From touchscreen phones to in-cabin AI: Neonode’s long search for scale
Long before the iPhone made a glass touchscreen feel inevitable, Neonode tried to build a phone around touch. The Swedish-designed N1 and N2 never became mass-market hits, but their optical interface left behind something more durable than the handsets: patents, sensing software and a technical lineage that Neonode is still trying to turn into a viable business.
Today’s Neonode is not a phone maker and is moving beyond the touch hardware that carried its zForce technology into printers, cars and industrial equipment. Its new bet is MultiSensing, camera-based software that interprets what happens inside vehicles and other environments. The transformation is technically coherent; the commercial evidence remains early.
Two histories came together under one name
Neonode’s corporate history needs careful wording because the listed US parent is older than the Swedish operation that created the brand. The public company began as SBE Inc., an American electronics business. In 2007 it combined with the Swedish Neonode business and adopted the Neonode name. That is why filings can trace the legal parent much further back while the recognisable product story begins with Swedish touchscreen phones in the 2000s.
The Neonode N1, N1m and N2 used an infrared optical system rather than the projected-capacitive touch technology that would dominate smartphones. Light emitters and receivers created a grid across the display; a finger interrupting the beams could be located without a conventional touch-sensitive layer over the screen.
The approach enabled gestures and finger control, but an inventive interface was not enough to win the phone market. Distribution, software, applications, manufacturing scale and timing mattered just as much. Neonode’s handsets became a footnote in mobile history while capacitive multi-touch ecosystems pulled far ahead.
zForce turned a failed device strategy into licensable technology
Neonode’s important decision was to stop treating the optical interface as something that belonged only in its own phones. From 2010 it licensed zForce to original-equipment manufacturers and automotive suppliers, allowing customers to embed the technology in their products. The company says licensees have shipped more than 95 million devices using its technology since then.
Optical sensing has practical differences from capacitive touch. It can work across varied surface materials, support larger or unusual shapes, and enable contactless gestures. That made it relevant to printers, office equipment, vehicle controls, kiosks and industrial systems. It did not replace capacitive touch in phones and tablets, where that technology’s scale, thinness and supplier base became formidable.
Neonode added a Touch Sensor Module business in 2017, manufacturing modules that incorporated its technology instead of collecting royalties only. The change offered more revenue per installation but also brought inventory, component sourcing and manufacturing risk. In December 2023, the company reversed course: it announced that the module operation would be phased out and that the underlying technology would be licensed to strategic partners.
The company is now putting zForce into maintenance mode
The retreat continued in September 2025, when Neonode moved zForce into maintenance mode. It stopped marketing the platform to new customers while committing to support existing licensees in office equipment, automotive, industrial, medical, military and avionics markets. This is a significant correction to older descriptions of Neonode as a broad optical-touch growth company.
Existing zForce royalties can continue as customers ship products, but future growth is meant to come from MultiSensing. Neonode’s 2025 annual report describes the company as a machine-perception software supplier focused mainly on automotive customers, not as a sensor-module manufacturer.
MultiSensing shifts the problem from touch to understanding
MultiSensing uses camera images, machine-learning models and synthetic training data to detect people, objects and actions. In a vehicle, possible uses include checking whether a driver is distracted or drowsy, detecting hands on the steering wheel, recognising seatbelt use, interpreting gestures and improving digital-mirror views. Neonode also presents computer-vision uses for self-checkout and loss prevention.
This is a larger market opportunity than adding touch to a display, but it is also intensely competitive. Neonode’s filing identifies Smart Eye, Cipia, Xperi, Seeing Machines, PUX and Jungo among competing automotive software providers. Carmakers and Tier 1 suppliers require strong performance across lighting conditions, occupants, camera placements and hardware platforms, followed by long validation and production cycles.
The wider trend is visible in software-heavy vehicles such as the Tesla Model 3, where cameras and software increasingly shape safety and the cabin experience. Neonode’s role would be far narrower than an automaker’s, but its business depends on being designed into those long-lived vehicle programmes.
Licensing can scale—but only after a customer ships
Neonode generally earns licence revenue per unit shipped by its customers. It also charges for non-recurring engineering work used to adapt and integrate the software. The model can produce attractive incremental margins once a design reaches volume production, because Neonode does not manufacture every unit. Before that point, however, engineers may spend years supporting evaluation and development without a guarantee of large royalties.
At the end of 2025, Neonode had 36 valid technology-licence agreements, but only nine licensees were shipping products containing its technology. Customer concentration is high: the annual filing named Seiko Epson, Alps Alpine, HP and a commercial-vehicle manufacturer as major revenue contributors. A delay or product change at one customer can therefore move Neonode’s results materially.
The 2025 patent payment should not be mistaken for operating momentum
Neonode assigned a portfolio of touchscreen patents to Aequitas Technologies in 2019 in exchange for a share of potential monetisation proceeds. Litigation followed against Samsung and Apple. A Samsung settlement in 2025 produced a $19.4 million gain for Neonode before a $3.8 million brokerage fee, leaving a net gain of roughly $15.5 million.
That cash materially strengthened the balance sheet, but it was not recurring product revenue. Neonode reported 2025 income from continuing operations of $8.0 million only because the patent gain outweighed operating losses; revenue from continuing operations was $2.1 million. The distinction is essential when assessing whether the underlying licensing business is improving.
What the latest quarter says
For the quarter ended 30 June 2026, Neonode reported $0.5 million in continuing-operations revenue, down 20.4% year on year. Licence revenue increased 9.2% to $0.4 million, while non-recurring engineering revenue fell sharply to $36,000. The company lost $2.1 million from continuing operations and used $1.9 million of cash in operations.
Management highlighted that MultiSensing licence revenue grew more than fivefold from a small prior-year base as an automotive customer increased production. That is directionally encouraging, but the absolute numbers remain small. At quarter-end, cash and receivables totalled $21.7 million, down from $25.8 million six months earlier. The complete figures are in Neonode’s Q2 2026 results.
On 6 August 2026, Nasdaq also notified Neonode that its shares had traded below the $1 minimum bid price for 30 consecutive business days. The notice did not immediately delist the shares; Neonode received until 2 February 2027 to regain compliance. It nevertheless shows how sceptically the market was valuing a business still consuming cash while awaiting larger licence volumes.
What buyers and investors should watch
The most useful signals are not demonstrations or evaluation projects. They are production design wins, the number of vehicles or devices shipped, recurring royalty growth, and evidence that MultiSensing can serve several customers without extensive custom engineering. Partnerships matter only when they shorten validation or open a route into production.
For automotive buyers, privacy, safety validation, processor requirements and performance across diverse drivers are as important as headline accuracy. For South African deployments, availability would probably depend on a vehicle maker, Tier 1 supplier or equipment partner embedding Neonode’s software; the company does not sell a general consumer driver-monitoring product directly.
A touchscreen pioneer searching for its second scalable act
Neonode survived its phone failure by separating the useful technology from the failed product. It then moved from licences into hardware modules and back out again. Its latest transformation applies the same instinct to computer vision: own the software, let customers manufacture the end product, and earn royalties when they ship.
The strategy is clearer than it was a few years ago. The outcome is not. MultiSensing has begun generating automotive licence revenue, but Neonode still needs far more production volume to cover its operating costs. The company’s history proves that good interface ideas can outlive the devices that introduced them; its future depends on whether those ideas can finally produce a repeatable business.
Pingback: VIA optronics AG history: origins, milestones and 2026 profile