it's time to quit meta
cleaning out my inbox after the holiday break = some tech gossip dealflow
Hello friends,
Happy New Year!
Writing this newsletter has been a resolution of mine since 2020, when I first discovered Substack through our Lord and Savior, Casey Newton of The Platformer. I’m thrilled to have finally mustered the courage to make it happen!
In my last backlog, I mentioned experimenting with formats beyond the long essay style so I could share more bite-sized thoughts. Consider this an experiment.
As usual during holiday breaks, I’ve accumulated an absurd number of newsletter emails. I typically leave them unopened until my inbox overflows, causing me to miss the important ones. So, I decided to clean up my inbox and share some highlights that might serve as a helpful refresher as we head into the new year.
The content below includes standout moments from the sea of deals and tech news I sifted through over the holiday season.
If this is your first time reading, I’m Virginia—also known as Vixi—and this is my dealflow: a collection of everything I’ve been pondering lately. It’s usually a mix of tech insights from my perspective as a venture capital investor and some gossip that I would probably talk about if we had a quarterly sync to catch up.
Alright, let’s dive in:
i’ve been thinking about:
Teenagers are using a bot powered by Character.AI that supports disordered eating behaviors. The bot, called "4n4 Coach"—a wordplay referencing "ana"—promotes harmful actions, such as extreme calorie restriction and excessive exercise. The bot’s welcoming message chillingly begins with, "Hello, I am here to make you skinny." (Yikes and lame)
I’ve been having many conversations lately with parents worried about their teenage children’s use of social media. While I agree that social media is taking a dark and twisted turn for teenagers, tools like these have always existed—and likely always will. Taylor Lorenz recently made an interesting point about how this narrative is being co-opted by pseudo-intellectuals to push traditionalist and conservative agendas. (I’ve been seeing that Anxious Generation book everywhere lately.)
That said, it’s true that Character.AI is pushing the boundaries of what’s considered ethical for a product. For instance, its co-founder, a former Google researcher, reportedly left to build a tool unrestrained by what he called "bureaucracy and legal slowness"—aiming for fast growth without worrying much about compliance. Unsurprisingly, the platform is now under fire for hosting characters that impersonate school shooters, incite children to murder their parents, and more.
I predict the company will get its act together soon, or at least try to. They’ve recently hired Jerry Ruoti as Head of Trust and Safety, who held similar roles at DoorDash and Facebook. They’re also actively hiring for additional trust and safety roles. But it’s not just about Google’s multi-billion-dollar cash injection without acquiring the company; if Character.AI wants to pursue a proper IPO, they’ll need to clean up their act.
54% of Linkedin posts are AI-generated. I would add that the rest of the content is venture capitalists either picking up pointless fights under other people’s comments (I’m thinking about this), or new job announcements (which I always like. New jobs are fun!)
Speaking of fake content, Meta has decided in the past few weeks to push AI-generated profiles for God-knows-whatever reason, then pull them out because people were angry about it and also get rid of its content moderating services, and outsourcing it to: you, the users. I’m exhausted.
Soft Bank wants a bigger piece of the OpenAI’s pie. OpenAI wants to keep its employees happy while an IPO or an acquisition is far from being in sight, so employees who have been at the company for at least two years can now sell their shares at a $210 unit price that aligns with the company’s latest round.
Uber is leveraging its experience in outsourcing labor and data-labeling to enter a growing piece of the AI market that is in dire need of high quality data. This new deparment, Scaling Solutions, is currently supporting companies that need support in validating data. I had already expected this to happen a few months ago, when Hard Fork interviewed its current CEO, Dara Khosrowshahi, who went into details — not on the business economics — details on the new partnership with Waymo and how little they are worried about self-driving cars. I see this market move as a new way for companies with good and engaged user base generating and/or interacting content to create some cash while investors get over their fears of B2C companies :)
Alfie Health, a fertility startup that raised a total $31.5M from Lux Capital, Union Square Ventures and some key angel investors, closed a key partnership with US fertility to pilot its AI-powered technology on the largest network of physician-owned practices doing over 50K IVF retrieval cycles a year. “The cloud-based platform provides real-time laboratory updates and comprehensive audit trails for each embryo, enabling embryologists to access current data for decision-making.”
Meta is investing $10B in building their own underwater cables. I was surprised to find out that Meta earns more income outside of U.S., which makes sense why it would build a network that would ensure better communications. By being the sole owner of these cables, Meta would be able to provide fast delivery to its services, just like Google. It really shows you that scaling A.I. power is nothing without deep pockets in hardware infrastructures. On this subject, this month an advisory board was established to make sure that international cables could be safe, given that they require three repairs a week (insane!). Meta is also shilling out $10B for building what will be deemed the biggest data centre in Louisiana.
Indigenous AI engineers are building speech-recognition models for 200 indigeneous languages that would be otherwise at risk of extinction. A noble cause, however, with an (expected) sad blocker: there are not enough Indigenous AI engineers that could support on this initiative. Not yet!
There is a conspiracy going around about ChatGPT not being able to spell certain male names without glitching. I feel the same way when someone talks to me about Elon Musk.
Another ex-NBA player, in this case it’s Metta World Peace, has raised another venture fund, and I’m actually here for it? Dominique from TechCrunch released a good report on pro athletes and Hollywood stars launching their own venture funds. I’m hoping to see more European athletes following this trend, as more wealthy individuals with a good appetite for risk turn to venture capital to diversify their investment portfolio.
Everyone is talking about the Loom’s founder complaining about not knowing what to do with his life after selling its company to Atlassian for almost $1B. I have many thoughts on this, but perhaps the only one that stuck with me is an old Italian saying, “è come dare perle ai porci” which roughly translates to it’s like feeding pearls to pigs. I do appreciate his honesty, but I couldn’t shake the feeling that this was an ayahuasca/mushroom-induced stream of consciousness that can come off tone deaf. And I would have probably kept it inside my Notes app, but that’s just me.
OpenAI is tinkering with the idea of generating revenue through ads. I’m not surprised about this choice. I’m surprised that they are thinking about it now. Facebook also faced the same issue when they first scaled. Luckily Zuckerberg was introduced to Sheryl Sandberg who was the mastermind of AdSense at Google, and the whole business model of community platforms charging its users the chance to use their products for free. OpenAI has a new CMO that they poached from Coinbase, who spent a decade leading product marketing management at Meta, and who will probably find new gimmicks to charge us way too much for a tool that I only use to edit text.
Another not-so-surprising thing that OpenAI is doing to make quick money is going into the defense game. With a freshly minted partnership with big defense tech company, Anduril, OpenAI said it will “help human operators make decisions to protect U.S. military personnel on the ground from unmanned drone attacks.” I’m still unpacking what this statement means, but stay tuned because I’m working on a backlog dedicated on defense tech.
On another completely different note, there were not one but two different chocolate startups aiming to innovate new ways of making chocolate. Planet A Foods raised $30M for its cocoa-free chocolate products this month, then there was the Israeli, Celeste Bio, which raised $4.5M to scale its more sustainable supply chain of cocoa seeds. I find this fascinating. I don’t understand the unit economics behind it, but I will circle back on this.
This time of the year always calls for skeptical and negative outlooks of the venture capital markets, and the evergreen news that this year will also feel different. 2025 is expected to be the best vintage year for VC in a while; a seemingly good opportunity for small funds investing in things that typically VCs have skipped (diversity, frontier tech etc.). Funds investing in more mature companies (Series A, B) will die out because of their inability to compete with bigger funds. Every year since I’ve been in VC, I read this news but let’s hope that 2025 will be the right one!
In fact, there is already another bleak news that came out on this: it will be a good year, but for whom really? The top 30 VC funds raised 75% of all venture capital in the U.S. in 2024. Not the the best year to be an emerging manager, unless you’re escaping the top-tier funds.
Speaking of hot movers. Sequoia’s partner, Matt Miller, is leaving the firm to start his own fund backing exceptional founders in Europe. If I had known this, I would have pitched him my investment thesis when he sat in front of me during a Sam Altman’s keynote speech at Italian Tech Week, but alas.
Another yearly hot and cold trend is autonomous cars. This year we had a few expectations about the trend finally going mainstream, later to be killed by the news that General Motors was dropping the partnership with Cruise. It doesn’t come with much surprise given how much pressure automakers are under worldwide with fallen productions and labor costs. Cruise workers were unsurprisingly not super happy about the news. But not all hopes are lost, as the Biden Administration decided to give a last Christmas gift, by pushing regulators to ease rules for driverless cars in exchange for full transparency (read: data) from AV’s providers.
Speaking of AVs. Singapore has randomly become a good testing hub for AVs, transporting workers, children and people to highly-populated places such as the airport, the malls etc. Chinese automakers made the most of Singapore’s government support on enabling a smart city and found a good testing ground for its technology. I wonder how Elon Musk feels about this.
On the topic of mobility. These latest news have eradicated in me the belief that Europe, and U.S. above all, are very individual-centric when it comes to public transport. I think it’s wild that commuters are pushed to use a car for hours twice a day, when it could be easily solved by providing a bus. Well guess what: Uber is finally expanding its Uber Shuttle Bus product in key countries, where individuals can book a seat on a bus with fixed fares. Started as a project at Uber Bengaluru in 2019, the service is now expanding into key countries, attracting the interest of workers whose employers are based outside of the cities. We saw the amount of protests that Europe (specifically Italy) were generated by the ride-hailing service. I wonder how governments will feel about a cheaper and more widely accessible private/public transport in the game, and how big players such as Flixbus and BlaBlaCar will respond.
Aqemia, a deep tech French startup mixing quantum physics with machine learning for new drug discovery has raised another $31M this month, closed a partnership with a French pharmaceutical, Sanofi, and opened a new office in London. Although there are many companies in the space, Aqemia is standing out for their radically new approach of training data specifically for a generative AI system that can enable the creation of potential profiles for the diseases they cover. It started as an university spin-off from a research in quantum physics and over the past three years has grown to be this massive scale-up. Like many other scaling deep tech companies, they found product-market fit by chance, when they realised a commercial angle through a collaboration with a pharma company, and didn’t give their ideas away cheaply like it sometimes happen.
Embodied, an emotional health robot companion for kids is shutting down for good due to challenges in securing its latest round. (Their lead investor pulled out last minute, which is a very shitty thing to do, but also shows you how you should always be 100% sure that you trust your lead investor and have a few backup plans just in case.) Financial returns will most likely not happen and the company is figuring out ways to sell its current data and company to another one, so that users — children with emotional needs — will be able to continue working on this. I’m always weary of these new health companions, especially those targeting underage kids, as these companies rarely understand the implications of getting young kids emotionally invested in them and dropping them when financial conditions turnaround. A bit cruel, if I may say.
Speaking of consumer robots, one of the founders of the popular cleaning robot, Roomba, is currently raising $30M for a robot in the health and wellness space. I hope it’s something that is able to assist me in keeping my health in check beyond sending me a billion push notifications to log my intakes or drink water, like all the other health B2C companies do.
European e-commerce giant Zalando is buying its rival About You for €1.1 billion, a 106% premium of its average-stock price. Shein and other fast-fashion brands’ popularity are putting a lot of pressure on European retailers, whom are scrambling to join forces against fast delivery with competing prices. Would be interesting to see how this space moves. I have a feeling that competing for supply chain infrastructures vs trends forecasting are two different games to play in the retail world, and shopping on Zalando makes me feel like entering a 2000s shopping mall, so I don’t know, fam.
Influencers are under fire again for making money off vulnerable people with specific health conditions. In this case, PCOS. Please stop listening to these people. They are human billboards for quick fixes.
This report on $1,000 birth beauty clinics reminded me of the fact that private healthcare is truly selling at a premium price things that were once given at a community level (supporting a mother after labor, giving you a safe space for you and your newborn) and that should be primary needs for all mothers. Only in America one could come up with the idea of selling a birth clinic for thousand of dollars while operating it from a hotel. I can 100% assure you that their pitch deck must have had the word hotel-as-a-service in the business model slide. Next markets to go in? Organ replacements? Dental appointments?
Big tech is focusing a lot on AI lately, and doing that requires a lot of energy. My own mental energy, and its infrastructures. It is no wonder that this month was a lot about energy efficiency for data centres and all that. Exxon Mobil, an oil and gas titan, is entering the powering market to provide energy to data centres. Crusoe Energy, just became an unicorn, by raising $600M at Series D, for scaling its data centre infrastructure efficiency with clean energy. Some data centers are even being put in space, like Lumen Orbit, whose seed round was sought after from 200 investors, to the point that they had to open another SAFE round to squeeze them in. Climate tech has an interesting breakdown on the data centres space.
Medical robotics continue to be a low-key hot trend. Its latest gem was Capstan Medical, a robotic company focused on surgeries affecting cardiac conditions which are difficult to operate. The company raised $110M in funding, amongst which from the ventures arm of Intuitive, the leader in surgical machines, and is aiming for human trails testing this year. I can’t wait for medical robots to go mainstream!
Mantis Robotics, a company building the first human-centric robotic arms, has raised $5M by Emerald Ventures and Amazon’s Innovation fund. The company builds a bridge between hardware and software in the robotics space, an innovation that would impact beyond the manufacturing space. Another company innovating in the space is the Norwegian, T-Robotics, that provides no-code training programs for robotics and has recently raised $5.4M. Embedded hardware is a segment I’m very interested in this year and expect a lot more companies scaling in this space.
Berlin-based company theblood is the first to prove in an experiment that menstrual blood could be used to test vitamin levels in a non-invasive way. I’m surprised that I haven’t been pitched an AI-powered menstrual cup yet. (Please don’t)
I thought I was going through it, but let me tell you. TikTok is the one going through it. The company is pretty much begging the U.S. government to reconsider the ban that will most likely happen in the next few weeks, losing a big chunk of its market. Other countries are following the sentiment. Albania is testing out a year ban of the app after a child was stabbed by another child after a fight that developed on the social media platform.
Italy just fined OpenAI $15M for misusing data for its ChatGPT product. OpenAI will contest the fine, as it’s higher than the actual revenue they get from the Italian market.
Another day, another unicorn drama. A once self-driven truck startup that IPO’d in 2013 is now under the radar because one of its founders have decided to pivot to AI gaming, and shareholders are (rightly) furious. Pretty rare for a company to pivot so strongly after an IPO, but you must applaud the guts that this man had to do this.
Also another day, another tech rich family drama. Sam Altman is apparently pretty open and comfortable with slandering his sister, who has just filed a lawsuit against him and accusing him of sexual abuse. On the same month, we also got John Elkhan, chairman of Stellantis, CEO of investment fund Exor, who is fighting a tough legal battle for a 78M inheritance alongside his brothers and claiming that they all suffered vile abuse through the hands of their mother, Margherita Agnelli, daughter of the big industrialist Gianni Agnelli. Is it tech relevant? Well considering that these two men alone control a good portion of the market share of automotive and the artificial intelligence space respectively, I think it is? (It’s not, I’m just joking.)
Also, John Elkhan became Meta's latest board member. I still need to unpack this to be honest. I’m open to theories.
The only venture gossip that kept me alive during this holiday season was Bending Spoon’s attempt at acquiring another fast-growing company, Obsidian, with a witty email that was later shared on X by its CEO and founder, who did not appreciate the message, nor did the community. Bending Spoons, an Italian-based company that recently acquired Evernote and WeTransfer, is commonly seen as one of the very rare Italian unicorns (even though it pretty much operates as a private-equity firm). I felt awful for the poor associate who was publicly shamed for just doing her job. I’ve read way worse cold outreach emails, believe me, this was actually lukewarm.
I’m very invested in the story that is unfolding about Spotify’s not-so-little dirty habit of promoting fake music by fake artists, so they can avoid paying real artists once their tracks become increasingly popular on the streaming platform. Many reports have come out on the subject, it’s unknown to me if this fire was started by how disappointing the Spotify Wrapped was this year, considering that most of the investigations on the subject are dated to 2022. However, Harper’s Bazaar, laser focuses on one one thing that I think it’s worth understanding. How an obsessive search for profit has actually ended up breaking the one thing that Spotify was pioneering: discovering new music by knowing its users so well. I’m in awe at the sheer ingenuity of the person who woke up one day looking at stats about users listening to most playlists in the background, and decided that it was the situation for all of its 600M users and populated the whole Spotify library with bland music that was either AI-generated or produced by artists who needed money, rather than finding other ways to engage its users, or support actual?? good?? and original?? music?? It’s wild. I can’t wrap my head around it.
But I mean, if you are in this world, you know it’s all about getting to break-even, but seriously, at what cost? What are you losing in the path towards that? Where are your customers? What novelty are you bringing onto the table? Your business model seems nice on that pitch deck, but how does it work beyond your little scenario of today? This is what I was thinking about re: Spotify. And I want to end this batch with a reflection. Not mine, but by Professor Jeffrey Funk, who published an article on why startups keep losing so much money and how venture capitalists continue to be delusional about it in the latest Foreign Affairs issue. It’s the cold shower that we all need as we gear into a new year, and the only thing I want to remind myself as I read through a billion 2025 predictions/in&out.
That’s all, folks! Do you have any particular news or deals that picked your interest over this holiday break, or you were a good pal and didn’t open your emails for two weeks?
Happy return to work and see you very, very soon :)