AI硬件创业者疯狂涌入手机的背面AI Hardware Entrepreneurs Are Flocking to the Backs of Phones


AI free token/AI免费积分
AI硬件创业者疯狂涌入手机的背面

快科技官方

关注

过去一年,AI 硬件创业者开始扎堆手机背面。

Plaud 的 AI 录音卡累计出货突破百万台。2026 年 6 月,Plaud 宣布 ARR 达到 1 亿美元、服务超过 200 万专业使用者;截至 9 月,公司最新披露的全球使用者数已超过 250 万。另一边,做磁吸电子纸屏的阅星曈完成近 5000 万元 A 轮追加融资,顺为、经纬、博裕和小红书等老股东继续跟投;极稚科技则称,其无源墨水屏手机壳已经卖出超过 14 万只,其中八成销往海外。

销量和融资之外,新的机型还在不断冒出来。Flowtica 在 IFA 展出集录音、电子纸和会议待办于一体的 Verso;取得线性资本天使投资的 NOOKLAB,准备推出兼容持续收音的 AI 随身屏 Moonphase Flip;极稚科技也计划用 Sloth 从静态墨水屏进入 OLED、麦克风和虚拟宠物。

大厂的动作更快。钉钉、出门问问、科大讯飞和 360 相继进入 AI 录音卡市面;荣耀、海信和影石,则分别从磁吸副屏、可拆卸电子纸和影像配件的方向试探手机背面的可能性。与此同时,华强北已经把基础录音卡的售价压到一两百元。

一个原本以手机壳、卡包和充电宝为主的配件位置,突然挤进了录音公司、电子纸创业者、消费电子团队、互联网大厂和手机厂商。它们给出的答案并不相同:有人想把它做成录音工具,有人放上一块低干扰显示屏,有人想做虚拟宠物,也有人试图把录音、表明和 Agent 全部装进去。

这些公司争夺的不只是一个新品类,更是手机上所剩不多的高频携带位置。一部手机的背面很难同时挂着录音卡、阅读屏、充电宝和 AI 宠物。当越来越多机型要求使用者把自己长期留在这里,它们之间就不再只是能力差异,而会形成直接的机型和战略竞争。

这也是手机背面今天值得被重新关注的原因。

MagSafe 早已让使用者习惯把配件吸在手机背后,但这一轮变化,是 AI 硬件开始借助这个成熟位置寻找自己的入口。Plaud 率先证明,独立 AI 工具可以在这里取得百万台规模和持续付费;显示屏机型的增长和融资,则让另一种可能浮现出来:手机背面或许不只负责采集信息,也可以成为 AI 向使用者返回结果的第二界面。

一场围绕几厘米空间的竞争,已经从配件生意进入 AI 硬件创业的中心。

一、手机背面的「隐藏金矿」

较早让市面意识到手机背面可以成为 AI 硬件入口的,是一张录音卡。

2023 年 6 月,深圳团队 Plaud 推出 Note:厚约 3 毫米、重 30 克,借助 MagSafe 吸在手机背后,按下按键即可录音,会议结束后在手机 App 里查看 AI 转写和摘要。它没有发明录音,也没有发明 AI 转写—— 讯飞这些公司已经做了很多年。它做对的事情是位置挑选:把一个独立硬件寄生在手机上,使用者不需要多带一个东西,不需要在开会前想起从包里翻出录音笔。

这个挑选缓解了独立 AI 硬件普遍面对的一道难题:如何让使用者每天带着它。

Plaud 的增长速度超出了大多数人的预期。

机型连续两年保持约十倍增长。6 月,Plaud 宣布其软件订阅业务的年度经常性收入(ARR)从 100 万美元增长到 1 亿美元(两年约 100 倍),设备出货量突破 200 万台—— 在达成这一里程碑的 AI 公司中,它是少有的以硬件为入口的公司,其余几乎都是纯软件。

Plaud 的成功引来了快速的模仿和围猎。2025 年以来,钉钉推出 DingTalk A1 录音卡(499 元起),出门问问发布 TicNote(999 元起),360、科大讯飞相继跟进,华强北白牌则把同类机型的售价打到 120 至 150 元。纯录音 + 转写的能力差异,在一年之内被压缩到几乎为零。

但 Plaud 真正验证的不只是「有人愿意为 AI 录音付费」,更是手机背面这个位置本身有价值:独立 AI 机型可以借助手机背面既有的磁吸位置,取得高频携带,并进一步建立付费需要。原本以防护、充电和装饰为主的配件位,开始具备成为 AI 硬件入口的可能。

这个判断很快又被另一批创业者从表明的方向继续验证。

极稚科技做了一款无源墨水屏手机壳 XTREM,利用手机 NFC 刷新电子纸,静态表明不耗电—— 使用者从 App 里选一张照片,把它留在手机背后。

机型售价 59 至 79 美元,有报道称其累计出货超过 14 万只,其中约 80% 销往海外。报道数据称,使用一个月后日均投图次数仍旧达到 4.8 次。使用者愿意反复经营这块原本几乎没有信息的区域,让它成为情绪表达和社交展示的一部分。

阅星曈则把这个方向推得更远。创始人胡宇沸在 2024 年创立阅星曈,做了一块 4.3 英寸的磁吸电子纸电纸书,重量约 75 克,吸在手机背面翻转即可阅读。

机型上市仅半年,月销量增长超过 10 倍,海外市面首月即破万台。今年 8 月,阅星曈完成顺为资本、经纬创投、博裕创投和小红书近 5000 万元 A 轮追加融资,资金首要用途就是提前锁定电子纸显示屏供应和产能。

融资的方向透露了一个关键信息:上游产能正在成为瓶颈。

全球超过 90% 的电子墨水薄膜由元太科技供应,2026 年 3 月元太宣布投资 50 亿新台币建设观音新厂,但一期要到 2027 年第二季度才落成、2028 年才量产。在产能释放之前,谁能提前锁定显示屏,谁才有资格谈规模。

到了 2026 年年中,一个此前只属于创业公司的位置,开始吸引手机厂商亲自下场。

荣耀在 600 系列上推出了可拆卸磁吸副屏,使用者称之为「电子吧唧」—— 可以表明照片、当自拍预览窗、控制补光,还可以取下来远程拍照。影石 Insta360 则在 3 月推出 Snap,一块 3.5 英寸的磁吸屏借助 USB-C 直连手机,应对后置镜头自拍时无法预览的难题,售价 499 元起。OPPO 也发布了 Bubble 潮玩自拍屏。海信在 8 月底发售 A10 Ultra,正面搭载墨水屏,背面则标配一块磁吸可拆卸彩色 LCD 副屏。

创业公司从录音和阅读两个方向验证了需要,资本在产能端押注,手机厂商则开始用自己的方式进入。几条线索在 2026 年交汇,手机背面不再只是配件市面的一部分—— 它开始具备成为独立硬件入口的条件。

但这也引出了下一个难题:当录音设备和显示屏设备争夺同一个位置时,这个品类接下来会往哪里走?

二、两条路线,和它们的交叉点

录音卡和阅读屏原本服务于不同的人群,但当它们都想长期留在手机背后,就开始成为彼此的竞争对手。使用者可以拆换,也可以叠加,但手机的厚度、重量和手感,限制了这个位置能容纳多少设备。

对创业者来说,难题随之变得具体:继续把单一能力做深,能否守住这个位置?如果对手在一块显示屏上补齐了录音能力,或者让录音卡直接表明摘要,使用者会不会换一个挑选?

一些团队已经跨过这条边界。

前段时间的 IFA 展会上,Flowtica 展出了 Verso:一块同时装有电子纸显示屏和麦克风阵列的磁吸设备。

它可以记录会议,再把摘要、行动项和日历表明在自己的显示屏上。使用者不必打开手机 App,也能看到一场会议留下了什么。机型预计售价 299 美元,计划于 2026 年第四季度交付。

类似的还有 NOOKLAB 推出的 Moonphase Flip,一块重约 54 克的磁吸 AI 随身屏。它兼容持续收音,可以从现实对话中识别人、项目和待办,再把一句总结和几项等待确认的事项留在显示屏上。使用者不需要重新打开 App 翻找会议记录,只要翻过手机,就能看到自己刚刚答应了什么、接下来需要处理什么。

按照 NOOKLAB 的设想,这块显示屏不负责承载更多信息,也很少主动推送,而是只留下 AI 判断后值得使用者注意的内容。团队把这种外观称为「低打扰的 AI 随身屏」:需要时进入使用者的注意力,平时则停留在余光里。

阅星曈也把「探索 AI 能力」列为融资后的投入方向。按照胡宇沸的设想,AI 可以在内容抵达显示屏之前完成整理,减少使用者寻找文件、导入和推送的操作。它并不一定要把所有能力都装进背屏,也可以借助云端、手机 App 与电子纸设备协同完成。

极稚科技则准备从展示走向互动。其下一代机型 Sloth 将静态墨水屏换成 OLED,加入麦克风、对话和虚拟宠物能力。第一代机型上用于展示照片的显示屏,到了这一代,开始依据使用者的交流发生变化。

沿着录音卡路线进入的企业也在增加表明能力。钉钉的 A1 Pro 已经为录音卡加上了彩色触控屏,可以表明状态和电量。不过,小型状态屏与可以阅读摘要、处理事项的完整显示屏,承担的任务并不相同。录音设备补上一块显示屏,并不代表它已经变成了背部信息终端。

推动这些团队向外扩展的,是单一能力越来越难做出差异。

对录音公司来说,转写和摘要正在成为常见能力。设备已经能把声音保存下来,下一步竞争自然会进入录音之后的工作:使用者怎样查看结果,如何从摘要中提取待办,又如何继续跟进没有完成的事情。显示屏给出了一种可能,让录音设备从采集工具进入后续流程。

对显示屏公司来说,难题在于表明什么。电纸书可以表明书籍,手机壳可以表明使用者挑选的照片;如果显示屏想主动给出信息,就需要理解使用者此刻关心什么。输入可以来自麦克风,也可以来自手机里的日历、消息和其他数据。无论采取哪种方式,显示屏都需要从被动承载内容,走向对内容开展挑选。

两条路线因而开始靠近:录音设备需要一个更直接的输出方式,显示屏设备则需要更多元的上下文。

Agent 让这种交叉有了新的用途。

当 AI 只负责整理一场会议,使用者在手机或电脑上打开摘要,任务就结束了。但如果它还要跟进会议里留下的事情,就需要记住谁答应了什么、截止时间是什么,以及事情后来有没有变化。

例如,录音中浮现「周五前把思路发给客户」,生成一条待办并不困难。要继续推进这件事,还需要判断任务属于谁、时间是否已经确认、思路有没有发出,以及什么时候提醒使用者最合适。

麦克风可以补充没有进入聊天记录的现场信息,Agent 负责理解和组织,显示屏则把少量需要确认或跟进的结果留在使用者眼前。这三种能力第一次有机会围绕同一项任务工作,而不只是被装进同一块硬件。

为什么还需要一块背屏,而不是继续使用手机主屏?区别不在于显示屏能表明多少内容,而在于信息以什么方式浮现。

手机主屏合适搜索、输入和处理复杂任务,却也装满了消息、应用和信息流。一次解锁,使用者面对的是无数个挑选。背屏更合适承载那些需要持续存在、又不值得专门打断使用者的信息:下一场日程、等待确认的待办,或者一项还没有兑现的承诺。使用者不必寻找应用,这些内容也不会在通知列表里迅速消失。

电子纸刷新慢、色彩有限,在这里反而形成了一种约束。它很难承载视频和不断滚动的信息流,只合适留下少量结果。NOOKLAB 将机型称为「低干扰 AI 显示屏」,瞄准的也是类似方向:采集语音和上下文,再把值得继续推进的事情放回使用者面前。

但显示屏并非所有录音情形的必需品。对录完访谈后集中在电脑上处理材料的人来说,更轻的设备、更长的电池续航和更可靠的拾音,可能比随时查看摘要关键。对只想阅读的人来说,麦克风、提醒和 Agent 也可能成为负担。

融合因而不会简单取代录音卡和阅读屏,更可能浮现的结果是机型分层。

所以这个市面接下来的竞争,更关键的难题可能是,哪支团队能让两种能力共同完成一项原来没有被做好、又足够频繁的任务。

三、最难的是让使用者接受

把录音、显示屏和 Agent 装进手机背面,工艺上已经没有太多悬念。更难的难题是:使用者愿意为这些能力承担多少额外的重量、电池续航焦虑和售价。

Plaud Note Pro 重约 30 克,可以达成最长 50 小时连续录音;Flowtica Verso 加入触控墨水屏、麦克风阵列和更大的电池后,重量达到约 98 克,最长电池续航为 72 小时。两款机型的能力与测试条件不同,却足以说明融合需要付出的代价:显示屏、录音和计算每增加一项,设备都会更接近一台独立硬件。

电子纸可以降低表明功耗,却无法消除录音和无线传输的耗电;磁吸应对了设备放在哪里,也没有自动应对充电难题。Verso 仍然使用 USB-C,其他录音卡则尝试无线充电或者手机反向供电。如果使用者需要频繁把设备取下来充电,或者它已经显著影响手机的握持感,「随手机携带」带来的长处就会迅速减弱。

售价会进一步改变使用者的预期。几十美元的手机壳,可以凭借外观、趣味性和新鲜感吸引使用者尝试;当一块融合设备卖到 299 美元,购买者会认真计算:它替自己减少了什么工作,节省了多少时间,又为什么不能直接使用手机?

能力越多,机型越不能依靠「配件」这个身份降低使用者预期。到了一两千元的价位,它已经是一台需要独立定义情形、独立证明价值的硬件。

更大的变量来自手机厂商和平台生态。

录音、转写和摘要正在进入手机平台,背屏也开始被整合进手机本体或官方配件。磁吸副屏、可拆卸 E Ink 背屏,以及自拍屏,都说明手机厂商和消费电子企业已经开始关注这个位置。相比创业公司,它们可以调用手机的电池、算力和平台数据,也具备更成熟的渠道。

独立设备仍然有自己的窗口。它可以跨手机品牌使用,也能从手机背面取下放到会议桌上,还可以围绕特定人群,外观专门的流程。大厂可以复制一块显示屏或者一个录音能力,却未必愿意为每一个小群体建立完整服务。

但如果创业公司无法围绕某类使用者建立足够深的机型习惯,它们验证出的需要反而可能被手机厂商吸收。这个窗口可以存在多久,取决于创业公司跑得多快,也取决于手机厂商愿意把平台能力开放到什么程度。

比硬件和竞争更难回答的,是使用者究竟为什么购买它。

手表找到了健康,Plaud 找到了录音。录音、显示屏和 Agent 汇合以后,同样需要一个足够清楚的情形,而「全天个人助理」还算不上答案。

另外,持续记录还会带来隐私边界的难题。设备可以录一整天,不等于使用者和周围的人愿意接受它在所有场合工作。清晰的录音提示、方便的启停方式,以及使用者对数据去向的控制,都是这类机型走出会议室的前提。

接下来真正需要观察的,是新鲜感过去以后,还有多少使用者继续使用、愿意为软件付费,并且没有把它从手机背面取下来。

手机背面从来不缺可以放进去的能力,真正缺的是一个足够关键的使用理由。接下来的竞争,就看谁能在「有用」与「无感」之间找到平衡:足够轻,让使用者忘记它的存在;足够明确,让使用者知道为什么需要它;也足够独立,可以在手机厂商的阴影下建立自己的价值。

谁能先做到这几点,谁才有可能定义手机背面的第一个新品类。

本文来源:https://finance.sina.cn/tech/2026-09-20/detail-inismzuz2586447.d.html
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。

Kuaitech Official

Over the past year, AI hardware entrepreneurs have begun flocking to the back of phones.

Plaud's AI recording card has shipped over one million units cumulatively. In June 2026, Plaud announced ARR reached $100 million with over 2 million professional users; by September, the company disclosed global users exceeded 2.5 million. Meanwhile, Yue Xing Tong, which makes magnetic e-paper screens, completed a nearly 50 million yuan Series A follow-on round, with existing investors Shunwei, Matrix, Boyu Capital, and Xiaohongshu continuing to invest; Ji Zhi Technology stated its passive e-ink phone case has sold over 140,000 units, with 80% shipped overseas.

Beyond sales and funding, new models keep emerging. Flowtica showcased Verso at IFA, an all-in-one device combining recording, e-paper display, and meeting task management; NOOKLAB, which received an angel investment from Linear Capital, is preparing to launch Moonphase Flip, an AI wearable screen compatible with continuous audio capture; Ji Zhi Technology also plans to evolve Sloth from a static e-ink screen into an OLED device with microphone and virtual pet capabilities.

Tech giants are moving even faster. DingTalk, Mobvoi, iFlytek, and 360 have sequentially entered the AI recording card market; Honor, Hisense, and Insta360 are probing the back-of-phone space from the directions of magnetic secondary screens, detachable e-paper, and camera accessories. Meanwhile, Huaqiangbei has driven the price of basic recording cards down to 100–200 yuan.

A position once dominated by phone cases, card holders, and power banks has suddenly been crowded by recording companies, e-paper startups, consumer electronics teams, internet giants, and smartphone manufacturers. Their answers differ: some aim to make it a recording tool, others mount a low-interference display, some want a virtual pet, and a few attempt to cram recording, display, and Agent capabilities all in one.

These companies aren't just competing for a new category—they're fighting for one of the last high-frequency carrying positions on a phone. A phone's back can't simultaneously hold a recording card, reading screen, power bank, and AI pet. As more devices ask users to keep them attached long-term, competition shifts from feature differences to direct product and strategic rivalry.

This is why the phone back deserves renewed attention today.

MagSafe had already accustomed users to snapping accessories onto the back of their phones, but this wave is different: AI hardware is leveraging that proven position to find its entry point. Plaud proved first that a standalone AI tool can achieve million-unit scale and recurring payments there. The growth and funding of display-oriented models reveal another possibility: the phone back may not only collect information but also serve as a second interface for AI to deliver results to users.

The competition over a few square centimeters has moved from accessory business into the center of AI hardware entrepreneurship.

I. The "Hidden Gold Mine" on the Phone Back

What made the market realize, relatively early, that the back of a smartphone could become an AI hardware entry point was a recording card.

In June 2023, Shenzhen-based team Plaud launched the Note: about 3mm thick, weighing 30g, it magnetically attaches to the back of a phone via MagSafe, records with a single button press, and lets users view AI transcription and summaries in the companion app after a meeting ends. It didn't invent recording, nor AI transcription—companies like iFlytek had been doing both for years. Its right call was location: attaching a standalone piece of hardware to the phone so users don't need to carry an extra device or dig through their bag for a recorder before meetings.

Image source: Plaud Taobao

That choice addressed a common challenge facing standalone AI hardware: how to get users to carry it daily.

Plaud's growth speed exceeded most people's expectations.

Revenue grew roughly tenfold year after year. In June, Plaud announced its software subscription ARR grew from $1 million to $100 million (roughly 100x in two years), and device shipments exceeded 2 million units—among AI companies hitting such milestones, it is one of the rare ones entering through hardware; the rest are almost entirely software.

Plaud's success quickly drew imitation and encirclement. Since 2025, DingTalk launched the DingTalk A1 recording card (from 499 yuan), Mobvoi released the TicNote (from 999 yuan), while 360 and iFlytek followed, and Huaqiangbei white-label products drove similar devices down to 120–150 yuan. The pure recording-plus-transcription capability gap shrank to near zero within a year.

What Plaud truly validated was not merely that “people are willing to pay for AI recording,” but that the position on the back of a smartphone itself has value: standalone AI devices can use the existing magnetic attachment position on the phone’s back to gain frequent carry, and then further build a paid need. Accessory positions originally dominated by protection, charging, and decoration have begun to offer the possibility of becoming an AI hardware entry point.

That judgment was quickly reinforced by another wave of entrepreneurs from the display direction.

Ji Zhi Technology made a passive e-ink phone case called XTREM, using the phone's NFC to refresh the e-paper—static display consumes no power. Users select a photo from the app and leave it on the back of their phone.

Image source: Ji Zhi Technology Xiaohongshu

Priced at $59–79, reports say cumulative shipments exceeded 140,000 units, with about 80% sold overseas. Data indicates that after one month of use, the average daily image update frequency remained at 4.8 times. Users are willing to repeatedly curate an area that previously held almost no information, making it part of emotional expression and social showcase.

Yue Xing Tong pushed this direction even further. Founder Hu Yufei founded the company in 2024, building a 4.3-inch magnetic e-ink reader weighing about 75g that attaches to the back of a phone and flips open for reading.

Image source: Yue Xing Tong

Within half a year of launch, monthly sales grew over tenfold, and overseas market sales broke 10,000 units in the first month. In August this year, Yue Xing Tong completed a nearly 50 million yuan Series A follow-on round led by Shunwei Capital, Matrix Partners, Boyu Capital, and Xiaohongshu; the primary use of funds is to secure ahead of time e-paper display supply and production capacity.

The direction of the funding reveals a key signal: upstream capacity is becoming a bottleneck.

Over 90% of the world's e-ink film comes from E Ink Holdings. In March 2026, E Ink announced a NT$5 billion investment to build a new Guanyin factory, but Phase I won't be completed until Q2 2027 and mass production won't begin until 2028. Before capacity ramps, whoever locks in display supply first earns the right to talk scale.

By mid-2026, a position that previously belonged only to startups began attracting smartphone manufacturers to enter directly.

Honor launched a detachable magnetic secondary screen on its 600 series, called "electronic badge" by users—it can display photos, serve as a selfie preview window, control fill lighting, and be removed for remote shooting. Insta360 launched Snap in March, a 3.5-inch magnetic screen that connects directly to the phone via USB-C, solving the problem of not being able to preview shots when using the rear camera. OPPO also released the Bubble trend selfie screen. Hisense launched the A10 Ultra in late August, featuring an e-ink front screen and a magnetic detachable color LCD secondary screen on the back as standard.

Startups validated demand from recording and reading directions, capital bet on the supply side, and smartphone manufacturers began entering in their own way. Several threads converged in 2026—the phone back is no longer just part of the accessory market; it is beginning to meet the conditions to become an independent hardware entry point.

But this raises the next question: when recording devices and display devices compete for the same position, where does this category go from here?

II. Two Routes and Their Intersection

Recording cards and reading screens originally served different audiences, but as both seek to stay on the phone back long-term, they become competitors. Users can swap or stack them, but the phone's thickness, weight, and feel limit how many devices that position can hold.

For entrepreneurs, the challenge becomes specific: can deepening a single capability defend this position? If a competitor adds recording to a display screen, or lets a recording card display summaries directly, will users switch?

Some teams have already crossed that boundary.

At a recent IFA trade show, Flowtica exhibited Verso: a magnetic device equipped with both an e-paper display and a microphone array.

Image source: Flowtica official website

It can record meetings and then display summaries, action items, and calendar entries on its own screen. Users don't need to open a phone app to see what a meeting produced. The device is expected to retail at $299, with deliveries planned for Q4 2026.

Similarly, NOOKLAB launched the Moonphase Flip, a magnetic AI wearable screen weighing about 54g. It supports continuous audio capture, can identify people, projects, and tasks from real conversations, and leaves a summary plus pending items on the display. Users don't need to reopen an app to search through meeting records—just flip the phone to see what they agreed to and what needs handling next.

Image source: NOOKLAB official website

According to NOOKLAB's vision, the display doesn't carry more information nor push proactively; it only surfaces content the AI judges worth the user's attention. The team calls this design a "low-disturbance AI wearable screen": it enters the user's awareness when needed and otherwise stays in the periphery.

Yue Xing Tong also listed "exploring AI capabilities" as a post-funding investment direction. Per Hu Yufei's vision, AI can organize content before it reaches the display, reducing the user's need to search for files, import, and push data. It doesn't have to cram every capability into the back screen—it can also collaborate with cloud services, phone apps, and the e-paper device.

Ji Zhi Technology is preparing to move from display to interaction. Its next-generation device, Sloth, will replace the static e-ink screen with OLED and add microphone, dialogue, and virtual pet capabilities. The screen on the first-gen device, used for showing photos, now changes based on user interaction.

Companies that entered along the recording card route are also adding display capabilities. DingTalk's A1 Pro has already added a color touch screen to its recording card to show status and battery level. However, a small status screen and a full display capable of reading summaries and handling tasks serve different purposes. Adding a screen to a recording device doesn't mean it has become a back-mounted information terminal.

What drives these teams to expand outward is that single capabilities are increasingly hard to differentiate.

For recording companies, transcription and summarization are becoming commoditized. Devices can already preserve audio; the next competitive frontier naturally shifts to post-recording workflows: how users view results, extract action items from summaries, and follow up on unfinished tasks. Displays offer one path, enabling recording devices to move from collection tools into downstream processes.

For display companies, the challenge is what to show. E-readers display books; phone cases display user-selected photos. If a display wants to proactively deliver information, it needs to understand what the user cares about right now. Inputs can come from a microphone, from the phone's calendar, messages, and other data. Regardless of approach, displays must evolve from passively bearing content to curating it.

The two routes are therefore drawing closer: recording devices need a more direct output method, and display devices need richer context.

Agents give this convergence a new purpose.

When AI only organizes a meeting, users open the summary on their phone or computer and the task is done. But if it also needs to follow up on things left from the meeting, it must remember who committed to what, what the deadlines are, and whether anything has changed since.

For example, if the recording reveals "send the idea to the client before Friday," generating a to-do item is straightforward. Pushing it further requires judging who owns the task, whether the deadline is confirmed, whether the material was sent, and when to remind the user.

The microphone supplements on-site information not captured in chat logs, the agent handles understanding and organization, and the display keeps a small number of results needing confirmation or follow-up right in front of the user. These three capabilities have their first chance to work together围绕 one task rather than merely being crammed into the same hardware.

Why still need a back screen rather than continuing to use the phone's main screen? The difference isn't how much content the display can show, but how information surfaces.

The phone's main screen is suited for searching, input, and handling complex tasks—but it's also cluttered with messages, apps, and feeds. One unlock and the user faces countless choices. A back screen is better suited for information that needs persistent presence without demanding dedicated attention: the next schedule, a pending item awaiting confirmation, or an unfulfilled promise. Users don't need to find an app, and this content won't quickly disappear from a notification list.

E-ink refreshes slowly and has limited colors—which here becomes a constraint. It can't carry video or endlessly scrolling feeds; it only holds a small amount of results. NOOKLAB calls its device a "low-disturbance AI display," targeting a similar direction: capturing voice and context, then returning what's worth pursuing to the user's view.

But a display isn't essential for all recording scenarios. For someone who concentrates on processing interview materials on a computer afterward, a lighter device, longer battery life, and more reliable pickup may matter more than随时 checking summaries. For those who only want to read, a microphone, reminders, and agents could become a burden.

Convergence therefore won't simply replace recording cards and reading screens; a more likely outcome is product tiering.

So the more critical question for this market's competition may be: which team can make two capabilities jointly accomplish a task that hasn't been well served and is frequent enough.

III. The Hardest Part Is Getting Users to Accept It

Fitting recording, a display, and agents into the phone back no longer presents major engineering mysteries. The harder challenge is: how much extra weight, battery anxiety, and price are users willing to bear for these capabilities?

The Plaud Note Pro weighs about 30g and supports up to 50 hours of continuous recording; the Flowtica Verso, with its touch e-ink screen, microphone array, and larger battery, weighs about 98g with up to 72 hours of battery life. While the two devices differ in capability and testing conditions, together they illustrate the cost of convergence: each addition of display, recording, and compute pushes the device closer to a standalone piece of hardware.

E-ink can reduce display power consumption but cannot eliminate the drain from recording and wireless transmission; magnetic mounting solves where the device sits but not how it charges. Verso still uses USB-C, while other recording cards experiment with wireless charging or reverse power from the phone. If users need to frequently remove the device to charge, or it significantly affects the phone's grip, the advantage of "carrying it with the phone" diminishes rapidly.

Price will further shape user expectations. A phone case costing tens of dollars can attract trial through appearance, novelty, and fun; when a converged device sells at $299, buyers will seriously calculate: what work does it save me, how much time does it save, and why can't I just use my phone?

The more capabilities it packs, the less a device can rely on the "accessory" identity to lower user expectations. At the 1,000–2,000 yuan price point, it has become hardware that needs to define its own scenarios and prove its value independently.

A bigger variable comes from smartphone manufacturers and platform ecosystems.

Recording, transcription, and summarization are entering phone platforms; back screens are also being integrated into phone bodies or official accessories. Magnetic secondary screens, detachable E Ink backs, and selfie screens all indicate that phone makers and consumer electronics companies have begun focusing on this position. Compared to startups, they can tap the phone's battery, computing power, and platform data, and possess more mature distribution channels.

Standalone devices still have their window. They can work across phone brands, can be removed from the phone back and placed on a meeting table, and can tailor experiences around specific user groups. Giants can replicate a display or a recording capability, but may not be willing to build complete services for every niche audience.

But if startups can't build sufficiently deep device habits around certain user segments, the demand they validate may instead be absorbed by phone manufacturers. How long that window lasts depends on how fast startups run and how much platform capability phone makers are willing to open up.

Harder to answer than hardware and competition is why users ultimately buy it.

Watches found health; Plaud found recording. After recording, displays, and agents converge, they still need a sufficiently clear scenario—and "all-day personal assistant" doesn't yet qualify as an answer.

Additionally, continuous recording raises privacy boundary issues. A device that can record all day doesn't mean users and the people around them will accept it operating in all contexts. Clear recording indicators, easy start-stop controls, and user control over data去向 are prerequisites for这类 devices to move beyond the conference room.

What needs closer observation next is: after the novelty fades, how many users continue using it, remain willing to pay for software, and don't remove it from the phone back.

The phone back has never lacked capabilities to put into it—it lacks a sufficiently compelling reason to use. The competition ahead comes down to who can find balance between "useful" and "imperceptible": light enough that users forget it's there, clear enough that users know why they need it, and independent enough to build its own value under the shadow of phone makers.

Whoever achieves these points first is the one most likely to define the phone back's first new category.

AI Hardware Entrepreneurs Are Flocking to the Backs of Phones

图片来源:Plaud 淘宝

图片来源:极稚科技小红书

图片来源:阅星曈

图片来源:Flowtica 官网

图片来源:NOOKLAB 官网

暂无评论,快来发表第一条评论吧!No comments yet. Be the first!

📮 需求咨询📮 Inquiry