Nuro

Magic AI

Nuro

For most of its life Nuro sold delivery-as-a-service. Retailers, restaurants and grocery chains paid Nuro to handle the last mile with autonomous vehicles, so instead of employing human drivers a partner simply used Nuro's robot fleet.

The economics are compelling on paper. A human delivery driver costs $15-25 per hour including wages, insurance, and vehicle costs.

A Nuro vehicle costs money to build and maintain, but once deployed, it operates nearly 24/7 with no driver wages, no tips, and no breaks. The breakeven point comes when the fleet reaches sufficient utilization in a market.

That was the plan, anyway. In 2024 Nuro changed course and stopped trying to run a delivery service itself.

The business now is licensing the autonomy stack — the Nuro Driver — to carmakers and mobility companies, which is a far cheaper way to grow than building and operating your own fleet. FedEx, Domino's, Kroger, Walmart and 7-Eleven had all tested or deployed the delivery bots before the pivot.

Magic AI

Magic is building an AI that works like a software engineer, not a chatbot. The goal is a model you can hand a whole code base and a task, and it writes, reviews, and debugs like a teammate.

Their edge is memory. In 2024 they showed off LTM-2-mini, a model with a 100 million token context window.

To put that in perspective, that is about 10 million lines of code the model can look at in one shot, by far the biggest of any model at the time. The plan is to sell this as an AI coworker to companies that write a lot of software.

As of the 2024 rounds, the product was still not for sale.

Nuro

Dave Ferguson and Jiajun Zhu were both principal engineers at Google's self-driving car project (later Waymo). Ferguson led the machine learning and motion planning teams.

Zhu worked on perception — teaching cars to see and understand the world. They were two of the most senior engineers in autonomous driving.

In 2016, they left Google with a contrarian insight: the hardest part of self-driving cars wasn't the technology. It was the stakes.

A robo-taxi carrying passengers needs to be essentially perfect — any accident could injure or kill someone inside and destroy public trust. But a delivery vehicle carrying groceries?

If it gets in a fender bender, the worst case is squished bananas.

By removing the human from the vehicle, Nuro eliminated the most complex variable in autonomous driving safety. Their vehicles are small, light (under 1,200 pounds), and slow (max 25 mph on surface streets).

If one hits a pedestrian, the impact energy is dramatically lower than a 4,000-pound car going 40 mph. This let them get regulatory approval years before full-size robo-taxis.

The first prototype, R1, looked like a cartoon car — about half the width of a normal vehicle, with no windows, no mirrors, and no driver's seat. Just cargo space and sensors.

They revealed it in 2018 and began testing in Scottsdale, Arizona with Kroger for grocery deliveries.

Magic AI

Eric Steinberger and Sebastian De Ro started Magic in 2022. Steinberger is an AI researcher who became convinced that the missing piece for a true AI software engineer was memory.

Chatbots forget what you told them a few thousand words ago. He wanted a model that could read your entire code base and remember all of it.

They raised a small seed in 2022, then a $23 million Series A in early 2023 led by Alphabet's growth fund CapitalG. From day one the pitch was not a chatbot.

It was an AI colleague that could handle real engineering work.

Nuro

Nuro's strategy is to become the default autonomous delivery infrastructure for major retailers. Rather than build a consumer-facing delivery app (competing with DoorDash and Instacart), Nuro provides the robotic fleet that powers deliveries for existing brands.

The partnership-first approach reduces go-to-market friction. Kroger, Walmart, and FedEx already have massive customer bases and order volumes.

Nuro provides the autonomous last-mile delivery layer. The retailer gets lower delivery costs.

Nuro gets guaranteed demand.

That was the original plan. Houston and Mountain View were the early markets, mapped city by city, each one needing its own regulatory approval before a single vehicle moved.

It was slow and it was expensive, and in 2024 Nuro gave it up in favour of licensing its self-driving software to companies that already build and operate vehicles.

Magic AI

Magic's strategy is research first, product later, which is both unusual and expensive. Instead of shipping an app and iterating in public, they raised huge sums to chase a technical breakthrough in model memory.

The headline move was the 100 million token context claim, which got the whole AI world talking. They also cut a deal with Google Cloud in 2024 to build two supercomputers, called Magic-G4 and Magic-G5, using NVIDIA chips, to train bigger models.

Star backers like Eric Schmidt served as validation. The bet is that if the tech is far enough ahead, the product and the customers follow.

Nuro

The technology still has limitations in complex scenarios. Nuro vehicles operate on surface streets at low speeds — they can't handle highways, heavy snow, or extremely dense urban environments like Manhattan.

This limits the addressable market to suburban deliveries in good weather, which is a large market but not the whole market.

The path to profitability is long. Building custom autonomous vehicles is extraordinarily capital-intensive.

Each generation of vehicle requires hundreds of millions in design, engineering, testing, and manufacturing. Nuro has burned through most of its $2.1 billion in funding and laid off 30% of staff in 2023, signaling the cash crunch is real.

Competition comes from multiple directions. Amazon is developing its own delivery robots (Scout, though paused).

Waymo could pivot to autonomous delivery, though Cruise is no longer a threat after GM shut it down at the end of 2024. Established delivery companies could partner with other autonomy providers.

And human drivers remain cheaper than robots at current scale — the economics only flip once Nuro has hundreds or thousands of vehicles per market.

Magic AI

Magic has raised close to half a billion dollars and, years in, still has no product the public can buy. That is the whole problem in one sentence.

A giant context window is a great demo, but rivals like Cursor, GitHub Copilot, and Cognition's Devin already have real users and revenue. Magic has to prove that more memory actually makes a smarter engineer, not just a bigger number on a slide.

And it has to do that while burning enormous sums on compute against competitors who are already shipping. The clock and the cash both keep running.

Nuro

Nuro R3 — the third-generation autonomous delivery vehicle, purpose-built with no passenger compartment. Larger cargo capacity than R2, improved sensor suite, and designed for commercial-scale manufacturing.

Nuro Autonomy Platform — the self-driving software stack including perception, prediction, planning, and control that runs Nuro's vehicles. Nuro Driver — the AI system that handles all driving decisions in real-time, combining lidar, cameras, radar, and thermal sensors.

Nuro Fleet Management — cloud-based tools for partners to monitor, dispatch, and manage Nuro vehicles across delivery zones.

Magic AI

The centerpiece is Magic's LTM, or Long-Term Memory, model family. In 2024 they unveiled LTM-2-mini, a model with a 100 million token context window, pitched as the key to an AI that can understand an entire code base at once.

The broader product is an AI software engineer, or coworker, that can write and change code across a whole project. The company also announced Magic-G4 and Magic-G5, two supercomputers built with Google Cloud and NVIDIA to train these models.

Much of it is still research rather than a shipping product.

Nuro

SoftBank Vision Fund led the massive $940 million Series B in 2019. Tiger Global led the Series C at an $8.6 billion valuation.

Greylock Partners and Gaorong Capital were early investors. Alphabet (Google's parent) invested through its venture arm.

Woven Capital (Toyota's investment fund) participated, reflecting automotive industry interest. The company has raised approximately $2.1 billion total.

Magic AI

Magic's backers are a murderers' row of AI insiders. It started with a $5 million seed round in 2022 and a $23 million Series A in early 2023 led by Alphabet's CapitalG.

Then in February 2024 came a $117 million round led by Nat Friedman and Daniel Gross, two of the most connected investors in AI. The blowout was August 2024.

Magic raised $320 million with checks from former Google CEO Eric Schmidt, CapitalG again, Atlassian, Jane Street, Sequoia, and Elad Gil. That pushed total funding to around $465 million, all for a company still building toward its first sellable product.