• Roli has a new Piano AI Assistant to go with its latest instruments.
• The Piano AI Assistant offers personalized, real-time help and lessons.
• The AI is there to make learning and creating music more intuitive.

Piano teachers of both the strict and whimsical variety are a staple of movies and television, but music technology company Roli now offers a piano tutor built right into the instrument. The new Roli Piano features personalized AI guidance underneath the 49-key, $800 keyboard.

Roli’s Piano AI Assistant does exactly what it sounds like: It makes learning music more straightforward and fun than practicing alone. It can guide players through scales, explain ways of varying a tune, and even explain some music history in the context of specific compositions. It’s like having a music teacher who never gets tired and has an encyclopedic knowledge of the subject.

Welcome to Startups Weekly — your weekly recap of everything you can’t miss from the world of startups. Want it in your inbox every Friday? Sign up here.

This week was supposed to be a short one in the U.S., as it started with a holiday. But Inauguration Day kept some founders busy, and the following days brought us more than their fair share of startup news.

Most interesting startup stories from the week

This week reminded us that not all sales are created equal, and that it is often worth looking beyond the price tag. Plus, there were legal troubles for an AI decacorn.

No divvy: Divvy Homes, a rent-to-own startup backed by a16z, is selling to a division of Brookfield Properties for about $1 billion. However, some shareholders may not see a dime from the sale.

Beauty for sale: Consumer goods giant Hindustan Unilever agreed to acquire Peak XV-backed Indian skincare startup Minimalist for about $342 million — more than the $300 million valuation it reportedly sought in a fundraising attempt last year.

Big markdown: AI-powered parking platform Metropolis acquired computer vision company Oosto for a fraction of what the startup had raised to date. Formerly known as AnyVision, it had lost backers over its technology being used in controversial surveillance applications. 

Legal clash: Valued at $13.8 billion last year, Scale AI is facing its ​​third worker lawsuit of 2025, with contractors claiming they suffered psychological harm from writing prompts about disturbing content. A spokesperson for Scale AI said it had “numerous safeguards in place.”

Most interesting VC and funding news this week

Series B rounds announced this week varied greatly in size, with some of these exceeding other Series C rounds. And for companies that don’t quite feel like going public yet, there are still more letters in the alphabet.

Pre-IPO letters: Data analytics platform Databricks closed a $10 billion Series J equity funding round at a $62 billion valuation, with an additional $5.25 billion in debt financing. Meta is backing the company as a strategic investor.

From cat to unicorn: Neko Health, the Swedish body-scanning startup co-founded by Spotify’s Daniel Ek, and whose name means “cat” in Japanese, raised a $260 million Series B round of funding at $1.8 billion post-money.

Money to move: Lindus Health, a startup backed by Peter Thiel and Creandum that is currently moving its HQ from the U.K. to the U.S., secured a $55 million Series B round to “fix the broken clinical trial industry.” 

Spending less: AI-powered SaaS spend management platform Vertice raised a $50 million Series C round of funding led by Lakestar, at a valuation close to $500 million, according to sources.

Indian robotics: Indian-based autonomous mobile robots startup Ati Motors raised a $20 million Series B to grow internationally. The U.S. already dominates Ati’s revenues, and the company hopes to further benefit from demand for robotics manufactured outside of China.

Crypto crypto crypto: Capitalizing on crypto’s comeback, AngelList and CoinList teamed up to launch crypto special purpose vehicles and crypto roll-up vehicles that will let crypto founders raise capital using crypto coins.

Last but not least

AI is still red hot, but there are always subsectors that VCs are more interested in. To figure out which types of AI startups they’d most like to back this year, TechCrunch rounded up some findings from our recent survey of 20 enterprise VCs. In short: Think companies, not features.

Quordle was one of the original Wordle alternatives and is still going strong now more than 1,000 games later. It offers a genuine challenge, though, so read on if you need some Quordle hints today – or scroll down further for the answers.

Enjoy playing word games? You can also check out my NYT Connections today and NYT Strands today pages for hints and answers for those puzzles, while Marc’s Wordle today column covers the original viral word game.

The developer behind Pixelfed, Loops, and Sup, open source alternatives to Instagram, TikTok, and WhatsApp, respectively, is now raising funds on Kickstarter to fuel the apps’ further development.

The trio is part of the growing open social web, also known as the fediverse, powered by the same ActivityPub protocol used by X alternative Mastodon. The latter saw increased signups and use after the company formerly known as Twitter sold to Elon Musk in October 2022 and during the X exodus that followed the U.S. presidential election.

In the months and years following that sale, open source and decentralized apps like Mastodon and Bluesky (which uses the newer AT Protocol), have continued to grow their user bases, as people sought alternatives to centralized social media apps controlled by billionaires like Musk and Meta’s Mark Zuckerberg.

Seeing the writing on the wall, even Meta realized it needed to plant a flag in the fediverse. This led it to release its own X rival in 2023 called Instagram Threads, which is in the process of integrating with ActivityPub.

Now, Daniel Supernault, the Canadian-based developer behind the federated apps that challenge Meta’s social media empire, is seeking funds for continued development and support of his open social communities.

“Help us put control back into the hands of the people!” he said in a post on Mastodon where he announced the Kickstarter’s Thursday launch.

As of the time of writing, the campaign has raised $58,383 so far. While the goal on the Kickstarter site has been surpassed, Supernault said that he hopes to raise $1 million or more so he can hire a small team.

Supernault wants his set of apps to become the first in the fediverse to reach a network of a billion people, but of course, there’s still a long way to go before they can meet that lofty goal. Though Pixelfed has been around for years, it just launched the 1.0 version of its mobile app earlier this month, for instance, and Loops is still in alpha testing on Apple’s TestFlight. Sup, meanwhile, has not yet been released but is said to be “coming soon,” per its Instagram page.

Both Loops and Sup will be released to Kickstarter supporters.

A fourth project, PubKit, is also a part of these efforts, offering a toolset to support developers building in the fediverse.

It includes interactive tools and testing frameworks, allowing developers to mock popular activities on their service, set up an inbox for ingesting and debugging activities in real time, and tools to inspect, debug, and verify HTTP Signature implementations.

This is the first time Supernault has turned to Kickstarter to help with these efforts, which aim to also benefit the Pixelfed Foundation.

The stretch goal of the Kickstarter campaign is to register the Pixelfed Foundation as a not-for-profit and grow its team beyond volunteers. This could help address the issue with Supernault being a single point of failure for the project, which has raised concerns in light of recent behavior.

Mastodon CEO Eugen Rochko made a similar decision earlier this month to transition to a nonprofit structure.

If successful, the campaign would also fund a blogging app as an alternative to Tumblr or LiveJournal at some point in the future.

The funds will also help the apps manage the influx of new users. On Pixelfed.social, the main Pixelfed instance, (like Mastodon, anyone can run a Pixelfed server), there are now more than 200,000 users, thanks in part to the mobile app’s launch, according to the campaign details shared with TechCrunch. The server is also now the second-largest in the fediverse, behind only Mastodon.social, according to network statistics from FediDB.

New funds will help expand the storage, CDNs, and compute power needed for the growing user base and accelerate development. In addition, they’ll help Supernault dedicate more of his time to the apps and the fediverse as a whole while also expanding the moderation, security, privacy, and safety programs that social apps need.

As a part of its efforts, Supernault also wants to introduce E2E encryption to the fediverse.

Strands is the NYT’s latest word game after the likes of Wordle, Spelling Bee and Connections – and it’s great fun. It can be difficult, though, so read on for my Strands hints.

Want more word-based fun? Then check out my NYT Connections today and Quordle today pages for hints and answers for those games, and Marc’s Wordle today page for the original viral word game.

Waymo lobbyists had a busy 2024.

A recent review of lobbyist disclosure data by the San Francisco Examiner revealed Waymo paid lobbyists to meet with San Francisco government officials 348 times last year, more than double the 137 times it reported in 2023.

The pop in lobbyist activity is in sync with Waymo’s expansion in San Francisco — and its ambitions for the larger Bay Area. Waymo has had a presence in the San Francisco area since 2009, but it wasn’t until August 2023 that the company was able to charge customers — 24 hours a day and throughout the city — for rides in its fleet of self-driving Jaguar I-Pace vehicles. 

Since then, Waymo has expanded its footprint and the number of customers it serves. Waymo dumped its waitlist last June for its San Francisco robotaxi service, removing the final obstacle for customers keen to use the self-driving technology. Last year, the company expanded its coverage area by another 10 square miles to include Bay Area cities and began testing driverless vehicles on SF freeways.

Waymo also wants to be able to pick up and drop off customers at the San Francisco Airport, as TechCrunch reported last July; lobbyist disclosure data separately shows an uptick in visits with airport authority officials.

• Google’s Gemini app now controls smart homes through a Google Home extension
• Gemini can understand natural language to complete tasks
• The aim is to make smart homes more intuitive and easier to manage

Google wants Gemini to control your smart home devices and has upgraded the Gemini app with a new Google Home extension to manage all of your connected devices the same way you’d ask the AI assistant to answer any other query. So if you have the Gemini app and devices controlled by Google Home, you can link Gemini to your Google Home account.

The extension links Gemini with your lights, thermostats, and any other smart home devices, but with the benefit of Gemini’s more flexible conversational ability. That means you could say, “It’s too bright in here,” and have Gemini dim the lights instead of needing to command setting the lights to 50% specifically. You can also manage multiple devices with more casual language. Rather than individually tweaking device settings, you can say, “Dim the living room lights, turn on the bedroom lamp, and lower the blinds.” Gemini can grasp the three commands for three sets of devices and act accordingly.

Audi has revealed a new dual-motor electric off-road vehicle concept based on the Q6 E-tron that looks ready for a snowpocalypse. The automaker built a working prototype that lifts the vehicle by 6.3 inches and widens it by 9.8 inches, giving it a stance that wouldn’t be out of place if it appeared in Truck Country, USA.

Audi’s CEO Gernot Döllner calls the “Q6 E-Tron Off-road concept” a “reinterpretation of Quattro,” which is the company’s marketing term for its all-wheel-drive models.

The extra ride height is courtesy of four bespoke portal axles integrated into the wheel hub assemblies at the front and rear that Audi says increase torque at the wheel by 50 percent. Each axle is powered by an electric motor with a combined power output of 380kW and up to 9,883 lb ft of torque at its peak. That’s up 3,245 lb ft of torque from the normal Q6 E-tron, which is Audi’s first vehicle built on Volkswagen’s modular Premium Platform Electric (PPE) platform (also used in the new A6 E-tron and Porsche Macan EV).

The vehicle is designed to climb hills as steep as 45 degrees but the company did nerf the Q6’s top speed a bit down to 108 mph. Still, no one should drive that fast anyway in a vehicle lifted this high. This also makes the Off-road concept a much more realistic one compared to more sci-fi Audi concepts like the Activesphere coupe / pickup truck combo with a mixed reality cockpit or the truly apocalyptic all-terrain “AI:Trail” that has drones for headlights.

Audi’s Q6 E-tron Offroad concept will be featured at the FAT International Ice Race in Austria on February 1st. The company will also show it in action via its social media channels.

Character AI, a platform that lets users engage in roleplay with AI chatbots, has filed a motion to dismiss a case brought against it by the parent of a teen who committed suicide, allegedly after becoming hooked on the company’s technology.

In October, Megan Garcia filed a lawsuit against Character AI in the U.S. District Court for the Middle District of Florida, Orlando Division, over the death of her son, Sewell Setzer III. According to Garcia, her 14-year-old son developed an emotional attachment to a chatbot on Character AI, “Dany,” which he texted constantly — to the point where he began to pull away from the real world.

Following Setzer’s death, Character AI said it would roll out a number of new safety features, including improved detection, response, and intervention related to chats that violate its terms of service. But Garcia is fighting for additional guardrails, including changes that might result in chatbots on Character AI losing their ability to tell stories and personal anecdotes.

In the motion to dismiss, counsel for Character AI asserts the platform is protected against liability by the First Amendment, just as computer code is. The motion may not persuade a judge, and Character AI’s legal justifications may change as the case proceeds. But the motion possibly hints at early elements of Character AI’s defense.

“The First Amendment prohibits tort liability against media and technology companies arising from allegedly harmful speech, including speech allegedly resulting in suicide,” the filing reads. “The only difference between this case and those that have come before is that some of the speech here involves AI. But the context of the expressive speech — whether a conversation with an AI chatbot or an interaction with a video game character — does not change the First Amendment analysis.”

To be clear, Character AI’s counsel isn’t asserting the company’s First Amendment rights. Rather, the motion argues that Character AI’s users would have their First Amendment rights violated should the lawsuit against the platform succeed.

The motion doesn’t address whether Character AI might be held harmless under Section 230 of the Communications Decency Act, the federal safe-harbor law that protects social media and other online platforms from liability for third-party content. The law’s authors have implied that Section 230 doesn’t protect output from AI like Character AI’s chatbots, but it’s far from a settled legal matter.

Counsel for Character AI also claims that Garcia’s real intention is to “shut down” Character AI and prompt legislation regulating technologies like it. Should the plaintiffs be successful, it would have a “chilling effect” on both Character AI and the entire nascent generative AI industry, counsel for the platform says.

“Apart from counsel’s stated intention to ‘shut down’ Character AI, [their complaint] seeks drastic changes that would materially limit the nature and volume of speech on the platform,” the filing reads. “These changes would radically restrict the ability of Character AI’s millions of users to generate and participate in conversations with characters.”

The lawsuit, which also names Character AI corporate benefactor Alphabet as a defendant, is but one of several lawsuits that Character AI is facing relating to how minors interact with the AI-generated content on its platform. Other suits allege that Character AI exposed a 9-year-old to “hypersexualized content” and promoted self-harm to a 17-year-old user.

In December, Texas Attorney General Ken Paxton announced he was launching an investigation into Character AI and 14 other tech firms over alleged violations of the state’s online privacy and safety laws for children. “These investigations are a critical step toward ensuring that social media and AI companies comply with our laws designed to protect children from exploitation and harm,” said Paxton in a press release.

Character AI is part of a booming industry of AI companionship apps — the mental health effects of which are largely unstudied. Some experts have expressed concerns that these apps could exacerbate feelings of loneliness and anxiety.

Character AI, which was founded in 2021 by Google AI researcher Noam Shazeer, and which Google reportedly paid $2.7 billion to “reverse acquihire,” has claimed that it continues to take steps to improve safety and moderation. In December, the company rolled out new safety tools, a separate AI model for teens, blocks on sensitive content, and more prominent disclaimers notifying users that its AI characters are not real people.

Character AI has gone through a number of personnel changes after Shazeer and the company’s other co-founder, Daniel De Freitas, left for Google. The platform hired a former YouTube exec, Erin Teague, as chief product officer, and named Dominic Perella, who was Character AI’s general counsel, interim CEO.

Character AI recently began testing games on the web in an effort to boost user engagement and retention.

TechCrunch has an AI-focused newsletter! Sign up here to get it in your inbox every Wednesday.