文 | 物理AI大观，作者 | 此间少年，编辑 | 卢旭成
“智元机器人赴港上市节奏已经放缓。”“5折出售手中老股，无人愿意接手。”“为限制老股流出，部分具身智能公司已经不同意股改。”“具身智能公司真正的融资压力，或将在11月后集中显现。”“投资人既不关注世界模型，也不看具身智能项目了。”
2026年下半年以来，关于具身智能估值泡沫的争议持续发酵，融资数据直观印证了这种变化：今年10月1日至15日，赛道融资共19起、金额49.34亿元，对比巅峰时期单月339.8亿元的融资总额，热度大幅缩水。如果再考虑不少披露交易其实是存量股权转让，赛道真实新增资金的萎缩，远比账面数字看上去更严峻。
可明明上半年，资本市场还在疯抢具身智能？
2026年上半年，投资机构害怕错过风口，具身智能拿到了当年自动驾驶的火热剧本。天使轮融资数亿元、数月内接连完成多轮融资、半年快速成长为独角兽，具身智能成了市场里的“吸金兽”。
IT桔子数据显示，2026年前8个月，机器人赛道融资额达1245.1亿元，是2025年全年的两倍。同样的狂热，也在海外上演。PitchBook数据显示，2026年上半年，机器人和物理AI企业全球融资334亿美元，已经超过2025年全年的287亿美元。
从害怕错过风口到如今变得谨慎，资本市场为何出现180度大反转？热潮褪去之后，具身智能行业将要面对哪些新压力？
433万元营收VS5亿美元估值：收入与估值的错位
36个交易日，宇树科技股价从开盘价1100元/股跌至404.03元/股，最大跌幅超63%，市值蒸发超2800亿元。宇树科技的上市，成为资本市场态度反转的导火索。
宇树之外，机器人零部件厂商梅卡曼德公开发售斩获约3835倍超额认购，刷新港股机器人赛道认购纪录。但上市首日直接破发，此后股价持续走弱。
接连的破发，让资本市场重新思考：此前资本到底是为具身智能的明星团队光环买单，还是看好赛道真实可兑现的长期增长？
当前具身智能独角兽、未来独角兽创始团队大致分为两类：
一类是以顶尖高校科研团队、教授专家为代表的学术派团队。曦诺未来团队汇聚清华、浙大等全球顶尖高校人才，极佳视界核心团队来自清北、中科院、中科大等院校。
另一类则是以大厂技术负责人、业务高管或核心研发人员为主导的产业迁移型创业。原力灵机、Sharpa创始团队分别来自旷视科技和禾赛科技，小鹏机器人脱胎于小鹏汽车。
明星团队+Demo+技术叙事，是支撑独角兽数十亿美元估值、未来独角兽数十亿元估值的关键，业内甚至出现依托明星团队背书的“PPT融资”现象。
但具身智能商业化落地节奏缓慢，是否撑得起这些高估值？
“机器人能表演、能参赛，和机器人能稳定进工厂干活，这是两码事。”安徽某制造企业质检经理张磊（化名）对物理AI大观说道。在他看来，工厂大规模部署机器人，核心要算清三本账。
一是经济账，只有比人力成本更低、效益更高，工厂才愿意复购。“假设一名产线工人一年薪酬七八万元，如果机器人价格更低、产出效益更高，老板自然愿意复购、逐步改造产线。反之，就很难推进。”
二是稳定性账，机器人需要可靠性强，适配复杂工况。“客户那边催着交货，工人可以加班赶工。一旦机器人突发故障，一两天才能修好，延误交付带来的经济损失由谁承担？”
三是责任账，权责边界必须清晰。“不少制造业生产讲究定岗、定人、定责，而具身智能机器人又不会‘背锅’。若机器人引发产线故障增多，责任最后落到部门负责人头上，轻则扣工资、扣绩效，重则可能被公司开除，让他们自然会排斥落地，而一线工人出于岗位被替代的担忧，也可能给机器人‘穿小鞋’。机器人真的能适应产业场景里，这套人情与权责的复杂现实吗？”
张磊的顾虑并非个案。人形机器人场景应用联盟数据显示，2026上半年国内公开人形机器人中标项目共218个，总规模17.23亿元；其中，教育科研机构采购占比55.5%，政府及国资平台占20.6%，来自工业企业的订单仅占21.1%。而工业订单里，绝大多数还停留在样机测试、展厅展示、实训试点，并非产线常态化采购。
订单总量有限，市场还呈现出高度集中的格局。SAG机构2026上半年统计数据显示，智元机器人与宇树科技，合计占据全球超七成市场份额；其余中小厂商上半年出货量均未破千台，头部与腰部企业的规模鸿沟已经拉开。
乐象科技联合CEO李元庆同样表示，无论推出哪款机器人，至少能卖出500台，大量对手会买来拆解逆向。但宇树有着十年积淀、消费四足产品线，叠加9.9万元的价格优势，让外界难以复制，赛道里仅有两三家同行出货破千台。
出货量不足，让具身智能收入与估值严重错位。公开数据显示，达闼机器人2025年前七个月营收仅140万元，净亏损高达8425万元，但巅峰时期估值达到223亿元。2025全年营收仅433万元的盈合机器人，巅峰估值突破5亿美元。这种错位，让资本市场对具身智能项目逐渐冷静。
这背后是，具身智能技术成熟度和企业真实需求之间，仍存在着巨大鸿沟。一边是当前具身智能产业链从上游核心零部件、整机制造，到下游场景适配、部署运维与客户付费，仍面临成本高企、场景碎片化、投资回报周期漫长等多重难题。
另一边却是，制造业既要具身智能攻克复杂环境下的感知、决策与精细操作难题，也要实现高可靠性、低部署成本、稳定正向的经济回报，更要建立清晰的故障责任划分、顺畅的人机协作机制。技术、成本、管理的三道门槛，让具身智能面临着长期的商业化压力。
估值承压，IPO遇阻
长期的商业化焦虑，让具身智能赛道衍生出两种特殊的交易模式：
一种是以数采中心为代表的循环交易模式。梅卡曼德创始人邵天兰公开批评这类模式：“现在很多‘攒局型’具身智能企业，包括北京、上海一些名气很大、估值很高、上过春晚的公司，大量利用所谓‘数采中心’，和地方、投资方、供应商开展关联交易，制造虚假、不可持续的收入。”
《金融时报》拆解了这套模式的运行方式：地方+机器人企业出资建设数采中心→数采中心向企业购买机器人用以数据采集和训练→采集到的数据，再卖给机器人企业，形成“地方建中心→中心采购机器人→采集数据再卖回企业”的循环交易。
另一种则是“既是股东又是客户”的交易模式。大型产业资本既参与投资或合资，又成为具身智能企业的客户，同时还可能向这家企业供应核心零部件，形成资本、产品、供应链和应用场景相互交织的关系。
当市场把重心放到具身智能商业化落地上，一连串疑问随之而来：企业手中的真实订单到底有多少？是否存在互相抬价、推高估值的情况？是否存在利益输送、关联交易？机器人进入企业场景后，稳定运行时长能达到什么水平？脱离股东客户后，企业能否持续开拓外部新客户？
这些质疑，让具身智能企业开始承受估值回调压力。《路透社》报道称，若剔除数采中心相关收入，部分机器人企业的估值可能下跌60%—70%，当前私募项目估值已经下调约30%—50%。
估值下调的背后，是市场难以看清企业的经营底色。若上市后股价持续走弱，风险将在二级市场暴露。监管也因此对机器人企业IPO审核趋严，企业上市压力显著上升。
《财联社》报道称，营收真实性、盈利稳定性、商业化落地能力，已经成为机器人IPO的核心审核导向。汇生国际资本总裁黄立冲提出，具身智能公司上市时，最难向资本市场解释的是：技术创造的价值，最终有多少能留在企业账上。模型能力提升、客户效率改善，并不等于企业利润同步增长。“客户是否愿意多付钱？公司有没有定价权？扣掉算力、数据处理、实施和人工复核之后，还剩多少？这些都要算清楚。”
Wind数据显示，港交所有超过50家具身智能赛道企业正在排队上市。虽然尚难以判断最终有多少企业能够顺利完成发行，但市场已经出现主动调整上市节奏的案例：原定10月28日挂牌、冲刺“港股工业具身智能第一股”的微亿智造，临时暂停上市进程；中坚科技也宣布终止H股上市计划。
IPO审核趋严、破发风险高悬，投资机构对具身智能赛道变得更加冷静。目前具身智能项目依旧高度依赖IPO退出。假设某标的以100亿估值入局，IPO发行10%流通盘，上市后市值冲高至300亿，6个月锁定期结束回落至150亿。账面收益看似50%，但叠加时间成本、流动性折价与经营不确定性，实际回报吸引力大打折扣。
资本市场的谨慎，又反向制约具身智能企业的IPO进程。以港股IPO认购为例，基石投资者认购股票后需要接受六个月的锁定期，优地机器人锁定期甚至长达九个月。
具身智能企业估值偏高，发行所需的基石资金从千万港元到几亿港元不等。公开数据显示，梅卡曼德共引入9家基石投资者，合计认购约1.86亿美元。
投中网数据显示，截至2026年10月，13家赴港上市的机器人企业中，有10家引入基石投资者，合计认购约48.61亿港元。但这10家里面，精锋医疗和华沿机器人的基石阵容分别浮亏1.52亿、4.57亿港元，合计亏损达6.09亿港元。
其他机构浮亏的前车之鉴、长达半年的锁定期、上市后的股价不确定性，后续具身智能IPO，机构是否愿意大额进场？投资人是否愿意承担破发带来的内部追责风险？种种不确定性，又是否会出现老股东不愿托底、新股东不敢进场的局面。具身智能企业的IPO融资及后续资本接续，压力被进一步放大。
热潮退去，具身智能的周期考验
解决融资压力的关键，仍是具身智能企业回归研发本身，加速商业化落地。2026年，赛道资金大多流向产业链中上游。据物理AI大观不完全统计，33家机器人独角兽中，29家集中在上中游，占比约88%；36家未来独角兽中，31家位于上中游，占比约86%。
艾瑞咨询数据显示，2026年全球机器人软件AI基础、具身智能大脑板块融资额分别达1670.93亿元、1625.52亿元，遥遥领先；下游行业应用、公共服务机器人等落地端融资仅十余亿元。资本大量涌向算法层，负责量产交付、场景验证的整机与落地赛道资金紧缺。
但当GPT-6 Astra已经能够完成3D建模、物理AI仿真等任务；Kimi K3.1 Pro可直接编写Blender、CAD脚本搭建3D模型，Astra正在持续冲击行业原有技术壁垒，资本还会继续押注具身智能的上层模型赛道吗？
对于专注具身智能模型、世界模型的企业而言，如果核心优势仅停留在环境理解、任务规划这类上层能力，就有可能被通用大模型覆盖替代。元戎启行CEO周光提到，GPT-6问世之后，宇树这类机器人本体公司反而更占优势，主打具身智能大脑、机器人大模型的企业将迎来巨大挑战。
与之相对，让机器人在真实物理环境中稳定完成复杂操作，并依靠真实交互数据持续提升任务成功率，才是更难被替代的竞争壁垒，这也进一步抬高了具身智能企业的研发门槛。
相比外部通用模型迭代带来的冲击，物理AI大观认为，更深层的研发风险来自企业内部。当前具身智能企业大致分为两类：一类是长期扎根真实场景，稳步打磨产品；另一类是被资本快速催熟的企业，长期维持高成本、高薪资模式。例如，盈合机器人曾被曝出耗资2800万元装修办公室、一次性发放500万元年终奖，引发投资方强烈不满。
一边是此前高薪优待、资本追捧的状态，另一边是融资收缩后必须降本增效、压缩开支。巨大落差之下，不少具身智能企业或将迎来研发人员离职潮，人才资源进一步向头部集中。
不止研发端，商业化落地环节同样存在人才适配难题。来自中科大的生物学博士朱阳（化名）提到，之前他师兄想让他加入一家具身智能企业，开出的薪资十分可观，但他最终还是选择留在学术界。
他解释，很多高校教授常年带着硕博学生，手握科研项目与经费，甚至能影响学生毕业，处在甲方位置，压力和责任边界清晰。可一旦进入企业，抛开环境差异不谈，最大的转变是身份：
在学校做科研，你是甲方；到企业做具身智能，就要成为服务客户的乙方，各类苛刻需求都要承接。做教授，核心是申请经费、花钱做研究；做企业，核心是赚钱，还要承担经营风险。这种身份与思维上的落差，很多科研出身的人很难适应。
具身智能依旧是长期重要的技术趋势。这轮资本冷静期，只是加速行业回归产业本质：技术不能止步于实验室，产品要经得起真实场景检验，企业最终要靠商业化能力证明价值。热潮终会褪去，只有真正创造产业价值的企业，才有机会穿越周期。
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
I'm a young man, I'm an editor.
“The speed at which the Jiwon robots go to the Hong Kong market has slowed.” “The old stock is sold at a discount and no one is willing to take over.” “In order to limit the flow of old stock, some smart companies have not agreed to a change of stock.” “The real financial pressure on smart companies will appear after November. “Investors are not interested in world models, nor in smart projects.”
Since the second half of 2026, disputes over smart valuation bubbles have continued to ferment, a change evidenced intuitively by financing data: from 1 to 15 October this year, track financing totalled 19 out of a total of $4,934 million, a significant decrease in heat compared to the total of $33,980 million a month during the peak period. If one considers that many of the disclosure transactions are in fact stock transfers, the decline of real new funds on the tracks is much more severe than the book figures.
But in the first half of the year, the capital market is still out of its mind?
In the first half of 2026, the investment agencies, fearing to miss the windfall, acquired the hot script of self-driving in the year. Angel Wheels finance hundreds of millions of dollars, multiple rounds of financing in months, and six months of rapid growth into unicorns, with smarts becoming “goldards” in the market.
IT orange data show that in the first eight months of 2026, robotic track financing amounted to $124.510 billion, twice as much as in 2025. The same frenzy, also performed abroad. The PitchBook data show that in the first half of 2026, US$ 33.4 billion in global financing for robotics and physics companies has exceeded US$ 28.7 billion throughout 2025.
From fear of missing the wind to being cautious today, why is there a 180-degree reversal in capital markets? What new pressures will the smart industry face when the tide fades?
Valuation of VS500 million received by $4.33 million: misalignment of income with valuation
On 36 trading days, the Waki technology share price dropped from $1,100 to $404.03 per share, with a maximum fall of over 63 per cent and evaporation of over $280 billion. The marketing of woo tree technology has become a trigger for the reversal of attitudes in capital markets.
In addition to Woo-ki, the robotic parts manufacturer, Mercamand, has publicly issued about 3,835 times oversubscribed and has updated the Hong Kong stock robot track. However, the first day of the market was a direct break, after which stock prices continued to weaken.
A series of break-ups that rethought capital markets: Before that, did capital pay for a smart team of stars, or did it look at the real and real long-term growth of the track?
The current intelligent unicorn and future unicorn founding team are broadly divided into two groups:
One is an academic team represented by the leading scientific team of higher education institutions, and by the expert professor. The Senor Future Team brings together the world's top high-schools, such as Zhejiang and Zhejiang, with the best-looking core teams coming from Qingbei, the Middle School and the Middle School.
Another category is the relocation of industries led by large plant technologists, business executives or core R&D personnel. The Elixir machine and Sharpa's founding team came from the technology of visual invisibility and from the technology of whaling.
Star team+Demo+ technology narratives are key to underpinning the billion-dollar unicorn valuation, the future unicorn valuation of billions of dollars, and even the “PPT financing” that depends on star team endorsement.
But is the slow pace of smart commercialization enough to sustain these high valuations?
"It's not the same thing that robots can perform, compete, and robots can stabilize into factories to do work." Zhang Zhang Zhang (alias) the manager of a manufacturing company in Anhui says to the physical AI view. In his view, the plant had a large-scale deployment of robots, and the core had to settle three books.
The first is economics, and factories are willing to repurchase only if it is cheaper and more cost-effective than manpower. “Assuming that a producer worker is paid $780,000 a year, if the robot is cheaper and more productive, the owner is naturally willing to repurchase and gradually re-engineer the line. On the contrary, it's hard to push."
The second is stability. The robot needs to be reliable and complex. “Customs are rushing for delivery and workers can work overtime. Who will bear the financial cost of the delay in delivery if the robot fails in a day or two?”
The third is accountability, with clear lines of authority and accountability. “Many manufacturing industries are targeted, man-made, accountable, and smart robots are not `back-to-back'. If robots cause an increase in production line failure, the responsibility ultimately falls on the head of the department, with less pay, less performance and less weight, they may be dismissed from the company, which will naturally exclude them, while first-line workers may be given “shoes” to robots for fear of being replaced by their jobs. Can robots really adapt to the complex reality of humanism and accountability in industry?"
Zhang Yi's concerns are not individual cases. Data from the Union for the Application of Man-Framework Sites shows that in the first half of 2026 there were 218 open-form robots in the country, with a total of RMB 1,723 million; of these, 55.5 per cent were procured by educational and scientific institutions, 20.6 per cent by the Government and the State, and 21.1 per cent were ordered from industrial enterprises. On the other hand, the vast majority of industrial orders are still in prototype testing, exhibition hall displays, hands-on piloting and are not routing.
The total volume of orders is limited and the market is also highly concentrated. According to SAG statistics for the first half of 2026, intellectual robotics and wild tree technology together account for over 70 per cent of the world ' s market; the remaining small and medium-sized producers did not lose 1,000 of their exports in the first half of the year, and the gap between the size of the head and the waist has widened.
CEO Lee Yuanqing, too, said that at least 500 robots could be sold, regardless of which one was launched, and that a large number of opponents would buy them to break the tide. However, with a 10-year-old, four-foot product line, with a price advantage of $99,000, it is difficult to replicate from outside, with only two or three of the tracks running out of thousands.
The low volume of deliveries is a serious mismatch between smart income and valuation. Public data show that the droids received only $1.4 million in the first seven months of 2025, with a net loss of $8.425 million, but the peak was valued at $22.3 billion. Only $4.33 million worth of condensed robots were collected throughout 2025, peaking at $500 million. This mismatch has calmed the capital market against smart projects.
Behind this is the huge gap between the maturity of smart technologies and the real needs of enterprises. On the one hand, the current smart industrial chain, from upstream core parts, mechanization to downstream adaptation, deployment of transportation and client paid, still faces multiple challenges, including high cost, fragmentation of the scene and long return-on-investment cycles.
On the other hand, the manufacturing industry needs to be smart enough to deal with the perceived, decision-making and operational challenges of complex environments, as well as to achieve high reliability, low deployment costs, stable positive economic returns and, more importantly, to establish clear lines of responsibility for failure and smooth human mechanisms. The three thresholds of technology, cost and management expose smarts to long-term commercial pressure.
Valuation under pressure, IPO blocked
Long-term commercial anxiety, which gave rise to two special trading patterns in a smart track:
One is a circular trade model represented by digital centres. The founder of Mercamand, Shao Tian Lan, has publicly criticized this model: "Many "silent" smart businesses now, including some in Beijing, Shanghai, with high reputations, high valuations, and late springs, have made extensive use of the so-called, ` Digital centres', and local, investor, supplier-related transactions to generate false and unsustainable income.”
The Financial Times has broken down the mode of operation of this model: local + robotics have funded the construction of digital centres to buy data from robotics for data collection and training and sell it to robotics enterprises, and a circular “local center for robotics for data collection and re-selling back to business”.
The other is the “both shareholders and customers” mode of dealing. Large industrial capital becomes both an investment or joint venture and a customer of a smart enterprise, and may supply the enterprise with core components, creating interlocking links between capital, products, supply chains and applications.
When the market puts its focus on smart commercialization, a series of questions follow: How many real orders do business have? Is there an increase in the value of each other? Is there a transfer of benefits, a related transaction? What level can robots reach when they enter the business scene? Can enterprises continue to develop new external customers after leaving shareholders?
These challenges have led to the start of valuation-back pressure on smart enterprises. Reuters reports that, if revenues associated with digital centres are excluded, the valuation of some robotic enterprises may fall by 60 per cent - 70 per cent - and that the current private sector project valuation has been reduced by about 30 per cent - 50 per cent.
Behind the downward revision of the valuation is the market ' s difficulty in seeing the business ' s background. If after-market prices continue to weaken, the risk will be exposed to the secondary market. As a result, controls have become more stringent on the IPO of robotics, and there has been a significant increase in the pressure on companies to market.
The Associated Press reported that realism, profitability and commercial landing capacity have become the core audit orientation of the IPO. The Managing Director of Swift International Capital argued that when smart companies were listed, it was the most difficult to explain to capital markets the value of technology creation and, ultimately, how much could be left on the corporate books. Improved modelling capacity and improved client efficiency do not equate to a synchronized increase in business profits. “Is the client willing to pay more? Do companies have pricing rights? How much remains after the deduction of computing power, data processing, implementation and manual review? That's what you need to figure out."
Wind data indicates that all Hong Kong enterprises with more than 50 furniture-based smart tracks are queuing to market. Although it is difficult to judge how many enterprises will eventually be able to successfully complete their distribution, there have been cases in which the market has proactively adjusted the market pace: On 28 October, China announced the termination of the H stock listing programme.
IPO audits have become more stringent and risk-prone, and investment agencies have become more calm about the smart track. Today, smart projects remain highly dependent on IPO exits. Assuming a target of 10 billion in value, IPO issues 10 per cent of its negotiable disks, the market value of which goes up to 30 billion after listing, and the six-month lock ends regularly to 15 billion. The booking gains appear to be 50 per cent, but the real return attractiveness is significantly reduced by a combination of time costs, liquidity discounts and operational uncertainties.
Caution in capital markets reverses the IPO process for smart firms. In the case of IPO subscriptions, a six-month lock period is required after a capitalist investor subscribes to the stock, and a nine-month period or more.
Smart enterprises are overvalued, with capital requirements ranging from HK$ millions to HK$ hundreds of millions for their distribution. Public data show that Mercamand introduced nine capital investors, with a combined subscription of approximately $186 million.
As of October 2026, 10 of the 13 robotics that were listed in Hong Kong had introduced capital investors, with a combined subscription of approximately HK$ 4,861 million. However, in 10 of these households, the cornerstones of the robotics and the elites suffered a loss of HK$ 152 million and HK$ 457 million, respectively, totalling HK$ 609 million.
Are institutions willing to enter the field in large amounts, given the pre-deficit experience of other agencies, the six-month-long lock-up, the uncertainty of stock prices after listing, and the subsequent intelligent IPO? Are investors willing to take on the risk of internal accountability as a result of a breakout? There is uncertainty as to whether there will be a situation in which old shareholders are reluctant to go down and new shareholders are afraid to come in. The pressure was further amplified by the succession of IPO financing and subsequent capital of smart enterprises.
The tide retreats, the smart periodic test.
The key to addressing financial pressures remains the return of smart enterprises to R & D itself and the acceleration of commercialization. In 2026, most of the money went to the upper part of the industrial chain. According to incomplete physical AI statistics, of the 33 robot unicorns, 29 are concentrated in the upper middle, accounting for about 88 per cent, and of the 36 future unicorns, 31 are located in the upper middle, accounting for about 86 per cent.
According to Irene Consulting data, in 2026, global access to the AIS base and smart brain block financing amounted to $167,093 million and $162,552 million, respectively, leading the way; downside financing for downstream industry applications and public service robots was only over a billion dollars. Capital surges to the algorithm, and there is a shortage of funds for the delivery of volume, the whole machine for scene validation and the landing track.
But when GPT-6 Astra has been able to complete the 3D modelling, physical AI simulations, etc.; Kimi K3.1 Pro can develop the 3D model directly with the Blender and CAD scripts, and Astra is continuing to hit the old technical barriers in the industry, and capital will continue to be invested in intelligent upper model tracks?
For enterprises that focus on smart models and world models, if the core advantage is confined to upper-level capabilities such as environmental understanding, mission planning, then it may be replaced by generic LLM coverage. The CEO Zhou of Yuanjun mentioned that after GPT-6 came into existence, the robotics of Woo-ki became more dominant, and that the big challenge would be for companies with smart brains and robots LLM.
By contrast, allowing robots to perform complex operations in real physical environments in a stable manner and relying on real interactive data for the continuous upgrading of mission success rate is a much more difficult competitive barrier to be replaced, further raising the research and development threshold for smart enterprises.
Compared with the impact of external generic models, the physical AI view suggests that deeper R & D risks come from within the enterprise. There are two broad categories of today ' s smart enterprises: those with long-term roots and steady grinding of products; and those with fast-tracked capital, which maintain high-cost and high-pay patterns over the long term. For example, the YCG robots were exposed to a cost of $28 million to renovate their offices and a one-time end-of-year award of $5 million, which provoked strong dissatisfaction among investors.
On the one hand, the former state of high pay and capital attraction, and on the other hand, the contraction of financing must reduce efficiency gains and reduce expenditures. With significant gaps, many smart enterprises or a wave of departures of R & D staff will lead to further concentration of human resources in the head.
Not only the research and development end, but also the problem of human suitability at the commercialization site. Dr. Zhuyang (alias) from the middle school mentioned that his brother had previously wanted him to join a furniture-smart enterprise with a very good pay, but he had eventually chosen to remain in academia.
He explained that many higher education professors, with their rich students on a perennial basis, were able to hold on to scientific projects and funding, and even to influence students to graduate, to be in the A position, and to have clear boundaries of stress and responsibility. Once in a business, however, and apart from environmental differences, the biggest change is identity:
In school, you're on the A side; when an enterprise is smart, it becomes the B side of the service, and all kinds of demanding demands are met. As a professor, the core is to apply for funding and to do research; as a business, the core is to earn money and to take on business risks. This difference in identity and mindset makes it difficult for many people of scientific origin to adapt.
Smartness remains an important long-term technological trend. This round of capital cooling-down only accelerates the return of industry to the nature of industry: technology cannot stop at a laboratory, products can withstand real scenery testing, and businesses ultimately rely on commercialization to prove value. The tide will eventually fade, and only businesses that truly create industrial value will have the opportunity to cross the cycle.
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.