Crypto is booming again. Bitcoin is near record highs, Walmart and Amazon are reportedly exploring stablecoins, Robinhood is tokenizing shares of public and private companies, and NFTsonce left for deadare stirring back to life. Even crypto-powered network states are inching toward reality.
But one star from the last bull run hasnt joined the rally: the metaverse.
Back in 2020 and 2021, the metaverse was the tech industrys favorite toyan immersive digital frontier where wed work, play, and shop. Facebook rebranded to Meta, VR headsets flew off shelves, and internet searches for metaverse jumped 7,200% in a single year. JPMorgan called it a $1 trillion opportunity that would likely infiltrate every sector.
Decentraland was among the breakout stars of the metaverse, a bustling virtual city with casinos, concerts, branded events from Dolce & Gabbana to Nestlé, and even a JPMorgan lounge.
But now, in 2025, the future of the internet looks far from the Ready Player One-esque revolution once promised.
Once-booming world hits record low
Built on the Ethereum blockchain and powered by its own cryptocurrency (dubbed MANA), Decentraland allowed users to buy and sell land as non-fungible tokens and customize avatars with tokenized wearables.
Blockchains, NFTs, and Web3 overall are similar,” says Matthew Ball, CEO of the venture capital firm Epyllion. “There are those who believe these technologies and/or systems will be needed to build the metaverse, others who say its the best way to do so, and some who believe that these technologies are irrelevant to the metaverse and in general.
While some saw Decentraland as a financial opportunity, most used it to socialize, attending concerts by artists like Grimes, Ozzy Osbourne, and Björk, or gambling in a virtual casino that reportedly generated $7.5 million in just three months.
But last year, reports surfaced claiming the once-thriving world had become a ghost town, with fewer than 50 active users during a 24-hour period.
Other blockchain-based worlds have also struggled to maintain engagement. The Sandbox, another decentralized platform, has attracted just 5.7 million usersin totalsince its founding in 2011.
Decentraland refuted the low usage claims, asserting that it had 8,000 daily active users instead.
Still, user sentiment paints a complicated picture. One Reddit user shared their early excitement but noted they “lost interest for a number of reasons,” citing poor user experience, low engagement, and abandoned projects. I wanted it to work, perhaps it still could, but not without an enormous overhaul to the overall design and leadership approach.
Others had similar experiences: Some made money and left, while others simply moved on. Made a shit ton of money from it,” one Reddit user said, “but when crypto winter came in 2021, that shit died. Another added: Used to log in during COVID and play in the casino for fun. There was never anyone around even then.
Crypto-based worlds taking the back seat
When asked for more recent figures, the Decentraland Foundationthe nonprofit overseeing the open-source platformreported 2.3 million unique visitors to platform properties and 24 million all-time unique visitors.
At Decentraland, we believe success in a virtual world isnt measured by how many people log in every day, Decentraland Foundation marketing chief Kim Currier tells Fast Company. Daily active users is a metric borrowed from social media and gaming platforms that are designed to keep people endlessly scrolling or grinding. Thats never been our goal.
Instead, the foundation emphasizes engagement. More than 274,000 friendships accepted, 19 million chat messages exchanged, and over 616 million Emotes [actions for your avatar like dance moves or a wave] expressed to date, Currier shared. These are signs of real presence and connection.
To improve user experience, the foundation has continued to build. In October, it made a major shift from a browser-based platform to a downloadable desktop version. It also removed the requirement for Web3 wallets, aiming to make the platform more accessible.
According to the foundation, the desktop version has been downloaded over 196,000 times. It added that third-party usage trackers can no longer reliably measure activity. Any numbers they share today are outdated and incomplete, a foundation spokesperson tells Fast Company.
Changing expectations
While blockchain-based metaverse platforms havent taken off as once expected, others have succeeded in offering immersive digital experiences without relying on VR or cryptocurrency.
Theyre thinking, like, way further in the future, Currier says. The misconception is that the metaverse is something where you put a headset on and youre fully immersed in a world that is photorealistic and completely different from your physical experience.
Many see the metaverse as VR, because when Zuckerberg renamed Facebook, their sole produt in market was a VR headset, Ball notes. Still, he adds, the technology isnt advanced enough yet, and its not essential for the metaverse to exist.
I think now we are in a place where people have more realistic expectations about whats physically possible and technologically possible, Currier says.
Todays metaverse may look more like platforms such as Roblox or Minecraftvast, user-driven spaces where social interaction is central.
The clearest market leader is Roblox, a virtual world platformand one which described itself as a metaverse company long before Facebook or Microsoft ever declared for the same ambitions, Ball says. He points out that during Roblox Corp.s IPO, the word metaverse appeared in its SEC filing 16 timesmore than it ever had before.
As the dust settles on early hype, the concept of the metaverse is also shifting in how its meant to fit into peoples lives.
At Decentraland, we really, firmly believe its not meant to replace your real life, Currier says. People are spending a little bit more time in the physical world with their real families, their friends, and thats okay.
Appliance-maker SharkNinja has a reputation for creating smart, viral appliances, from a frozen slushy-maker to an LED cryo-mask to an indoor-outdoor fan with an ingenious cooling mist attachment. Key to SharkNinjas success is its ability to create both ultra-functional products (vacuums, air fryers, blenders) and ones that take consumers by surprise, especially on social media. And despite the challenges posed by President Trumps ever-evolving trade deals, the company continues to grow: Net sales were up 16% year-over-year in Q2 2025.
I spoke to SharkNinja CEO Mark Barrocas about where his teams find the company’s next big idea, how quickly it can move from idea to being in stores, and which category SharkNinja’s expanding into in September.
SharkNinja has been expanding its product portfolio. The company went from selling blenders and vacuums to the beauty category, launching hair stylers and more recently an LED face mask. How do you decide which categories to enter?
Our innovation comes from identifying known or unknown consumer problems. You could be a consumer products company and you could build your product roadmap off of a core technology. If you do that, it will lead you only to the places that core technology is applicable to.
We think there’s an endless number of consumer problems to be able to solve, which is why over the last 17 years we’ve gone into 37 different product categories. On any given day we’re focused on making robots, on skincare, on outdoor cooking, on haircare, on cleaning, on air purification, on fans.
[Photo: SharkNinja]
Does that mean you have customers coming to you and saying, I want something to curl my hair and I also want a slushy machine?
We have a team of consumer insights researchers that are constantly looking at online reviews, negative sentiments, social media sentiment. We have people that are in consumer homes every day, they’re in restaurants, they’re in commercial environments. And we’re constantly mining the next problems to be able to solve.
On any given quarter, we will run things called hack weeks where we’ll set up a team of 8 to 10 people and we’ll have them go and hack on a particular problem or idea that they have. That could range from the consumers doing something outside the home that they’re not able to do inside of the home. The consumer used to have to go to a med spaso we developed an LED infrared cryo mask for them to be able to do that at home.
SharkNinja’s planning on introducing 25 new products this year. How long does it take between having an idea for a product and putting it on the market?
We launch 25 new products a year. Thats not a new color or a new knob. That’s a new, ground-up, from-scratch product. We’ll start with a pipeline of anywhere from 50 to 60 ideas. Ideas will eventually weed themselves out because the technology is something we might not have right now, the product might be too expensive, or the consumer might not be ready for it.
We just launched our first FDA-approved product, the Shark Cryo Globe mask. That took two and a half years for us to develop. We take anywhere from 12 to 16 months to bring technology we already have to market. If it’s a completely new technology, we think that two and a half years is about the time for us to be able to do that.
Whats an example of a product that consumers werent ready for?
In New York, they passed a law around composting. That’s a new thing for the majority of New Yorkers, and it’s not something that exists really in lots of other areas of the country today. We think that as consumers start composting, there’s going to be some real challenges. They’re sitting there with a plastic little bin full of fruit flies and dealing with the smell of six days worth of food scraps.
Our team got really motivated and excited about wanting to solve that problem. The challenge is that the problem is not yet well understood by the consumer. They’re going to need to compost for a year. It’s going to have to get beyond just New York. It might be something that is a great idea, but we might be too early on the cycle for it right now.
Maybe as we get into 2026 and 2027, we pull it back out of the roadmap and say, now is the time where the consumer will really be able to listen to this story.
SharkNinja CEO Mark Barrocas [Photo: SharkNinja]
What’s an example of a product that you wanted to make, but the tech isn’t there?
We believe there are lots of unmet needs in the lawn care space, but from a technology standpoint right now, we don’t have a solution that the world needs us for. Its important to recognize, when we go into a category, that we have to answer the question: Why does the world need us? Why does the consumer need us? Are we doing it cheaper? Are we bringing something to the consumer that they’re not able to get anywhere else?
If we can’t answer that question, then it’s not a category for us to expand into.
Is the same team of engineers coming up with all of these products across categories?
We have 1,300 engineers around the globe. They’re based in Boston, London, and Asia. And I think one of the exciting things about being an engineer and working at SharkNinja is that you can find your next job at SharkNinja. If you’re an engineer that’s working on haircare andyou’re interested in thinking about outdoor cooking, you can go on the outdoor cooking team, you can switch to the robotics team, you can switch to the slushy team or the air fryer team or the air purifier team at SharkNinja.
We’re constantly cross-pollinating ideas across the global engineering teams. There’s mechanical engineers, electrical engineers, software engineers, people working on IoT (internet of things). Were bringing all of those pieces together to bring as much technology as possible into a product that could sell for on average for $199 to $239.
How have you been affected by tariffs?
That process started for us way back in the first Trump administration. When tariffs went into effect five years ago, 35% of our products were tariffed at that time coming into the United States at 25%. Prior to that, 100% of our production was made in China. Four and a half years ago, we started making our first product outside of China, and by the end of Q2, we moved primarily to Southeast Asia, Vietnam, Thailand, Cambodia, Indonesia, Malaysia.
There really isn’t any necessarily safe haven other than manufacturing the products in the United States, but four and a half years ago is when we started making our first product outside of China. By the end of this month, we’ll be able to make nearly 90% of our production outside of China. By the end of the year, we will have nearly 100% of our production outside of China.
So has it had an impact? Of course, it’s had an impact, but I think that we’ve got a highly diversified supply chain. I think we’ve developed a really high-quality, fast-turn, low-cost supply chain that’s been a real competitive advantage for the business.
Have you passed any costs onto consumers?
We have had to pass some costs onto consumers. We’ve tried everything possible to keep that to an absolute minimum.
So much of the way we think about our business is what we call affordable, accessible innovation. We really want to make products for almost everyone. You could buy a Ninja product for $49; you could buy one for $999. You could buy a Shark product for $59 or buy one for $899.
What’s so core to the Shark or Ninja brands is you could be a Walmart shopper and be a SharkNinja consumer. You could be a Sephora shopper and be a SharkNinja consumer. I don’t think there’s another brand out there that has as large a socioeconomic group of consumers as us.
At least on TikTok, there’s a widespread perception the Shark hair products are an affordable dupe for Dyson products. I actually think that’s probably helped you sell quite a few units. How do you respond to those comments?
In every industry, a brand has to find their white space in the market. For us, that white space is market leading performance, great quality, and great value. I think that the products that we bring to market are much more versatile.
You bring up our haircare. We looked at products in the market and we saw that products were really single-use products. You had to buy a hairdryer, you had to buy a styler, you had to buy a brush. We don’t think the consumer wants to have four different products to be able to do their hair. We saw that unmet need and we developed a product called the Shark FlexStyle.
If you have a great experience with one of our products, you’re willing to try us as you go into [a different] category. I think that really helped us get into the beauty business. [Consumers thought] they had a great experience with a Shark vacuum or a Shark air purifier or a Shark steam mop. Let me now try them in haircare.
You have products across categories, price ranges, and retailers. How do you market them?
We sell our products in every major retailer in every major market. We’re one of the most searched brands on Amazon. We have a robust direct-to-consumer business. We want to be relevant wherever the consumer chooses to shop for our products.
It’s our job to create consumer demand through viral marketing, and ultimately, it’s up to the consumer to decide where they want to shop for the product, whether they want to shop at a brick-and-mortar store or online or they want to go direct-to-consumer. We just want to be relevant wherever they choose to shop.
On the marketing side, I think what’s so interesting is we were a company 16 years ago that only marketed [via] long-form infomercials. I mean, my partner and I at the time didn’t have any money. We would run 30-minute TV infomercials. Fast forward to today and we have seven times more social media engagement than our nearest competitor. Our products not only go viral on social media, but they generate a tremendous amount of user-generated content.
What’s the next category you want to get into?
In September we’ll be launching a new outdoor category. Weve publicly stated that we feel like we can enter two new product categories a year as we move forward.
Morgan Stanley projects the space economy will hit $1.8 trillion by 2035. Yet most companies still dont have a strategy for it. Last quarter alone, multiple space startups secured seven-figure funding rounds. NASAs Artemis program also hit a major milestone. This is a very real, trillion-dollar economy forming in real time, and the window to get in early is closing fast.
Critically, space is the first truly infinite domain of commerce. Unlike Earth-based markets, it offers endless potential for new infrastructure, new services, and new economies to emerge. And while many are distracted by AI or supply chain chaos, the next massive growth platform is quietly taking shape above our heads.
Its not just rockets and rich guys. Its agriculture, R&D, and your next competitor
When most people think of space, they picture billionaires in zero gravity or cinematic sci-fi. But the reality is that its becoming a critical infrastructure layer for future business, just as the internet did 30 years ago. The parallels are striking. In the early 1990s, many leaders saw the internet as a novelty. Today, its the backbone of the global economy. Space is on the same path, but with even broader implications.
From biotech to agriculture to cybersecurity, space is already critically transforming industries. Biotech companies like Redwire and Varda Space Industries are leveraging microgravity for drug development and bioprinting, enabling breakthroughs not possible in Earth-bound labs.
Agriculture firms are using real-time satellite imagery to optimize water use, detect crop stress, and improve yield forecasting. Platforms like Planet Labs and Descartes Labs are making precision agriculture scalable and climate-resilient.
Cybersecurity providers have also been looking beyond Earth for years, as satellite networks become part of critical infrastructure. Companies like SpiderOak are pioneering zero-trust security models for space assets.
The market is heating up and capital is flowing in
In 2023 investors poured $12.5 billion into space startups globally, despite broader tech market pullbacks. Startups focused on in-space manufacturing, small satellite constellations, and launch technologies are leading the charge. The Artemis program is unlocking new lunar and deep space opportunities, while commercial players like SpaceX and Rocket Lab are slashing launch costs.
Public-private partnerships are expanding rapidly. The Department of Defense is investing in space-based logistics and mobility. NASA is funding space-based solar power and commercial space stations. Private equity firms are acquiring ground infrastructure and launch supply chains. The smart money is building and pivoting, quickly.
Four ways to make space part of your growth strategy
Meanwhile, too many companies and policymakers remain tethered to earthbound thinking. The point of tapping into this market isnt to become a space company. Most companies wont build satellites or spacecraft. Instead, theyll find new ways to leverage the unique conditions and infrastructure of space to improve their products, services, and operations on Earth. Or theyll take existing products and services and fid ways to adapt them for space.
Heres how to ensure youre positioned to lead in this rapidly expanding trillion-dollar market:
Start thinking like a space vertical. You dont have to build rockets to benefit from space, but you do need to understand how the industry works. Study space value chains and learn where your products, services, or capabilities might fit in. Even surprising players can break into the market. For example, a small watchmaker in Albuquerque was tapped to build components for space-bound hardware. This wasnt because they were in aerospace, but because they solved a unique precision problem.
Tap into satellite-enabled insights to optimize operations. If youre not already, look into ways your business can uniquely leverage satellite data for actionable insights, whether its tracking supply chain bottlenecks, improving precision in agriculture, or identifying untapped markets with geospatial analysis.
Codevelop with space startups tackling niche challenges. Consider partnering with startups innovating in microgravity manufacturing, on-orbit servicing, or space-based energy solutions. Explore shared R&D that aligns with your industrys specific needs, like 3D-printed components or novel materials.
Train your teams on space-driven opportunities. Upskill your workforce by collaborating with universities or space-focused research institutions. Equip your employees to identify how advancements like quantum communications or hypersonic transport could create revenue streams in your sector.
The cost of waiting? Irrelevance.
Its time to adapt, innovate, and lead. The companies that embrace space as a critical business opportunity will not only future-proof themselves but also define the next chapter of economic history. Every major technological revolution has created winners and losers. Space will be no different.
On a hot, oppressively muggy summer day in a city like New York or Atlanta, when you crank up the AC, it might not feel like its working well. Thats because conventional air conditioners arent optimized to deal with humidity. Your AC will run longer as it tries to deal with both heat and moisture in the airand if the humidity stays too high, your home can feel clammy or sticky even if the temperature is dropping. Because humidity makes the air feel hotter, you might not feel much cooler even as your electric bill climbs.
But what if you could save 90% of the cost of your air-conditioning electric billsand actually be cool during a sweltering summer? That’s the promise of a new kind of AC technology that deals with humidity more effectively; its just coming out of testing and into commercial development. Though the technology exists, you’re going to have to wait (but not too long) before you can have it in your home.
The innovations, from a handful of startups and larger companies, can save huge amounts of energy and provide more effective temperature control. As the planet gets both hotter and more humid, new tech can help more people afford to stay cool. It also can help the grid so blackouts are less likely in a heat wave. And with less energy use, it can help tackle the cooling paradox: the fact that the growth of conventional air-conditioning is a major source of emissions, forcing us to rely on ACs even more.
The problem with legacy AC
New technologies take different approaches to solving the same challenge. The air-conditioning problem really is a humidity control problem, says Russ Wilcox, CEO of Trellis Air, an air-conditioning startup spinning out tech that was originally developed by Harvard researchers.
Standard air conditioners remove humidity and cool the air at the same time. When hot, wet air passes over cold coils inside the machine, it condenses, like beads of water on a cold drink. But because the system’s main goal is to cool, on a very humid day, you need to turn the temperature way down to remove enough humidity to try to feel comfortable.
The AC has to run longer, guzzling more energy. Rooms can end up either cool and clammy or too cold. In some large spaces, like a movie theater, overcooling sometimes means that the heat comes back on, despite the fact that its sweltering outside.
There are more than a billion air conditioners in use now, responsible for a carbon footprint thats around twice as large as that of aviation, and around a third of the electricity they use is for dehumidification. Emissions are also quickly growing as more people buy air conditioners. By 2050, the number of units in use around the world is expected to triple, causing emissions from air-conditioning to potentially double to 2 billion tons of CO2 per year.
How the new tech works
Cutting-edge AC tech deals with humidity separately from temperature. Trellis, for example, uses a membrane to filter water vapor out of the air before cooling it, an approach that is far more efficient than a typical air conditioner that expends energy cooling both the air and the water inside it.
That gives us a huge edge in energy for dehumidification, Wilcox says. And we do it with an engineered plastic film, which means its a pretty passive, simple, reliable, potentially very cost-effective way of dehumidifying.
Blue Frontier, a startup that has raised more than $36 million from investors including Bill Gates’s Breakthrough Energy Ventures, uses a salt-based desiccant to store energy that can later be used when electricity prices are high. During peak hours, the system uses the desiccant to remove humidity. (It’s like a battery, but instead of storing electricity it stores drying power.) The technology can reduce electricity use by 50% to 90%.
Transaera, which raised an $8.2 million round of seed funding in November 2024, uses a type of material called a metal-organic framework (or MOF) with a microscopic tinker-toy-like shape.
MOFs are “really powerful because they allow us to target a specific moleculeyou make the pores just the right size for that molecule to go in, says Ross Bonner, cofounder and CTO of the Massachusetts-based startup. In our case, we have tuned them for water. Transaera uses the material to coat a substrate, and then can add it to a standard air conditioner. Depending on the climate, it can cut energy use by around 40%.
AirJoule, another startup, uses metal-organic frameworks along with waste heat to efficiently dehumidify and cool the air (and to produce pure water that can be used elsewhere). Data centers are target customers. Industry veteran Carrier has partnered with the startup to incorporate the tech into its own equipment.
Two large AC manufacturers, Chinas Gree and Japans Daikin, have developed super-efficient air conditioners that use different sensors and controls for humidity and temperature.
Its really smart design and smart controls, and the ability to sense and respond to real-time conditions, that enables them, says Ankit Kalanki, who works on the carbon-free buildings program at the nonprofit RMI. The designs from Gree and Daikin also use the most efficient components possible, from heat exchangers to compressors. Both companies won the Global Cooling Prize, a contest that launched in 2018 to encourage innovation in air conditioner design.
Proven tech
Over the past few years, the companies have been proving that the technology works. RMI recently partnered with Gree and Daikin to test their units in real-world conditions in India. They rented seven apartments in a city outside Mumbai and pitted the new designs against the most efficient ACs and mini-splits currently on the Indian market, looking at how much power it took to stay below 80 degrees Fahrenheit and 60% humidity.
Earlier this year, after nine months of testing under different weather conditions, they published the results: The new tech cut energy use by 60%.
Transaera began testing a prototype of its tech on a large commercial building in Houston last year. Our approach was, okay, we have this technology, weve proven it out in the lab, we want to put it through its paces and really see if it can perform and do what we say it can do, says Bonner. So we found the most punishing climate that weve been in.
Last summer, when they went on the roof of the building for the installation, they measured the surface temperature: 140 degrees. It was so hot that the installers had to wear knee pads so they didnt burn themselves. After months of testing in Houstons ultra-humid weather, where a typical summer day might have a heat index of 110 degrees, the AC has been saving even more energy than projected. Now Transaera is working with a manufacturer to make a full-size prototype for testing.
The path to market
If you need a new window air conditioner, you can’t yet go to the store and buy one of the new designs. So far, the first product to come to market is a commercial one. Blue Frontier launched a 15-ton “dedicated outdoor air system” (or DOAS) unit earlier this year.
Selling first to commercial customersfrom medical centers to schools to restaurantsmans that the company can have the biggest climate benefit with each unit it sells. “The conventional technology DOAS are the ‘gas guzzlers’ of the industry,” says Daniel Betts, founder and CEO of Blue Frontier. The standard tech of this type is very inefficient and energy-intensive. Blue Frontier’s version also offers energy storage so the units can run for four to six hours with little electricity use; that lets building owners make better use of renewable electricity and lower electric rates at certain times of day.
The technology can also be used in smaller residential units, but that will come later. “It’s just a matter of picking a market entry strategy that makes sense to us and that helps our community the most,” Betts says.
Gree and Daikin, the Global Cooling Prize winners, initially aimed to bring products to market in 2025. Their technology is ready, says RMIs Kalanki, who is working with the companies on commercialization. But it isnt likely that the ACs will be in stores this year. From a technical feasibility standpoint, I think that has been proven through the testing, he says. So its more about commercial viability now.
One challenge, Kalanki says, is that international standards for residential air conditioners dont yet measure the energy used to remove humidity. Were working very closely with the international standards organization to really bring dehumidification into the conversation, he says. This needs to get reflected so we can reward the products in the right way and industry has the right target to design for.
Though customers will be able to save money over time on electric bills, the up-front cost of the units will be higher, making them a little more challenging to sell. Institutional buyers, who purchase in bulk, could help jump-start the market, Kalanki says, noting, “That demand signal is going to be very critical for manufacturers to make those early investments.
More commercial options are likely to be available sooner. AirJoule plans to be on the market next year. Transaera is now working with a manufacturer that will be able to produce its commercial units at scale after the current pilots end, and is already in conversations with customers. The technology has the biggest advantage against conventional products in a commercial application, Bonner says, but the company also plans to later make residential ACs. (It’s already made a viable prototype.)
Trellis, which launched last year, is at an earlier stage and hasn’t yet started testing prototypes. The process will take time. “I think we have a lot of ambition of how we can manufacture this cost-effectively in the supply chain and make a robust product,” says Wilcox, noting that the team previously worked together on the development of the screen for the Amazon Kindle. “But we also appreciate that it takes some years to really make something robust enough to ship all around the world.”
The startups recognize the urgency of their work, as the need for ACs and their impact continues to grow. “I’m always impatient,” says Bonner. “We can go faster, and we need to go faster. And we have a responsibility to future generations to make the difference that we know we need to.”
When it comes to designing a safer football helmet, Jason Neubauer knows what he’s up against. “You can make a very safe helmet that ranks No. 1 for performance,” he says. “But if the players don’t like the way they look in it, it really doesn’t matter. You’re not going to protect anyone.”
What Neubauer and his team at Schutt Sports have built with the F7 Pro could be one of the safest helmets ever produced. It’s certainly one that players are gravitating toward.
Since the F7 Pro launched this spring, it has become the fastest-adopted helmet in NFL history. While Aaron Rodgers may not be a fan, an All-Pro roster that includes Justin Jefferson, Ja’Marr Chase, CeeDee Lamb, and Travis Hunter (the two-way star taken No. 2 overall in this year’s draft) will kick off the 2025 season donning the sleek F7 Pro, which earned a top-5 ranking in the NFL and NFL Players Associations rigorous 2025 helmet performance testing.
New helmet models don’t really peak in their adoption rate until about the third year, Neubauer says. It takes a while for players to get comfortable with a new look.
Neubauer has spent more than 25 years developing sporting goods, focusing on extreme sports before shifting to football helmet design in 2016. Hes one of the masterminds behind the F7 Pro, whose innovations fundamentally reimagine how helmets protect against the defining threats of modern footballthe high-speed collisions that make highlight reels, and the thousands of smaller impacts that accumulate over a career.
[Photo: Schutt Sports]
What the NFL does right
The NFL has caught plenty of flak forwell, almost everything. But its historical approach to player safety is high on the list. Over the past decade, however, the league has done what any smart company does in the digital age: dive into data.
To better understand how and when head and neck injuries occur, the NFL compiles detailed reports on every head and neck injury sustained in practice or games. The league tracks actual on-field impacts using sensors and cameras, documenting the speed, location, and type of every hit. This information is then shared with helmet manufacturers so engineers can discern what they’re building to protect against.
This data-driven approach is working. Preseason concussions dropped by more than half from 2017 to 2024, from 91 to just 44.
This granular data informed the F7 Pro’s overall design philosophy. The information showed not just where and how hits occur, but also the timing, force distribution, and frequency patterns that traditional helmet design hadnt accounted for. Armed with this data, Neubauer’s team could optimize protection at a fundamental levelrethinking everything from materials to architecture rather than just adding more padding.
[Photo: Schutt Sports]
F7 Pro innovation: 3D-printed lattice
Traditional football helmets work like old-school steel car bumperssolid structures that conduct the full force of impact, transferring it to the passenger. The F7 Pro works more like a modern car bumper. Its outer shell uses a custom material blend designed to flex under impact, while a layer of 3D-printed lattice beneath the shell does the real work.
Research from Tulane University’s physics department found that a large defensive lineman hitting a quarterback generates impact forces equivalent to a car hitting a brick wall at 18 to 20 mph.
You’ve got two guys who are 220 pounds running at a very fast rate and hitting each other, Neubauer says. You can’t get rid of that energy, so you need to slow it down to the slowest rate you can to minimize the forces on the brain.
The F7 Pros lattice does exactly that with a network of microscopic shock absorbers, each smaller than a pencil tip, all working together to distribute an impact across thousands of tiny columns that buckle and bend in controlled sequences.
“The physical nature of buckling and bending is what’s slowing the impact down, so you don’t feel all the force at once, but you feel it over that offset distance, Neubauer explains.
Instead of each hit unfolding as one blunt force, the impact is more like a controlled demolition.
[Photo: Schutt Sports]
F7 Pro appeal: Lighter, sleeker, safer
By using 3D printing to integrate various functional elements into a unified design, the lattice eliminated eight separate plastic components that traditional helmets required. The result is a seamless design that is lighter on players’ necks while enhancing protection, something traditional manufacturing couldn’t achieve.
Schutt developed its lattice technology in-house rather than licensing existing solutions. Out there in the world right now, there are quite a few different lattice technologies that companies could choose from, Neubauer says. It’s literally like a drop-down menu. That would have been a lot easier for us to do. But we found that we were able to get a better-performing, lighter-weight result by doing it ourselves.
Weight reduction is critical because players’ heads and necks endure thousands of impacts over a season, and every ounce of helmet weight adds to fatigue and long-term neck strain. But it also allows for a sleeker profile, addressing something equally important: The helmet looks damn good.
Players have to want to wear it, and when stars like Jefferson and Chase sport the low-profile design in prime time, other players notice. It’s functional vanity at its finestsafety technology that doesn’t make you look like you’re wearing a fishbowl.
[Photo: Schutt Sports]
A new era of customization
In 2021, helmet manufacturer Viciswhich had been acquired by Certor Sports, Schutt’s parent companyintroduced the first position-specific helmet, the Trench model, designed for linemen. It focuses on protecting against the thousands of smaller hits that accumulate from play after play in the trenches. Schutt followed up with quarterback models that prioritize back-of-head protection because quarterbacks are frequently slammed to the turf when sacked and can’t brace themselves.
The F7 Pro’s variants optimize protection based on real impact data. And as the data gets more intuitive, new position-specific helmets will likely enter the market, with Schutt and Vicis leading the way.
Its OctoFit system lets players customize foam pod combinations based on their unique head shapes. So a process that once required custom ordering and waiting for delivery now occurs in the locker room in real time.
Its AiR-Lock system is activated by a small push button located on the back of the helmet. Remember the old Reebok Pumps? The AiR-Lock is similar. Players can pump their helmets up for a tighter game fit, then release pressure to be more comfortable in practice or during walk-throughs, adjusting helmet security without using tools or having to leave the field.
This combination of position-specific protection with real-time fit adjustment represents where helmet design is heading: equipment that adapts to how players get hit based on how they experience the game, while catering to their individual comfort preferences.
The future of protection
Virginia Tech has an independent helmet testing lab that serves as the industry’s safety standard, evaluating helmets and assigning star ratings that guide consumers. When the university updated its protocols in July 2025, 77% of helmets that previously received five-star ratings were downgraded (from 26 to just 6), signaling that safety standards are evolving at every level.
And as the NFL helmets evolve, high school and youth gear will follow. Schutt is partnering with national youth and varsity organizations to gather impact data similar to what the NFL provides, studying how younger players get hit and what protection works best for developing bodies.
“The types of impacts that kids aged 8 take are very different from an NFL athlete,” Jeremy Erspamer, CEO of Certor Sports, says. “And we as helmet manufacturers need to understand that and develop technologies that specifically keep players at each level safe.”
Schutt is set to launch a new youth helmet this fall, according to Erspamer, which will also be five-star rated.
The number of concussions in the NFL decreased 17% from 2023 to 2024, reaching a historic low last season, while preseason concussions fell more than 50% from 2017 to 2024. The F7 promises to continue that momentum in 2025 toward a safer game for players at all levels.
We believe it’s the best helmet out there at the elite level, Erspamer says. But what we also know is that in three years, we’re going to have even better technology. So we’re excited about where we are, but we’re even more excited about where we continue to go.
After years of chasing user growth, Bumble founder and CEO Whitney Wolfe Herd now wants low-quality users off her service. Soon, people with grainy profile pictures or lacking a bio may be forced to leave the app unless they improve.
Our product is people. The quality of someone’s experience, how they engage, find what they’re looking for, and monetize depends on the quality of who and what they encounter on the platform, Wolfe Herd explained on the companys earnings call on August 6. She outlined a strategy to create a healthier app ecosystem that involves categorizing users based on the quality of their profiles and pushing out people who are degrading the experience of Bumble.
The almost 10-year-old-app has struggled in recent years to regain the momentum it had during the pandemic. In March, Wolfe Herd stepped back into the role as CEO, after ceding it to former Slack executive Lidiane Jones just over a year earlier. Wolfe Herd is now overhauling the app, reorienting it away from a growth-at-all-costs strategy and toward one that prioritizes a higher-quality user base and real connections. As part of that, Bumble is starting to separate the wheat from the chaff.
Approve, Improve, and Remove
On the earnings call, Wolfe Herd said the app will use AI and human moderators to sort users into three categories: Approve, Improve, and Remove. Approve users, according to Wolfe Herd, are ones who have adequately filled out their profiles, complete with multiple photos, and offer a clear picture of who they are. Improve users have incomplete profiles that could be refined to get more matches.
Remove users are bots, scammers, people with multiple profiles, and those who have violated user agreement terms. Theyre also people whose profiles are inadequately filled out and who refuse to improve them. Roughly 10% of Bumble users fall into the Remove category, while the majority fall under the Improve category.
“They’re not bad people,” Wolfe Herd said of the Improve members. “They’re not nefarious members. They have no clue how to build a profile. These could be extraordinary people that when you meet them in real life you’re like, How are you single? But when you look at their Bumble profile, they have one photo, they’re wearing a ski mask, and they have no bio. There’s no chance for them on our product in that construct.”
Bumble is launching new features to help. The company recently introduced a dating app concierge that offers AI and human advice to help users fill out their profiles. This month, it plans to unveil an AI-powered coaching hub that will give users tips and actionable steps to make their profile more attractive and engaging. If they follow those tips, they may end up in the Approve category. If not, they could get the boot. (The app is also improving its identity verification tools to get bad actors off the platform.)
A return to quality
Fewer better is always going to win when it comes to connection and relationships, Wolfe Herd explained while laying out the strategy. If you were to swipe through 100 people [who] you never wanted to meet, you would walk away feeling very, very disappointed. But if you were to go through even just 5 or 10 or 15 . . . very high-quality profiles, and everyone was actually quite interesting to you, you would feel very, very compelled to return.
Wolfe Herd is banking that getting more users into the Approve category will eventually help the company make money. She noted that Approve users monetize at approximately double the rates of people in the Improve category. The average revenue per user on Bumble has dropped 15% since 2021. For the second quarter of 2025, total revenue decreased 8% year over year, and paying users decreased 11% compared to the same period last year.
Everything we build is grounded in real-world outcomes, not endless engagement, Wolfe Herd said on the companys Q2 earnings call. As part of this, the company is winding down the digital marketing that it began during the pandemic. That practice attracted a slew of new users who diluted the apps dating pool and diminished the user experience, according to Wolfe Herd.
The Hinge playbook
In this regard, Bumbles approach mirrors Hinges playbook. Though parent company Match Groups second-quarter revenue was flat overall, Hinges revenue jumped 25% to $167.5 million, beating analysts expectations. Match Group CEO Spencer Rascoff said the company plans on investing $50 million in part to fund Hinges geographic expansion to Latin America and Europe.
Hinge has succeeded largely by creating an experience that focuses on showing users quality matches and getting them off the app to form lasting connections. To do this, the app has introduced friction in its user experience, putting limits on swipes and implementing penalties for ghosting to improve user behavior. Hinge is also using AI to refine its matching algorithm and launch new features like notifications to encourage users to keep up conversations with matches they are interested in. Hinges monthly active users rose nearly 20% in the first half of 2025.
In getting users to improve their profiles, Bumble is taking another cue from Hinge, which has a more extensive sign-up process than other dating apps. It requires users to answer a series of questions and sets a minimum requirement for the number of photos uploaded to their profile before theyre able to swipe.
In their Q2 earnings calls, both Rascoff and Wolfe Herd suggested that building algorithms and features that foster lasting connections is key to attracting younger users. Tinder has launched double-date options and college-specific features to help foster connections in lower-pressure environments. Bumble is doubling down on Bumble BFFa separate app to help users find new friendsand plans to launch more off-platform experiences to increase serendipitous connections.
A lot of the exact product solves that we are so maniacally focused on right now are specifically the issues Gen Z has with online dating, Wolfe Herd said on Bumbles earnings call. For example, they don’t want to feel like they swipe endlessly through people they’re not interested in. They don’t want to feel judged. They don’t want to feel rejected. They don’t want to feel like they’re talking to someone that is not actually who they say they are.
All sorts of consumer-facing companies seem excited about the potential to use artificial intelligence to set prices. There’s only one problemconsumers hate it.
So-called dynamic-pricing strategies offer the possibility of tweaking prices according to changing external circumstances, shifting demand, even individual consumer situations. If combining the practice with cutting-edge technology sounds alarming, youve identified the problem companies are facing: how to talk about their AI-driven pricing plans without scaring customers, sparking backlash, and drawing terrible press and even legislative scrutiny.
As soon as you talk about dynamic pricing, there is immediately repulsion by consumers, says Stephan Liozu, chief value officer at Zilliant, a pricing management and optimization software firm. There’s no positive first impression in the consumer.
This has long held true for examples like simple price gouging (jacking up prices when consumers have limited options, such as for gas or ice after a natural disaster), but sophisticated tech-driven iterations are just as reviled. Theres a lack of understanding of why this is done, so the consumer thinks its all about profit, Liozu says. And in a way, it is.
In a world where backlash can be spread rapidly on social media and amplified by the traditional press, thats a dynamic brands cant shrug off.
The latest example involves an industry already known for opaque pricing strategies. In an earnings call in July, the president of Delta Air Lines noted that the carrier is experimenting with ways to use AI to generate a price thats available on that flight, on that time, to you, the individual. Delta already uses AI to drive personalized pricing on 3% of its flights, he added, and aims to bump that to 20% by the end of the year.
When this was reported, it sparked a rush of criticism from consumer advocates and politicians. Arizona Senator Ruben Gallego called it predatory pricing. Admittedly, the critics had little if any specific knowledge of the factors Deltas AI pricing procedure actually entails, but thats sort of the point of the criticism. In the absence of transparency, its easy to invent Black Mirror-like scenarios.
Think about one popular hypothetical: AI deduces youre going to a funeral and will pay more than usual for a flight. Delta has denied that anything like this has happened or is planned, and insists its pricing is based on interpreting market conditions, not individual data.
Dynamic pricing isn’t new
Dynamic pricing isnt exactly new; youve experienced it if youve ever purchased off-season theme park tickets or traveled in peak holiday windows. That doesnt always mean outright price gouging or unfair price manipulation. But the idea of turning to sophisticated and inscrutable technology to bolster the practice can feel unnerving, perhaps especially as it works its way into everyday categories from fast food to retail.
Liozu, the pricing expert, has argued that ultimately dynamic pricing is just a tool, and its impact is shaped by human decisions: People decide what data goes into the algorithm. People choose what variables it prioritizes. People determine the thresholds for price changes and approve the pricing strategies the algorithm supports. But AI critics and enthusiasts alike tend to focus more on the power of the algorithms.
Theres a bigger fear (and likelihood, Liozu suggests) that the practice will spread across all sorts of categories, especially online. A January Federal Trade Commission report on what the agency termed surveillance pricing found that details like a persons precise location or browser history can be frequently used to target individual consumers with different prices for the same goods and services.
That would be a big change: Set prices have been routine since they replaced individual bargaining in retail scenarios in the late 19th century. And Delta is not the first to face a backlash in response to disclosing pricing experiments.
Last year, Wendys endured a wave of criticism after its CEO mentioned during an earnings call that the burger chains new digital menu boards might enable dynamic pricing and day-part offerings along with AI-enabled menu changes and suggestive selling. He did not use the term surge pricingassociated with ride-share servicesbut much of the news coverage of his statements did. Eventually Wendys explicitly said it had no plans to raise prices at busy times.
The Delta incident brought even more critical scrutiny, and that may be because the airline industry is already associated with highly variable pricing that can seem to consumers like a black box. Add AI to the mix, and its like tossing a heavy blanket over the black box, leaving customers feeling powerless.
As Gallego and other senators argued in a statement, AI-fueled pricing will likely mean fare price increases up to each individual consumers personal pain point. Delta denied that, too, in a statement reported by Reuters assuring the senators (and fliers): There is no fare product Delta has ever used, is testing or plans to use that targets customers with individualized prices based on personal data.
It’s not clear how much scrutiny will continue from the FTC under the Trump administration, which has signaled a light regulatory touch on AI. But that is unlikely to resolve the fears and pushback of many consumers.
Liozu argues that companies moving toward AI-driven dynamic pricing need to operate with more transparency and communication, thinking in terms of explainable AI. That means they also need to tweak their actual beavior: Do pricing research around what is too much for consumers, right? How frequently do you change your prices? We need to figure out in dynamic pricing what’s the floor and what’s the ceiling after which its considered unfair, he says.
Most important, Liozu adds, companies need to communicate that to consumers. Gallego complained that Delta is telling their investors one thing, and then turning around and telling the public another. Companies dont seem to have figured out how to talk to consumers about AI-guided pricing strategies in ways that show they care about the consumer impact. As Liozu puts it: You need to be able to tame your algorithms.
AI skeptics have found a new way to express their disdain for the creeping presence of artificial intelligence: through slurs.
Out on the streets and in stores, people have begun harassing robots they encounter in the wild. (Anyone else feel a bit sorry for the robot?) Online, the internet has revived a Star Warsinspired insult, clanker, with Google Trends data showing a spike in searches for the term in early June.
@semdenpriv original sound – semdenpriv
POV: Me at the clanker rally in 2088, one TikTok user joked. Keep your oily soulless clanker hands away from my delicious human food, another X user wrote in response to a clip of Elon Musks Optimus robot dishing out popcorn at the Tesla Diner (not a sentence I ever thought Id write).
Keep your oily soulless clanker hands away from my delicious human food https://t.co/DXF7JNKD0W— EckhartsLadder (@EckhartsLadder) July 20, 2025
The term has also been picked up by politicians. Sick of yelling ‘REPRESENTATIVE’ into the phone 10 times just to talk to a human being? Sen. Ruben Gallego (D-AZ) posted on X last month. My new bill makes sure you dont have to talk to a clanker if you dont want to.
Sick of yelling REPRESENTATIVE into the phone 10 times just to talk to a human being? My new bill makes sure you dont have to talk to a clanker if you dont want to. pic.twitter.com/9aUv478gSP— Ruben Gallego (@RubenGallego) July 30, 2025
While some direct their insults at the technology itself, others target those using AI systems. On one thread, suggestions for users of the xAI chatbot Grok included Grokkers, Groklins, and Grocksuckers. Meanwhile, on TikTok, someone coined sloppers for people becoming increasingly overreliant on ChatGPT.
@intrnetbf shoutout to Monica. Incredible command over the English language original sound – intrnetbf
The trend reflects a broader mood. Concerns about AI among U.S. adults have grown since 2021, according to the Pew Research Center. More than half (51%) say they are more concerned than excited about the technologys rise, with worries ranging from AI taking away jobs to chatbot addiction.
Still, some see embracing new slurseven those aimed at robotsas problematic, especially when they echo existing racial slurs or stereotypes.
@thebrookboys This bout to be the biggest fear for all Dads in year 2050 #meme #clanker #robo Bell Sound/Temple/Gone/About 10 minutes(846892) – yulu-ism project
Others simply fear theyll regret their words later. As one X user wrote: I dont want to have to look a robot in the eye in fifty years and be like, you dont understand it was a different time
star wars did give us a slur for robots (clankers) but i dont use it bc i dont want to have to look a robot in the eye in fifty years and be like you dont understand it was a different time— anna !!! (@frogs4girls) July 20, 2025
In the late 2010s, cultured meat was everywhereand yet nowhere. From Reddit to major magazine covers, articles touted the latest advances in “lab-grown meat,” promising cruelty-free, environmentally friendly steaks at your local supermarket. The hype was palpable.
One 2019 report predicted cultured meats would halve the number of cows on the planet by 2030, disrupting the world’s oldest industry by delivering ethical meat with negligible environmental impact that tasted identical to traditional meatand at a fraction of the price.
[Photo: Vow]
That promise of rapid disruption terrified conventional animal agriculture stakeholders. Under pressure from these livestock constituents, lawmakers in multiple states have banned this new protein source entirely. Florida and Alabama passed bans in 2024, with more states following. Indiana imposed a manufacturing moratorium with steep fines, Nebraska prohibited its production and sale, and Montanas governor signed legislation to ensure consumers could “continue to enjoy authentic meat.”
In June, a Texas ban became law, with the state’s agriculture commissioner touting the “God-given right” to pasture-raised meateven though the vast majority of what Americans eat comes from industrial feedlots.
But here’s the irony: Lawmakers are fighting a version of cultured meat that never materialized. Today, while you can eat cultured meat at more than 60 venues in Singapore and Australia, and cultured seafood at two restaurants in the U.S. at the time of this writing, it’s far from the rapid disruption that was forecasted. More than a decade after the world’s first cultured hamburger was announced, the hype has virtually disappeared.The reality of how and why this all transpired is complicated. However, we would argue that what we’re witnessing isn’t industry failure, but the natural evolution of a transformative invention finding its true market fit. Cultured meat technology works; what needed adjustment were the timelines and business models that promised too much, too quickly, and to replicate conventional meats that people already enjoy en masse.Rather than viewing this as a setback, some in the industry are discovering something potentially more valuable: sustainable, scalable pathways to market that don’t require displacing existing agriculture but can grow alongside it. As the industry turns the page to a new chapter, once uncertain regulatory pathways are now established in multiple countries.
[Photo: Vow]
The technology itself continues to advance. Production yields are improving, costs are declining, and new species beyond traditional livestock are proving viable for cultivation.More importantly, early market success demonstrates genuine consumer appetite. In Singapore, where cultured meat has been available the longest, restaurants report strong repeat customers and growing demand. In Australia, where cultured meat became available at dozens of restaurants in recent weeks, initial sales and demand for the items are taking off.
Forged Cultured Japanese Quail Whipped Pate [Photo: Vow]
This suggests cultured meat purveyors arent just scratching a theoretical itch, but delivering real value and excitement that consumers recognize and seek out.This reality is leading to a strategic pivot that may actually benefit both the industry and consumers: innovation over imitation. Rather than trying to perfectly replicate a chicken wing or rib-eye steakproducts that traditional animal agriculture already produces and consumers are accustomed tocompanies that are finding success are creating entirely new culinary experiences that excite chefs and diners alike.
Forged Cultured Japanese Quail Foie Gras [Photo: Vow]
Take Japanese quail, a species that demonstrates cultivated meat’s unique advantages. Traditional quail foie gras is impossible to produce commerciallythe birds are so petite that conventional methods are prohibitively labor-intensive, and the production process itself remains controversial.
Japanese quail, however, proves remarkably well-suited for cultivation technology, enabling the creation of previously undoable delicacies like foie gras, whipped pâté, and even edible tallow candles.
Forged Cultured Japanese Quail Tallow Candle [Photo: Vow]
And Vow can make a lot of it. The company recently completed the largest cultured meat harvest in history: more than one metric ton of quail. And it projects it will have the capacity, by the end of 2025, to harvest up to 130 metric tons annually.
While that’s still minimal compared with the 12.29 million metric tons of beef American farmers produced in 2023 and 2024, it is proof that cultured meat can offer consumers genuinely new choices and advance consumer acceptance. Its an illustration of how the industry can position itself as expanding culinary possibilities while avoiding potential conflicts with traditional agriculture.Rather than letting politicians dictate what should be on our plates in order to protect incumbent industries, we should trust consumers to decide for themselves. When given the freedom to choose, consumers are embracing these innovations as exciting additions to culinary experiences, the evidence suggests. Thats a decision best left to diners, not lawmakers.
As a partner at Theory Ventures, a VC firm built around deep technology and market research, I spend my days swimming in information: academic papers, market reports, interview notes, and written analyses. Our job is to synthesize these data points into a nuanced perspective to inform our investment decisions.
Reading the hype online, its tempting to think you can just delegate anything to AI. But for something so critical to our job, we dont just want it done, we need it to be excellent. How much can AI really do for us?
In this piece, I will share:
How we structure instructions to get the best analysis out of an AI model
Where I critically intervene and rely on my own thoughts
How you can get an AI to mirror the way you write
When relying on an LLM you often get something that only seems good at first glance: often the AI has missed details, or an important nuance. And for this core part of my job, decent isnt enoughI need output to be excellent.
This AI accuracy gap creates a painful cycle where you spin in circles, trying to re-prompt the system to get what you want until youre essentially left rewriting the entire output on your own. In the end, its unclear if AI actually helped at all. The more effective approach is understanding how you (the human) do the thinking and leave writing (i.e., formatting and synthesis) to the LLM. This simple separation is what elevates AI-augmented workflows from decent to exceptional.
Heres an example of how we build those kinds of workflows at Theory Ventures, and how you can too. Well illustrate an example with the automation of our internal market research reports.
Step 1: Define the thinking process
Prepare a document with very detailed instructions on the underlying analysis/construction you seek to achieveclearly outline the context & goals, then dive into all of the details on how you deconstruct a broad analysis: the specific questions you would ask, follow-up sub-questions, how they should be answered with data, and key callouts or exceptions.
You can use an AI assistant to help you generate a first draft of this, sharing completed documents and asking it to deconstruct the analysis. But these instructions are critical, so its important to finish writing it by hand and continue to update it over time when you tweak your analysis.
Example analysis instructions included in the prompt (note: the full instructions will typically be 2 to 10 or more pages long)
Analyze the underlying market structure: Is it fragmented or consolidated? Why? (e.g., high specialization needs, regulatory barriers, network effects, legacy tech debt). How is fragmentation changing over time, and is it different across market segments?
Use the following data sources and analyses: . . .
Evaluate key market dynamics: What are the typical switching costs? How prevalent is tech debt? What are the typical sales cycles and buyer behaviors? How do incumbents maintain their position (moats)?
Use the following data sources and analyses: . . .
Step 2: Lay out your human-led analysis
Provide your primary analysis, along with raw notes and instructions to the AI. We set our systems up so they require the user to provide their key takeaways and analysis to guide the system towards whats most importanthighlighting areas to focus on, key opportunities, and potential concerns. These are typically four to five detailed bullet points of two to four sentences each. This is the crux of the analysis and should therefore never be AI-generated.
Example key takeaways provided to the system:
This market has historically been small and fragmented without major software providers. We expect it will grow dramatically, primarily through currently automating labor spend and consolidating a set of point solutions. The underlying demand for this capability will also increase with XYZ challenges. We feel very confident in these two growth levers.
Theres substantial concentration at the upper end of the market. Major platforms control around X% of the market and have all invested heavily in their own technology. But below the top-n largest players, there is a healthy cohort of medium-large buyers that have the scale to need this solution but dont want to build it. We think this is sufficient to build a sizeable company, although market concentration and build versus buy remains a key long-term risk.
Step 3: Run an interactive Q&A to hone the analysis
This dialogue is the most interesting and fun step: Have the system generate questions to clarify the contours of your analysis. Based on the primary analysis, along with the notes and general instructions, the system asks questions about things that either werent clear or had conflicting information/instructions. This helps sharpen the analysis and gives the user the opportunity to share more of their thought process and guidance.
Example Q&A:
Q from the AI: You said that major platforms have invested heavily in this technology, but conversations with some of those companies indicated an excitement to buy. Do you think that will be common, or were they exceptions?
A from the human: Good point. I do think that many of them will buy eventually, but because theyve built a lot of technology internally they are more likely to need a new platform only for certain components, versus buying an end-to-end system. And the very largest companies (top three to five) will build everything in-house.
Step 4: Share past work to match tone, not ideas
Use previous examples of your work to replicate tone and style only after the scaffolding work is done. Most people skip immediately to this step, but we found (and research shows) that providing finished examples is most useful just to match tone and writing style, as opposed to shaping the analysis itself.
In researching the best AI-native products, weve seen that practically all of the work goes into defining the thinking and analysis portion of the problemdetailed instructions, guidelines, orchestration, and toolingso the AI system knows what it should do and just executes on it.
At Theory Ventures, weve started to mirror the same system by developing highly-constrained, human-in-the-loop workflows that direct the analysis, leaving the LLM to execute basic information extraction and synthesis. Thats how weand our AI systemshave started working smarter. Not by asking AI to think for us, but by helping it think better.
