Every morning, millions of people put small white capsules in their ears, and not a single one is connected to a phone by a wire. Behind that gesture lie thirty years of engineering decisions, beginning with an irritation: too many wires on a Swedish engineer’s desk.
By the late 1980s, the mobile phone was still a luxury item and weighed as much as a small book. But even then, a web of cables was beginning to grow around it. Headset — one wire. Computer connection — another, through the RS-232 serial port, a standard developed back in the 1960s for communicating with teletypes. Peripheral synchronization — a third. Every cable needed its own connector, its own protocol, its own length. If you worked at an Ericsson office in Lund, southern Sweden, you were literally sitting in a nest of wires, and every new device added another loop to that nest.
Nils Rydbeck, Ericsson’s chief technology officer, began thinking as early as 1989 that short cables between devices were an engineering absurdity. Long-range radio communication already existed: cellular networks worked, pagers beeped in pockets. But no one was seriously tackling radio communication at arm’s length — one meter, two, ten. It seemed too small a problem for a serious company. Why build a radio bridge where you could simply plug in a connector? Around the same time, Swedish physician and inventor Johan Ullman received two patents — in 1989 and 1992 — for a wireless headset and hands-free communication. The idea was in the air, but it had no industrial implementation: those two inventions would later form the basis of the project’s requirements.
The problem with RS-232 went deeper than inconvenience. A serial port is a one-lane road with strict traffic rules. Data-transfer speeds were limited, every device required a separate physical connection, and the connector itself — the bulky trapezoidal DB-25 or the somewhat more compact DB-9 — fit no conceivable idea of mobility. In a world where phones were getting smaller every year, dragging around a tail of cables with connectors the size of a postage stamp looked increasingly ridiculous.
In 1994, a project began within the Ericsson Mobile Communications division under the working name MC Link. Tord Wingren was responsible for the specification, while two engineers — Jaap Haartsen and Sven Mattisson — led the technical development. The task was modestly framed: create a cheap, low-power, compact radio link capable of replacing the cable between a mobile phone, a headset, and a computer. No global standard, no headphones for runners — just get the wires off the desk.
Haartsen and Mattisson chose the 2.4 GHz band — the so-called ISM (Industrial, Scientific and Medical) band, a range of frequencies that can be used without a license in most countries around the world. This was a fundamental choice, with very concrete consequences. An unlicensed band meant any manufacturer could build such a radio module into its device without negotiating with regulators in every country. But microwave ovens, industrial heaters, and — a little later — Wi-Fi networks already operated in the same band. The 2.4 GHz airwaves were, and remain, a communal kitchen: noisy, crowded, and someone is always reheating soup.
To survive this chaos, the engineers used frequency-hopping spread spectrum — FHSS. The transmitter and receiver jumped in sync between 79 channels, making 1600 hops per second. Interference on one channel? No problem: a fraction of a millisecond later, the device is already on another. It is like having a conversation in an overcrowded café where the speakers move to a new table every second: eavesdropping is difficult, and random noise at the next table does not have time to ruin the sentence.
Transmitter power was deliberately kept low — around one milliwatt in the basic class. Low power limited the range to a few meters, but that was exactly what was needed: communication at arm’s length, with minimal battery drain. The architecture was built on a “master–slave” principle: one device, the master, controlled a network to which up to seven active slaves could connect. This miniature network was called a piconet — from the prefix “pico,” meaning something unimaginably small. I admire the honesty of that name: the engineers did not pretend they were building global infrastructure. They were building a radio network the size of a desk.
A technology created inside a single company has a ceiling. Even if Ericsson released a phone with a radio module and a headset with the same module, it would remain a branded toy. For the RS-232 cable to truly die, a phone from one manufacturer had to be able to talk to a laptop from another and headphones from a third. And that required a standard. A standard is when competitors sit down at one table and come to an agreement.
On May 20, 1998, five companies announced the creation of the Bluetooth Special Interest Group (SIG). On Ericsson’s side were the engineers who had started MC Link. On IBM’s side was Adalio Sanchez, who saw wireless peripherals as the future of laptops. On Intel’s side was Stephen Nachtsheim, who understood that processors were meaningless without convenient ways to connect devices. Nokia wanted its phones to talk to the world without wires. Toshiba brought experience in portable electronics. Five companies, four countries, three continents — and all of them were direct competitors in the market.
Agreeing on compatibility is not a technical task but an almost diplomatic one. Every participant wanted the standard to account for its own architecture, its own use cases. Ericsson, naturally, pulled the blanket toward mobile phones. IBM and Intel were thinking about computers. Nokia — about its ecosystem. But everyone understood the arithmetic: a proprietary standard from one company is doomed to a niche, while an open standard from five companies means tens of millions of devices immediately. The SIG was structured so that any manufacturer could join. By December 1999, more than 1,200 companies already belonged to the SIG — in less than two years since its launch. Openness proved a more powerful weapon than any technical specification.
The MC Link project had a technical name unfit for marketing. When the time came to name the standard, Jim Kardach of Intel proposed the word Bluetooth. The story was suggested to him by Sven Mattisson: in Toronto pubs, after their presentation had received a cool reception at a HomeRF meeting, Mattisson casually retold Frans G. Bengtsson’s historical novel The Long Ships — his own little lecture on Vikings. Back home, Kardach ordered some nonfiction on Scandinavian history and, in the pages of a book by Gwyn Jones, came across a photograph of the Jelling rune stone depicting Harald Bluetooth (Harald Blåtand), a 10th-century king who united Norway and Denmark. The metaphor fit the task exactly: the program was meant to bring together the PC and telecommunications industries — just as Harald had brought together his lands.
A caveat is needed here. Harald Bluetooth did not “unite Scandinavia” in the sense in which we understand the unification of states today. He was a medieval ruler who, through force and diplomacy, gathered the Danish lands and part of Norway under his hand. It was a metaphor, not a one-to-one historical parallel. But the metaphor worked — and the name, intended as temporary and working, remained.
The logo was created by combining two Younger Futhark runes: ᚼ (Hagall, corresponding to H) and ᛒ (Bjarkan, corresponding to B) — Harald Blåtand’s initials. The result was the distinctive symbol, resembling a stylized snowflake with two vertical strokes. I like that the logo of a 21st-century technology contains symbols from an alphabet a thousand years old — and no one finds that strange. The runes simply became part of a visual language seen by billions of people on their phone screens, without the slightest idea of Harald and his teeth — by the traditional interpretation, the nickname referred to a darkened, diseased tooth: the Old Norse blár meant not only “blue” but also “dark, black,” and a chronicle from the late 12th century glosses the epithet literally: “a tooth bluish or black.”
The Bluetooth 1.0 specification came out in 1999 and was, to put it mildly, raw. Compatibility problems between devices from different manufacturers, difficult pairing, unstable connections — early users got a technology that promised freedom from wires but in practice demanded patience and a thick manual. Version 1.0B fixed some of the problems, but real stability came only with 1.2.
The first commercial Bluetooth devices appeared in 2001. The Ericsson R520m and T39 phones allowed users to connect a wireless headset — bulky, with an ear hook, resembling a prosthetic from a B-grade science-fiction film. The IBM ThinkPad A30 got a built-in Bluetooth module, making it possible to synchronize a phone book without a cable. By today’s standards — nothing. By 2001 standards — a trick that inspired mistrust: “Does it really work? The data won’t get lost?”
Version 1.0 offered throughput of up to 732.2 kbit/s: enough for voice and small files, but hopelessly insufficient for high-quality music or photo transfers. (Later, in version 1.2, practical speeds reached 721 kbit/s with a more stable link — exactly what the first users had been missing.) The system of profiles — sets of rules for specific use cases — gradually grew new specifications: the headset profile (HSP), the file transfer profile (FTP), the advanced audio distribution profile (A2DP), which appeared later and made stereo sound possible. Each profile was a separate agreement on exactly how devices would talk in a particular situation: sending a business card is one thing, streaming music another, controlling a car’s multimedia system a third.
Car integration, incidentally, became one of Bluetooth’s most visible applications in the mid-2000s. The HFP (Hands-Free Profile) made it possible to talk on the phone through a car’s speakers — and lawmakers in many countries immediately made use of it, banning phone conversations with a handset in one’s hand and effectively making Bluetooth a mandatory feature for drivers.
The turning point came in 2010 with the Bluetooth 4.0 specification, which incorporated Bluetooth Low Energy (BLE) — a protocol developed by Finland’s Nokia under the name Wibree and then folded into the common standard. BLE was essentially a separate technology within Bluetooth: a different packet architecture, a different approach to power consumption, different use cases. Where classic Bluetooth was designed for continuous data transmission — voice, music, files — BLE was designed for devices that wake for a fraction of a second, send a tiny packet of data, and go back to sleep. A fitness band sending heart-rate data once a second. A temperature sensor in a warehouse. A beacon in a store telling your phone about a discount on the shelf beside you.
BLE made an entire category of devices possible that had previously been impractical. A coin-cell battery powering a sensor for a year or two is not science fiction, but a direct consequence of a radio module consuming microwatts instead of milliwatts. The Internet of Things — a term that had already become a cliché in presentations by the 2010s — gained a real physical foundation. Smart locks, key trackers, medical monitors, industrial sensors — all of it runs on BLE, and most users do not even distinguish it from “regular” Bluetooth.
Version 5.0 (released in 2016) and the subsequent 5.x versions increased range and throughput, and added the ability to broadcast to several receivers at once. The LE Audio specification, based on the new LC3 codec, promised better sound quality with lower power consumption and — perhaps more importantly — the ability to transmit audio to an unlimited number of devices. Imagine an airport where flight announcements stream directly into your headphones in your own language. Technically, this is already possible. The question is infrastructure and the willingness of airports.
And now — the cautious counterfactual question we began with. What if Rydbeck had not thought about those extra wires in 1989? What if Haartsen and Mattisson had gone off to do something else in 1994? What if the five companies had not sat down at one table in 1998?
The wireless world would have emerged anyway, of course. Radio physics belongs to no single company. IrDA (the infrared port) already existed and might have evolved into something more powerful. Zigbee, Z-Wave, proprietary radio protocols — alternatives existed and still do. Wi-Fi eventually learned to operate at low power consumption levels (Wi-Fi HaLow). The need for wireless peripherals would not have gone away — it was dictated by the physics of shrinking devices.
But the absence of Bluetooth would have changed not so much the fact of wireless communication as its architecture. Bluetooth SIG is a model in which competitors agreed on an open standard, and any manufacturer could release a compatible device. Without that precedent, we might live in a world of proprietary ecosystems: Sony headphones work only with Sony phones, a Dell keyboard only with Dell laptops. Apple, with its W1/H1/H2 chip, is already moving in that direction, layering a proprietary system on top of Bluetooth — but Bluetooth itself remains the common foundation, and AirPods can still connect to an Android phone, even if they lose some functionality.
The balance between an open standard and a proprietary ecosystem is a pendulum, and Bluetooth set its initial position closer to openness. Without that initial push, the pendulum might have swung the other way. The world of wearable electronics, smart homes, and low-power sensors might have looked different — more fragmented, with incompatible islands of devices, with the need to buy adapters (adapters again!) between ecosystems.
I think about how this whole story began with irritation — with wires tangling across a desk in a Swedish office. Not with a visionary speech about the future, not with a presentation to investors. With an everyday inconvenience that one person noticed, and several others decided to spend years of their lives on. Today, nearly five billion Bluetooth chips are shipped in a single year (4.7 billion already in 2021). Harald Bluetooth’s runes flicker on screens in Tokyo, Nairobi, and Buenos Aires. And RS-232 quietly lives out its days in industrial controllers and old server rooms — where wires bother no one.