CodeCanyon

For years, the U.S. Environmental Protection Agency’s (EPA) Energy Star program has helped consumers save a collective $40 billion in annual energy costs. Now, the Trump administration wants to wind it down, according to a report from CNN. The Energy Star program, which has a budget of $32 million, is a public-private partnership that works with appliance and electronics manufacturers to certify energy-efficient products while also helping consumers find rebates to lower the purchase cost. “Eliminating the Energy Star program would directly contradict this administration’s promise to reduce household energy costs,” Paula Glover, president of the nonprofit coalition Alliance to Save Energy, said in a statement. The program delivers a 350-to-one return on investment, she added. Energy Star was created in 1992 under President George H. W. Bush, and it was reauthorized in 2005 under President George W. Bush, placing oversight for the program under the EPA and the Department of Energy. The program’s signature yellow labels appear on appliances and electronics for sale throughout the U.S., informing consumers of how much they’ll spend on electricity or natural gas throughout a year of typical use. Energy Star saves the average U.S. household about $450 on their energy bills each year.

Tech executives have long talked about how AI is going to revolutionize the advertising industry. In particularly, Meta CEO Mark Zuckerberg has been quite vocal about how exactly he wants his company to lead the transformation. Speaking onstage at Stripe’s annual Sessions conference in San Francisco on Tuesday, Zuckerberg laid out his plans to automate the entire ad industry with a black-box, end-to-end AI ad tool. A key component of such a product would entail putting thousands of AI-generated “test” ads in front of Facebook, Instagram, and Threads users, said Zuckerberg. “The basic end goal, here, is any business can come to us, say what their objective is — we get new customers to do this thing, or sell these things — tell us how much they’re willing to pay to achieve those results, connect their bank account, and then we just deliver as many results as we can,” he explained. “In a way, it’s kind of like the ultimate business results machine. I think it’d be one of the most important and valuable AI systems that gets built.” Zuckerberg first described this hypothetical machine on Ben Thompson’s Stratechery podcast last week, and — if built as Zuckerberg envisions it — it would have huge implications for the ad industry. During his Sessions appearance, Zuckerberg posited that while creative ad agencies would continue to exist were Meta to deploy this AI, small businesses might not “have to start off with the creative” and Meta could simply handle all of their advertising operations. In fact, Zuckerberg asserted during Sessions that Meta’s ad tools, several of which have generative AI capabilities, are sophisticated enough already that the company doesn’t even recommend that customers specify the demographics they’d like to target. Meta’s tools can find interested users better than human marketers can, claimed Zuckerberg. The next logical step, he says, is trying to apply this data-driven optimization to the creative side. “We’re gonna be able to come up with, like, 4,000 different versions of your creative and just test them and figure out which one works best,” said Zuckerberg. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW While this sort of solution may appeal to businesses, it’s an open question as to what AI ad testing will do to Meta’s platforms from a user experience point of view, considering they’re already brimming with generative AI slop. Meta has experimented with AI-generated images and comments in feeds, as well as AI chatbots users can interact with. The company more recently launched a social portal and dedicated app that spotlights generative AI content. Now, it appears yet another category of generative AI — ads — is poised to gunk up Meta’s social media ecosystem. If it wasn’t clear before that users are the product on Meta’s platforms and advertisers the customers, it’s about to become crystal. While testing AI-generated ads may deliver value for companies, it means users will have to suffer through yet more slop. The ad industry isn’t likely to take kindly to Zuckerberg’s vision, either. There’s been significant backlash over the ethics of using generative AI in creative fields. In October 2024, more than 11,000 creators signed an open letter condemning the use of human-generated art to train AI systems. Creators have also filed lawsuits against companies developing AI art tools, such as Midjourney and Stability AI. To be fair, there are just as many creators and advertising executives who believe AI tools won’t threaten their livelihoods anytime soon. Johnny Hornby, founder of the ad agency The&Partnership, on Tuesday published an op-ed arguing that creating successful branding campaigns is still a uniquely human task. In any event, it seems that Zuckerberg has a pretty clear idea of how he wants to automate the ad industry and fill Meta’s feed with AI — whether ad agencies or users like it or not.

Photo and video editing platform VSCO on Wednesday launched an AI-powered collaborative moodboard to expand how its products are used by photographers and artists. Called Canvas, the moodboard lets you import and edit your photos using the standard VSCO editing tools — so you can tune settings like shadows, brightness, exposure, temperature, tint, grain, blur, vibrance and hue. The AI chops come into play when you want to generate images using text prompts. The moodboard also lets you select parts of an image and use a “region prompt” menu to have AI recreate those parts with text prompts. The weights of the region prompt can be adjusted with a slider to generate different versions. There’s also a variation button that, as it says on the tin, makes the AI create variations of a generated image. This also comes with a slider to control how close to the original image the generated image is. Image Credits: Screenshot by TechCrunchUsers can share the moodboard with other people in a project and create different iterations of an idea. “Photographers, who often work alone, use Google Slides or Pinterest to create a vision for a project that they want to show the clients. We thought there could be a better tool that was designed for ideation with creators in the front and center of it,” VSCO’s CEO Eric Wittman told TechCrunch. The moodboard feature uses tech from an image editing startup called Facet that VSCO acquired last year. Facet had raised over $13 million in funding before it was acquired. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Image Credits: VSCOThis is VSCO’s first time implementing AI features into its products, and the company says it has seen positive traction. VSCO said more than 84% of content was generated using AI during Canvas’ test phase. The company plans to enable users to search VSCO and import photos uploaded onto the platform into the moodboard. The launch comes days after Adobe unveiled its own moodboard that has AI-powered image generation and editing features. Startups like Visual Electric, Cove and Kosmik have also tried to build whiteboards and moodboards to help people collaborate on ideas. Canvas is available to all users, but paying subscribers will get extra credits for prompting and generating images using the AI model.

Games drove the creation of GPU processors back in the 1990s, so it’s only fitting that artificial intelligence — the technology that GPUs are used to power nowadays — is making its way into nearly every aspect of video game design. In keeping with that trend, on Wednesday a startup called Sett — which is building AI agents to build and run mobile games — is emerging from stealth with $27 million in funding. The funding was raised in two tranches, the most recent of which was a $15 million Series A, led by Bessemer Venture Partners. Saga VC, Vgames and Akin Babayigit — the founder and former head of the UK-based games unicorn Tripledot, who now heads VC firm Arcadia Gaming Advisors — also invested. Earlier, Sett had raised $12 million in seed funding from F2, Bessemer, and some gaming industry leaders as angel investors. (In a case of uncanny timing, sources tell me that AppLovin, a would-be competitor of Sett’s, is today announcing the sale of its gaming assets to Tripledot. That deal, for $800 million — not $900 million as AppLovin previously estimated — is set to be publicly confirmed later today around AppLovin’s Q1 earnings. More on that below.) Up to now, Tel Aviv-based Sett has taken the same approach to ‘stealth mode’ as a lot of other B2B startups. Since being founded in 2022, it’s been under the radar, honing product-market fit and nurturing its early customer base. Today, that customer list features Zynga, Scopely, Playtika, SuperPlay, Rovio, Plarium, Candivore and Unity. It announced a website five months ago, but now that it’s fully out of stealth, Sett is still not putting its pedal to the marketing metal. It says it has over 100 gaming studios on a waiting list to be onboarded, and so the plan is to use the new funding to hire engineers and AI specialists. As for the product, the focus is on what CEO Amit Carmi — who co-founded the company with CTO Yoni Blumenfeld — believes is one of the biggest pain points in the mobile gaming business: Getting noticed. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW “Gaming is one of the most competitive industries in the world,” he told TechCrunch in an interview. “There are a lot of players, but you actually have more games than people. It’s pretty easy to build games, but almost impossible, statistically, to make a game that is successful.” (R-L) SETT’s CTO Yoni Blumenfeld and CEO Amit Carmi. Image Credits: SettImage Credits:SettCompanies spend a lot on user acquisition marketing to improve those chances of success, he continued, but typically it’s very expensive to build and place that content. On average, approximately $29 billion is spent to make around $100 billion in revenue, according to research from AppsFlyer. Sett’s solution is an AI agent for game publisher marketing. Extensive user-level tracking is a thing of the past on iOS, so the focus is now on what Carmi describes as “creative content” — in-game and marketing streams of interactive moments built on the aesthetics of the game that aim to draw in users to try out new games, or to play them more. These “playable” ads and marketing efforts are very catchy at the moment, but they can be very expensive and time-consuming to create, akin to building new versions of the game. That is where Sett sees an opportunity. What humans previously had to code, place and measure from the ground up can now be built using Sett, the startup claims, 15 times faster and 25 times cheaper. Arcadia’s Babayigit, from his time at Tripledot, knows first-hand how important marketing is for helping games stand out and get played. He described the idea as a “no brainer” in an interview. “It’s just a phenomenal team and an incredibly talented group of people.” The opportunity that Sett is targeting is also one that has been proven out. The gaming studio assets that Sett’s competitor AppLovin is selling to Tripledot for $800 million were built out in the first place, we understand, in large part to train the AI models that AppLovin now uses across a wide range of ad and marketing tools, including the creation of its own playable ads for customers by way of SparkLabs. Now that the AI models and wide networks of users are established — AppLovin has a market cap of $103 billion, despite a lot of short seller noise — the game studios are no longer core to AppLovin. Meanwhile, AppLovin has its sights set on a much bigger prize: It’s one of the companies that has publicly stated it will bid to buy the global business of TikTok. How much AI is too much? There is a big question mark over all the AI services that have the capacity to take over an increasing number of functions previously carried out by humans. How much is too much? Is there even a “too much”? Companies like Agave are already putting some AI into the creative process, and arguably, once the genie is out of the bottle, that could be it. Carmi said while he believes that you will eventually be able to build AI agents to develop and market games end to end, this may not be where Sett settles. “We believe it’s actually a bigger opportunity than what we’re doing now. This is the reason why we built our game engine and the agentic layer in a way that it generates code and enables us to enter all of what we’re doing basically to the game itself,” he said. “The vision of Sett is really taking both the marketing content and in-game content for now.” “I don’t think the genesis is to replace ‘all aspects’ of game design and execution,” Babayigit said. “I don’t even know if that’s possible right now, since to compete in a very crowded area, the bar is SUPER high, so you need to make a game in which the details are SUPER SUPER important. But what I do know is that this team is operating with real technology behind them, so if anyone can make certain parts of game production and distribution automated, it’s them.”

Rubrik co-founder Soham Mazumdar, who left in 2023, has a new data startup called WisdomAI. The company offers AI data analytics that can deliver business insights with structured, unstructured, and even “dirty” data, meaning data not cleaned of typos or errors. Working with data where and how it is, that’s essentially the holy grail for enterprise business intelligence software and why Coatue led the giant seed round of $23 million. Madrona, GTM Capital, The Anthology Fund, and others also participated. Rather than asking a data analytics team to run reports, business managers can ask WisdomAI questions and drill into the details. Mazumdar gives an example of a chief of revenue wanting to know, “How am I going to close my quarter?” WisdomAI’s answer would offer a list of pending deals the team should focus on, along with the information on what’s delaying each one, such as the list of the questions each customer is waiting on. “You can get the CRO to literally see all the way down to this last level of detail through our platform with, like, five key strokes, as opposed to a process which involves five individuals, including some analysts, and a whole lot of time,” Mazumdar told TechCrunch. That’s just one example of the type of questions WisdomAI hopes to answer. Another early customer is an oil and gas company that has thousands of workers in the field using WisdomAI to ask questions about production, tapping into data from everything from stored documents to telemetry. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Obviously, every business analytics tool already available — and a host of startups — are also offering AI-powered natural language prompts. WisdomAI stands out for the pedigree of the founders — all previously worked with Mazumdar at Rubrik. But the platform’s superpower is its accuracy, even against messy data, Mazumdar says. It can find answers in structured data like databases as well as unstructured data stored in files. Equally importantly, WidsomAI won’t deliver hallucinations. Most enterprises are pursuing AI app accuracy by focusing on the data used to train their AI models, as well as model size, prompt engineering, and, perhaps, real-time retrieval techniques like retrieval-augmented generation (RAG). Yet they still run the risk of fabricated answers. WisdomAI uses GenAI in the query formation — not in the creation of answers. “Ultimately, GenAI can hallucinate. What we use GenAI to do is to write small little programs … that can query these different systems,” Mazumdar says. So if WidsomAI’s model hallucinates, all it will do is write a fake query that fails to retrieve data. The data itself – the answer to the question – won’t be fabricated. WisdomAI claims ConocoPhillips, Cisco, and Descope as early customers and has customers who are working with major cloud data storage services like Snowflake, Google’s BigQuery, Amazon’s Redshift, Databricks, and Postgres. It can be trained on any data storage system by studying the query language through query logs and other sources, Mazumdar says.

Few markets are moving as rapidly as China’s automotive sector. There, new models are rolled out in as little as 18 months, putting tremendous pressure on legacy Western automakers, which need four-plus years to go from concept to sales floor. “With the increasingly short development cycles in China, it’s driving a huge amount of cost and time focus,” Ian Campbell, co-founder and CEO of Breathe Battery Technologies, told TechCrunch. “In both geographies, in the East — in China and Asia — and in the West as well.” Much of that focus has been centered around batteries — the components that can make or break electric vehicle sales. Automakers are forced to predict where the market will be a few years out, but those forecasts don’t always pan out given how quickly the EV landscape is evolving. Making changes to physical components can be expensive and unpredictable, which is why Campbell’s startup has been trying to give batteries more flexibility via software. Breathe has developed a suite of tools that Campbell said helps automakers and others get the most out of their batteries. The startup recently raised a $21 million Series B led by Kinnevik Online AB, the company exclusively told TechCrunch. Lowercarbon Capital and Volvo Cars Tech Fund participated. The new funding will help Breathe continue to push its software earlier in the battery development process. The company currently has four products: Design, Model, Map, and Charge. Charge was Breathe’s first offering, and it optimized charging strategies to speed refilling or increase the longevity of a battery. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Though battery manufacturing is tightly controlled, no two cells that roll of the line are 100% identical. As a result, some might generate more heat during fast charging, while others might be able to withstand more charge and discharge cycles than their peers. Chinese mobile phone maker Oppo was the first to adopt it, and the software cut charging time by 27%. On the automotive side, Volvo has Breathe’s code installed on its forthcoming ES90 sedan, helping it to charge 10% to 80% in 20 minutes. In essence, Breathe’s software lets them make the most of each cell given its individual quirks. The startup’s other offerings help automakers and electronics companies design and predict how their batteries will perform years down the line, letting them determine where to invest development resources. For example, if a new chemistry is lower cost and looks to have a longer lifespan, then designers may decide to let it charge a little faster at the expense of some of that longevity. “They want to understand what room they have and what will happen when they make trade offs throughout the development program of their battery system,” Campbell said. To do that, Breathe has built a lab in London where it can run a range of tests on batteries its customers are interested in using. In as little as four weeks, it has enough to ship the customer a model (called Breathe Model) that can simulate likely future performance. After that, the cells stay on in the lab, contributing more data so that Breathe can eventually ship the customer its Map product, which augments simulated data with more real world results, Campbell said. The Design product will round out the suite when its released in the coming months, providing a customers with set of software tools to speed — you guessed it — battery design. The goal is to reduce the amount of “brute force lab testing” needed to bring a battery to market, Campbell said. He likens Breathe’s software tools to those used in the semiconductor industry, which have helped companies like Apple and Nvidia work closely with foundries like TSMC to implement their processor designs in silicon. “We want to try and do for batteries what we’ve seen the simulation software from Cadence and Synopsis do so effectively in semiconductor design,” he said.

Netflix announced on Wednesday that it’s currently testing a short-form video feature, signaling that even a streaming giant with over 300 million subscribers is concerned about losing viewers’ time spent on mobile to apps like TikTok, YouTube Shorts, and Instagram Reels. The company debuted the feature at its first-ever product and tech event, where it also revealed several other upcoming plans for the service. These plans include a redesign of the TV homepage and improved real-time recommendations. Netflix’s new mobile-only vertical feed allows users to easily scroll through clips of its original titles. Within this feed, users can tap on buttons to either watch the entire show or movie immediately, save it to their “My List,” or share it with friends. Of note, the clips are curated from the “Today’s Top Picks for You” section, rather than being chosen from Netflix’s entire library. This approach makes it specifically tailored to each user, ultimately encouraging viewers to watch the full shows. The experiment rolls out globally in the coming weeks on iOS and Android devices. It’ll appear for users as a tab on the in-app homepage. The introduction of this feature comes at a pivotal time, as competition among platforms for viewer attention intensifies. Audiences increasingly favor quick entertainment, leading to a shift in traditional viewing habits. As a result, even large players like Netflix are adapting to retain and attract subscribers. This also follows President Trump’s second extension of the deadline for the TikTok ban. The latest test follows a trend among other streaming services that are trying out similar features. Tubi had its entry into short-form video last year with its “Scenes” feature. More recently, Peacock launched curated vertical video playlists earlier this year that not only feature short clips from TV series and films but also sports and news content. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Netflix is no stranger to experimenting with short-form video content. In 2021, the platform rolled out a TikTok-inspired feature called “Fast Laughs,” which focused on funny clips. However, this new test aims to reach a broader audience beyond just comedy fans and will be more personalized. In terms of Netflix’s new homepage update, one small change implemented is that the shortcuts for “Search” and “My List” were moved to the top of the page, where they were previously located on the left-hand side, making them easier to access. Netflix also has new “callout” badges that help viewers discover titles. For example, if users are searching for the latest Emmy-winning content, these badges will be prominently displayed on the title cards featured on the homepage. Additionally, Netflix improved its real-time recommendation system. Now, when subscribers search for content, it considers factors such as the trailer a viewer recently watched or the actors they are looking up. For instance, if they give a thumbs up to the popular TV series “Wednesday,” the system will quickly adjust their homepage to display similar recommended titles.

After hinting at a new AI-powered search experience during its recent earnings call, Netflix officially unveiled the feature at its tech and product event on Wednesday. This new search experience will utilize OpenAI’s ChatGPT to provide users with a conversational discovery experience. Users can enter their preferences using natural phrases like “I want something funny and upbeat” or even more detailed requests, such as “I want something scary, but not too scary, and maybe a little bit funny, but not haha funny.” The feature is set to roll out this week to iOS users as an opt-in beta. Some subscribers in Australia and New Zealand have already had access to it, as reported by Bloomberg last month. Other competitors are also leveraging generative AI for search. For instance, Amazon has an AI voice search experience on Fire TVs that responds to open-ended inquiries about TV shows and movies. A closer comparison is Tubi’s ChatGPT-powered search tool, which answered content-related questions and suggested movies based on a user’s specific request. However, Tubi later discontinued the feature, probably because of low adoption. It remains to be seen whether Netflix’s new feature will face similar challenges. Additionally, at the tech and product event, the company mentioned plans to use generative AI to update title cards in subscribers’ preferred languages.

Kapor Capital’s managing partner Ulili Onovakpuri said yesterday that she is leaving the firm. Onovakpuri started as a principal at Kapor Capital more than a decade ago, and rose through the ranks to become managing partner, and eventually (with Brian Dixon) took over the reins from the firm’s co-founders Mitch Kapor and Freada Kapor Klein. In her LinkedIn goodbye post, Onovakpuri said she had co-raised a $126 million fund and backed more than 70 companies during her time at Kapor. “I’ll truly miss it,” she wrote. “This isn’t a goodbye to investing or to funding the founders building critical solutions. But it is a purposeful pause,” she wrote, noting that she didn’t have any plans just yet. She did say, however, that the venture world hasn’t seen the last of her. “But my inbox has definitely seen the last of ‘just wanted to follow up on the deck I sent you,” she said. “At least for now.”

As “vibe coding” gains in popularity and tech companies push devs in their employ to embrace generative AI tools, a platform that scans for vulnerabilities in AI-generated code has raised a fresh round of funding. Ox Security, which models risk across both AI- and human-produced code, on Wednesday announced that it closed a $60 million Series B. The round was led by DTCP with participation from IBM Ventures, Microsoft, Swisscom Ventures, Evolution Equity Partners, and Team8, and it brings Ox’s total raised to $94 million. Neatsun Ziv and Lior Arzi founded New York- and Tel Aviv-based Ox in 2021. Software and IT engineers by trade, the pair met at Check Point, where they worked on the security firm’s threat prevention product lines. Ox’s platform, which TechCrunch last profiled in 2022, is aimed at both security teams and developers, offering tools to scan code in applications and secure a company’s broader supply chain. Ox can model threats and even recommend fixes, assisting with code reviews and generating executive reports that highlight breaches and possible reasons they occurred. “Over the past year, AI has significantly transformed software development,” Ziv told TechCrunchw. “While these tools accelerate development for both experienced developers and beginners, they often lack the critical thinking and judgment needed to catch subtle security flaws … Ox frees up developers’ time, allowing them to focus on innovation, while simultaneously improving the organization’s overall security posture.” Ox provides tools to secure code, including AI-generated code.Image Credits:Ox SecurityZiv claims that Ox is analyzing over 100 million lines of code daily for around 200 customers, including eToro, SoFi, and two of its investors, Microsoft and IBM. “Our customer base spans from Fortune 10 companies to small- and medium-sized businesses,” Ziv said. “We also count military and government entities as clients, as well as federal agencies.” Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW According to Ziv, Ox’s new capital will be put toward growth and expansion as the 150-employee startup competes for market share against rivals such as Snyk, Veracode, Synopsis, and Checkmarx. Ox is generating around $10 million in annual recurring revenue — a figure Ziv anticipates will double by the end of the year — and plans to be cash-flow-positive within the next 2-3 years. “We want to position ourselves for long-term success and this way we can focus on scaling and reaching our bigger goals,” ZIv said. “We’ve seen significant growth in revenue, and received offers that give us the opportunity to make a leap forward. We felt it was the right time to take this step for the company.”

Mitsubishi Motors is in talks with Taiwan-based Foxconn’s electric vehicle division to source an OEM electric car model that the Japanese automaker plans to sell in Australia and New Zealand in the latter half of 2026. Mitsubishi said on Wednesday that it had signed a memorandum of understanding with Foxtron Vehicle Technology to develop the EV, and would hold further discussions to reach an agreement. Foxtron is a joint venture between Foxconn and Taiwan-based Yulon Motor Co, which assembles and makes cars for Nissan. Mitsubishi and Nissan are also tapping their three-way alliance with Renault to augment their EV lineups in North America. Mitsubishi said it will start selling a rebadged version of the Nissan Leaf electric car in the second half of 2026 in North America, while Nissan would sell a rebranded plug-in hybrid electric vehicle model made by Mitsubishi, also in North America next year. Mitsubishi said it would also similarly sell rebranded EV models made by Renault in Europe.

Amazon Web Services is building a new AI-powered code generation tool codenamed “Kiro,” Business Insider reported, citing internal documents it had viewed. The tool can use prompts and existing data to generate code in “near real-time” by connecting with AI agents, the report said. The tool is said to have web and desktop apps, multimodal capabilities, and can be configured to work with third-party AI agents as well, Business Insider reported. Kiro can also create technical design documents, flag potential issues, and optimize code, the report said. The company already has an AI-powered coding assistant called Q Developer, which is akin to GitHub Copilot. Amazon was mulling launching Kiro towards the end of June, but those plans might have changed, Business Insider reported. AI-powered coding tools are a hot property in tech right now. Cursor maker Anysphere has reportedly raised funding at a $9 billion valuation. Its rival Windsurf is reportedly close to being acquired by OpenAI in a $3 billion deal.

OpenAI sees itself paying a lower share of revenue to its investor and close partner Microsoft by 2030 than it currently does, The Information reported, citing financial documents. The news comes after OpenAI this week changed tack on a major restructuring plan to pursue a new plan that would see its for-profit arm becoming a public benefit corporation (PBC) but continue to be controlled by its nonprofit division. OpenAI currently has an agreement to share 20% of its top line with Microsoft, but the AI company has told investors it expects to share 10% of revenue with its business partners, including Microsoft, by the end of this decade, The Information reported. Microsoft has invested tens of billions in OpenAI, and the two companies currently have a contract until 2030 that includes revenue sharing from both sides. The deal also gives Microsoft rights to OpenAI IP within its AI products, as well as exclusivity on OpenAI’s APIs on Azure. Microsoft has not yet approved OpenAI’s proposed corporate structure, Bloomberg reported on Monday, as the bigger tech company reportedly wants to ensure the new structure protects its multi-billion-dollar investment. OpenAI and Microsoft did not immediately return requests for comment.

Marathon Venture Partners, a venture firm in Athens that prides itself on being “day one partners to Greek tech partners,” just closed its newest fund with €75 million in capital commitments, according to partner Panos Papadopoulos. The vehicle brings the firm’s total assets under management to €175 million — a meaningful amount for an eight-year-old, seed-stage investor in Greece and a reflection, too, of some sizable exits. Among them was the sale last year of Marathon’s portfolio company Augmenta to CNH, a maker of farm machinery and construction equipment in a cash deal that valued Augmenta at $110 million. Marathon also sold some of its shares in Hack the Box, a cybersecurity upskilling and talent assessment platform, to the investment firm Carlyle in a secondary transaction. We chatted with Papadopoulos ahead of an in-person sit-down with him as part of TechCrunch’s first StrictlyVC evening in Athens on Thursday, May 8, a night that will also include a deep dive with Greece’s prime minister, Kyriakos Mitsotakis. What we wanted to know — and what the central questions will be Thursday night — is: why Greece, and why now? Greece has historically seen less venture investment than other European countries. What, if anything, has changed locally that enabled you to raise a €75 million fund when global fundraising has become more challenging? For starters, Marathon I is a top percentile performer globally in [realized returns]; we built a portfolio that captured the current zeitgeist well before, for example, AI-assisted scientific research, robotics or defense became the norm. What is your firm’s thesis and how does this newest fund’s thesis differ given the extended timeline we’re seeing for exits globally? We are backing founders who do something hard in important markets. It can be hard because it requires unique knowledge, like a research PhD, or high agency, meaning understanding of a regulated or overlooked industry like power grid management. And we’re going to continue doubling down on our fast-growing community, which has been accumulating experience and expertise, along with ambition. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Greek startups have traditionally faced challenges scaling beyond the domestic market. How are you evaluating a company’s international growth potential in this environment where capital efficiency matters more than rapid expansion? I beg to differ. Greek startups leverage local talent to serve leading global customers and markets from day one. Across our portfolio there is virtually no revenue coming from the domestic market. But they are serving the best part of Fortune 500. At the same time, capital efficiency and team grit are second nature to our community. We’re seeing fewer IPOs globally and extended holding periods for venture-backed companies. How did this affect your conversations with your limited partners about expected timelines and returns? We don’t need decacorns for our fund economics to work. We invest early on, maintain substantial equity positions, and keep our fund sizes small. These provide for various opportunities for meaningful returns, including secondaries and strategic M&A, well before an IPO. We did secondaries back in 2021 when most of the market was promising infinite holding times. In our culture, cash is king. It seems that many others forgot it. Many European VCs are emphasizing deep tech and AI. Is Marathon taking a similar approach, or do you see different opportunities specific to the Greek ecosystem? Of course we all are, but the definition of deep tech is stretched and means many different things to different people. We are not focusing on any specific sector per se – instead, we are focusing on people changing their sectors. We were perhaps the first generalist VC to invest in defense before the Ukraine war. Greek founders have historically received less funding than counterparts in Berlin, Paris, or Stockholm. Are you seeing valuations for Greek startups that reflect this discount, and does this create opportunities for better returns? In our experience, this is not about geography or price. We are backing founders in non-consensus opportunities that most VCs would ignore. We move fast with conviction and we don’t ask who else is investing. These might sound like table stakes; they still are not. Given the challenging global exit environment, how are you advising your portfolio companies about strategic alternatives like secondary sales or acqui-hires? We work with our portfolio companies toward default alive scenarios. Starting from there, all options are on the table. We see founders truly want to run their companies for the long term. We believe a secondary sale can actually help towards that, and most often we are supportive of such scenarios. The EU has emphasized supporting startups through various funding mechanisms. How important is non-dilutive capital from these sources to your portfolio companies compared to five years ago? We welcome any such initiative. We advise, however, our portfolio founders not to waste time on non-market related activities. How has Greece’s improved macroeconomic situation affected both your fundraising process and the quality of startups you’re seeing? It’s always good when you are not making the press headlines, but what we do is less relevant to local macro. When it comes to the talent front, I would say truly based on naive empiricism that, if there is any correlation, that is inverse. Adversity is the mother of all invention. Many American VCs have pulled back from European investments. Has this created more opportunities for local funds like Marathon, or has it made syndicating deals more challenging? It is definitely a different market but also creates increased opportunity for European investors. I do not think the flood of capital in 2021 truly changed the opportunity for European companies. We must always count on ourselves and be aligned with founders for the long term.

It’s 2025, but business cards are still in vogue — just visit any conference or industry expo and you’ll end up with a pile that’s likely to be discarded sooner rather than later. But as smartphones have become our repositories of information and contacts, people are understandably keen to try out digital alternatives to business cards. Blinq, a startup out of Melbourne, bet that trend would take off when it started off as a hobby project in 2017, offering a digital business card app with a QR-code widget. Today, the company is making off with a bag of gold: It now has more than 2.5 million users — both individual customers and across 500,000 companies in the U.S., Canada, the U.K., and Australia. Off the back of that progress, the startup has now raised a $25 million Series A funding round led by Touring Capital. Returning backers Blackbird Ventures and Square Peg Capital also participated in the round, as did new investor HubSpot Ventures. “[The Blinq’s QR] was a simple, personal way to share who you are, and it worked well between iPhone users. But it wasn’t until late 2019 when most Android devices caught up on QR scanning, and adoption started to grow,” Jarrod Webb, CEO and founder of Blinq, told TechCrunch. “Then came COVID — QR codes went mainstream, in-person meetings became more intentional, and Blinq’s focus on making those moments seamless and memorable started to take off.” The startup has taken the B2C2B route ever since. The app lets users create several customized digital business cards for different needs and connect with contacts using them. The app can also automatically capture details and sync them with CRM systems such as HubSpot or Salesforce by using QR codes, email signatures, NFCs, short links, or video call backgrounds. Blinq is used by individuals, small businesses, and global enterprises, and 80% of its customer base is located in the U.S., Webb said. Its team has scaled from five employees based in Melbourne to 67 across Sydney, Melbourne, New York, and San Francisco, supporting its product development and go-to-market efforts. “Every time someone uses Blinq, they’re introducing it to someone new. And further, we see more frequent usage by active users the longer they’re on the platform,” Webb said. “That built-in virality drives organic growth and keeps our customer acquisition costs low. On the business side, companies pay per seat. As more employees adopt the product, teams grow organically, creating expansion revenue over time.” Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Blinq competes with several companies providing similar digital business card services, including Mobilo, Popl, Wave, and Wix. Of course, the app also has to contend with social networking platforms like LinkedIn, landing pages, and services like Linktree. But Webb feels Blinq is more suited to building relationships and provides users more ways to follow up and engage with new contacts. Webb sees digital business cards as more than just an end point. “They’re our wedge. Because when you are the trusted tool at the moment a relationship begins, you earn the right to shape what follows. We’re focused on giving people everything they need to turn first impressions into real momentum — from dynamic, context-rich profiles to smart ways to stay top of mind. That means expanding into new markets, deepening our presence with businesses and enterprises, and continuing to evolve how people connect in a world that’s changing fast.”

Instacart has launched a new drinks and snack delivery app called Fizz, the company announced on Tuesday. The 21+ app is built for friends and family and is designed to make it easier to stock up for a party or get-together. The host of a party can invite guests into their Fizz cart to allow everyone to choose what they want to bring to the gathering. The app aims to eliminate the headache of splitting the bill, so everyone pays for what they add to a cart. “It’s becoming more and more common when people gather that they share in contributing to what it takes to put together a party, but then they do this kind of bill-splitting thing,” Daniel Danker, Instacart’s chief product officer, told TechCrunch. “Afterwards, they open up Venmo, and they kind of divide up the cost. And while we think that that’s great, certainly better than a more manual way of doing it in the past, we actually think there’s an opportunity to just completely eliminate that step and make it even more effortless.” With the launch of Fizz, Instacart is diversifying its revenue streams and offerings, all while appealing to new types of customers. The social, party-planning aspect of Fizz enables the company to tap into a younger demographic that may not be using Instacart. As the host on Fizz, you can create a party cart and drop the link into your group chat. Guests can add to the party cart even if they don’t have the Fizz app downloaded. You decide when to place the full order and get it delivered immediately or scheduled in time for your party for a flat $5 delivery fee. Guests don’t have to pay for a delivery fee or any other sort of fee. Image Credits:InstacartUnlike Instacart, which allows you to place orders across grocery, restaurant, retail, and express partners, Fizz is focused on a specific category: drinks and snacks. This includes things like beer, seltzer, non-alcoholic drinks, chips, dip, healthy snacks, or party supplies. When you buy drinks, you earn “Snack Bucks,” which can be redeemed for discounts on snacks. The items in a party cart will be shopped from a nearby grocery store on Instacart’s platform and delivered by an Instacart shopper. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Danker said Instacart developed the idea for Fizz after recognizing that the way people gather and socialize is evolving. For instance, the older generations would pay for everything when hosting an event and choose what they’re going to serve. However, the younger generation is doing things a bit differently, as Gen Z and millennials tend to feel a need to share in the cost of a gathering, Danker notes. Plus, they want to ensure that everyone’s needs are considered, whether it’s someone who doesn’t drink or someone who follows a gluten-free diet. Fizz aims to address this evolving landscape in an affordable way, he says. While the app is designed for groups, Danker believes there’s an opportunity for Fizz to be used by individuals who may prefer the new app over Instacart. “Instacart is designed for the weekly recurring shop and then filling in orders in between,” Danker said. “The Fizz experience is a very lightweight experience. You don’t even have to choose a store up front. It’s $5 flat, and you get your order delivered. You don’t have to have a membership, where most customers that use Instacart, they come back every week. So the membership makes a ton of sense. Fizz, no membership required, really lightweight, really simple concept. You’re earning Snack Bucks as you shop. So, it’s designed for a different demographic and different audience than tends to use Instacart.” Image Credits:InstacartAs part of the launch, event planning app Partiful has integrated Fizz directly into invites for hosts and guests. You and your guests will be able to tap the Fizz link to get started on a party cart. Everyone on the event page can see the whole cart right within Partiful. “With our new Fizz integration, we’re solving another big logistical headache: stocking up for a gathering and making sure everyone’s tastes are accounted for — bringing your whole event to life in a just a few taps,” Partiful CEO Shreya Murthy said in a statement to TechCrunch. The launch of the new app comes a year after Uber shut down alcohol delivery service Drizly after acquiring it in 2021. Although Drizly and Fizz share some similarities, Instacart is taking a different approach by incorporating a social aspect in its new app, as it aims to capture the attention of Gen Z and millennials to drive its success. Plus, while Drizly was focused on alcohol delivery, Fizz features snack and party supplies. Fizz complies with local laws and jurisdictions regarding how alcohol sales are handled. The app is available in Alabama, Arizona, California, Connecticut, Florida, Georgia, Hawaii, Idaho, Illinois, Iowa, Kentucky, Louisiana (some parts), Massachusetts, Michigan, Minnesota, Mississippi, Missouri, Nebraska, New Jersey, New Mexico (some parts), New York, Nevada (some parts), North Carolina, Ohio, Oregon, Texas, Tennessee, Virginia, Washington, and Wyoming.

YouTube continues to dominate the podcasting space, but Spotify is actively working to close the gap. The latest development: a new feature that lets users see how many times an audio-only or video podcast episode has been actively listened to or watched. The streaming giant revealed its new podcast metric called “plays” on Tuesday, making it possible for users to see which podcast episodes are most popular. This is the first time a podcast metric like this will be viewable for creators and users. The plays will appear next to a podcast episode throughout the app, including the home page, episode page, and show page. It’s also available for creators on Spotify for Creators and Megaphone. With this new metric for podcasts, Spotify aims to encourage users to explore podcasts they may not be familiar with, especially if they see that other listeners highly favor these episodes. For creators, this information sheds light on which episodes resonate most with audiences and, more importantly, allows them to benchmark their performance against competitors. Spotify’s announcement follows its first-quarter earnings, revealing a gain of 5 million premium subscribers, totaling 268 million. This marks the second-highest total ever and the largest first-quarter increase since 2020.

Uber is expanding its presence in Turkey with its latest acquisition. The ride-hail and delivery giant has acquired an 85% controlling stake in Trendyol Go, the online meal and grocery delivery business based in Istanbul, for about $700 million in cash. The deal gives Uber instant market share for Uber Eats in the country, where Uber currently only operates a ride-hail service. Uber’s acquisition announcement comes a day before the company is scheduled to report its first-quarter earnings. And with consumer spending projected to fall in 2025 due to President Trump’s tariffs, investors will want reassurance that Uber is still poised for growth, even as the company recently pulled out of its $950 million bid for Delivery Hero’s Foodpanda business in Taiwan. Uber has been laying the groundwork for expansion in other arenas, too. It seems like every few weeks, Uber announces a new autonomous vehicle partnership across its ride-hail, delivery, and freight verticals. Trendyol Go is the food delivery arm of parent company Trendyol Group, a fashion and retail e-commerce platform, which is majority owned by Alibaba. The company said in a filing that it delivered more than 200 million orders in 2024, generating $2 billion in gross bookings, which was up 50% from the previous year. There are more than 90,000 restaurants and 19,000 couriers on its marketplace. The acquisition is expected to close in the second half of 2025. When it does, users will be able to keep using the Trendyol Go app, and Uber will “introduce key capabilities over the coming years from Uber Eats,” the company said. TechCrunch has asked Uber for more information. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Uber’s latest acquisition comes as it struggles for market share back home in the U.S. due to competition from DoorDash. In February, Uber filed suit against DoorDash, alleging anticompetitive behavior. DoorDash has urged the courts to dismiss the case. Earlier today, DoorDash announced two major acquisitions to expand its presence in Europe: The U.K.’s Deliveroo for $3.86 billion and SevenRooms for $1.2 billion.

The autonomous navigation market — where ships, guided by AI, steer themselves, resulting in fuel and time savings — is projected to sail past $11 billion by 2028. As a result, companies in this space are pushing on an open door. The latest is Orca AI, which closed a Series B funding round of $72.5 million led by Brighton Park Capital. Existing investors Ankona Capital and Hyperlink Ventures also participated. The London-based company has now raised over $111 million, including a $23 million funding round last year. So what drove the new round? In a word: defense. Founded in 2018 by CEO Yarden Gross and CTO Dor Raviv, Orca AI applies AI-powered decision-making and autonomous capabilities to ships based on a marine visual dataset of over 80 million nautical miles. By employing AI in navigation, it’s possible to significantly reduce collisions and allow crews to focus attention on other aspects of the voyage. “The main business still is in the commercial sector. We already have collaborations and POCs,” Gross told TechCrunch. “But we see opportunities in defense coming from navies around the world around autonomy,” he added, “where they want more cost-effective assets that can operate more efficiently with less human intervention. We’ve already signed the first contract in the defense field, deployed on a navy ship.” Orca’s growth is also benefiting from the expansion of Starlink, which allows real-time data to be transmitted to Orca AI for mapping routes, traffic monitoring, and sharing critical information. “Starlink enables us to collect data at scale directly from the ship sensor. We see that as a huge opportunity,” Gross said. The company claims that a 2024 analysis of Orca AI’s alerts system showed a 54% reduction in close-encounter events leading to an average of $100,000 savings in fuel per vessel per year. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Other companies working on autonomous navigation at sea include Avikus (subsidiary of Hyundai HD) and Sea Machines.

Elon Musk may have a sizable fan base, but residents of the upscale Austin suburb of West Lake Hills, Texas, are unimpressed by their celebrity neighbor, reports The New York Times. Instead, a $6 million home Musk purchased in 2022 has become the center of a battle after his team erected an unauthorized 16-foot chain-link fence, installed a metal gate, and mounted outward-facing cameras. “I call that place Fort Knox,” says Paul Hemmer, a Tesla owner and retired real estate agent who lives across the street and serves as president of the neighborhood homeowners association. Musk’s visibly armed security personnel and their vehicles have disrupted the tranquil street, and the prospect of Musk’s return from Washington has some worried over more than his failure to obtain construction permits. “If you follow him at all in the news, he’s always guilty of building stuff and then asking for permission later,” Hemmer complained in a planning meeting. For Hemmer, the billionaire’s proximity may come at the steepest price. Per the Times, Musk’s security team once reported Hemmer to police, claiming he was naked in the street. Hemmer, who has flown drones over Musk’s house looking for ordinance violations, countered he was on his own property in his underwear.

Employer.com has acquired MainStreet.com for an undisclosed amount, the latest fintech startup to get snapped up by the workforce management company. In a post on X, Employer.com chairman and co-founder Jesse Tinsley said the two companies were “merging forces to simplify business back office solutions into one powerhouse platform.” Tinsley confirmed the acquisition to TechCrunch. MainStreet, a San Jose, California-based startup founded in 2019, built a business around helping startups uncover research and development tax credits. The startup generated revenue by taking a cut from the pool of credits. MainStreet had some success in its first year, crossing the $1 million ARR run rate threshold and helping the average client save $51,000. In 2021, MainStreet’s revenue crossed $15 million, per industry newsletter Not Boring. Signs of potential trouble appeared in 2022 when MainStreet laid off about 30% of its staff, citing “an incredibly rough market.” At its prime in 2021, MainStreet was valued at $500 million. The company was said to have closed on a financing in 2022 at a $200 million valuation. It’s unclear what MainStreet’s balance sheet looked like immediately prior to this acquisition, although Tinsley told TechCrunch in an interview the company was profitable. In total, MainStreet raised about $75 million in known venture capital from investors such as SignalFire, Tusk Ventures, Shrug, Moxxie Ventures, Weekend Fund, Gradient Ventures, Sound, and SV Angels. One of MainStreet’s investors introduced the company to Employer.com, according to Tinsley. MainStreet’s 15-person team will be joining Employer.com as part of the transaction, which has about 500 employees across all its companies. With the acquisition, Employer.com is valued at just north of $700 million, Tinsley said. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW The San Francisco-based company has been on a shopping spree recently. In late 2024, Employer.com announced it was acquiring Bench, a VC-backed accounting startup that left thousands of customers locked out of their accounts after it suddenly shut down, in a fire sale. Last week, Bench conducted a round of significant layoffs. And in January, Employer.com had offered to acquire Level, a fintech startup that abruptly shut down after failing to find a buyer, but that deal didn’t go through. “When we originally started Employer.com and then bought Bench, the overarching theme… is basically automating an end-to-end platform for the G Suite for the business back office,” he told TechCrunch in an interview. Buying MainStreet is in line with that goal, Tinsley said. In late January, Tinsley and Employer.com was reportedly teaming up with YouTuber MrBeast and others to save TikTok by submitting an all-cash bid for the app, according to a report in Bloomberg. It’s unclear what happened to that alleged buyout attempt, although Tinsley publicly confirmed in March that he was part of that $30 billion bid. This story was updated post-publication to reflect MainStreet’s accurate funding amount.

Individuals may work closely with AI agents as they become increasingly prevalent in the workplace. According to a report by Boston Consulting Group, the market for AI agents is expected to grow at a 45% compound annual growth rate over the next five years. Just like human employees, AI agents could be onboarded to learn various roles, access company information and business contexts, and integrate into workflows. In addition, unlike traditional automation tools, AI agents have the potential to constantly adapt and improve their operations. Relevance AI, a San Francisco- and Sydney-based startup developing an AI agent “operating system” to enable businesses to build teams of AI agents, has raised $24 million in Series B funding led by Bessemer Venture Partners. Returning investors King River Capital, Insight Partners, and Peak XV also participated, bringing Relevance’s total raised to $37 million. The company did not provide its valuation. The fundraising comes about a year and a half after the startup closed its Series A. Relevance says that it has experienced rapid growth, with 40,000 AI agents registered on its platform in January 2025 alone. Customers include Qualified, Activision, and Safety Culture. Relevance has to compete with players in the AI agent space like Retell, Qeen.ai, SmythOS, Gooey.AI, Cykel AI, and Microsoft. The 5-year-old company sees agent builder platforms, vertical agent software, and agent engineering frameworks as its competition, according to co-CEO and co-founder Daniel Vassilev. “[W]e’ve even seen incumbents like Salesforce make a big bet on the value of agents,” Vassilev told TechCrunch in an interview. “[Relevance] enable[s] training the agent to really specialize in the niche workflows of their organization. We’re also tool- and model-agnostic, allowing our customers to leverage their entire tech stack across their entire business rather than just a single vendor’s ecosystem.” Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Relevance says that it’ll use its new funding to further enhance the product capabilities of its AI agents and provide support to customers in its primary markets of Australia and the U.S. Vassilev has relocated to San Francisco to open Relevance’s office and build up its go-to-market team there. The company says it has 80 people across its San Francisco and Sydney offices today, up from 19 employees in 2023. Coinciding with its Series B funding, Relevance is introducing two new features on its platform. “Workforce” is a no-code multi-agent system designed to help non-technical professionals and engineers build specialized teams of agents to collaborate like human employees, completing complex processes from start to finish. “Invent” is a tool that lets users create AI agents using text prompts.

In 2017, Raghav Gupta set out to solve a personal problem: He wanted easy access to the home-cooked meals he grew up eating without having to spend time cooking or spend money on takeout or hiring a private chef. He turned to robotics, which led him to found the startup Posha. Posha, a former TechCrunch Startup Battlefield company, builds countertop robots that make meals using computer vision. Users scroll through a list of recipes, select the one they want, add the proper amounts of the requested ingredients, and the machine makes the meal from there. The process is designed to be customizable and forgiving, Gupta told TechCrunch, so the machine allows people to make substitutions, and Posha still works if a user doesn’t measure their ingredients perfectly. “It’s like a coffee machine for food,” Gupta said. “So when you want to drink a cup of coffee, you choose a brew of coffee on your coffee machine. You put beans, sugar, and milk in different containers. You tap brew, and out comes a cup of coffee. Posha does something similar, but for food.” A coffee machine is a good, but not perfect, comparison to Posha, as Posha requires a bit more labor than a coffee maker. While Posha does a substantial amount of the work by cooking these meals, consumers still play an active role in shopping for ingredients and prepping everything that goes into the device. Chopping, especially, can take up a fair amount of a recipe’s cook time. Gupta agreed that some people are just not going to go for a solution that still requires them to do some of the cooking. He said that Posha has found the most success thus far with customers who like to cook two to six times a week anyway and are looking to lighten the load a few of those evenings. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW “These people are already spending an hour in the kitchen every single day, deciding what to eat, shopping for ingredients, cooking a meal, [and] cleaning up afterwards,” Gupta said. “We help them shave off at least 70% of this time, so they now end up spending only about 10 to 20 minutes every single day.” Posha, formerly known as Nymble, originally started out as a robotic arm, Gupta said, but the company’s time in Bosch’s accelerator program prompted them to change course. They learned consumers didn’t want something that moved around their kitchen or that would be hard to clean. The company has kept in close contact with its early customers ever since. “We have been super focused and super obsessed with customers from day one,” Gupta said. “We don’t use Zendesk to chat with them; we have WhatsApp conversations with over 100 of our customers. Most customers know me personally. I moved to the U.S. in the middle of the pandemic, just to be close to my customers.” That system can’t scale, but clearly works for Posha for now. Gupta said that, so far, Posha has mainly relied on word-of-mouth marketing for the $1,750 direct-to-consumer countertop device. Posha recently raised an $8 million Series A round led by Accel with participation from existing investors, including Xeed Ventures; Waterbridge Ventures; and Binny Bansal, the co-founder of Flipkart; among others. Gupta said that Posha will use the funding to continue to develop the product. In particular, the company wants to add more recipe options and the ability for people to suggest recipe ideas and have generative AI turn those ideas into instructions and add them into the device quickly. The company launched its Posha robots in January 2025 and has since sold out of its first batch — and is taking preorders for its second. “If you look at your microwave, your dishwasher, your refrigerator, at some point in time, these devices were countertop devices,” Gupta said. “They became so indispensable over time in consumer homes that builders started installing these devices in your homes. We feel Posha will have the same fate very soon.”

Agree.com says its AI-powered e-signature platform is different from competitors because it includes invoicing and payment processing. That’s why the company might have a shot at tackling the industry Goliath, Docusign. Because the startup makes its money from transaction fees for any money movement facilitated by its platform, Agree.com has made e-signatures free to all users. And now it’s raised a $7.2 million seed round, the company tells TechCrunch exclusively. Founded in February 2024, Agree also raised $3 million in a pre-seed round of funding last year led by Sheel Mohnot, general partner at Better Tomorrow Ventures. This latest financing was oversubscribed and led by Tyler Hogge at Pelion Venture Partners, according to Agree.com co-founder CEO Marty Ringlein. Funding for the raise only took two weeks, according to a source familiar with the transaction. Agree.com uses AI on top of optimal character recognition (OCR) software so that it can auto-detect and label all of a contract’s input fields and signature blocks. Its technology can also identify and extract “any and all” payment terms to dynamically generate invoices. “At the end of almost every signature, someone has to pay someone money,” Ringlein told TechCrunch. “We combine what has historically been a disjointed and fragmented workflow to make signing better and payments faster.” Ringlein believes that because of its multitasking approach, Agree.com can potentially replace traditional e-signature software and invoicing and accounts receivable tools such as Bill.com. Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW “Agree extracts every character, indentation, semicolon, and hyphen to not only understand the type of contract being signed, but make it fully editable and collaborative with commenting, redlining, and version control,” Ringlein told TechCrunch. Although it primarily competes with Docusign, Agree’s business model is a fintech company through B2B payments. So far, its trajectory seems promising. In its first three months, after launching in early September 2024, it hit 10,000 users. Seven weeks later, it doubled to over 20,000 users. Today, it has over 25,000 users, including ad networks such as Beehiiv and Product Hunt, B2B SaaS startups such as Rho and TaxGPT, and enterprise sales teams like Brico and Thoropass, it says. Agree offers a premium offering for larger teams that charges a traditional monthly SaaS fee per seat. It also will monetize invoicing and billing logic on transaction volume. Presently, Agree has seven employees, including co-founders Will Hubbard (COO) and Evan Dudla (CTO). All of the founders have launched and sold multiple previous startups. Ringlein, for instance, previously sold design agency nclud to Twitter in May 2012 for an undisclosed amount. In 2016, Ringlein, Dudla, and Agree’s CPO Mike Dick sold a startup called nvite to Eventbrite. In 2020, that trio also sold Gather to Brex. Hubbard started his first company, air-quality monitory startup ChemiSense, as a junior at UC Berkeley. He ran it for about six years and sold it to Kaiterra in 2019. Hubbard then started his next company, Niche (verticalized community marketplaces), shortly thereafter, and it was acquired by Opera Event in 2020. More recently, Hubbard and Ringlein also started early-stage venture firm Adventure Fund, which has invested in the likes of Mercury and Beehiiv. As for the growth plan for Agree, Pelion partner Tyler Hogge told TechCrunch that “the smartest way to get massive adoption would be to use e-signature as the wedge, give it away for free, and make it impossible for incumbents to reply.” Hogge added that Agree’s “business model is truly unique: free software, monetized through invoicing and payments.” Blank Ventures also participated in the seed round, along with angel investor Gokul Rajaram. All existing backers, including Better Tomorrow Ventures, 8-Bit Capital, Sophia Amoruso’s Trust Fund, Hustle Fund, Everywhere Ventures, Singh Capital Partners, and Firsthand VC doubled down on their investment. While the company primarily operates in the United States today, it intends to expand internationally later this year, starting with the United Kingdom, Canada, and Australia.

NewLimit, a startup that aims to increase how long people can live a healthy life by genetically programming their cells, has raised a $130 million Series B led by Kleiner Perkins. New investors Nat Friedman, Daniel Gross, and Khosla Ventures also invested, as did returning backers Founders Fund, Dimension Capital, Elad Gil, Garry Tan, Patrick Collison, and others. The fundraise comes two years after the company raised a $40 million Series A. Founded over four years ago by Coinbase CEO Brian Armstrong (pictured above), former GV partner Blake Byers, and stem cell professor Jacob Kimmel, NewLimit claims it has made significant progress toward developing treatments that can restore youthful characteristics to aged cells. Kimmel told TechCrunch the company has discovered three prototype medicines that reprogram liver cells. NewLimit’s lab experiments, he said, demonstrate this rejuvenation can restore the cells’ ability to process fat and alcohol. The company measures its progress by contrasting how cells from a younger individual and an older person respond to the substances. An older liver cell that has gone through NewLimit’s epigenetic reprogramming behaves more like a younger cell, Kimmel said. Promising early results notwithstanding, NewLimit is still a few years away from starting human trials. In the meantime, the company wants to continue developing new anti-aging medicine using an AI model and then testing the most promising of these machine-generated drugs in its laboratory. “What the AI model allows us to do is run all those experiments in simulation and then only follow up on the most promising subset,” Kimmel said, adding that the data points from the experiment are then used to retrain the AI model in a process known as “lab in a loop.” Techcrunch event Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 BOOK NOW Other noteworthy biotech startups developing drugs that aim to increase lifespan include Retro Biosciences, which raised $180 million from OpenAI CEO Sam Altman two years ago and is reportedly raising a $1 billion Series A; and Altos Labs, which launched with $3 billion in 2022 and is backed by Jeff Bezos.