Connect with us

Tech

Carbon Robotics built an AI model that detects and identifies plants 

What is and isn’t a weed that needs to be eliminated in the field is determined by the eyes of the farmer — and now, increasingly, by a new AI model from Carbon Robotics.

Seattle-based Carbon Robotics, which builds the LaserWeeder — a robot fleet that uses lasers to kill weeds — announced a new AI model, the Large Plant Model (LPM), on Monday. This model recognizes plant species instantly and allows farmers to target new weeds without needing to retrain the robots.

The LPM is trained on more than 150 million photos and data points collected by the company’s machines across the more than 100 farms in 15 countries where the robots currently operate. The model now powers Carbon AI, the AI system that serves as the brains inside the company’s autonomous weed-killing robots.

Paul Mikesell, the founder and CEO of Carbon Robotics, told TechCrunch that prior to LPM, every time a new type of weed would show up on a farm — or even the same type of weed in different soil or with a slightly different appearance — the company would have to create new data labels to retrain its machines to recognize the plant.

This process took about 24 hours each time, Mikesell said. Now, LPM can learn a new weed instantly, even if it’s never seen it before.

“The farmer can live in real time and say, ‘Hey, this is a new weed. I want you to kill this,’ and that was something that had never been done before,” Mikesell said. “There’s no new labeling or retraining because the Large Plant Model understands, at a much deeper level, what it’s looking at and the type of plant.”

Mikesell said that the company, which was founded in 2018, started developing this model shortly after it began shipping its first machines in 2022. Mikesell has years of experience building these types of neural networks from previous roles at Uber and working on Meta’s Oculus virtual reality headsets.

Techcrunch event

Boston, MA
|
June 23, 2026

This new model will reach the company’s existing systems through a software update. From there, farmers can tell the machine what to kill and what to protect by selecting photos that the machine has collected in the robot’s user interface.

Carbon Robotics has raised more than $185 million in venture capital from backers including Nvidia NVentures, Bond, and Anthos Capital, among others. Now, the company will look to continue to fine-tune the model as the machines continue to feed the LPM new data.

“We have over 150 million labeled plants now in our training set,” Mikesell said. “We have enough data now that we should be able to look at any picture and decide what kind of plant that is, what species it is, what it’s related to, what its structure is like, without having ever even seen that particular plant before, because we have so much data going into the neural net.”

source

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Tech

Skyryse lands another $300M to make flying, even helicopters, simple and safe

Skyryse, an El Segundo, California-based aviation automation startup, has raised more than $300 million in a Series C investment, pushing its valuation to $1.15 billion and into unicorn territory.

The round, which was announced Tuesday and led by Autopilot Ventures, provided a multimillion-dollar accelerant for the startup as it nears the end of a lengthy Federal Aviation Administration certification process for its flight control system. The capital will also be used to integrate its operating system, known as SkyOS, across numerous aircraft, including U.S. military Black Hawk helicopters.

Other investors in this round include Fidelity Management & Research Company, ArrowMark Partners, Atreides Management LP, BAM Elevate, Baron Capital Group, Durable Capital Partners, Positive Sum, Qatar Investment Authority, RCM Private Markets Fund managed by Rokos Capital Management, and Woodline Partners. The startup, which was founded in 2016, has raised more than $605 million in equity capital.

Image Credits:Skyryse

Skyryse has made inroads with investors as well as the U.S. military, emergency medical service operators, law enforcement, and private operators for its simplified flight system. The startup has stripped out dozens of mechanical flight controls like gauges and switches and replaced them with a system containing several flight computers that automate the more complicated and dangerous aspects of flying.

This is not a fully autonomous system; a pilot must still handle the operations. But it’s designed to automate the trickiest aspects of flying, enhance the skills of pilots, and improve safety.

That simplicity and ease of operation — a literal swipe of the finger on a touchscreen — has won over companies like United Rotorcraft, Air Methods, and Mitsubishi Corporation, which have contracts with Skyryse to integrate SkyOS on a variety of helicopters and airplanes.

Skyryse started building and testing its system on helicopters, one of the most unstable aircraft to operate. But the idea is that SkyOS can be applied to any aircraft. That initial operating system, known as Skyryse One, automates takeoff and landing and fully automates hover and engine-out emergency landings. The company has since integrated the operating system on Black Hawk helicopters.

Techcrunch event

Boston, MA
|
June 23, 2026

Skyryse has made progress on its certification process with the Federal Aviation Administration. Last year, the FAA granted final design approval for the company’s SkyOS flight control computers. Skyryse must now complete formal flight testing and verification to achieve full certification.

source

Continue Reading

Tech

Homeland Security is trying to force tech companies to hand over data about Trump critics

The Department of Homeland Security has been quietly demanding tech companies turn over user information about critics of the Trump administration, according to reports.

In several cases over recent months, Homeland Security has relied on the use of administrative subpoenas to seek identifiable information about individuals who run anonymous Instagram accounts, which share posts about ICE immigration raids in their local neighborhoods. These subpoenas have also been used to demand information about people who have criticized Trump officials or protested government policies.

Unlike judicial subpoenas, which are authorized by a judge after seeing enough evidence of a crime to authorize a search or seizure of someone’s things, administrative subpoenas are issued by federal agencies, allowing investigators to seek a wealth of information about individuals from tech and phone companies without a judge’s oversight.

While administrative subpoenas cannot be used to obtain the contents of a person’s emails, online searches, or location data, they can demand information specifically about the user, such as what time a user logs in, from where, using which devices, and revealing the email addresses and other identifiable information about who opened an online account. But because administrative subpoenas are not backed by a judge’s authority or a court’s order, it’s largely up to a company whether to give over any data to the requesting government agency.

Administrative subpoenas are not new; the use of these self-signed demands by Trump officials to seek identifiable information about people who are critical of the president’s policies has raised alarm.

Bloomberg reported last week that Homeland Security sought the identity of an anonymous Instagram account called @montocowatch, which says its goal is to share resources to help protect immigrant rights and due process across Montgomery County in Pennsylvania. This comes amid an ongoing federal immigration crackdown across the United States, which has drawn widespread protests and condemnation. Homeland Security lawyers sent an administrative subpoena to Meta demanding it turn over personal information of the person who runs the account, citing a non-Homeland Security employee who claimed to receive a tip that ICE agents were being stalked.

The American Civil Liberties Union, representing the account owner, said there was no evidence of wrongdoing and that recording police, sharing that recording, and doing so anonymously is legal and protected under the First Amendment. Homeland Security withdrew its subpoena without providing an explanation.

Techcrunch event

Boston, MA
|
June 23, 2026

The ACLU called the subpoena “part of a broader strategy to intimidate people who document immigration activity or criticize government actions.”

Bloomberg reported the effort to unmask the @montcowatch account was not an isolated incident, referencing at least four other cases where Homeland Security officials used administrative subpoenas in efforts to identify the people running Instagram accounts publishing content critical of the government. Those subpoenas were also withdrawn after the account owners sued to block the attempt.

Tech companies have in recent years published transparency reports that detail how many government demands for data they receive. But most do not break out how many judicial and administrative subpoenas they receive over a period of time, even though the two kinds of demands are fundamentally different.

When asked by TechCrunch, Meta spokesperson Francis Brennan did not say if Meta provided Homeland Security any data involving @montcowatch or if the company was asked to provide information about the account another way. 

A new report by The Washington Post on Tuesday found that an administrative subpoena was also used to seek information from Google about an American retiree within hours of him after sending a critical email to Homeland Security’s lead attorney Joseph Dernbach. The retiree’s home was later visited by federal agents inquiring about the email.

The Post described the retiree as someone critical of Trump during his first term, who attended a No Kings rally last year, regularly attends gatherings and protests, and wrote criticisms to lawmakers, all actions protected under the First Amendment. 

Within five hours of emailing the Homeland Security lawyer — who was named in an article about the case of an Afghan the U.S. was trying to deport and whose email address is listed on the Florida Bar’s website — the retiree received an email from Google, according to The Washington Post. The email notified him that his account had been subpoenaed by the Department of Homeland Security.

The subpoena demanded to know the day, time, and duration of all his online sessions, his IP address and physical address, and a list of each service he used, and any other usernames and identifiable information relating to his account, such as his credit card, driver’s license, and Social Security numbers.

Two weeks later, Homeland Security agents were on his doorstep, asking him questions about the email that he sent to Dernbach, which the agents conceded broke no laws.

Google spokesperson Katelin Jabbari told TechCrunch the company pushes back against overbroad or improper subpoenas, “as we did in this instance,” referring to the subpoena referenced in The Washington Post’s reporting. 

When asked by TechCrunch, Homeland Security assistant secretary Tricia McLaughlin would not say why the U.S. was seeking information about people who have been critical of the Trump administration and accounts documenting ICE activity, or say for what reason the subpoenas were withdrawn.

“HSI has broad administrative subpoena authority under 8 U.S.C. § 1225(d) and 19 U.S.C. § 1509(a)(1) to issue subpoenas,” said McLaughlin, referring to Homeland Security Investigations, an investigative unit within ICE.

Not all companies are able to hand over data about their customers. For instance, information that is end-to-end encrypted and can only be accessed by obtaining a person’s phone or devices. That said, many companies are still able to provide large amounts of information about their users, including where they log in, how they log in, and from where, which may allow investigators to unmask anonymous accounts.

End-to-end encrypted messaging apps, like Signal, have long championed how little data it collects about its users. The messaging app responds to occasional legal demands by stating that it is unable to produce user data that it does not have to begin with.

The reliance on U.S. tech giants is another reason why European countries and ordinary consumers are seeking to rely less on American tech giants, at a time when chief executives and senior leaders at some of the largest U.S. tech companies are overtly cozying up to the Trump administration.

source

Continue Reading

Tech

2026 plans: What’s next for Startup Battlefield 200

TechCrunch Startup Battlefield 200 is the flagship early-stage startup competition held at TechCrunch Disrupt in San Francisco, spotlighting the world’s most promising young companies. Each year, thousands of applicants from around the globe vie for one of 200 slots in the cohort, which provides unparalleled visibility, access to top-tier investors, and opportunities to grow. 

Early-stage startups from around the world are joining the Startup Battlefield email list to be the first to know when applications open this month. Don’t miss a beat if you’re ready to join the action. 

What is Startup Battlefield? 

Valentina Agudelo Vargas
Image Credits:Kimberly White / Getty Images

Startup Battlefield 200 is a highly selective startup launch program designed to prepare early-stage startups for visibility, investment, and scale. Each year, TechCrunch hand-selects 200 companies to gain unmatched exposure, credibility, and opportunities to scale faster than they ever imagined:

  • Free exhibit space at TechCrunch Disrupt for all three days
  • Exclusive access to masterclasses and curated networking
  • Direct exposure to top-tier press and investors 
  • A chance to pitch live at Disrupt
  • A shot at the main Disrupt Stage to compete for a $100,000 equity-free prize and the iconic Disrupt Cup 

Alumni who’ve entered Startup Battlefield 

Startup Battlefield’s track record is among the strongest in the industry — alumni represent 1,700+ companies that have collectively raised over $32 billion and include well-known names such as: 

  • Mint.
  • … and many others that have gone on to significant exits and growth.
TechCrunch Disrupt 2025 Startup Battlefield
Image Credits:TechCrunch

What to expect in 2026 

In 2026, Startup Battlefield 200 will once again bring together a global cohort of high-potential early-stage startups building across AI, climate, health, fintech, and beyond.

Applications will open in the middle of this month and close in mid-June. Selected startups will be notified around September 1, when the virtual preparation program begins — designed to help founders refine their story, sharpen their pitch, and get ready for the pitch of a lifetime on the TechCrunch Disrupt stage.

Ready to join the battle this year? 

For founders and investors alike, Startup Battlefield 200 remains one of the most influential global early-stage startup stages — a place where breakout companies get seen, funded, and scaled.

Techcrunch event

Boston, MA
|
June 23, 2026

Don’t miss a moment of the action: Join the Startup Battlefield mailing list to be the first to know when applications open next month.

TechCrunch Disrupt startup Battlefield presentation
Image Credits:Kimberly White / Getty Images

source

Continue Reading