Rembrand
Here's a company that has a real shot at getting big, but that I (Alex) worry about a bit. Remembrand uses generative AI to insert branded materials (images, etc) into videos. The gist is that online users hate ads, and view them as a pain in the backside. Also, ad blocking is only becoming more and more popular. So, Remembrand wants to integrate advertising into videos for brands. You can't skip it if it's native, right? On one hand, good that tech is solving a problem. On the other hand, more ads that are harder to avoid? Sounds like a future IPO. And gnashed teeth.
Unitary
Unitary is at the forefront of AI-powered content moderation, providing digital platforms with tools to manage online content more effectively. As regulatory scrutiny on social media and digital platforms intensifies, demand for Unitary’s solutions is expected to rise. However, the fast-moving AI landscape presents challenges, as emerging technologies could reshape the market and introduce new competition.
Music AI
Consumers love music, and love AI. And consumer AI music companies are doing well. But what about using AI for audio production work? Music AI's tech that allows for stem separation, mastering, and other critical and human-driven tasks today in audio production is super cool. Music AI has built consumer tech (Moises) and tools for pros. Combined with a fresh new $40 million round, and we're super bullish on the company's vision, and product mix.
Decart
Decart is innovating with real-time AI-generated video games, representing a leap forward in interactive content creation. With applications across gaming, education, and simulation, the company is tapping into a massive TAM. Decart’s rapid product velocity and early traction, showcased in a live demo on the show, position it as a potential financial and creative outlier.
Qodo
Code-generating startups and tools are a dime a dozen — but what Qodo wants to do, per TechCrunch, is “as much focused on generating code as on generating the tests and code quality in general.” Put another way the company has a focus on quality code, and deep integration into developer workflows. Killer!
HiddenLayer
There are many ways to make money in the current AI craze. Selling chips? Big business. Building AI models? Big business. Helping companies apply those models? Big business. HiddenLayer is a wager that AI security will be big business as well. Companies know that they need to work to keep ceratin data from AI contexts, but HiddenLayer reckons that the security risks of the AI era go far deeper. Given TWIST's general view that cybersecurity remains an industry far from being 'figured out,' and how much AI folks want to use, seems that HiddenLayer could be sitting on a future goldmine.
LangChain
LangChain offers tools to help companies build applications that use LLMs. It offers both open-source tools to help people 'chain' LLMs together (LangChain), and tools for building AI agents. It's working, too. THe company claimed that it 10x'd revenue in 2024, for example. While major model companies get a lot of press, the companies taking new AI tech and making it work inside of companies are doing quite well, too.
DevRev
AI For Devs
Customer support and success
The gist behind DevRev is that it wants to bring customer support and product development together. From a very high level, the pitch makes sense. You want to build what your customers need and want, so melding customer support (inputs) and development (outputs) is very reaonable. The proof will be in the pudding of course, but in terms of taking an AI-first approach to rethinking big planks of the modern software business cycle is a welcome effort. Famous founders and big-name backers don't hurt, either.
Rivos
Still under wraps after raising a quarter-billion-dollar Series A, Rivos is building chips and hardware using an open-source framework. The goal is “industry-leading power efficient, high performance, secure server solutions” using “workload-defined hardware.” More when it unveils its plans; for now, it’s richly-backed by big names and tackling a massive, massive market.
Etched
What if you made a chip that was designed just to run transformers, the technology that powers modern LLMs? That’s Etched’s goal: Beat the GPU gang at their own game by going specialized. A lot of money is backing the company. If transformers retain their present-day centrality to the AI game, Etched could be the next Nvidia.
Hyperbolic
Hyperbolic is another AI inference startup, but with a notable twist. It offers inference as a service and raw GPU compute. It is also working on a marketplace where folks with spare GPUs can rent them to others. If the concept scales, we could get off-peak GPU pricing into the market. Anything that makes AI cheaper is potentially massive, so we're bullish on the company (and its market, let's be honest).
Nscale
Neoclouds are the rage because they offer access to GPU clusters that customers might not be able to purchase the ingredients for on their own. Given compute demand for AI model training and inference, there's oodles of demand. And companies like CoreWeave are taking that demand to the bank. So too might Nscale, which is busy expanding its "pipeline of greenfield data centres across Europe and North America from 300MW to 1.3GW," it said when announcing a nine-figure round last year. The company bets that by being vertically integrated it can offer compelling service at a lower cost, and risk profile.
AI models? Hot. AI-powered coding tools? Hot. AI agents? Hot. What if we put all those together? It would look something like Cognition. Known for its ‘Devin’ AI software engineer, Cognition is a bet that AI is going to shake up how humans write computer code, for good. That’s, ahem, a lot of TAM.
Together AI
Building AI models is exceptionally competitive. As is the inference wars. While hyperscaled cloud platforms like what Microsft and Amazon offer are popular, smaller companies are also working on offering inference-as-a-service. Together AI is one such player, with claims of speed and low costs. It's growing quickly, and recently (reportedly) crossed the $100 million ARR threshold. Put Together on the IPO watch list.
Jina AI
Jina AI develops advanced AI models for search functionalities, allowing for more intuitive and efficient data retrieval across various applications. Their technology could have the potential to transform how information is accessed and utilized
Physical Intelligence
Today's AI models can solve huge math problems with ease. But, PI argues, getting an AI robot to fold "a shirt or cleaning up a table requires solving some of the most difficult engineering problems ever conceived." Its solution is to build 'general-purpose robot foundation models' to help bring artificial *physical* intelligence to the market. Backed by Khosla, OpenAI, Sequoia, Thrive, and Lux, a lot of folks are betting that PI can crack the issue. If it can, then we could be in for a future in which we never have to fold clothes again. Which sounds small, but is actually rather profound.
Humanoid robots with a general focus. The company’s robot, Apollo, is pitched by the company for manufacturing, logistics, and retail work environments. Apptronik says that its robot will help with labor crunches and reduce the pain of human employee turnover.
The MenteeBot claims to be an “AI-based robot that you can mentor.” The company mentions both household and commercial use cases for its bot, including notes that it can interpret multi-part instructions via voice.
Humanoid robots that are “general-purpose,” the company claims. Its Phoenix robot uses the company’s in-house “AI control system” called Carbon. Sanctuary, therefore, is a good example of the fusion of AI and hardware that modern robotics companies must master. It’s also Canadian!
We love ADUs, and companies that want to build more of them, more quickly. That puts Samara onto our list alongside Cover, and others.
What if a company shipped you a microhouse or ADU in a compact format that you could expand in an hour, and install in a day? That’s what BOXABL is offering, and it even raised a pile of equity capital from crowdfunding to build out its service.
Industrial-scale painting is not an artisanal task. That makes it a great place for robotics. And what PaintJet has built is just that, in spades.
Partiful
Breakout consumer social? In 2024? Yep. Despite limited capital formation via venture capital, Partiful has been seeing rising search demand, and just won Google’s app of the year award.
Bluesky
Bluesky is a neat company. We don't have many social media plays in the TWiST500, because there are only so many breakout players. Farcaster in crypto, sure, but what else fits the bill. Bluesky does, thanks to a massive ramp in its user base in recent months. Social media ownership is getting more political, which means that the market may suppport more total social plays. Bluesky is said to be fundraising as we add it to the TWiST500, so it seems that at least some VCs agree with our thesis here.
Unrivaled
Women's professional sport has seen a renaissance in the last few years, with the NWSL and WNBA putting up growing attendance and audience metrics. Now, Unrivaled wants to build on the momentum by launching a new, off-season women's basketball league. With tens of millions of dollars raised, it's a legitamate startup. And as anyone who has seen the worth of large professional leagues before knows, when they work, they work.
Kalshi
Kalshi was one of two companies that brought wagering on the world to the public consciousness during the 2024 American elections. Along with Polymarket, Kalshi found a new way to combine investing -- hedging, perhaps -- and gambling. The mix turned out to be very popular. A recent partnership with Robinhood means that the company's distribution is improving, and that it is finding new ways to engage after the conclusion of election season.
Farcaster’s distributed social network is notable for quickly signing up tens of thousands of users, and the company raising $150 million at a $1 billion post-money valuation. Jason pointed out that that company’s valuation/user metric is expensive; but both Jason and Alex agreed that this is a fun binary wager. Either Farcaster builds the next great social network, or it doesn’t. That puts the adventure in venture capital.
Monad Labs is not a fun new NFT project that is going to spam your X feed. Instead, it’s building “a blockchain protocol that streamlines Ethereum throughput via parallel transaction processing technique.” Or as we might gist that down, it’s working on making critical blockchains faster. Given that blockchains are often slow and expensive, it’s no small problem to tackle.
Stably
Stably offers a stablecoin service to help other companies get their own stable up and running. We have a general view here at TWiST that stablecoins are going to be huge. Ergo, the on-ramps are going to be both busy, and valuable.
Resistant AI
Cybersecurity
AI (vertical)
As financial systems digitize, fraudsters adapt. Resistant AI uses cutting-edge technology to safeguard against synthetic fraud and other digital threats. Their tools ensure financial integrity in an era of evolving challenges.
Witness AI
Cybersecurity
AI (vertical)
Some companies want to build AI models, tools, and features at a break-neck pace. Some companies, however, need to move a little bit more carefully. For those firms, Witness offers "the guardrails that make AI safe" to quote its own material. Given that we have a thesis that enterprise AI adoption will be a constantly upward tilting curve, companies like Witness could build huge businesses helping the enterprise get its AI shop in order. And with just $27.5M raised, Witness is no OpenAI or Figma in being fully sorted out. We're watching this one closely.
LynxTech
AI to combat fin fraudFinancial fraud is a trillion-dollar problem, and LynxTech’s AI solutions are here to fight back. With real-time fraud detection and adaptive learning systems, LynxTech is tackling one of the most pressing issues in the financial sector, making it a clear standout.
Castelion
Defense-tech startup, has raised $100 million to develop hypersonic missile systems intended for the U.S. military. The company plans to expand its manufacturing facilities in California and Texas and aims for its first hypersonic test flight this summer, with operational readiness expected by 2027
Alloy Automation
From Alex in the newsletter: "Another company I’ve had my eye on for a long time, Alloy started as a business promising ”complex automation made easy, and with no code.” Since then it’s developed an ecommerce focus, and most recently, moved upmarket. Often when we add a company to the TWiST500, we have revenue growth figures to benchmark against. Not so in this case, but having watched the company and its products mature, I’m taking a flyer on them here."
Aklymi
Enterprise data management
AI (vertical)
Turning unstructured data into actionable insights is a massive challenge for finance professionals. Alkymi’s vertical AI solution makes it easy to extract, analyze, and act on data, setting them apart in a rapidly growing niche.
Enterprise SaaS
Construction
A great example of vertical SaaS for an industry, PermitFlow is taking on a key sticking point in construction — permitting — and turning it into tuned software.
Synctera
Banking-as-a-Service has seen its share of challenges, but Synctera is proving there’s room for innovation. By connecting banks with fintechs, Synctera simplifies integration and empowers financial innovation, making it a standout in the sector.
Workpay
Workpay recently raised its largest round to date, some $5 million. But what’s more impressive is what it has managed with a very modest capital base. The African-focused HR tech company now has over 1,000 customers in 20 countries, and is on track to double in size this year. A few more of those and we could be looking at another African tech IPO.
While many of the constructiontech startups that we picked for the TWiST500 are using robots to build, Cuby has built robots that build. Micro factories are the name of the game here, which are highly-automated to boot. This is a startup with a big vision making big bets.
Xona Space Systems
There are a number of GPS-like systems in the world. Four or five, depending on how you want to count. And they are owned by nation-states or blocs. Enter Xona, which wants to build one that is super-accurate and privately owned. It will feature 258 satellites when complete, and is one hell of a vision. We suspect that the company will require more capital, so keep an eye out for that.
OrbitFab
Imagine you find yourself in orbit, but your tank is running low. What do you do? Look for a filling station, of course. That's what OrbitFab hopes, at least. The company is building in-space refueling tech for orbital devices. This includes building a phsyical interface to accept in-orbit fuel, and softawre to coordinate the satellites in question. Given that we expect the in-orbit economy to greatly expand in the coming decades, OrbitFab's TAM should only expand and expand.
Turion Space
Want to launch a satellite but don't want to design it yourself, or build the software required to run it? Great news, Turion Space wants to do all that hard stuff for you. The company is a hybrid bet on hard space tech (satellites, moving them about) and soft space tech (robust software to manage in-orbit operations.) The company is doing so well that it landed a $32.6 million contract with the DoD for "multi-payload satellites and real-time command and control." Let's see how those work out.
Starfish space
SATELLITE ECONOMY, Y'ALL. Starfish Space is innovating in the satellite economy by developing technology for in-orbit servicing and debris management. With the growing reliance on satellite infrastructure for communication, navigation, and observation, Starfish is tackling critical challenges in orbital sustainability. Its strategic focus on foundational satellite service could signal long-term relevance in the space industry.
Digantara
Digantara is not a household space name. But the company's space material tracking technology has a number of interesting applications -- defense, insurance, commercial, etc -- and it has joined the United States Space Forces’s Prime Fusion (Space RCO) Pilot Accelerator along with Andruil. While launch technologies (SapceX, Blue Origin, etc) get a lot of press, keeping things safe once we get them to space is a big deal as well.
Whisper Aero
Whisper Aero is advancing quiet propulsion technology and possibly redefine noise standards and coulld be big for eVTOLs, drones, and aviation. With increasing demand for sustainable and efficient aerospace solutions, Whisper Aero’s unique approach addresses both environmental concerns and operational efficiency. Early market adoption and rising partnerships signal potential for rapid growth and a significant long-term impact