In the first week of November, Temple, the company that Deepinder Goyal built after stepping down as chief executive of Zomato’s parent company, Eternal, changed what it had earlier claimed to be selling. Temple 01 is a screen-less optical sensor that weighs less than two grams and less than an inch wide. It sticks to the side of the head between the ear and tape with medical-grade tape. It claims to track the heart rate and something called a ‘Brain Flow’ index and keeps a score called Entropy.Temple plans to add a voice AI companion which picks up the voice of the user in the absence of both a camera and a microphone. Goyal calls it “humanity’s last personal device.” The first batch opened its pre-orders on November 1 at Rs 85,000 with every device being made to order and delivered in approximately three months. The central promise made by Temple is that a sensor on the side of the head of the wearer can show how well the brain is supplied with blood in real time and how hard the body is working. It has been pitched as a way to see how sleep, exercise, posture and caffeine affect the brain over days and weeks.On October 31, the company published a heart-rate study. The study was funded, run and analysed by itself. Some doctors on X called the claims made for the brain unscientific. The hepatologist Cyriac Abby Philips, popular as ‘The Liver Doc’, compared it to the fallen medical tech company Theranos. On November 5, Goyal responded with a post addressing concerned Indian doctors to clarify that Temple was a fitness tracker and a wellness device, and in his own words, “not a medical device.” The next day, he apologised saying that he “should not have bucketed all Indian doctors into one single box,” clarifying that his anger was reserved for a few ‘influencer’ doctors. On November 8, Goyal also posted a 10-point note saying that Temple was made for the world and not for India with the latter to account for under 1% of sales.TechCrunch reported in February that Temple had raised $54 million in a ‘friends-and-family’ round at a post-money valuation of about $190 million. Goyal claims that the company spent $30 million on development with two-third of that being his own.The independent studies that would settle the question of whether Temple works or not have not been undertaken. No social and legal arrangement lets a question like this to be settled by a 14-day returns policy. Temple sits between wellness and medicine, is a gadget and a data collector, and is also a personal device and a luxury good. On each line, the boundary is first drawn by the seller, then disputed online and left alone by the state.Temple has made three different kinds of claims and the evidence behind each of them is different.HeartThe first claim is about the heart. It cited a study dated October 31 which compared Temple with five flagship wearables. The study reported 43 healthy adults between the ages of 20 and 35 running on a treadmill wearing six devices all at once that included Temple, the Apple Watch Series 12, the Fitbit Air, the Whopp MG, the Garmin Forerunner 970 and the Oura Ring 5. All the readings were compared with a Polar h10 chest strap that reads the electrical signal of the heart, and is commonly used as a reference. On an average, Temple’s reading was 1.18 beats per minute away from the strap’s and had the smallest gap of the six, the study purportedly found. Apple Watch was 1.94 and Oura 5.21. But it must be noted that the test covered only treadmill exercise in a lab, overlooking sleep, daily life, other sports, older people or those who were not perfectly healthy. The study was conducted and analysed by Temple, and is yet to be independently replicated or peer-reviewed. Goyal has invited independent researchers to repeat the work that Temple does by making an offer to “loan Temples to any qualified researcher” with some reports suggested that he has also offered to fund studies of the kind. Heart rate is also the least contested thing Temple measures, and leading with it lent the company credibility for the claims that are harder to test.BrainThe second claim is about the brain. Temple sits over the superficial temporal artery and claims to report a proprietary Brain Flow Index that is meant to track changes in cerebral circulation. The idea grew out of what Goyal calls his Gravity Ageing Hypothesis, which proposes that the lifelong effort of moving blood up to the brain against gravity may shape the way we age, but, as Goyal himself admitted on November 20, 2025, it is “part of the hypothesis, not proven yet”. One of the company’s preprints compared the index with transcranial Doppler readings of blood velocity in the middle cerebral artery in 23 young, healthy adults, and reported a correlation of about 0.79 under the test conditions. The signal is relative and normalised, and an optical sensor at the temple can also pick up blood moving through the scalp. The study was posted on medRxiv in September 2026, however, it is not peer-reviewed yet. As per the disclosures, Temple financed the lead academic while one co-author is a consultant working for Temple. The authors admit that further studies are needed to show the signal is specific to the brain. The doctors’ objection is that the superficial temporal artery lies outside the skull and supplies the scalp as well as face, while the brain is fed by the internal carotid and vertebral arteries. Philips, for one, has said the device “can only measure blood flow and flushing to the face in humans.”The whole bodyThe third claim is about the whole body with Goyal saying that the main point of Temple is Entropy. A score ranging from 1 to 250, according to the company, shows how hard the body is working to shed heat. A low score would mean deep rest with a high one indicating peak exertion. In his November 8 note, Goyal acknowledged that nothing on Entropy, which is a proprietary score trademarked by Temple, has been published yet.The voice features rest on a further claim that the motion sensors pick up vibrations of the wearer’s own voice, so the device hears its wearers and nobody else in the room, which Temple calls ‘roomdeaf’ audio. No one outside the company has tested it.Goyal has said that different audiences buy at different levels of proof and told sceptics that they were free to wait. Earlier, he argued that a scam would be discovered within a few days, since Temple offers 14-day, no-questions-asked returns and that nobody spends $30 million to run a scam which pays nothing.Photo: Rens Dimmendaal & Banjong Raksaphakdee / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/Tiers of proof – and customersIn Biocapital, the anthropologist Kaushik Sundar Ranjan studied the biotech industries of the United States and India. He described how value in the life sciences is produced from promises before it is produced from products. The future is sold first and the evidence is expected to catch up. Goyal’s ladder of proof states that logic too. Standards of scientific proof become tiers of customers. The early buyer pays for the device and, with it, for the climb from preprint to peer review, and they carry the risk of a rung break.The returns policy has an even deeper limit. A two-week trial can tell a buyer whether the device is comfortable, whether the app crashes and whether the battery lasts for a week. It cannot tell them whether a number called Entropy corresponds to anything in their body, because they have no way of verifying it. Nobody can feel or verify immediately a wrong reading of their cerebral circulation. As per law, “we are not a medical device” is a regulatory act and a description. In India, medical devices are regulated under the Drugs and Cosmetics Act, 1940 and the Medical Devices Rules, 2017. Since 2020, the government has covered all medical devices instead of a notified list. If a product falls within the ambit depends on its intended use, which as per the Rules is tied to what the manufacturer claims the device is for, and the definition of a medical device is wide enough to include the investigation of a physiological process. Earlier this year, the Central Drugs Standard Control Organisation finalised guidance on medical device software which carves out “general wellness” software. It is rather interesting to take note of how such software might claim to maintain a general state of health and might even link a healthy lifestyle to a lower risk of chronic disease as long as it does not name a disorder or a pathological condition. However, software which measures or reports physiological values to diagnose or monitor does not count as wellness. When Temple is held against this, its position then depends entirely on the way the organisation words the product. A device which estimates brain circulation, produces a whole-body strain score and as per Goyal, gathers signals predicting health outcomes for years to come can fall within the wellness category as long as no one in the organisation names a disease since the seller writes the classification. There seems to be no public statement from CDSCO about Temple for now.Now, how different would things be if the regulator took the road not taken? The device would have to be assigned with a risk class ranging from A to D, and require a licence. The makers of the machine would need a quality management system which the Rules prescribe and for a function with no predicate device, the Rules generally expect a clinical investigation, approved by the regulator – unlike in this case where the corporation funded and tested on their own. It would need a stamp of approval before a licence is granted. The new software guidance adds duties on cybersecurity and post-market surveillance along with an algorithm change protocol which sets out the rules for what a company must do when its algorithm changes after approval, and that last duty is what matters. As per the manifesto of the device, it will always remain a project that is driven by ‘research’ and changes with its users, but a product that is designed to never settle down is precisely what an algorithm change protocol is for. And, as a wellness device, Temple is under no such protocol.Selling healthThe sociologist Andrew Abbott in The System of Professions argues that different professions compete with one another to gain control over particular problems which he labels as ‘jurisdiction.’ In the same way, who gets to diagnose, advise or treat is settled not just in a court of law, but rather, in the workplace and in public opinion. The quarrel on X is a jurisdictional contest over who gets to interpret the body: the physician, whose authority rests on clinical training and a licence, or the founder-engineer, whose authority rests on sensors, capital and an audience.While Temple is the current example that we look at, the same moves run through the entire industry. The clearest case of a regulator moving the line comes from the United States. In July 2025 the Food and Drug Administration sent a warning letter to Whoop, which makes a screenless fitness band, saying its Blood Pressure Insights feature was a medical device function. In January 2026 the FDA rewrote its general wellness guidance so that non-invasive wearables could estimate blood pressure, blood oxygen, glucose and heart-rate variability for wellness purposes. In June it dropped the Whoop warning letter after the company changed the feature and its labelling. Nothing about blood pressure changed in those 11 months, but the policy did. The traffic also runs the other way. Ultrahuman, a Bengaluru company in which Goyal has invested, sells the M1, which pairs a continuous glucose monitor, originally Abbott’s FreeStyle Libre, a regulated medical device made for people with diabetes, with an app pitched at “metabolic fitness” for people who do not have diabetes. Here a medical-grade sensor like Abbott’s FDA-cleared FreeStyle Libre is repackaged and sold to those without diabetes as a tool for ‘metabolic fitness,’ where the hardware remains medical, but then, the label is wellness.Gabit, founded by the former Zomato co-founder Gaurav Gupta and also backed early by Goyal, sells a smart ring that it says tracks more than 150 health markers, and has expanded into nutrition and skincare. In other words, the readings that the ring generates is then utilised to further recommend products which the same company sells, thereby, the blurring of lines between health and wellness is not restricted to expensive devices alone.Budget smartwatches sold in India for a few thousand rupees have long advertised blood oxygen and even blood pressure readings. Gizmore’s Gizfit 910 Pro was launched at Rs 2,499 with “SPO2, BP Monitoring” listed among its health features. The readings, which are estimates, are drawn from an optical sensor. The European Society of Hypertension’s working group noted in 2022 that cuffless devices should not be used to evaluate or manage hypertension in clinical practice owing to their accuracy not being validated, and as per Eugene Yang, a cardiologist at the University of Washington, these devices are inaccurate when it comes to measuring blood pressure or diagnosing hypertension. Staying wellOn the AI side, the US company Friend sells an always-listening pendant marketed as a companion, which critics have attacked for its constant audio monitoring. Temple defines itself against devices like this, and its manifesto opens by complaining that a microphone takes in everyone in the room. Privacy has become a product feature, which tells us how little the law currently guarantees it.Each device moves a little more of the work of staying well from public institutions onto individuals, and routes it through a purchase and, often, a subscription. Temple’s privacy policy, last updated on October 31, opens with a passage about the body as a record kept since birth, which until now nobody could read back to you. It is a good description of what self-tracking promises, and of the relationship it sets up, with the body as the text and the company as its reader.Photo: Marcin Wilkowski / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/The policy lists what the device measures: heart rate and its variability, respiratory rate, movement, steps, Brain Flow, energy expenditure and sleep stages. Users can add their height, weight, sex, food, mood, menstrual cycle and free-text journal entries, and the company keeps AI assistant chat histories. The policy says that inferences derived from all this are personal data too. The manifesto adds what is coming: chewing detection, snoring patterns, sleep position, clenching of the jaw during the night, REM sleep and core body temperature. The company makes strong promises. It says it will never sell health data, never use wellness data for advertising and never give insurers or employers access. It asks for separate consent before processing wellness data, and again before sharing it with research partners.Some of the fine print matters more. De-identified wellness data can be used to improve Temple’s algorithms and shared with research collaborators, and it can be kept after a user deletes her account. Data may be processed on servers outside India. Third-party generative AI providers are listed among the service providers whose tools may use a user’s prompts and device metrics, though the policy says they will not train on them. The manifesto, for its part, says conversations are end-to-end encrypted and never reach a model company. The two documents may be describing different features. A buyer cannot tell from them which promise covers what. And the policy, written a week before the voice companion was announced, does not list speech or voice among the things the device measures.The law of the landIndia’s Digital Personal Data Protection Act, 2023 has been notified along with its Rules, but the provisions that matter most here, on notice, consent, the rights of data principals, security safeguards and breach reporting, take effect only in May 2027. Temple 01 is due to ship in early 2027. Until May, the older regime under Section 43A of the Information Technology Act and its 2011 rules on sensitive personal data applied, and those rules single out health conditions, medical records and biometric information. When the DPDP Act’s main provisions take effect, Section 43A will be omitted, and the new Act has no separate category for sensitive data. Heart rate, voice and jaw movement will be personal data in the same way an email address is. Temple’s own policy shows what that gap looks like. In its annex for European users it describes wellness data as sensitive and notes that the European GDPR requires a second, separate legal condition before health data can be processed. In India its promise of separate consent is a company commitment, not a statutory duty. The India annex tells users they can complain to the Data Protection Board, and adds in brackets that this applies once the provision is in effect.Temple, Goyal says, is made for the world. Priced at Rs 85,000, Goyal expects India to account for under 1% of sales and has asked people not to back the company out of nationalism. It is a candid answer, and it raises the question of who the humanity in ‘humanity’s last personal device’ is meant to be.Pierre Bourdieu’s Distinction showed how tastes and bodily habits mark class, and how what looks like personal preference does the work of social sorting. Health practices are among the strongest of these markers, because they are carried on the body. Temple is worn on the temple, about the most visible place a wearable can go. The cricketer Shubman Gill was reported to be wearing one during an India-New Zealand match. The footer of the company’s website says it pays nobody to wear or promote the device. Perhaps visibility is part of what is being bought.Temple may turn out to measure what it says it measures. Its founder has money, patience and engineers. Some of what now looks like hype may become a real contribution. But right now nobody with public authority is in a position to find out whether that will be true.Ankush Pal is a sociologist and writer interested in urban spatiality, informal labour, public transportation, and temporality.