Using data from Crunchbase and PitchBook, TechCrunch tracked down the VC-backed startups that became unicorns so far this year.
Read more AI health records startup Torch will be using its tech for OpenAI’s new ChatGPT Health, according to co-founder Ilya Abyzov.
Read more We tried Amazon’s new AI wearable Bee. It’s not for pro users yet, but more features are expected this year.
Read more Amazon explains where its wearable Bee fits in and whether it will merge with Alexa.
Read more Meta is ramping up its efforts to build out its AI capacity — Zuckerberg said the company intended to drastically expand its energy footprint in the coming years.
Read more This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.
I decided to go to CES kind of at the last minute. Over the holiday break, contacts from China kept messaging me about their travel plans. After the umpteenth “See you in Vegas?” I caved. As a China tech writer based in the US, I have one week a year when my entire beat seems to come to me—no 20-hour flights required.
CES, the Consumer Electronics Show, is the world’s biggest tech show, where companies launch new gadgets and announce new developments, and it happens every January. This year, it attracted over 148,000 attendees and over 4,100 exhibitors. It sprawls across the Las Vegas Convention Center, the city’s biggest exhibition space, and spills over into adjacent hotels.
China has long had a presence at CES, but this year it showed up in a big way. Chinese exhibitors accounted for nearly a quarter of all companies at the show, and in pockets like AI hardware and robotics, China’s presence felt especially dominant. On the floor, I saw tons of Chinese industry attendees roaming around, plus a notable number of Chinese VCs. Multiple experienced CES attendees told me this is the first post-covid CES where China was present in a way you couldn’t miss. Last year might have been trending that way too, but a lot of Chinese attendees reportedly ran into visa denials. Now AI has become the universal excuse, and reason, to make the trip.
As expected, AI was the biggest theme this year, seen on every booth wall. It’s both the biggest thing everyone is talking about and a deeply confusing marketing gimmick. “We added AI” is slapped onto everything from the reasonable (PCs, phones, TVs, security systems) to the deranged (slippers, hair dryers, bed frames).
Consumer AI gadgets still feel early and of very uneven quality. The most common categories are educational devices and emotional support toys—which, as I’ve written about recently, are all the rage in China. There are some memorable ones: Luka AI makes a robotic panda that scuttles around and keeps a watchful eye on your baby. Fuzozo, a fluffy keychain-size AI robot, is basically a digital pet in physical form. It comes with a built-in personality and reacts to how you treat it. The companies selling these just hope you won’t think too hard about the privacy implications.
Ian Goh, an investor at 01.VC, told me China’s manufacturing advantage gives it a unique edge in AI consumer electronics, because a lot of Western companies feel they simply cannot fight and win in the arena of hardware.
Another area where Chinese companies seem to be at the head of the pack is household electronics. The products they make are becoming impressively sophisticated. Home robots, 360 cams, security systems, drones, lawn-mowing machines, pool heat pumps … Did you know two Chinese brands basically dominate the market for home cleaning robots in the US and are eating the lunch of Dyson and Shark? Did you know almost all the suburban yard tech you can buy in the West comes from Shenzhen, even though that whole backyard-obsessed lifestyle barely exists in China? This stuff is so sleek that you wouldn’t clock it as Chinese unless you went looking. The old “cheap and repetitive” stereotype doesn’t explain what I saw. I walked away from CES feeling that I needed a major home appliance upgrade.
Of course, appliances are a safe, mature market. On the more experiential front, humanoid robots were a giant magnet for crowds, and Chinese companies put on a great show. Every robot seemed to be dancing, in styles from Michael Jackson to K-pop to lion dancing, some even doing back flips. Hangzhou-based Unitree even set up a boxing ring where people could “challenge” its robots. The robot fighters were about half the size of an adult human and the matches often ended in a robot knockout, but that’s not really the point. What Unitree was actually showing off was its robots’ stability and balance: they got shoved, stumbled across the ring, and stayed upright, recovering mid-motion. Beyond flexing dynamic movements like these there were also impressive showcases of dexterity: Robots could be seen folding paper pinwheels, doing laundry, playing piano, and even making latte art.
CAL SPORT MEDIA VIA AP IMAGES
However, most of these robots, even the good ones, are one-trick ponies. They’re optimized for a specific task on the show floor. I tried to make one fold a T-shirt after I’d flipped the garment around, and it got confused very quickly.
Still, they’re getting a lot of hype as an important next frontier because they could help drag AI out of text boxes and into the physical world. As LLMs mature, vision-language models feel like the logical next step. But then you run into the big problem: There’s far less physical-world data than text data to train AI on. Humanoid robots become both applications and roaming data-collection terminals. China is uniquely positioned here because of supply chains, manufacturing depth, and spillover from adjacent industries (EVs, batteries, motors, sensors), and it’s already developing a humanoid training industry, as Rest of World reported recently.
Most Chinese companies believe that if you can manufacture at scale, you can innovate, and they’re not wrong. A lot of the confidence in China’s nascent humanoid robot industry and beyond is less about a single breakthrough and more about “We can iterate faster than the West.”
Chinese companies are not just selling gadgets, though—they’re working on every layer of the tech stack. Not just on end products but frameworks, tooling, IoT enablement, spatial data. Open-source culture feels deeply embedded; engineers from Hangzhou tell me there are AI hackathons every week in the city, where China’s new “little Silicon Valley” is located.
Indeed, the headline innovations at CES 2026 were not on devices but in cloud: platforms, ecosystems, enterprise deployments, and “hybrid AI” (cloud + on-device) applications. Lenovo threw the buzziest main-stage events this year, and yes, there were PCs—but the core story was its cross-device AI agent system, Qira, and a partnership pitch with Nvidia aimed at AI cloud providers. Nvidia’s CEO, Jensen Huang, launched Vera Rubin, a new data-center platform, claiming it would dramatically lower costs for training and running AI. AMD’s CEO, Lisa Su, introduced Helios, another data-center system built to run huge AI workloads. These solutions point to the ballooning AI computing workload at data centers, and the real race of making cloud services cheap and powerful enough to keep up.
As I spoke with China-related attendees, the overall mood I felt was a cautious optimism. At a house party I went to, VCs and founders from China were mingling effortlessly with Bay Area transplants. Everyone is building something. Almost no one wants to just make money from Chinese consumers anymore. The new default is: Build in China, sell to the world, and treat the US market like the proving ground.
Emissions from air freight have increased by 25% since 2019, according to a 2024 analysis by environmental advocacy organization Stand.Earth.
The researchers found that the expansion of cargo-only fleets to transport goods during the pandemic — as air travel halted, slower freight modes faced disruption, but demand for rapid delivery soared — has led to a yearly increase of almost 20 million tons of carbon dioxide, making up 93.8m tonnes from air freight overall.
And though fleet modernization and operational improvements by freight operators have contributed to ongoing decarbonization efforts, sustainable aviation fuel (SAF) looks set to be instrumental in helping the sector achieve its ambitions to reduce environmental footprint in the long-term.
When used neat, or pure and unblended, SAF can help reduce the life cycle of greenhouse gas emissions from aviation by as much as 80% relative to conventional fuel. It’s why the International Air Transport Association (IATA) estimates that SAF could account for as much as 65% of total reduction of emissions.
For Christoph Wolff, CEO of the Smart Freight Centre, “SAF is the main pathway” to decarbonization across both freight and the wider aviation ecosystem.
“The great thing about SAF is it’s chemically identical to Jet A fuel,” he says. “You can blend it [which means] you have a pathway to ramp it up. You can start small and you can scale it. By scaling it there is the promise or the hope that the price comes down.”
At at least twice the price of conventional jet fuel, cost is a significant barrier hindering broader adoption.
And it isn’t the only one standing between SAF and wider penetration.
Bridging the gap between a concentrated supply of SAF and global demand also remains a major hurdle.
Though the number of verified SAF outlets has increased from fewer than 20 locations in 2021 to 114 as of April 2025, according to sustainability solutions framework 4Air, that accounts for only 92 airports worldwide out of more than 40,000.
“SAF is central to the decarbonization of the aviation sector,” believes Raman Ojha, president of Shell Aviation. “Having said that, adoption and penetration of SAF hasn’t really picked up massively. It’s not due to lack of production capacity, but there are lots of things that are at play. And book and claim in that context helps to bridge that gap.”
Bridging the gap with book and claim
Book and claim is a chain of custody model, where the flow of administrative records is not necessarily connected to the physical product through the supply chain (source: ISO 22095:2020).
Book and claim potentially enables airlines and corporations to access the life cycle GHG emissions reduction benefits of SAF relative to conventional jet fuel even when SAF is not physically available at their location; this model helps bridge the gap between that concentrated supply and global demand, until SAF’s availability improves.
“To be bold, without book and claim, no short-term science-based target will be achieved,” says Bettina Paschke, vice president of ESG accounting, reporting and controlling at DHL Express. “Book and claim is essential to achieving science-based targets.”
“SAF production facilities are not everywhere,” she reiterates. “They’re very focused on one location, and if a customer wants to fulfil a mass balance obligation, SAF would need to be shipped around the world just to be at that airport for that customer. That would be very complicated, and very unrealistic.” It would also, counterintuitively, increase total emissions. By using book and claim instead, air freight operators can unlock the life cycle greenhouse gas emissions reduction benefits of SAF relative to conventional jet fuel now, without waiting for supply to broaden. “It might no longer be needed when we have SAF product facilities at each airport in the future,” she points out. “But at the moment, that’s not the case.”
At DHL itself, the mechanism has become central to achieving its own three interconnected sustainability pillars, which focus on decarbonizing logistics supply chains, supporting customers toward their decarbonization goals, and ensuring credible emission claims can be shared along the value chain.
Demonstrating the importance of a credible and viable framework for book and claim systems is also what inspired the 2022 launch of Shell’s Avelia, one of the first blockchain-powered digital SAF book and claim solutions for aviation, which expanded in 2024 to encompass air freight in addition to business travel. Depending on the offering, Avelia offers freight forwarders the opportunity to share the life cycle greenhouse gas emissions reduction benefits of SAF relative to conventional jet fuel across the value chain with shippers using their services.
“It’s also backed by a physical supply chain, which gives our customers — whether those be corporates or freight forwarders or even airlines — a peace of mind that the SAF has been injected at a certain airport, it’s been used and environmental attributes, with the help of blockchain, have been tracked to where they’re getting retired,” says Ojha.
He adds: “The most important or critical part is the transparency that it’s providing to our customers to be sure that they’re not saying something which they can’t confidently stand behind.”
Moving beyond early adoption
To scale up SAF via book and claim and help make it a more commercially viable lower-carbon solution, its adoption will need to be a coordinated “ecosystem play,” says Wolff. That includes early adopters, such as DHL, inspiring action from peers, solution providers such as Shell, working with various stakeholders to drive joint advocacy, and industry associations, like the Smart Freight Centre creating the required frameworks, educational resources, and industry alignment.
An active book and claim community made up of many forward-thinking advocates is already driving much of this work forward with a common goal to develop greater standardization and consensus, Wolff points out. “It helps to make sure all definitions on the system are compatible and they can talk to one another, provide educational support, and [also that] there’s a repository of transactions so that it can be documented in a way that people can see and think, ‘oh this is how we do it.’ There are some early adopters that are very experienced, but it needs a lot more people for it to get comfortable.”
In early 2024, discussions were held with a diverse group of expert book and claim stakeholders to develop and refine 11 key principles and best practices book and claim models. These represent an aligned set of principles informed by practical successes and challenges faced by practitioners working to decarbonize the heavy transport sector.
Adherence to such a framework is crucial given that book and claim is not yet accepted by the Greenhouse Gas (GHG) Protocol nor the Science Based Targets Initiative (SBTi) as a recognized model for reducing greenhouse gas emissions — though there are hopes that might change.
“The industrialization of book and claim delivery systems is key to credibility and recognition,” says Wolff. “The Greenhouse Gas Protocol and the Science Based Targets Initiative are making steps in recognizing that. There’s a pathway that the Smart Freight Centre is very closely involved in the technical working groups for [looking]to build such a system where, in addition to physical inventory, you also pursue market-based inventories.”
Paschke urges companies not to sit back and wait for policy to change before taking action, though. “The solution is there,” she says. “There are companies like DHL that are making huge upfront investments, and every single contribution helps to scale the industry and give a strong signal to the eco-space.”
As pressure to accelerate decarbonization gains pace, it’s critical that air freight operators consider this now, agrees Ojha. “Don’t wait for perfection in guidelines, regulations, or platforms — act now,” he says. “That’s very, very critical. Second, learn by doing and join hands with others. Don’t try to do everything independently or in-house.
“Third, make use of registries and platforms, such as Avelia, that can give credibility. Join them, utilize them, and leverage them so that you won’t have to establish auditability from scratch.
“And fourth, don’t look at scope book and claim as a means for acquiring a certificate for environmental attributes. Think in terms of your decarbonisation commitment and think of this as a tool for exposure management. Think in terms of the bigger picture.”
That bigger picture being a significant sector-wide push toward faster decarbonization — and turning the tide on emissions’ steep upward ascent.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff. It was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.
This content is produced by MIT Technology Review Insights in association with Avelia. Avelia is a Shell owned solution and brand that was developed with support from Amex GBT, Accenture and Energy Web Foundation. The views from individuals not affiliated with Shell are their own and not those of Shell PLC or its affiliates. Cautionary note | Shell Global
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
Introducing this year’s 10 Breakthrough Technologies
It’s easy to be cynical about technology these days. Many of the “disruptions” of the last 15 years were more about coddling a certain set of young, moneyed San Franciscans than improving the world. Yet you can be sympathetic to the techlash and still fully buy into the idea that technology can be good.
We really can build tools that make this planet healthier, more livable, more equitable, and just all-around better. And some people are doing just that, pushing progress forward across a number of fundamental, potentially world-changing technologies.
These are exactly the technologies we aim to spotlight in our annual 10 Breakthrough Technologies list. These are 10 technologies that we believe are poised to fundamentally alter the world, and they’re a matter of hot debate across the newsroom for months before being unveiled. So, without further ado… Here’s the full list.
Do you think we’ve missed something? You have until April to cast your vote for the 11th breakthrough!
Why some “breakthrough” technologies don’t work out
—Fabio Duarte is associate director and principal research scientist at the MIT Senseable City Lab.
Today marks the 25th year the MIT Technology Review newsroom has compiled its annual 10 Breakthrough Technologies list, which means its journalists and editors have now identified 250 technologies as breakthroughs.
A few years ago, editor at large David Rotman revisited the publication’s original list, finding that while all the technologies were still relevant, each had evolved and progressed in often unpredictable ways. I lead students through a similar exercise in a graduate class I teach with James Scott for MIT’s School of Architecture and Planning, asking them what we can learn from the failures.
Although it’s less glamorous than envisioning which advances will change our future, analyzing failed technologies is equally important. Read about why that is.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 Iran has almost completely shut its internet down
Which makes it very hard for the world to witness its government killing people. (AP)
+ The shutdown is chillingly effective, and likely to last. (The Guardian)
+ President Trump is considering military strikes against Iran. (NYT $)
2 ICE is gaining powerful new surveillance capabilities
It’s purchased tools that give it the ability to track individuals across entire neighborhoods. (404 Media $)
+ The ICE shooting shows why reality still matters. (The Verge $)
+ It’s time for Apple to reinstate ICEBlock. (Engadget)
3 Malaysia and Indonesia have blocked access to Grok
They are the first in the world to ban the AI tool, which is being used to make explicit non-consensual deepfakes. (BBC)
+ How Elon Musk’s platform unleashed a torrent of abuse upon women and girls. (The Guardian)
4 Silicon Valley’s billionaires are panicking over a proposed 5% wealth tax
Poor dears. (Wired $)
5 Meta signed a deal with three nuclear companies
It’s becoming a favored power source for tech companies as their AI ambitions grow. (TechCrunch)
+ Can nuclear power really fuel the rise of AI? (MIT Technology Review)
6 AI has a memorization problem
The fact it reproduces copyrighted work shows it might not work the way its makers claim. (The Atlantic $)
+ DeepSeek is poised to release a new flagship AI model. (The Information $)
7 Here’s the stuff from CES you might actually consider buying
It’s always a bit of a gimmick fest—but these items made their way onto reporters’ wishlists. (The Verge $)
+ On the flipside, you absolutely should not purchase anything on the ‘worst in show’ list. (The Register)
8 How WhatsApp took over the world 
It’s used by more than three billion people every month—nearly half the global population. (New Yorker $)
9 AI music is here to stay
Love it or hate it, it’s only going to play a bigger role going forward. (Vox)
+ It’s complicating our definitions of authorship and creativity in the process. (MIT Technology Review)
10 We’re crying out for better experiences online
The question is: who will give them to us? (WP $)
Quote of the day
“Things here are very, very bad. A lot of our friends have been killed. They were firing live rounds. It’s like a war zone, the streets are full of blood. They’re taking away bodies in trucks.”
—An anonymous source in Iran’s capital Tehran tells the BBC how the government is cracking down on protests.
One more thing
This startup is about to conduct the biggest real-world test of aluminum as a zero-carbon fuel
Found Energy aims to harness the energy in scraps of aluminum metal to power industrial processes without fossil fuels. Since 2022, the company has worked to develop ways to rapidly release energy from aluminum on a small scale.
Now it’s just switched on a much larger version of its aluminum-powered engine, which it claims is the largest aluminum-water reactor ever built. Soon, it will be installed to supply heat and hydrogen to a tool manufacturing facility in the southeastern US, using the aluminum waste produced by the plant itself as fuel.
If everything works as planned, this technology, which uses a catalyst to unlock the energy stored within aluminum metal, could transform a growing share of aluminum scrap into a zero-carbon fuel. Read the full story.
—James Dinneen
We can still have nice things
A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line or skeet ’em at me.)
+ I enjoyed this heartwarming story of love across the generational divide 
+ It turns out you can cook an egg in an air fryer. The big question is—should you?
+ Of course Japan has a real-life Pokémon Fossil Museum.
+ If you haven’t already, make 2026 the year you get a hobby. Your life will be all the richer for it.
In February 2025, cyberattackers thought to be linked to North Korea executed a sophisticated supply chain attack on cryptocurrency exchange Bybit. By targeting its infrastructure and multi-signature security process, hackers managed to steal more than $1.5 billion worth of Ethereum in the largest known digital-asset theft to date.
The ripple effects were felt across the cryptocurrency market, with the price of Bitcoin dropping 20% from its record high in January. And the massive losses put 2025 on track to be the worst year in history for cryptocurrency theft.
Bitcoin, Ethereum, and stablecoins have established themselves as benchmark monetary vehicles, and, despite volatility, their values continue to rise. In October 2025, the value of cryptocurrency and other digital assets topped $4 trillion.
Yet, with this burgeoning value and liquidity comes more attention from cybercriminals and digital thieves. The Bybit attack demonstrates how focused sophisticated attackers are on finding ways to break the security measures that guard the crypto ecosystem, says Charles Guillemet, chief technology officer of Ledger, a provider of secure signer platforms.
”The attackers were very well organized, they have plenty of money, and they are spending a lot of time and resources trying to attack big stuff, because they can,” he says. “In terms of opportunity costs, it’s a big investment, but if at the end they earn $1.4 billion it makes sense to do this investment.”
But it also demonstrates how the crypto threat landscape has pitfalls not just for the unwary but for the tech savvy too. On the one hand, cybercriminals are using techniques like social engineering to target end users. On the other, they are increasingly looking for vulnerabilities to exploit at different points in the cryptocurrency infrastructure.
Historically, owners of digital assets have had to stand against these attackers alone. But now, cybersecurity firms and cryptocurrency-solution providers are offering new solutions, powered by in-depth threat research.
A treasure trove for attackers
One of the advantages of cryprocurrency is self custody. Users can save their private keys—the critical piece of alphanumeric code that proves ownership and grants full control over digital assets—into either a software or hardware wallet to safeguard it.
But users must put their faith in the security of the wallet technology, and, because the data is the asset, if the keys are lost or forgotten, the value too can be lost.
”If I hack your credit card, what is the issue? You will call your bank, and they will manage to revert the operations,” says Vincent Bouzon, head of the Donjon research team at Ledger. “The problem with crypto is, if something happens, it’s too late. So we must eliminate the possibility of vulnerabilities and give users security.”
Increasingly, attackers are focusing on digital assets known as stablecoins, a form of cryptocurrency that is pegged to the value of a hard asset, such as gold, or a fiat currency, like the US dollar.
Stablecoins rely on smart contracts—digital contracts stored on blockchain that use pre-set code to manage issuance, maintain value, and enforce rules—that can be vulnerable to different classes of attacks, often taking advantage of users’ credulity or lack of awareness about the threats. Post-theft countermeasures, such as freezing the transfer of coins and blacklisting of addresses, can lessen the risk with these kinds of attacks, however.
Understanding vulnerabilities
Software-based wallets, also known as “hot wallets,” which are applications or programs that run on a user’s computer, phone, or web browser, are often a weak link. While their connection to the internet makes them convenient for users, it also makes them more readily accessible to hackers too.
“If you are using a software wallet, by design it’s vulnerable because your keys are stored inside your computer or inside your phone. And unfortunately, a phone or a computer is not designed for security.” says Guillemet.
The rewards for exploiting this kind of vulnerability can be extensive. Hackers who stole credentials in a targeted attack on encrypted password manager application LastPass in 2022 managed to transfer millions worth of cryptocurrency away from victims in the subsequent two or more years.
Even hardware-based wallets, which often resemble USB drives or key fobs and are more secure than their software counterparts since they are completely offline, can have vulnerabilities that a diligent attacker might find and exploit.
Tactics include the use of side-channel attacks, for example, where a cycbercriminal observes a system’s physical side effects, like timing, power, or electromagnetic and acoustic emissions to gain information about the implementation of an algorithm.
Guillemet explains that cybersecurity providers building digital asset solutions, such as wallets, need to help minimize the burden on the users by building security features and providing education about enhancing defense.
For businesses to protect cryptocurrency, tokens, critical documents, or other digital assets, this could be a platform that allows multi-stakeholder custody and governance, supports software and hardware protections, and allows for visibility of assets and transactions through Web3 checks.
Developing proactive security measures
As the threat landscape evolves at breakneck speed, in-depth research conducted by attack labs like Ledger Donjon can help security firms keep pace. The team at Ledger Donjon are working to understand how to proactively secure the digital asset ecosystem and set global security standards.
Key projects include the team’s offensive security research, which uses ethical and white hat hackers to simulate attacks and uncover weaknesses in hardware wallets, cryptographic systems, and infrastructure.
In November 2022, the Donjon team discovered a vulnerability in Web3 wallet platform Trust Wallet, which had been acquired by Binance. They found that the seed-phrase generation was not random enough, allowing the team to compute all possible private keys and putting as much as $30 million stored in Trust Wallet accounts at risk, says Bouzon. “The entropy was not high enough, the entropy was only 4 billion. It was huge, but not enough,” he says.
To enhance overall safety there are three key principles that digital-asset protection platforms should apply, says Bouzon. First, security providers should create secure algorithms to generate the seed phrases for private keys and conduct in-depth security audits of the software. Second, users should use hardware wallets with a secure screen instead of software wallets. And finally, any smart contract transaction should include visibility into what is being signed to avoid blind signing attacks.
Ultimately, the responsibility for safeguarding these valuable assets lies on both digital asset solution providers and the users themselves. As the value of cryptocurrencies continues to grow so too will the threat landscape as hackers keep attempting to circumvent new security measures. While digital asset providers, security firms, and wallet solutions must work to build strong and simple protection to support the cryptocurrency ecosystems, users must also seek out the information and education they need to proactively protect themselves and their wallets.
Learn more about how to secure digital assets in the Ledger Academy.
This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.
This content was researched, designed, and written by human writers, editors, analysts, and illustrators. This includes the writing of surveys and collection of data for surveys. AI tools that may have been used were limited to secondary production processes that passed thorough human review.