OrCam

2010: same founders, a completely different problem

OrCam was founded in 2010 in Jerusalem by Amnon Shashua and Ziv Aviram, the same pair of founders who had set up Mobileye eleven years earlier. The link between the two companies is not only biographical but truly technological: both deal with the question of how to make a small camera understand what it sees, and draw a useful conclusion from it in real time. The difference is in who receives the conclusion. In a car, the system brakes or warns; here, the user is a blind or visually impaired person, and the conclusion is read aloud into their ear. This move from a car to a person sounds modest, but it changes every engineering constraint along the way: there is no car battery, no room for hardware, no carmaker's production line assembling the product. Instead, there is a person who has to attach the device to their glasses and go out into the street with it. The meaning of a mistake is completely different too. In a car, a system that detects an obstacle that is not there causes needless braking; in a wearable device, a device that reads out the wrong name of the person standing in front of you creates social embarrassment that the user is the one to bear. The accuracy bar here is not lower; it is simply different.

The device: what it actually did

The main product, OrCam MyEye, was a small wearable device that attached magnetically to the arm of a pair of glasses and combined a tiny camera with computer vision and a speaker. It could recognize printed text and read it aloud, recognize the faces of familiar people and announce who was in front of you, read barcodes and identify objects. The user activated it with a simple gesture, pointing a finger at the text or object, and the device read the content directly into their ear, with no screen and no need for an internet connection. The MyEye 2.0 model, launched in December 2017, weighed only 22.5 grams. The device supported a range of languages, including Hebrew, English, French, German and Spanish, and more were added later. It is important to understand what this means in practice for the user: being able to read a restaurant menu, a sign at a station or a letter that came in the mail, without asking another person for help, is a much bigger daily difference than a list of features suggests. It is worth noting that this was not a single product but a family. Alongside MyEye, the company also developed OrCam MyReader, which focused on reading text aloud, and OrCam Read, a portable reading device also aimed at people with reading difficulties who are not blind at all.

The funding rounds, the valuation, and the IPO that never happened

Over the years OrCam raised money on a scale that shows the confidence the field enjoyed. As early as 2014, Intel Capital invested about $6 million in it, out of a total Intel investment of about $15 million up to that point. In March 2017 the total raised reached $41 million at a valuation of $600 million; in February 2018 another $30.4 million was raised at a valuation of $1 billion; and in March 2021 another $50 million at a peak valuation of $1.5 billion. At that time an IPO in New York was also planned, at a valuation said to be above $2.5 billion, and it never happened. At its peak the company employed more than 400 people. Again, it is worth separating the numbers: the valuation is an estimate investors agreed on, not money that came in. The actual money came to tens of millions of dollars, not billions. The fact that the IPO did not happen is not a technical detail. A company planning an IPO builds a spending plan that assumes the money will indeed arrive, and when it does not, it is left with a cost structure built for a different size.

Who actually buys a device like this

This is the heart of the business difficulty, and it deserves to be explained without softening. A dedicated device like this is expensive to make and therefore expensive to buy, while the population that needs it is relatively small and, for the most part, neither young nor well-off. As a result, the assistive device market does not behave like an ordinary consumer market: the purchase is often made through a third party (a health fund, an insurer, a rehabilitation organization or a government agency) rather than directly by the user. OrCam did work through such channels; among other things, in 2016 it signed an arrangement with the State of California's rehabilitation agency for eligible residents. But a channel like that adds a layer of approvals, public budgets and long decision times between the product and the person who needs it. The result is a hard combination: a target audience that is small to begin with, multiplied by a slow sales pace, against hardware development costs that do not come down. The gap between the value to the user and the payoff for the investor is one of the well-known problems in assistive technology in general. It is also the reason a large part of the field relies on public funding, on nonprofits, and on large companies that develop it as part of a broader product rather than as a business that stands on its own.

July 2024: the company shuts down its glasses line

In July 2024 OrCam announced it would stop developing its vision products and sell off the remaining stock. The reason it gave deserves attention, because it is unusually candid: according to the company, technological progress in image processing using language models makes further development of the vision products unnecessary. In other words, a smartphone with a multimodal model can now do much of what the dedicated device did, and for free. The rounds of layoffs told the same story: 62 employees in October 2022, about 16% of the workforce; another 50 in February 2024; and 100 in June 2024, about half of those who remained. By early 2025 the company had fewer than 50 employees. Big Tech 50, a publicly traded partnership that holds shares in private companies and reports their estimated value, cut its valuation of OrCam after the announcement from $333 million to $31 million. Instead, the company turned to developing hearing devices, under the name OrCam Hear. For people who already owned the device, the announcement that the line was closing was especially hard: an assistive device is not a gadget you can swap out, and people had built a whole daily routine around it. The new product, OrCam Hear, first shown at the CES trade show, is close in concept to a smart hearing aid (AI Hearing Aid): earbuds, a phone adapter and an app that learn each speaker's voiceprint within seconds and let the user choose whom to listen to, up to four speakers at once.

What this teaches about assistive technology

OrCam's story illustrates a familiar pattern, in an especially harsh version. A product that won worldwide recognition, highly experienced founders with a proven success, and funding at valuations of $1 billion and more, and yet it did not become a sustainable business. The reasons add up: a market too small to repay an investment in hardware, a high price that limits access precisely for those who need it most, and dependence on slow public funding channels. Then a fourth, ironic reason was added: the same progress in AI that the product had pioneered ended up making a dedicated device less necessary, because a multimodal model on a phone now does much of what the device did. There is good news in this for users (the capability has become much cheaper and more available) and hard news for companies. The turn to hearing devices, a much larger and more established market, looks like an attempt to draw exactly this conclusion. Still, the other side deserves to be said too. The device worked, people used it, and the proof that computer vision can serve a blind person in daily life remains true even after the company that proved it left the field.