Sandbar closed a $23 million Series A in March 2026, led by Adjacent and Kindred Ventures.

Prior to the Series A, Sandbar emerged from stealth in November 2025 with $13 million in seed funding. True Ventures and Upfront Ventures led the seed round, with Betaworks also participating.

Total funding raised across both rounds stands at $36 million.

Stream is a two-part system: a physical ring worn on the dominant-hand index finger and a companion iOS app. The ring serves as the input device, and the app is where spoken input becomes structured notes, searchable memory, and an ongoing conversational workspace.

The core interaction is intentional and physical. You hold the ring's glass touchpad, speak, and release. The microphone only activates while the touchpad is held, it is not always listening. The ring confirms a successful capture with a haptic vibration, so the user knows the thought was recorded without looking at the phone. If earbuds are connected, Stream can also speak back in real time. If not, it saves the note silently and acknowledges with a buzz.

The ring also handles media controls: a tap pauses or plays audio, a double tap skips to the next track, and a swipe adjusts volume. The hardware itself has an aluminum outer surface, a resin inner surface, a glass touchpad, Bluetooth connectivity, an IMU for gesture input, and a linear resonant actuator for haptics. It is water resistant to rain and handwashing and comes in sizes 5 through 13.

The app organizes captured input into three layers. Notes is where you edit and organize individual thoughts. Convo is the real-time conversational layer, most useful when earbuds are connected. Memory is the longitudinal layer where ideas accumulate and connect over time, letting users ask the AI about things they captured days or weeks earlier without remembering exactly when or how they said them.

Inner Voice generates responses in a voice partially derived from the user's own voice, powered by generative voice models from ElevenLabs. The design intent is to make the interaction feel like internal thinking rather than talking to a separate assistant with its own distinct persona.

Early usage data from demo testing showed that roughly 20% of interactions were note creation, 20% were single one-off queries, and 60% were back-and-forth conversation. That breakdown suggests the product's center of gravity is iterative thinking support, working through an idea, prepping for a meeting, exploring a concept while walking, rather than passive transcription or simple dictation. The target user is someone who thinks while moving: founders, writers, students, commuters, and people who already use voice dictation but find phone-based input too slow or too public.

The system uses a split-compute architecture: the ring handles capture and haptics, the connected iPhone handles companion functionality and some local processing, and the cloud handles the heavier AI workloads. Sandbar uses multiple large language models depending on the task. Audio of ring interactions is not saved. Speech is transcribed into text that lives in the app.

Batch 1 sold out, and Batch 2 is scheduled to ship in summer 2026. The product is currently iOS-only and ships in the U.S.

Sandbar uses a vertically integrated B2C hardware-plus-subscription model. The ring serves as the acquisition mechanism and the app as the retention and monetization layer. Neither works as well without the other, which is the structural logic of the business.

The hardware sale at $249-$299 funds customer acquisition and creates physical habit formation. Once a user has built the reflex of holding the ring to capture a thought, switching to a phone-based alternative involves real friction cost, even if a competitor app could replicate some features. The ring creates behavioral lock-in that a pure software product cannot.

The subscription layer captures recurring value from the users who engage most deeply. The free tier covers unlimited notes, which keeps the product accessible and reduces the barrier to hardware purchase. The $10/month Pro tier gates unlimited conversational AI interactions, the most compute-intensive and differentiated feature set, behind a paywall that aligns pricing with the actual cost driver. That structure means Sandbar is not charging for storage or basic transcription. It is charging for the ongoing AI relationship, which is where the product's differentiation sits.

The three-month free Pro period for preorder customers is a deliberate onboarding investment. It gives new users enough time to build the conversational habit before the billing decision arrives, which may improve attach rates relative to a model that charges immediately.

The cost structure is heavier than a pure SaaS notes app. Sandbar owns the full stack: ring firmware, BLE and sensor integration, mechanical and electrical design, iOS app development, backend infrastructure, and machine learning including model behavior tuning, memory architecture, voice expressivity, and inference optimization. That vertical integration raises fixed costs and execution complexity, but also means Sandbar controls the interaction quality end-to-end, which matters in a product where latency, mic quality, and conversation design directly determine whether the habit forms.

Because the ring relies on a connected iPhone rather than carrying its own screen, cellular radio, or heavy onboard compute, the bill of materials is more manageable than more ambitious standalone AI hardware. That helps on both unit economics and battery life, which Sandbar rates as all-day.

The flywheel, if it works, runs like this: hardware gets the ring on a user's finger, daily use generates a growing personal memory corpus, richer context makes conversational interactions more useful, higher utility drives Pro conversion and retention, subscription revenue funds model improvements and new features, and better software makes the next hardware batch more compelling to new buyers.

The biggest structural question is Pro attach rate. If users treat Stream primarily as a quick capture tool rather than an ongoing conversational partner, the subscription ceiling will be lower than the product's design implies. The 60% conversational usage figure from early testing is the clearest early signal that Sandbar is moving toward the higher-value outcome.

The most direct form-factor competitor is Pebble Index 01, which targets the same core job, capturing a fleeting thought from the finger with a press and a whisper, but from a different strategic angle.

Pebble preorders at $75 and retails at $99, carries no subscription requirement, processes on-device, works offline, and supports both iPhone and Android. For price-sensitive early adopters and users who prioritize privacy and simplicity over AI depth, Pebble is a credible alternative at roughly one-third the upfront cost and zero ongoing fees.

Sandbar's advantage over Pebble is experience depth: conversational note interaction, longitudinal memory, Inner Voice personalization, earbud-mediated real-time conversation, and media controls. But that depth also means Sandbar carries more execution risk on latency, model quality, and ongoing value justification, whereas Pebble can compete by being good enough and cheap.

Plaud is the strongest pure-play commercial competitor in terms of current distribution. Its NotePin is a wearable AI note-taker for in-person conversations, and the company claims over one million users across more than 170 countries.

Plaud supports 112 languages, speaker diarization, custom vocabulary, thousands of output templates, and compliance certifications including SOC 2, HIPAA, GDPR, and ISO standards. That enterprise-readiness posture is a structural advantage for regulated buyers, legal, medical, and compliance-sensitive teams, where Sandbar's current consumer positioning and U.S.-only, iOS-only availability creates procurement friction.

Plaud's threat is less about ring form factor and more about distribution maturity. If the category consolidates around reliable capture, structured output, and admin controls rather than intimate conversational UX, Plaud's scale and workflow orientation could outweigh Sandbar's interaction design edge.

The most consequential competitive dynamic is not any single startup rival but the accelerating acquisition of the category by large platforms.

Amazon acquired Bee in July 2025. Bee's wearable and app are aimed at ambient personal AI that learns from conversations and context over time, priced at $49.99 for the Pioneer device. Amazon's ownership gives Bee a path to distribution through Alexa, Amazon devices, and Prime bundling that no independent startup can match on cost of customer acquisition.

Meta acquired Limitless, which had been one of the highest-profile memory-wearable startups. Limitless has stopped selling the Pendant to new customers and is supporting existing users through 2026. That removes an independent competitor from the market near-term but signals that Meta is integrating memory capture capabilities into its broader AI wearables strategy, including Ray-Ban smart glasses and EMG wristband input research.

Oura, which Sacra tracked growing from roughly $130M in revenue in 2022 to an estimated $1B in 2025, acquired Doublepoint for gesture technology in a move that suggests smart rings themselves may become more interactive. With over 5.5 million rings already on fingers globally, Oura has installed-base leverage if it adds lightweight voice interaction or contextual AI features. Sandbar's form-factor lead is vulnerable if the dominant ring platform adds enough input capability to make a dedicated voice ring feel single-purpose.

OpenAI's acquisition of Jony Ive's io hardware studio in May 2025 adds another platform-level threat: a purpose-built AI hardware device from the company that makes the underlying models Sandbar and its competitors depend on.

Sandbar also competes with phone-native behavior more than it competes with hardware peers. ChatGPT voice mode, Siri, and Google's assistant stack can all absorb spoken note-taking without requiring a new device. Otter.ai, Fireflies.ai, Fathom, and Grain cover transcription and summarization for meeting contexts. Gong and Outreach, which Sacra has tracked at roughly $285M and $250M in ARR respectively, show how call intelligence can consolidate into existing workflow systems at enterprise scale, without requiring any new wearable at all.

The infrastructure layer underneath all of these products is also commoditizing. Deepgram and ElevenLabs, the latter of which powers Sandbar's own Inner Voice feature, show how speech-to-text and text-to-speech are becoming building blocks rather than moats. That lowers the barrier to launching a voice wearable but also makes differentiation harder for everyone in the category.

The most immediate TAM expansion is separating the Stream app from the ring requirement. As of March 2026, the app requires a Stream Ring for first-time setup, which caps the addressable market at people willing to buy a $249–$299 wearable.

Sandbar has indicated it is considering opening app access to users who do not own the ring. If it does, Stream can compete directly in the larger market for voice journaling, AI note capture, and personal knowledge management, a market currently served by Otter.ai, Notion AI, and phone-native dictation, without requiring a hardware purchase at entry. The ring would remain the premium, lowest-friction experience, but the app could build a broader top-of-funnel and a separate subscription revenue stream.

Stream is currently iOS-only and ships only in the U.S. Moving to Android is likely the highest-ROI TAM expansion available because it broadens geographic reach, price-band accessibility, and device compatibility without requiring a new hardware category.

English-speaking markets outside the U.S., Canada, the U.K., Australia, and parts of Western Europe, are logical near-term targets because the core use case does not depend on local merchant acceptance, healthcare reimbursement, or language localization. Full international expansion would follow as hardware operations, support infrastructure, and app readiness mature.

Sandbar has said it does not believe in walled gardens and plans to support exports to apps like Notion. Extending that into native integrations with task managers, calendar systems, and CRM tools would let Stream become the lowest-friction voice input layer for larger workflows rather than a destination app that users must remember to revisit separately.

The deeper strategic opportunity is moving up the value chain from capture hardware into personal memory infrastructure. The ML hiring Sandbar has done, covering model behavior tuning, memory architecture, synthetic data generation, voice expressivity, and inference optimization, points to a product that accumulates personalized context over time and becomes more valuable the longer it is used. If Stream becomes the place where a user thinks, retrieves prior ideas, and plans future work, it starts to look less like a gadget and more like a persistent cognitive layer, a category that sits closer to AI productivity platforms than to smart rings, and one with materially larger pricing power and retention dynamics.

Sacra's broader research on voice and workflow suggests that voice capture becomes most valuable when it triggers downstream actions across apps rather than remaining a standalone recording utility. Zapier-style automation, reminder creation, and calendar integration are natural next steps that would let Sandbar capture more of the value it creates for users who currently have to manually move captured thoughts into their actual work systems.

Sandbar describes itself as an interface company rather than a ring company, and its hiring spans firmware, mechanical, hardware program management, and ML roles tied to novel device form factors and mass production. Stream Ring is framed as the first wedge into a broader hardware-software platform, with plausible adjacency into other low-friction input interfaces where voice, gesture, and private AI interaction matter. The specific form factors are not disclosed, but the company's framing and hiring profile both point to a multi-device roadmap rather than a single-SKU hardware business.

Hardware execution: Sandbar is heading into a limited summer 2026 hardware launch with 15 employees and manufacturing operations still being built out. Consumer electronics shipping at quality, on time, and with acceptable return rates is operationally demanding, and a poor first-batch experience could damage trust before the software platform reaches enough scale to stand on its own.

Platform absorption: Amazon, Meta, and OpenAI are all building or acquiring AI wearable capabilities at a scale that exceeds Sandbar's resources. If any of these platforms bundles voice capture, memory, and conversational AI into a device users already carry or wear, Alexa earbuds, Meta glasses, or a future OpenAI hardware product, Sandbar's standalone economics become structurally harder regardless of the quality of the ring interaction model.

Subscription ceiling: Sandbar's recurring revenue depends on users engaging with Stream as a high-frequency conversational product rather than a niche note-capture tool. If the habit does not form at the depth the product requires, or if users find the free tier sufficient for their actual usage patterns, Pro attach rates will disappoint and the business will skew toward a one-time hardware sale with limited software monetization upside, compressing lifetime value and making the unit economics difficult to sustain at scale.