The warehouse AMR market has already seen significant consolidation: 6 River Systems was acquired and then divested by Shopify, Fetch was absorbed by Zebra, and several smaller AMR vendors have exited. Locus's $2 billion valuation ¹⁰ was set in November 2022, at the peak of venture enthusiasm for robotics. The subsequent tightening of venture capital markets, combined with the technical gap between claimed and demonstrated autonomous picking performance, creates a scenario in which Locus faces pressure to either demonstrate the Array's commercial viability at scale or accept a strategic acquisition at a valuation below its 2022 peak. This is editorial inference, not a disclosed fact.

Competitive comparison

Robot	Maker	Autonomy	Conf.
iRobot Roomba Combo 10 Max	iRobot	Autonomous	0.90
Mobile ALOHA (Stanford)	Stanford University	Teleoperated	0.90
1X NEO	1X Technologies	Remote-Assisted	0.90

---

10Geopolitical Context and Constraints

Supply Chain Nationalism and Reshoring

The broader geopolitical context for Locus Robotics is, on balance, favourable. The reshoring of manufacturing and fulfilment capacity to North America and Western Europe — driven by supply chain disruptions during 2020–2022, US-China trade tensions, and the CHIPS and Science Act — increases the volume of warehouse and distribution centre construction in Locus's core markets. More domestic fulfilment capacity means more potential deployment sites.

However, reshoring also intensifies competition for warehouse labour in the short term, which is the primary demand driver for warehouse automation. If labour markets tighten further in the United States, the economic case for AMR deployment strengthens. If immigration policy changes increase the available labour pool, the urgency of automation investment may moderate. The net effect is uncertain and depends on policy trajectories that are outside the scope of this report.

China Exposure

The dossier does not disclose Locus's hardware supply chain in detail. This is a material unknown. If the company sources sensors, motors, or compute hardware from Chinese manufacturers — as is common across the robotics industry — it faces exposure to US export controls, tariff escalation, and potential supply disruptions. The MWPVL review ⁹ and Robot Report coverage ¹⁴ describe the hardware at a functional level (omnidirectional base, vision system, robot arm, NeuraGrasp gripper) but do not identify component suppliers.

The Nexera Robotics acquisition — the source of the NeuraGrasp gripper technology ¹⁴ — is a Canadian company, which reduces one dimension of geopolitical risk. But the broader supply chain question remains publicly undisclosed.

European Operations and Regulatory Environment

The GEODIS deployment spans both US and European sites ¹¹. European warehouse operations are subject to the EU Machinery Regulation (replacing the Machinery Directive 2006/42/EC), which came into force in 2023 and applies to autonomous mobile robots operating in shared human environments. Compliance with this regulation is a prerequisite for commercial deployment in EU member states. The dossier does not confirm CE marking or EU Machinery Regulation compliance status for the Locus Array specifically; the Origin and Vector fleet, being older products with established European deployments, presumably carries the necessary certifications, but this is not explicitly stated.

The EU AI Act, which classifies certain AI systems used in safety-critical environments as high-risk, may impose additional conformity assessment requirements on autonomous manipulation systems like the Locus Array if they are deployed in EU warehouses. The timeline for AI Act enforcement (2025–2027 for different risk categories) overlaps with the Array's commercial rollout period. This is a regulatory risk that is not publicly acknowledged in the company's communications.

Labour Relations and Social Licence

Warehouse automation is politically sensitive in jurisdictions with strong organised labour. The United States, the UK, and Germany have all seen public debate about the displacement effects of warehouse robotics. Locus's human-assisted model (Origin/Vector) is, in this context, a more politically defensible proposition than fully autonomous picking: it augments workers rather than replacing them. The Locus Array's "90% labor reduction" claim ³, if realised at scale, would attract scrutiny from labour unions and potentially from legislators in jurisdictions with worker protection frameworks.

The company's public communications emphasise productivity gains and worker safety rather than headcount reduction, which is a standard industry posture. Whether this framing survives contact with large-scale Array deployments that demonstrably reduce warehouse headcount is an open question.

---

11The Hype, the Real and the Ugly

This section applies systematic scrutiny to the gap between what Locus Robotics claims and what the available evidence supports. The company operates in a sector where marketing language routinely outpaces demonstrated capability, and the dossier contains several specific conflicts that warrant direct examination.

The Real: What the Evidence Actually Supports

Human-assisted AMR productivity gains are plausible. The Forrester TEI report ⁵, while commissioned by Locus and therefore not independent, uses a methodology (Total Economic Impact) that Forrester applies consistently across vendors, and its figures are corroborated in structure — if not in precise magnitude — by the MWPVL independent review ⁹. A bot-to-picker ratio of approximately 3.5:1, combined with reduced walking time for human associates, is a credible mechanism for the 2x–3x productivity claim in the worker-assist model. This is the most defensible of the company's performance claims.

The GEODIS deployment is real and substantial. 1,000 robots across 14 global sites ¹¹ is a verified, named-customer deployment at a scale that few warehouse robotics vendors have publicly documented. This is not a pilot or a press release partnership; it is a multi-year, multi-site operational commitment from a major 3PL.

The RaaS pricing model is coherent. The triangulation of $1,100 per robot per month (SAP marketplace ⁷), $950 per month (2019 Forrester model ⁵), and $35,000 all-in per robot (MWPVL ⁹) is consistent with a RaaS model that has evolved over time and varies by contract structure. None of these figures is implausible.

The NeuraGrasp gripper acquisition is a credible technical investment. Acquiring Nexera Robotics to obtain a purpose-built adaptive gripper ¹⁴ is a more defensible path to autonomous picking capability than attempting to develop gripper technology in-house. The gripper's claimed adaptability across shape, surface, material, porosity, and weight ¹⁴ is consistent with the requirements of a high-SKU-variety warehouse environment.

The Hype: Claims That Outrun the Evidence

"Fully autonomous fulfillment" and "100% SKU coverage." These claims ³ are the most aggressive in the company's portfolio and the least supported by independent evidence. The Locus Array is a newly launched product ¹⁴ that has won industry awards ¹² but has not been independently benchmarked at production scale. The 100% SKU coverage claim is particularly striking: no robotic picking system in commercial deployment has demonstrated reliable autonomous handling of the full SKU range of a high-variety warehouse, including irregular, fragile, soft-packaged, and very small items. This claim should be treated as aspirational until independently verified.

"90% reduction in picking and putaway labor." This figure ³ appears in the context of the Locus Array's autonomous model. If the Array achieves 64 picks per hour ¹⁸ — even accepting that this benchmark is not Array-specific — and a human achieves 1,300 picks per hour, then replacing a human with an Array robot reduces throughput by approximately 95%, not labour by 90%. The arithmetic only works if the Array is deployed in a context where 24/7 operation compensates for lower per-hour throughput, or if the benchmark figure is not representative of the Array's actual performance. The company has not published its own picks-per-hour figures for the Array.

"24/7 continuous autonomous operation." This is a theoretical capability of any robot that does not require sleep, but it elides the practical realities of maintenance downtime, battery charging cycles, error recovery, and the human oversight required when the system encounters edge cases. The documented operational model includes human co-presence in the warehouse ⁹, which is inconsistent with the implication of fully unattended operation.

The Ugly: Structural Concerns

The performance gap is large and publicly documented. The Reddit benchmark ¹⁸ — 64 picks per hour for the best AI model versus 1,300 for a human — is a community-sourced figure, not a peer-reviewed study, and it is not specific to the Locus Array. But it is the only independent quantitative throughput figure in the dossier, and it points to a gap of approximately 20:1 between current AI picking capability and human performance. Even if the Array outperforms the benchmark by a factor of five, it would still be operating at roughly one-quarter of human throughput. This is not a gap that marketing language can close; it requires demonstrated engineering progress.

The valuation was set at peak market conditions. The $2 billion valuation ¹⁰ was established in November 2022, during a period of elevated venture valuations for robotics companies. The subsequent correction in growth-stage valuations, combined with the technical challenges of autonomous manipulation, creates a structural tension between investor expectations and near-term commercial reality. This is not a claim that the company is in distress — the dossier contains no evidence of financial difficulty — but it is a risk factor that any serious investor or partner should weigh.

The autonomy narrative conflates two different products. The company's marketing materials ¹²³ present a unified narrative of autonomous fulfilment that spans the Origin/Vector fleet and the Locus Array. In practice, these are fundamentally different products with different autonomy levels, different competitive positions, and different risk profiles. The conflation serves the company's marketing interests but obscures the commercial reality for potential customers evaluating the system.

Claim tracker

Autonomous picking speed of the best AI model benchmarked on a real warehouse task is ~64 picks/hour, versus ~1,300 picks/hour for a humanSupported

An independently conducted community benchmark [18] tested four robot AI models on a real industrial picking task and found the best model achieved ~64 picks/hour; the ~1,300 picks/hour human baseline was also measured in the same study — though the benchmark does not specifically test the Locus Array hardware, leaving a gap in direct product-level validation.

GEODIS deployed 1,000 LocusBots across 14 global warehouse sites (U.S. and Europe) over 24 monthsSupported

A GEODIS press release [11] — an independent customer announcement — directly confirms the expanded agreement to deploy 1,000 LocusBots across 14 sites in the U.S. and Europe; this is a customer-issued statement, though actual operational outcomes (throughput, labor savings) at those sites are not independently audited.

The legacy Origin/Vector AMR model requires human associates to perform the physical pick; robots navigate and carry totes but do not manipulate itemsSupported

The MWPVL independent consultant review [9] and the Forrester TEI report [5] both explicitly describe the operational model as robots guiding human associates to pick locations while workers perform the physical pick, with a bot-to-picker ratio of ~3.5:1; this is further corroborated by the official picking page [4].

Locus Array hardware features an omnidirectional base, vision system, robot arm with NeuraGrasp AI-powered gripper, vertical reach up to 10 feet, and centimeter-level precision near double-deep shelvingUnknown

Robot Report [14] (independent trade press) and the official site [1] corroborate the hardware description including the NeuraGrasp gripper from the Nexera Robotics acquisition, but no independent third-party test has verified the claimed 10-foot reach, centimeter-level precision, or gripper adaptability across shape/surface/material in a production environment.

Locus Robotics has raised over $330M in venture funding at a ~$2B valuation (as of Series F, November 2022)Supported

The Series F press release [10] and Nasdaq Private Market listing [8] independently confirm the $117M Series F round bringing total funding over $330M at a ~$2B valuation; however, this reflects a 2022 snapshot and no subsequent funding round or updated valuation has been reported in the dossier.

Locus Robotics robots are safe for human-shared environments and require no human-exclusive safety zones, including deployment on mezzaninesUnknown

The MWPVL independent consultant review [9] corroborates the vendor's claim that no human-exclusive zone is required, but this assessment appears to be based on vendor-provided information rather than independent safety certification or regulatory audit, and no third-party safety standard compliance documentation is cited in the dossier.

---

12Future Scenarios

The following scenarios are editorial inferences constructed from the available evidence. They are not predictions, and they are not endorsed by Locus Robotics. They are intended to frame the range of plausible outcomes for a sophisticated reader.

Scenario A: The Array Delivers — Autonomous Picking at Scale (Probability: Low to Medium)

In this scenario, the Locus Array achieves commercially viable autonomous picking throughput — defined here as greater than 400 picks per hour at acceptable error rates — within 24 to 36 months of its commercial launch. The NeuraGrasp gripper's adaptive capability proves sufficient for the SKU variety of major 3PL clients. GEODIS or another named customer publicly confirms Array deployment at scale, providing the independent validation that is currently absent.

If this scenario materialises, Locus Robotics would be positioned as one of a small number of vendors with a proven end-to-end autonomous fulfilment system, combining fleet navigation (Origin/Vector), autonomous manipulation (Array), and unified orchestration (LocusONE). The $2 billion valuation would be defensible, and an IPO or strategic acquisition at a premium would become plausible.

The conditions required for this scenario are demanding: the AI picking models must improve substantially beyond the current state of the art, the gripper must handle the long tail of difficult SKUs reliably, and the system must demonstrate acceptable uptime in 24/7 production environments. None of these conditions is impossible, but none is currently demonstrated.

Scenario B: The Human-Assist Model Sustains the Business — Array Remains a Niche Product (Probability: Medium to High)

In this scenario, the Locus Array's autonomous picking performance remains insufficient for high-throughput fulfilment at competitive cost, and the company's commercial base continues to rest on the Origin/Vector human-assisted fleet. The Array finds a niche in specific use cases — overnight replenishment, low-velocity SKU picking, or environments where labour availability is the binding constraint rather than throughput — but does not displace human pickers at scale.

This is a viable business scenario. The GEODIS deployment demonstrates that the human-assisted model has real commercial traction. The RaaS pricing model generates recurring revenue. The SAP EWM integration provides enterprise sales channels. The company could sustain and grow this business without the Array ever achieving its most ambitious claims.

The risk in this scenario is valuation: a business built on human-assisted AMRs, competing against Fetch/Zebra, Geek+, and others, is unlikely to support a $2 billion valuation on a revenue multiple basis. A down-round or strategic sale at a reduced valuation becomes more probable.

Scenario C: Market Consolidation — Acquisition by a Larger Logistics or Technology Player (Probability: Medium)

The warehouse AMR market is consolidating, as noted in Section 9. Locus Robotics's combination of a large installed base, SAP EWM certification, a RaaS revenue model, and the Array's autonomous manipulation technology makes it an attractive acquisition target for several categories of buyer: large logistics equipment manufacturers (Dematic, Vanderlande, Honeywell Intelligrated), enterprise software companies seeking to extend into physical automation (SAP itself, Oracle), or large 3PLs seeking to internalise their automation capability (DHL, FedEx, UPS).

The GEODIS relationship ¹¹ is particularly notable in this context: GEODIS is a subsidiary of SNCF (the French national railway), which has the financial capacity for a strategic acquisition. Whether GEODIS has acquisition interest is not publicly disclosed.

An acquisition would likely be at a valuation below the $2 billion Series F mark, given market conditions since 2022, but would provide liquidity for investors and a path to scale for the technology.

Scenario D: Technical Stagnation and Distress (Probability: Low but Non-Negligible)

In this scenario, the Array fails to achieve commercial viability, the human-assisted AMR market commoditises further, and Locus faces difficulty raising additional capital at a valuation that satisfies existing investors. The company would face pressure to restructure, reduce headcount, or seek a distressed sale.

This scenario is not supported by any specific evidence in the dossier — there is no indication of financial distress, customer churn, or operational failure. But it is a non-negligible risk for any venture-backed robotics company operating at a $2 billion valuation in a market where the core technical challenge (autonomous manipulation at human-competitive throughput) remains unsolved.

Scenario Probability Summary

---

13What to Watch: A Live Monitoring Checklist

The following indicators are the most informative signals for tracking Locus Robotics's commercial and technical trajectory. They are ordered by analytical priority.

Technical Performance

• Locus Array picks-per-hour data from independent sources. The single most important missing data point in this report. Any peer-reviewed paper, independent consultant review, or named-customer case study that quantifies Array throughput at production scale should be treated as high-priority intelligence. The current benchmark of 64 picks per hour ¹⁸ is not Array-specific; Array-specific data would substantially revise the autonomy verdict.

• NeuraGrasp gripper performance on difficult SKUs. Soft packaging, very small items (under 50mm), fragile goods, and items with irregular surfaces are the known failure modes for robotic grippers. Any independent report of gripper failure rates or SKU exclusion lists would be significant.

• LocusONE AI platform updates. The orchestration software is the least publicly documented component of the system. Any technical publications, patent filings, or conference presentations describing the fleet management algorithms would illuminate the scalability of multi-robot coordination.

Commercial Signals

• New named customer announcements, particularly for the Locus Array. The GEODIS deployment ¹¹ is the only large-scale named deployment in the dossier. A second major named customer, especially one confirming Array deployment, would substantially strengthen the commercial case.

• Contract renewal or expansion announcements from existing customers. Renewals are more informative than new deployments because they indicate sustained operational satisfaction rather than initial adoption.

• SAP marketplace reviews or case studies. The SAP marketplace listing ⁷ is a sales channel that may generate customer reviews or case studies over time. These would provide independent (if potentially curated) evidence of deployment outcomes.

• System integrator activity. The named SI partners — Hy-Tek, VARGO, and Zion Solutions Group ⁷ — are indicators of channel health. If additional SIs certify on the Locus platform, it signals growing market confidence. If existing SIs reduce their Locus practice, it signals the opposite.

Financial and Corporate Signals

• Series G funding round or IPO filing. The Series F closed in November 2022 ¹⁰. Depending on the company's burn rate and revenue trajectory, a Series G or IPO filing would be expected within the 2024–2026 window. The terms of any new round — particularly the valuation relative to the $2 billion Series F mark — would be a direct indicator of investor confidence.

• Executive leadership changes. CEO transitions, CFO appointments, or departures of key technical leaders are often leading indicators of strategic pivots or financial pressure.

• Acquisition activity. Any announcement of Locus acquiring additional technology companies (as it did with Nexera Robotics) would indicate continued investment in autonomous manipulation capability. Conversely, any announcement of Locus itself being acquired would be a major strategic event.

Regulatory and Standards Signals

• EU Machinery Regulation compliance confirmation for the Locus Array. As noted in Section 10, the Array's EU deployment requires regulatory compliance that has not been publicly confirmed.

• EU AI Act classification of the LocusONE platform. If the European Commission classifies warehouse manipulation AI as high-risk under the AI Act, Locus would face additional conformity assessment requirements for EU deployments.

• OSHA guidance on autonomous manipulation in shared workspaces. US regulatory guidance on human-robot collaboration in warehouse environments is evolving. Any new OSHA standards or enforcement actions in this area would affect the operational model for all vendors in the sector.

Research and Publication Signals

• Academic or industry publications on NeuraGrasp or LocusONE. The dossier contains no research publications from Locus Robotics [research count: 0]. Any peer-reviewed publication describing the company's AI or gripper technology would be the first independent technical validation of its core IP.

• Patent filings. Patent applications in robotic grasping, multi-robot fleet management, or warehouse AI from Locus Robotics or its employees would indicate the direction of technical development.

---

14Sources and Methodology

Methodology

This report was produced using a structured evidence-assessment framework that distinguishes between four categories of information:

Verified Facts are statements supported by regulatory filings, official product documentation, named-customer confirmation, peer-reviewed or primary research, or corroboration from multiple independent sources. These are cited with bracketed numerals and treated as the evidentiary foundation of the report.

Company Claims are statements made by Locus Robotics in its own communications — website copy, press releases, product documentation — that have not been independently verified. These are identified as such in the text and in the claim-tracker table in Section 11.

Editorial Inferences are reasoned conclusions drawn from the available evidence. They are clearly labelled as such and represent the analytical judgement of the report's authors, not established facts.

Unknowns are material questions that are not publicly disclosed. The report identifies these explicitly rather than filling gaps with speculation.

The research dossier underlying this report was gathered on 21 June 2026 and contains 20 numbered sources across official company materials, commerce listings, news coverage, and community sources. The dossier contains zero peer-reviewed research publications from or about Locus Robotics, which is itself a material finding: for a company claiming advanced AI-driven autonomous manipulation, the absence of any published technical research is notable and limits the depth of technical assessment possible in this report.

The autonomy verdict of "Supervised-Autonomous" (confidence 0.62) reflects genuine uncertainty about the Locus Array's operational status. The legacy Origin/Vector fleet is clearly a human-assisted system. The Array claims full autonomy but lacks independent production-scale validation. The verdict will require revision as Array deployment data becomes available.

Source List

¹ Automated Warehouse Robots | Warehouse Robotics Solutions — https://locusrobotics.com/

² Boost Warehouse Productivity with Locus Solutions — https://locusrobotics.com/solutions

³ Autonomous Fulfillment Solutions for Optimized Productivity — https://locusrobotics.com/solutions/autonomous-fulfillment

⁴ Robotic Picking Solutions for Optimized Productivity — https://locusrobotics.com/solutions/warehouse-picking

⁵ The Total Economic Impact™ Of Locus Robotics — https://www.thenewwarehouse.com/wp-content/uploads/2019/12/Total-Economic-Impact-of-Locus-Robotics_June_2019F.pdf

⁶ Locus Robotics Stock $37.74 | How to Buy, Valuation, Stock Price, IPO — https://notice.co/c/locusrobotics

⁷ Locus Robotics, Inc. | Locus Autonomous Mobile Robots for SAP EWM — https://www.sap.com/products/scm/partners/locus-robotics-inc-locus-autonomous-mobile-robots-for-sap-ewm.html

⁸ Sell or Invest in Locus Robotics Stock Pre-IPO — http://www.nasdaqprivatemarket.com/company/locus-robotics

⁹ Locus Robotics - Independent Consultant Review — https://www.mwpvl.com/html/locus_robotics_-_independent_consultant_review.html

¹⁰ LOCUS ROBOTICS ANNOUNCES $117 MILLION IN SERIES F FUNDING, BRINGING ITS VALUATION CLOSE TO $2 BILLION — https://www.prnewswire.com/news-releases/locus-robotics-announces-117-million-in-series-f-funding-bringing-its-valuation-close-to-2-billion-301688540.html

¹¹ GEODIS Signs Expanded Agreement with Locus Robotics to Deploy 1,000 LocusBots at Global Warehouse Sites | GEODIS — https://geodis.com/us-en/newsroom/geodis-signs-expanded-agreement-locus-robotics-deploy-1000-locusbots-global-warehouse

¹² Press Release Archives - Locus Robotics — https://locusrobotics.com/blog/news-type/press-release

¹³ Locus Robotics: Latest News & Press Releases — https://locusrobotics.com/news

¹⁴ Locus Robotics launches Locus Array for fully autonomous fulfillment — https://www.therobotreport.com/locus-robotics-launches-locus-array-for-fully-autonomous-fulfillment

¹⁵ What are the top companies for robotics? - Reddit — https://www.reddit.com/r/robotics/comments/1fo6gom/what_are_the_top_companies_for_robotics

¹⁶ With current advances in robotics, robots are capable of kicking very... — https://www.reddit.com/r/singularity/comments/1pfop23/with_current_advances_in_robotics_robots_are

¹⁷ Anybody who says that there is a 0% chance of AIs being sentient is overconfident — https://www.reddit.com/r/ArtificialInteligence/comments/1iolln3/anybody_who_says_that_there_is_a_0_chance_of_ais

¹⁸ [Project] I benchmarked 4 robot AI models on a real industrial task. The best one does 64 picks/hour. A human does 1,300. — https://www.reddit.com/r/robotics/comments/1s8tt6j/project_i_benchmarked_4_robot_ai_models_on_a_real

¹⁹ Top-down vs. swarm intelligence in robots — https://www.reddit.com/r/IsaacArthur/comments/1qov663/topdown_vs_swarm_intelligence_in_