---
title: "Tutor Intelligence $75M: Cassie, Sonny & Robot Pricing"
description: "Tutor Intelligence's $75M a16z-led Series B puts Cassie and Sonny warehouse robots in focus. Actual hourly fees, setup costs, data flywheel and competitive alternatives."
date: 2026-11-03T11:17:01.595Z
source: https://venturecapitaltracker.com/tutor-intelligence-75m-series-b-robot-pricing-cassie-sonny-2026
---

# Tutor Intelligence $75M: Cassie, Sonny & Robot Pricing

> Tutor Intelligence's $75M a16z-led Series B puts Cassie and Sonny warehouse robots in focus. Actual hourly fees, setup costs, data flywheel and competitive alternatives.

**Tutor Intelligence announced a $75 million Series B on November 2, 2026, led by Andreessen Horowitz (a16z), with Union Square Ventures and Neo returning.** The Watertown, Massachusetts robotics company's cumulative funding reached a reported **$117 million**. It did **not** announce a valuation or revenue figure. The money is intended to help train robot foundation models and manufacture and deploy more robotic workers. [Read the company's issuer-signed funding statement](https://www.businesswire.com/news/home/20261102314852/en/) and [a16z's investment thesis](https://a16z.com/announcement/investing-in-tutor-intelligence/).

What makes this more than another physical-AI round is the actual commercial offer. Tutor leases robot *labor*, not just hardware: its Cassie machine handles pallets and cases, while its two-armed Sonny targets smaller ecommerce items. Operators at real warehouses can use remote human help whenever the robots encounter tasks they cannot finish reliably. That produces useful work now and training examples for later models. The advertised entry rate starts at **$14 per robot-hour**, although the detailed pricing terms introduce meaningful setup costs that a headline rate alone hides.

## Funding, investors and measurable facts

| Item | What is established |
| --- | --- |
| New financing | **$75M Series B**, announced November 2, 2026 |
| Lead investor | **Andreessen Horowitz** |
| Returning investors | **Union Square Ventures** and **Neo** |
| Prior round | Tutor announced a **$34M Series A** led by USV in December 2025 |
| Total money raised | **$117M** reported cumulative, including the latest round; not the amount of the new raise |
| Valuation and ARR | **Not disclosed** |
| Product | Cassie for palletizing/case picking; Sonny for finer each-picking |
| Company | Founded by MIT alumni Josh Gruenstein and Alon Kosowsky-Sachs; Massachusetts-based |
| Customer payments | Hourly fees for palletizing; quotes tied to completed picks for case/each picking |
| Future goal | More than 1,000 robots **under contract by end-2027**, an investor-reported target, not the current deployed fleet |

Sources: [Tutor's funding announcement](https://www.businesswire.com/news/home/20261102314852/en/), [its earlier company blog](https://tutorintelligence.com/blog) and [a16z partners Raghu Raghuram, Martin Casado and Jamie Sullivan](https://a16z.com/announcement/investing-in-tutor-intelligence/). Another [contemporary independent account, The Next Web](https://thenextweb.com/news/tutor-intelligence-75m-series-b-a16z-robot-workers), separately reported the funding, operating model and absent valuation.

## What the robots actually do: a warehouse buyer's first visit

Imagine a packaging factory with cartons leaving a conveyor and pallets waiting to be stacked. An employee ordinarily watches the end-of-line flow, lifts mixed cartons, follows a pallet pattern and changes the setup when products change. Tutor proposes to take over that defined job, rather than claiming a general humanoid can replace every warehouse role.

1. **Confirm the task and constraints.** A production lead gives Tutor throughput, carton sizes, line layout, shift hours and safety requirements. For Cassie, Tutor advertises palletizing, depalletizing and case picking. Its [Cassie product documentation](https://tutorintelligence.com/cassie) lists a roughly 50-pound payload limit, a 16-hour mobile-battery figure, 5–18-plus cases per minute depending on configuration and a typical two-day on-site installation. These are **vendor specifications**, not an independently guaranteed rate for every customer.
2. **Receive the mobile robot and agree an operation.** Cassie can be positioned near a line with power and network access. Tutor says an ordinary palletizing deployment needs no mandatory warehouse-management-system integration; larger mobile picking jobs can use its WMS-facing software and EVE I/O hub. A no-code tablet lets warehouse staff start and change pallet patterns.
3. **Run a real shift.** The robot recognizes cartons, picks and stacks them. In a complex case, Tutor's remote **human tutors** can supervise, correct or directly operate it. The site does not need to assume autonomy is perfect. Tutor claims remote service, equipment maintenance and model updates are included in the use-based rate.
4. **Monitor output and exceptions.** Supervisors use Tutor Portal for live streams and performance data; operational metrics should include cases actually handled, failed grips, human interventions, downtime and whether the robot blocks line capacity. A production deployment is a commercial installation, not proof that every advertised benchmark has been reached.
5. **Pay for measured work and expand selectively.** Factory palletizing is quoted by operating hour; case or each picking is quoted per completed pick. The buyer should compare the fully loaded cost of the specific work, not just robot hourly rate versus employee wages.

Tutor's [Cassie technical page](https://tutorintelligence.com/cassie), [case-picking documentation](https://tutorintelligence.com/case-picking) and [pricing terms](https://tutorintelligence.com/pricing) support this workflow. It is an **illustrative purchaser walkthrough**, not VCT's own hands-on deployment.

## Tutor robot pricing: $14–18/hour does not mean zero upfront cost

The company's [public rate sheet](https://tutorintelligence.com/pricing) is more informative than most robotics funding articles:

| Product/service | Published economics | Limitations |
| --- | --- | --- |
| Cassie palletizing/de-palletizing | **$14–18 per robot-hour** | Rate depends on scope and actual agreement |
| Minimum hourly usage | **80 hours/month** | At listed rates, $1,120–$1,440/month *in usage charges* at that minimum, before taxes and other fees |
| Implementation | **$10,000 flat fee** | Upfront outlay, not amortized into $14/hr |
| Refundable deposit | **$8,000** | Refundability does not mean no cash tied up |
| Case/each picking | Individually quoted per pick, with a stated **10–50% savings ambition** against current cost/pick | A vendor proposal, not a published universal price or independent realized-saving study |
| Ownership alternative | Cassie available for outright purchase with annual services | Purchase price and full annual service fee are not published |

The minimum hourly arrangement implies **$18,000 of implementation/deposit cash commitments** before usage charges, although the deposit is described as refundable. At 80 hours/month, a buyer should not call the service a $1,120 **all-in first month**. This is a key diligence correction to simplified coverage that describes the offer as no upfront expenditure.

A comparison should also examine whether overtime shifts can use the robot, what failure or interruption hours count as billable, minimum utilization, warehouse safety and insurer requirements, whether teleoperation creates data/privacy obligations, and how much in-house supervision is still needed. Tutor's **10–50% cost-per-pick improvement** is a published marketing claim; contract-level evidence and comparable baseline accounting would be needed to verify it.

## Cassie versus Sonny: do not mix their capabilities

**Cassie** is a mobile, single-arm system built around cases and pallets. Its documented capabilities include recognizing a mixed assortment of cartons, pallet pattern changes, site mobility and WMS-facing picking workflows. A published customer example is **Productiv's Dallas facility**, a selected reference on Tutor's [Cassie page](https://tutorintelligence.com/cassie); it is not an independently randomized comparison with human labor.

**Sonny** has two arms and is positioned for finer individual-item handling in ecommerce fulfillment. Its vision-language-action research is partly developed in **Data Factory 1 (DF1)**, which Tutor describes as **100 Sonny robots used for data collection**. Those **100 research robots must not be counted as 100 paying customer deployments**. [Tutor's April 2026 engineering explanation](https://tutorintelligence.com/blog/building-a-100-robot-data-factory-toward-factory-ready-ai) describes human-supervised data collection, the Ti0 model, feedback labeling and a test environment in which a failure encountered by one robot can be found across the fleet more quickly.

The distinction matters to a prospective customer: an impressive video of item manipulation is not a service-level agreement for live fulfillment. Before buying, ask which robot model is commercially available for this exact workflow, what task exceptions require a tutor, how the work is priced and what site-specific throughput has been demonstrated.

## The AI training flywheel: useful research or expensive human labor?

The investor's thesis is unusually concrete. [a16z argues](https://a16z.com/announcement/investing-in-tutor-intelligence/) that Tutor began with paying jobs and remote human operation, used those operational trajectories to train models, then redeployed improved software onto more robots. In principle:

**Paid warehouse job → human intervention on edge cases → labeled real-world robot demonstrations → model training → fewer interventions → more jobs at better economics.**

But two ratios matter much more than the size of any demo model: **successful autonomously completed tasks per paid operating hour**, and **human tutor minutes required per robot-hour**. Without those, investors cannot tell whether more installed robots improve gross margin or simply require more remote operators. The company has not published independently audited unit economics, contribution margins or intervention rates. Vendor and investor descriptions should not be misread as causal proof.

a16z also reported that **nearly one-third of customers receiving Tutor robots during the previous 12 months were existing clients ordering additional robots**. That is a potentially interesting expansion signal, but the underlying cohort count, revenue retention and deployment uptime are undisclosed. The reported goal of **1,000-plus contracted robots by end-2027** remains forward looking.

## Compare actual alternatives before a factory signs

| Approach | Where it fits best | Key trade-off |
| --- | --- | --- |
| **Tutor Cassie / Sonny** | Vendor-managed case/pallet tasks and developing each-pick automation | Human tutor dependence and exact all-in costs need diligence |
| [Boston Dynamics Stretch](https://bostondynamics.com/products/stretch/) | Automated mobile case handling, particularly unloading and case-picking workflows | Evaluate job-specific integration, capacity, availability and pricing rather than robot appearance |
| [Locus Robotics](https://locusrobotics.com/why-locus-robotics/robots-as-a-service/) | Large-scale assisted order picking, transport and warehouse orchestration with RaaS billing | The subscription model is **not unique to Tutor**; tasks and human/robot division differ |
| **Fixed automation / ASRS** | Highly repeatable high-volume operations and well-defined storage layouts | Larger infrastructure commitment and process-change risk |
| **Conventional staffing or third-party logistics** | Highly variable, irregular work or sites where automation cannot clear safety/ROI thresholds | Labor availability, shifts, injury risk and performance variability |

This is a workflow comparison, **not a benchmark ranking**. Tutor's advantage would be demonstrated in the particular buyer's throughput, flexibility and total cost—not simply through the existence of remote tutors.

## What the $75M really needs to prove

Tutor has an identifiable product, published commercial terms, a customer operating story and an engineering approach that bridges supervised work and model training. That is substantially more actionable than a generic robotics fundraising claim.

The critical unresolved questions are current deployed versus research robot count, actual paid utilization, customer concentration, real labor savings after implementation/deposit and supervision, teleoperation-data governance, and whether improving autonomy reduces human costs quickly enough to finance fleet growth. Neither a16z's optimism nor 100 DF1 research robots settle those questions.

**Bottom line:** this Series B finances a company trying to sell useful warehouse output **today** while learning better robotic autonomy **tomorrow**. Its real competitive test is the invoice and the operational scorecard at a warehouse, not the $75M funding figure alone. See [Tutor Intelligence's VCT startup profile](/startup/tutor-intelligence) and related [industrial AI investment analysis](/industrial-ai-startups-factory-automation-2026).

**By:** [Venture Capital Tracker](https://venturecapitaltracker.com/editorial-policy)

**Editorial note:** AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.