The U.S. 4K Blu-ray player market has become fairly easy to understand. Panasonic still sells excellent conventional players, Magnetar will happily sell you something substantially more ambitious, and almost everyone else has either left the building or is standing suspiciously close to the exit.

Japan apparently did not get that memo.

Panasonic’s new DMR-4X1010 is not merely an Ultra HD Blu-ray player. It combines a 10TB internal hard drive, 11 television tuners, automatic multi-channel recording, Ultra HD Blu-ray playback and recording, cloud storage support, remote viewing and extensive program management inside one component.

In other words, Japan gets a 10TB home video command center, while America gets another notification asking whether we are still watching.

Related Reading:

This Is Not Just a 4K Blu-ray Player

Calling the DMR-4X1010 a disc player misses most of the point.

Panasonic has equipped its flagship new DIGA recorder with 10TB of internal storage and 11 tuners covering Japan’s terrestrial digital, BS/CS satellite and BS/110-degree CS 4K broadcasts. Eight tuners are dedicated to automatic recording, two can switch between automatic and conventional recording duties, and one is reserved for conventional recording.

With the appropriate recording settings, Panasonic says the DMR-4X1010 can automatically retain as many as eight HD channels for approximately 28 days, while an additional allocation can capture either one 4K channel for approximately 13 days or two additional HD channels for roughly the same period.

Older recordings can then be automatically deleted as space is required.

10TB Is Only the Beginning

The internal hard drive can store approximately 390 hours of native 4K broadcast material at 33Mbps, according to Panasonic. Longer-duration compression modes stretch that dramatically further, although image quality naturally becomes part of the trade-off.

Users can also archive compatible recordings to recordable Blu-ray media. The DMR-4X1010 supports BD-RE discs up to 100GB and BD-R media up to 128GB, along with recordable DVD formats.

That combination of enormous local storage and removable physical media already distinguishes the DIGA from virtually anything available to U.S. home theater buyers.

Panasonic then added the cloud.

Cloud Recording Without Surrendering Everything to Streaming

Panasonic describes one of the DMR-4X1010’s major new features as Cloud Recording, although the terminology requires some explanation.

The machine records television programming locally first. Selected recordings can then be copied to Microsoft OneDrive or Google Drive, allowing users to free space on the internal hard drive, protect important recordings from a hardware failure and access stored programs remotely.

Panasonic says as much as 10TB of recorded programming can be stored in the cloud, subject, obviously, to whatever storage plan the user has purchased from Microsoft or Google.

Cloud-based recordings can also be viewed from a smartphone through Panasonic’s DIGA Anywhere app. Panasonic specifically cautions that cloud services can change, suffer outages or disappear entirely, so the feature should not be confused with guaranteed permanent archival storage.

Physical media just quietly cleared its throat.

The DMR-4X1010 also introduces playlist editing, allowing users to assemble favorite portions of recordings without deleting the original program, along with custom program tags for sorting material by performers, music or other criteria.

That sounds almost quaint until you remember how often a show disappears from a streaming service because somebody in a conference room changed a licensing agreement.

And Yes, It Still Plays 4K Blu-ray Discs

For all of the recording technology packed inside, the DMR-4X1010 remains a serious disc machine.

It supports Ultra HD Blu-ray, Blu-ray, Blu-ray 3D, DVD-Video and CD playback, along with BD-R/RE and multiple recordable DVD formats. Panasonic also includes its 4K Real Chroma Processor, HDR brightness adjustment and 4K direct chroma upconversion.

High-resolution audio file playback includes WAV up to 32-bit/384kHz, AIFF up to 32-bit/384kHz, FLAC up to 24-bit/192kHz, ALAC up to 32-bit/192kHz and DSD up to 11.2MHz.

Connectivity includes two HDMI outputs, three USB ports, Ethernet and built-in Wi-Fi. There are no analog, optical or coaxial audio outputs, which tells you fairly clearly how Panasonic expects the DMR-4X1010 to be used.

Panasonic DMR-4X1010 Specifications

• Product Type: 4K Blu-ray recorder/DVR
• Internal Storage: 10TB HDD
• Tuners: 11 total
• Automatic Recording: Up to 8 HD channels for approximately 28 days
• Additional Automatic Recording: Up to 1 4K channel or 2 HD channels for approximately 13 days
• 4K Native Recording Capacity: Approximately 390 hours at 33Mbps
• Cloud Services: Microsoft OneDrive, Google Drive
• Maximum Supported Cloud Recording Storage: 10TB
• Disc Playback: Ultra HD Blu-ray, Blu-ray, Blu-ray 3D, DVD-Video, CD
• Recordable Blu-ray: BD-RE up to 100GB; BD-R up to 128GB
• Video Processing: 4K Real Chroma Processor, HDR brightness adjustment, 4K direct chroma upconversion
• Hi-Res Audio: WAV, FLAC, AIFF, ALAC, DSD
• HDMI Outputs: 2
• USB: 3; 1x USB 2.0, 2x USB 3.0
• Networking: Ethernet, built-in Wi-Fi
• Power Consumption: Approximately 54 watts
• Dimensions: 430 x 66 x 239 mm (16.9 x 2.6 x 9.4 inches)
• Weight: Approximately 4 kg (8.8 pounds)
• Power: AC 100V, 50/60Hz
• Japan Price: ¥434,500 including tax through Panasonic’s online channel
• Availability: Japan, late October 2026
• U.S. Availability: Not announced

The Bottom Line

The Panasonic DMR-4X1010 is unique because it combines an 11-tuner DVR, 10TB of local storage, recordable Blu-ray support, Ultra HD Blu-ray playback, cloud archiving, and remote viewing in one component. Very few manufacturers are still treating physical media, broadcast television, and local storage as parts of the same home entertainment system.

What is it missing? For North American buyers, the obvious answer is compatibility. The DMR-4X1010 is built around Japanese broadcast standards, uses B-CAS, runs on 100V power, and has no U.S. version announced.

That makes this a product Japanese home theater and physical media enthusiasts get to enjoy while the rest of us admire it from afar. Panasonic has built exactly the kind of all-in-one video machine many enthusiasts still want.

Price & Availability

Panasonic announced the DMR-4X1010 in Japan on September 14, 2026, with availability scheduled for late October 2026 at ¥434,500.

A less expensive DMR-4X410 version (¥209,000) reduces internal storage to 4TB while retaining 4K tuner capability and many of the same new cloud and organizational features.

There has been no announcement regarding U.S. or European availability.

Related Reading:

Financial markets have always relied on timely information, and the drive for timeliness has always adapted to the latest technology. From clipper ships transiting the oceans to telegraph wires connecting cities to fiber-optic cables conducting trades in microseconds, traders have embraced any advantage to get the most up-to-date information. Indeed, the history of finance is really a story about how fast you can move information and who controls the interface.

It’s only natural that people also figured out a way to profit by supplying that market intel. In 1841, for example, the Mercantile Exchange (predecessor to Dun & Bradstreet) began selling proprietary business information to its U.S. clients. The following decade, Paul Julius Reuter began selling news services and stock price information. To supplement the company’s telegraph dispatches, he sent pigeons between Aachen, Germany, and Brussels; each bird carried a cylinder containing slips of paper with that day’s stock prices. In 1867, an inventor named Edward Calahan introduced the first telegraphic ticker-tape machine, which spooled out stock price information in near real time; Thomas Edison improved upon the design with his patented version in 1871.

The Dow Jones Industrial Average debuted in 1896 as an index of 12 key businesses listed on U.S. stock exchanges. It included gas, oil, coal, and electric companies, as well as enterprises dealing in leather, rubber, and tobacco. Messengers delivered quotes from the trading floor to brokerage offices, while stock tickers kept investors informed of prices. By the time New York City held its first official ticker-tape parade, in 1919, telegraphy in Western Europe and the United States had become the chief means for quick transmission of vital stock information.

In 1960, the first paperless financial service debuted, when Quotron introduced its electronic screens for displaying market quotes. Over the next two decades, other companies rolled out similar innovations for distributing financial news and data.

So when Michael Bloomberg decided to enter this well-established industry in 1981, the big question was: How would his new company stand out?

The Birth of the Bloomberg Terminal

Bloomberg had cofounded Innovative Market Systems (IMS) after being fired from the investment bank Salomon Brothers. Landing on his feet with his US $10 million equity payout and joined by former Salomon colleagues Thomas Secunda, Duncan MacMillan, and Charles Zegar, Bloomberg pursued his belief that Wall Street would pay a premium for specialized financial data. He’d earned an electrical engineering degree from Johns Hopkins University and an MBA from Harvard, and he’d built computerized financial systems for Salomon. IMS focused on developing a computer terminal that not only provided up-to-date information but could also do instant quantitative analysis based on historical data.

Michael Bloomberg believed Wall Street would pay a premium for access to specialized financial data. Karjean Levine/Getty Images

At the time, most financial data still circulated through telephone calls, printed price sheets, and specialist publications, and analysis involved a fair amount of gut instinct guided by human expertise. Companies such as Reuters and Dow Jones provided subscription-based services for access to business news. But traders still had to assemble information from multiple sources and perform their own calculations and analysis.

IMS proposed an integrated system with a single interface. Its Market Master terminal consisted of a monochrome CRT monitor, a custom keyboard, and a communications/controller unit that connected to the company’s private network. At launch, it provided only U.S. government bond prices and bond-calculation tools, but the dream was much bigger: a dedicated terminal that would sit on a trader’s desk and run different market scenarios, produce yield curves, and support investment calculations.

IMS initially had just one client, Merrill Lynch, which invested $30 million (about $110 million today) in exchange for a 30 percent stake in the company and exclusive rights to the terminals for five years; Merrill waived that right in 1984. The first 22 Market Master terminals were delivered to Merrill in 1982, in the middle of a global recession. The timing was fortuitous. Worldwide, stock markets were transitioning to electronic trading, and the U.S. Federal Reserve was allowing more freely floating interest rates. Bond prices were more volatile, and investors were eager to figure out how to value them accurately. Bloomberg’s specialized financial terminals provided the data and the analytical tools to process and comprehend those sweeping changes.

Five years after its launch, IMS rebranded as Bloomberg LP and expanded its clientele, and the Market Master became known as the Bloomberg Terminal.

How Did the Bloomberg Terminal Work?

The Bloomberg Terminal’s keyboard was designed with traders and analysts in mind. The function keys were color-coded and given labels specifying their usage, so that users didn’t have to remember. The original keyboard, affectionately referred to as “the Chiclet,” was hand assembled. A cable ran from the keyboard to the Bloomberg Controller, which had a dedicated phone line to connect to a local hub. The internet wasn’t commercially available yet, so the company basically built its own closed network, with centralized computers that maintained large databases and performed most of the calculations. Commands entered on the keyboard sent a request to the hub, which processed the information and sent back the result.

The Chiclet keyboard for the Bloomberg Terminal was introduced around 1983. Although it looks like a generic keyboard, its function keys were finance-specific hot keys.National Museum of American History/Smithsonian Institution

Hot keys let traders easily call up data on government securities, corporate debt, and currency markets, among other things. A series of keystrokes would pull up other historical and real-time data, run an analysis, or place a trade. Learning how to use the terminal and digest the vast amount of information, which was presented mostly in tabular form, became a rite of passage for users.

In 1990, Bloomberg added a trackball to the keyboard, which helped the user navigate the multiple windows and menus typically displayed on screen. Two years later, the keyboard gained a built-in speaker, to support multimedia information; this design also included telephone, headphone, and microphone jacks. One of the most popular features was Instant Bloomberg, which allowed users to chat directly with fellow Bloomberg Terminal users over the proprietary network. By 1996, Bloomberg had keyboards that supported 23 different languages. In the early 2000s, the company began incorporating biometric authentication for terminal login, via a fingerprint reader on the keyboard.

As the company’s business model evolved, the Bloomberg Terminal added services well beyond its initial offerings. In 1990, for example, worried that Dow Jones would stop providing access to its news stories, Bloomberg set up its own news service. It recruited Wall Street Journal reporter Matthew Winkler to oversee a dozen reporters; their stories on market and securities movements used graphs and calculations that served as advertisements for the terminal’s capabilities. These days, the Bloomberg news empire includes Bloomberg Businessweek, Bloomberg Radio, and Bloomberg Television.

Bloomberg’s subscription-based financial model included the leasing of a Bloomberg Terminal with its specialized keyboard and other hardware, access to a dedicated private network, and a suite of services. In 1999, a subscription to a single Bloomberg Terminal cost $1,600 per month with a minimum two-year contract and a discount on each additional terminal. Today the annual price is upwards of $32,000 (trending a little below inflation). In 1995, the company launched a suite of “Open Bloomberg” software products that ran on the customer’s own PC; five years later, it stopped leasing dedicated terminals. Current customers also have access to mobile applications that allow terminal functions to run on phones and tablets. Today, “Bloomberg Terminal” has come to refer to the integrated data, analytics, news, communications, and trading environment.

The Legacy of the Bloomberg Terminal

Although the shift away from dedicated terminals was a logical response to the rise of the internet and publicly available market data, it altered the material culture of financial work. For nearly two decades, Bloomberg Terminals commanded an aura of power and financial prowess. They were emblems of market mastery, with a brand that was distinct from other office computers. With Open Bloomberg, users were no longer tied to a single desk or a fixed set of monitors.

And so, cast-off Bloomberg Terminals found their way into museum collections. They’re a physical embodiment of the ethereal nature of financial markets, and a manifestation of mathematical calculations, network infrastructure, and business culture.

The Bloomberg keyboard used by “Bond King” Bill Gross has his login and password taped on the front.National Museum of American History/Smithsonian Institution

The Smithsonian Institution’s National Museum of American History has a number of Bloomberg keyboards in its collection, but my favorite is object number 2014.0012.02, which was used by “Bond King” Bill Gross during the 1990s and 2000s at Pacific Investment Management. Gross had cofounded PIMCO in 1971 and built it into a $2 trillion bond investment firm. I especially love that Gross taped his login and password directly on his keyboard, which makes the object more relatable. I may never know what it’s like to manage billions in assets from a Bloomberg Terminal, but I absolutely understand the trial of remembering my passwords.

Part of a continuing series looking at historical artifacts that embrace the boundless potential of technology.

An abridged version of this article appears in the October 2026 print issue as “The Keyboard That Moved Markets.”

Boston Dynamics spent this morning showing Atlas a new pair of hands, and the first thing that stands out is a missing digit. There is no pinky. Four fingers sit on a palm a little larger than an average human hand, and the assembly is a little stronger than the gripper it replaces. Engineers reached that count with a blunt test: they taped a pinky to a ring finger and spent a day trying to work. The missing finger never blocked the jobs they cared about. Chief technology officer Zack Jackowski put the lesson in plain shop language. The best part is no part.

Older Atlas grippers, called GR2, had only seven degrees of freedom and were primarily meant to close around an object. The transition to newer models, such as the GR3, takes a significant leap to 13. Add in a thumb with four joints that can meet any other finger, as well as three joints on each of the other fingers that allow them to spread out and aim with precision. Every joint is powered by the same type of motor, which is snugly tucked inside the finger, eliminating the possibility of frayed external wires as the hand moved. This allows the hand to read the force of touch directly through the motors. The thick touch sensors on the fingertips and palm can detect even the smallest of touches that the motors may not be sensitive to.

Clips from the latest video demonstrate how beneficial all of those extra joints are. A thumb can slide along another finger’s length or width when gripping a small object. Gripping with the thumb and any of the digits is also an option, and a 3-point grasp can be used to hold a tool steady while yet allowing you to move it about. Atlas has already demonstrated excellent performance by pressing triggers on drills, torque drivers, grinders, nail guns, and welding torches. If a grasp begins to slip, the fingers can recover without dropping the object being held. Atlas has a payload rating of over 100 pounds, and we’ve seen Boston Dynamics use the robot to walk with a loaded small fridge on its back, so these hands better be up to the task.

Alberto Rodriguez, who leads Boston’s robot behavior team, insists that hand design is all about giving up one thing to acquire another. Of course, this implies that strength, durability, affordability, repairability, and sensing are all competing for the same limited amount of space. Many other businesses are attempting to produce something that closely resembles a human hand, including tendons flowing through soft fingers. Those hands can look quite convincing in a movie, but they’re also fragile, prone to straining or cracking, and expensive to construct in large quantities. Atlas is willing to give up the cosmetic appearance in order to achieve something that functions.

Rodriguez claims that the hand design is not horrible, but it does mean that the fingers can travel farther than a human hand would ever allow. Even so, he believes that by practicing in simulation, we can discover new ways to use those extra motions, and that because the hand is near enough in size to a human hand, we can still learn a lot from watching a human perform a task, even if it’s just the general outline of what has to happen.

The other part of the issue is getting the simulation correct. The motors are simple enough that we can replicate them fairly realistically, and training sessions can alter a variety of factors like as torque, friction, item form, and how something would drop on your palm. That way, when we come to the actual thing, we can be confident that the hand will continue to perform as planned. On the other hand, the motors’ quick readings provide the majority of the feedback we require, with the pads detecting even the tiniest touches. Swapping out a failing motor is as straightforward as swapping out an entire unit, so we can keep the hand going even if one of its components fails.

After adapting Stephen King stories like Gerald’s Game, Doctor Sleep, and The Life of Chuck, horror filmmaker Mike Flanagan has taken on the author’s breakout novel in Prime Video’s Carrie series. Many people are already familiar with the story: a shy teenager with telekinetic powers faces vicious bullying from her peers until she unleashes psychic vengeance on Prom Night.

While Carrie has been adapted into multiple feature films, most notably by Brian De Palma in the ’70s, Flanagan expands the original novel into an eight-episode Prime Video miniseries set in the present day. At this point, one may wonder why another Carrie adaptation is needed. But with its layered story, realistic characters, and timely depiction of school bullying, Flanagan’s Carrie delivers one of the best Stephen King adaptations ever.

Carrie expands King’s story and characters

Fans already know Carrie White as a girl brought up by her abusive, fanatical mother. However, Flanagan’s series builds upon that as we follow Carrie from childhood to adolescence with a new story. In the show, Carrie, played by Summer H. Howell, had never left her house and was homeschooled by her paranoid, conspiracy-based parents. On her first day of public school, she enters a whole new world, having never used a smartphone or even a computer. Because Carrie has grown up scared of the outside world, her journey to high school becomes even more terrifying for her and the audience.

The series doesn’t just change Carrie. This eight-episode story spends much of its time exploring the lives and struggles of Carrie’s mother, her classmates, and the adults who knew her in the town of Chamberlain. Even villains like Margaret White (Samantha Sloyan) and Chris Hargensen (Alison Thornton) evoke more sympathy because of their nuanced characters and troubled backgrounds. We then see how the two of them became abusive, mean-spirited people, showing how bullying can begin not at school, but at home.

Overall, the town of Chamberlain feels more like a character that changes throughout the series. The scope and depth of the story only heighten Chamberlain’s tragedy when Carrie destroys the town and kills many of its residents.

Flanagan delivers a fresh, faithful interpretation of Carrie

From Oculus to The Haunting of Hill House to The Fall of the House of Usher, Mike Flanagan has often told his horror stories in a nonlinear perspective, playing with audiences’ perception of time and reality. While this makes for unconventional retellings of classic stories, it fits perfectly with Carrie. King’s novel actually presents its narrative through in-universe books, news stories, and investigative reports about Carrie and her rampage on Prom Night, featuring testimonies from students, teachers, and locals who knew her.

As a result, Flanagan’s series tells the story much like a true-crime docuseries, with Kate Siegel’s character interviewing Prom Night survivors and exploring psychic phenomena throughout human history. It actually feels like we’re learning things about the story that weren’t covered before in previous adaptations. Since general audiences have seen multiple Carrie films told the same way, Flanagan’s series stands out with a more distinctive approach.

The series explores adolescence, bullying, and violence in the digital age

By setting King’s narrative in the present day, Flanagan’s Carrie expands on the source material by tackling issues affecting modern American youth. The show addresses problems like cyberbullying on social media, with Carrie becoming the center of a vicious meme account attacking everyone at her school. Sparking debates about misinformation and freedom of speech, this new storyline unleashes a firestorm of controversy that nearly destroys Chamberlain even before Prom Night.

Carrie also acts as a parable for mass shootings in schools, which have become more common in the decades since King published his novel. This can make Carrie’s psychic rampage at her school hit much closer to home with audiences. However, that makes King’s story as relevant today as it was 50 years ago.

Carrie ultimately shows how schools have become much more dangerous in modern America. At the same time, the show conveys how adults have failed the younger generation by mishandling parenting and education, making Chamberlain’s destruction seem even more inevitable.

Overall, Flanagan’s adaptation of Carrie makes some drastic changes to King’s original narrative. This can throw off fans of the latter, but the series justifies its existence by delivering a unique reimagining that tackles issues affecting modern schools. Though the story leaves the door open for another season, it preserves the horror and tragedy that made King’s story a groundbreaking classic.

Mike Flanagan’s Carrie premieres on Prime Video on October 7.

It’s disheartening how much power gets generated and then promptly lost as it travels through grid networks. This leaking of electricity happens when it vanishes as heat as well as when it is pilfered by thieves and nonpaying customers.

More than half of the countries that track these metrics lost at least 10 percent of their electricity in 2023, according to the World Bank. Losses topped 20 percent for 24 of those nations. Two countries lost more than half of what they generated.

With numbers this high, cutting down on losses makes sense. Electricity demand is rising beyond what many grid operators can supply; reducing waste would help meet some of that demand without having to build new power plants. Plus, when the power comes from fossil fuels, any loss means emitting even more greenhouse gases into the atmosphere. And electric losses hit the bottom lines of power providers, which ultimately pass those costs on to everyone else.

The trouble is, reducing electricity losses is a hard and expensive process that takes a long time. Typically, the less maintained the grid infrastructure, the more electricity that’s lost. And the more fragile the region’s law enforcement and government, the more prevalent the power theft. Natural disasters and war make things worse.

Fixing a power grid requires a systemic approach across many sectors. There’s no one technology that will solve the problem. At the outset, the obstacles to success may feel insurmountable. Equipment across entire grid networks must be updated. Multiple arms of government must agree to reforms and coordinate to ensure power providers are set up to succeed. Regulations must be written or revised, investments made, cultures changed.

The city of Delhi did all those things. Over the past 25 years, it cut its electricity losses from about 50 to 5 percent. How the city pulled off that impressive feat is the focus of “The Epic Comeback of Delhi’s Power Grid” by Mini Shaji Thomas, an electrical engineer at the university Jamia Millia Islamia who has lived in Delhi since the 1990s. She gives us a view from the inside—as a resident and a power systems expert.

Delhi is a shining example, but some other regions have significantly lowered electricity losses over the last quarter century too. The country of Georgia went from losses of over 16 percent in 2002 to about 8 percent in 2023. In Singapore, losses dropped from 6.6 percent to a nearly nonexistent 0.2 percent over the same time period.

Global Electricity Theft Crisis

But there are many parts of the world where electricity losses remain a problem or have gotten worse. In Jamaica, where power theft is rampant, losses have hovered between 21 and 28 percent for years. Argentina’s losses nearly doubled between 2015 and 2023, going from an all-time low of about 12 percent to an all-time high of nearly 24 percent. The main problem: Transmission and distribution companies lacked the capital to maintain and upgrade their networks, which left equipment operating under stress. A delay in the installation of smart meters has allowed thieves to more easily siphon power and tamper with meters.

Thomas says she hopes her account of Delhi’s grid comeback will serve as a blueprint for others. It’s possible to replicate the sweeping changes Delhi made, she says. But it “requires a concerted effort from all stakeholders, customers, the utility, the government, and their employees.”

The AI startup Photon is so sure that agents will eventually come to replace mobile apps that it held a funeral for the latter — yes, a real funeral in a church, with speeches and everything. Photon, which helps developers build agents that work over iMessage and WhatsApp, now has $4.5 million in seed funding to develop its products. The raise comes after the company signed up over 40,000 developers grew its revenue 10x in just four months, it says.

Photon co-founder and CEO Daniel Tian said the “app funeral,” held in San Francisco on September 17, was also a developer day for the startup, featuring panels from Vercel, Stripe, and OpenAI. It also included an actual coffin for app icons.

The move to host an app funeral was obviously somewhat tongue-in-cheek, but Tian believes the app era is coming to an end. “Honestly, it’s going to take a while, but I think, directionally, it’s inevitable,” he said of the shift from apps to agents.

The thesis behind the startup, as you may have guessed, is that users no longer want to install new apps. Instead, they want access to agents that work through the messaging apps they already use, like iMessage and WhatsApp.

Meta’s Muse, whose app has soared to No. 1 on the app stores, is putting that theory to the test. But even if Muse comes to dominate the consumer AI space, Tian sees a market for serving developers who want their applications to work over messaging services. That’s why Photon is offering a package of tools, including a unified API, an extensible channel framework, a command-line interface (CLI), and an observability suite. Combined, they allow developers to build and operate an agent experience that works over iMessage, WhatsApp, Telegram, SMS or RCS, email, voice, and more.

The idea for Photon emerged from the time co-founders Tian and Ryan Zhu, the CTO, spent building consumer apps as students, often at hackathons. Every time they finished a new app, they hit the same bottleneck: getting people to discover it. By building an agent that worked within iMessage, they realized they could reach people through an app those people already use every day: Apple’s Messages.

Tian, for instance, had built a bot that would text his friends back as him over iMessage. When he and Zhu made the technology open source on GitHub, thousands of developers adopted it. So they went all in. Tian dropped out of UPenn’s M&T program, and Zhu left high school, later becoming a visiting student at MIT’s Media Lab.

Today, the open-source version of Photon still accounts for 98% of its use.

This April, the founders decided to create a managed platform and sell access to it via subscription tiers. This hosted version of the product offers 99.95% uptime, which is much higher than what you’d get if you were just running the open-source version on your local Mac for a hobby project. It’s also SOC 2 Type II and HIPAA-compliant, which opens the door to healthcare agent use cases.

To date, the startup’s platform has attracted over 40,000 developer sign-ups and now includes paying clients. (Any usage beyond 10 users requires moving out of the free tier to one of the three paid tiers, we’re told.) The company isn’t sharing revenue figures but said that, in addition to the 10x revenue growth, it’s seeing less than 3% churn, and its messaging volume has grown by 5x in the last month alone.

Customers using Photon include Corgi Insurance; social introductions app Boardy; Gen Z dating platform Ditto; business banking platform Rho; finance AI assistant Fliptexts; AI email client Slashy; and others building agents for fintech, consumer AI, and other vertical use cases.

Photon’s technology integration partners also include companies such as like Vercel (Eve and ChatSDK) and Nous Research, whose Hermes agent uses Photon as its default iMessage layer. Photon is also the layer under Tencent’s QClaw and NanoClaw. Plus, Photon has seen framework and infra integrations across LangChain, Mastra, Convex, Render, Railway, and Telnyx, among others.

Vercel is also one of Photon’s new backers. The startup’s seed round was co-led by Gradient and A* with participation from Vercel, HongShan, Z Fellows, Llama Ventures, Karman, and other angel investors.

Across both the hosted and open-source versions, the technology reaches millions of end users, Tian estimated.

While today, Photon focuses on agents connecting with humans, Tian believes there’s potential to grow into an expanding market that involves connecting agents to other agents.

“I think, in the future, agents will be able to find you the agent that can do the right things — so it’s going to be the A-to-A communication layer,” Tian said. For instance, if a person was chatting with their AI assistant making travel plans, the agent could pull in other agents when needed, like a flight-booking agent or hotel-booking agent. “It’s been pretty powerful to observe the whole A-to-A-to-P [agent-to-agent-to-person] space,” he added.