1. The Democratization of AI
1. AI的普及化
在当下的科技领域中，人工智能（AI）无疑是最具颠覆性与影响深远的趋势之一。
与2017年相比，当今的AI应用率已提升至约当年的2.5倍。
值得强调的是，已有超过50%的企业在至少一个关键业务功能中融入了AI的力量。还是那句话“所有行业都可以用 AI 重新做一遍”。
目前人工智能正大步跨入金融、医疗、制造、零售等众多行业领域，影响范围正不断扩张。
更重要的是，AI早已不再是大型企业的特权，各类组织都可轻松挖掘其潜力。
在开源AI方案的助力下，加之成本与系统复杂度的显著降低，AI的普及化进程正如火如荼地展开。
其中，最具代表性的案例莫过于OpenAI——这家支持非营利与商业研发的机构正是ChatGPT背后的推手，为AI的全面平民化提供了生动的范例。
这家公司当前的估值已高达1500亿美元，并预期在今年（2024）年实现高达10亿美元的营收目标。
当OpenAI在2022年11月推出ChatGPT时，它让整个世界为之震撼。
这款聊天机器人可通过自然语言指令，创造出多样化且流畅的对话文本，从而彻底刷新了人们对AI潜能的想象。
在面世后的短短两个月内，已有超过1亿人使用了ChatGPT，这一惊人的数字见证了AI走向大众的宏大趋势。
继ChatGPT之后，谷歌又推出了自己的AI聊天机器人Gemini。
微软同样推出了整合OpenAI技术的必应（Bing）聊天机器人。
就连Meta也不甘落后，正专注于Llama这一开源大型语言模型，意图在这场AI竞赛中占据一席之地。
虽然这些技术仍处于早期阶段，但大型语言模型（LLMs）无疑有望为企业运作方式带来颠覆性的变化。
举个例子，客服人员可以借助LLM在数秒内回应客户咨询；企业无需额外聘请文案人员，也可轻松创建个性化的营销与教育内容；开发者则能依托其快速编写复杂代码，而企业高层则可利用其高效分析数据。
各类组织正以多种方式将AI融入自身业务中。
普华永道（PwC）的一项调查显示，那些正在采用AI的商业领袖的首要目标包括：通过自动化提升生产力、由数据驱动的洞察优化决策、以及为客户带来更出色的体验。
以Eleos Health为例，这是一家致力于为治疗师提供AI驱动自动化平台的公司。
他们的CareOps Automation解决方案是一款语音式AI平台，能够在治疗会谈的背景中悄然记录、自动数字化所有对话内容，并即时识别出潜在的干预点，同时迅速产出会话总结。
目前，该公司已宣布累计融资金额达到6800万美元。
再举个例子，制造业和仓储领域正不断探索AI的全新应用方式。
在仓库管理方面，AI的图像识别能力可实现库存的自动化监控，不仅能在库存不足时及时发出警示、提醒人力介入，还能将生产进度或供应链延迟等因素纳入考虑。
在制造工厂中，AI平台更是“智慧工厂”趋势的重要推手。这些系统可进行预测性维护，减少资源浪费，并有效提升工人作业环境的安全性。
$200 billion was invested in Cleantech companies last year alone.
(That's up 70% YoY)
In fact, clean technology has gained so much momentum that more than 25% of all VC dollars now flow to cleantech companies.
Industry experts suggest that more funding and interest is on the way due, in part, to the Inflation Reduction Act.
The law includes loans, grants, and tax incentives aimed at encouraging the private sector to devote more dollars and time to cleantech.
The leader of Breakthrough Energy Ventures estimates this act will lead to the creation of between 300 and 1,000 new companies.
Many of these companies could be in the green hydrogen industry.
Hydrogen is the most abundant element on Earth and burning it doesn’t release CO2, giving it great potential as a source of green energy.
The green hydrogen market is expected to grow at a CAGR of 61% through 2027 and surpass $7 billion in value.
The Hydrogen Council estimates that around $700 billion in hydrogen-focused investments will be needed in order for the world to reach net-zero emissions by 2050.
They report that 680 large-scale hydrogen projects had been announced as of May 2022 — an increase of 160 projects compared to 2021.
In addition, the US Department of Energy announced $47 million in available funding for clean hydrogen technologies in early 2023.
Plug Power is one of the largest companies producing hydrogen fuel cell systems in the United States.
The company opened a hydrogen fuel cell manufacturing plant in Slingerlands, New York, in 2022.
On the heels of new partnerships with TC Energy and Nikola, the company’s stock price is up nearly 30% as of 2023.
Non-US investments in hydrogen are also on the rise.
A hydrogen-powered passenger train debuted in Germany in September 2022. The train can run more than 600 miles without having to refuel and the top speed is 86 miles per hour.
Germany plans to transition 2,500 to 3,000 of its trains to hydrogen fuel in the coming years.
But, while green initiatives — like planting trees and switching to hydrogen-powered vehicles — can reduce carbon emissions over time, many experts believe the impact of these efforts alone will be too little too late.
According to the Brookings Institution, global greenhouse gas emissions hit 58 gigatons in 2022. That’s the largest amount ever recorded.
It’s well-documented that carbon emissions are one of the largest drivers of modern climate change.
In order to directly eliminate some of these emissions, cleantech leaders are using what’s known as carbon capture technology to make immediate progress toward reducing and even reversing emissions.
The process involves working with super-emitters — like power plants and concrete manufacturing facilities — to capture carbon molecules when they would normally be released into the air.
Carbon capture can effectively remove up to 90% of emissions from the air released at power plants and industrial facilities.
From there, carbon capture companies isolate and extract the carbon through a variety of chemical processes before reselling it or depositing it deep into the earth, where it can be transformed back into stone.
PwC reports that funding for carbon capture technology in the first three quarters of 2022 was nearly double that of total funding for 2021.
在过去的一年中，新能源科技企业共获得了2000亿美元的投资资金。
（这一数字较上一年增长了70%）
事实上，新能源的势头已然高涨，如今超过四分之一的风险投资资金正源源不断地流入这一领域。
行业专家预测，受《降低通货膨胀法案》（Inflation Reduction Act）的推动，新能源领域的资金与关注度将进一步攀升。
这项法案提供了贷款、补助与税收优惠，旨在鼓励私营部门投入更多资金与精力到新能源领域中。
Breakthrough Energy Ventures的负责人预计，此举将催生约300至1000家新兴企业。
其中，很多可能将活跃于绿色氢能产业。
氢作为地球上最为丰富的元素，其燃烧过程并不产生二氧化碳，赋予其作为绿色能源来源的巨大潜能。
绿色氢能市场有望在2027年前以约61%的年复合增长率快速扩张，总市值将突破70亿美元。
根据氢能理事会（Hydrogen Council）的估算，要在2050年前实现全球净零排放，约需7000亿美元的氢能相关投资注入。
他们的报告显示，截至2022年5月，已宣布的大型氢能项目数量达到680个，较2021年增加了160个。（目前我找了只有22年数据）
此外，美国能源部在2023年初宣布，将为能源氢能技术提供4700万美元的可用资金支持。
Plug Power是美国最大的氢燃料电池系统生产企业之一。
该公司于2022年在纽约州的Slingerlands建立了一家氢燃料电池制造厂。
伴随着与TC Energy和Nikola等企业达成全新合作关系，该公司的股价在2023年已上涨近30%。
与此同时，非美地区对氢能的投资也在不断上升。
德国于2022年9月推出了一列以氢为动力的客运列车，该列车无需中途加氢就能行驶600多英里，最高时速可达86英里/小时。
德国计划在未来几年内将2500至3000列列车转向使用氢燃料。
然而，尽管植树造林、普及氢能车辆等绿色举措的确能在长期内逐步降低碳排放，许多专家却认为，这些努力的影响可能仍然不足以及时扭转局势。
根据布鲁金斯学会（Brookings Institution）的数据，2022年全球温室气体排放量已达580亿吨，创下历史新高。
毫无疑问，碳排放已成为现代气候变化的最主要推手之一。
为直接减少部分排放量，一些新能源领域的领军者正利用所谓的“碳捕获”技术，以求在减排甚至逆转排放的进程中取得立竿见影的成效。
这一过程的核心在于与“超级排放源”进行合作——如发电厂和水泥制造厂——在碳分子原本会排放到大气中时将其捕获。
通过碳捕获技术，电厂和工业设施排放的碳有望减少多达90%。
碳捕获企业利用多种化学工艺将捕获的碳分离提取，然后将其再次出售，或是深埋入地层，使其重新转化为坚石。
普华永道（PwC）的报告显示，2022年前三个季度，碳捕获技术所获投资额几乎是2021年全年总额的两倍。（目前市面上也只有22年的数据参考）
3. Tech for Early Disease Detection Advances
New tech solutions are focused on identifying diseases earlier on.
This one innovation could help save countless lives.
For example, if ovarian cancer is caught while it’s still localized, the five-year survival rate is 93%.
A similar increase in life expectancy is seen for patients with melanoma. There’s a 99% five-year survival rate for localized cancer, but only a 32% survival rate after the cancer has spread.
Unfortunately, certain types of cancer, like those that affect the pancreas, lungs, and ovaries, are notoriously difficult to catch early.
In an effort to increase survival rates, researchers and tech industry startups are dedicating an increasing amount of resources to cutting-edge technologies designed to detect cancer cells earlier in development.
Take pancreatic cancer, for example.
A company called Biological Dynamics has developed a lab-on-a-chip test that looks for bio-markers specific to pancreatic cancer. The test is entering human trials now.
So far, the company has raised $125 million in funding and hopes to build on the success of the pancreatic cancer test in order to launch tests for lung cancer and ovarian cancer.
Using AI for early diagnosis is another trend in the health tech space.
Researchers at MIT are using AI models to assess lung cancer risk in patients.
The tool was trained on six years worth of lung scans from patients in the United States and Taiwan. From these scans, the tool learned to identify and categorize patterns.
It’s best at predicting lung cancer that will occur within one year, but is also capable of predicting the disease up to six years in advance.
And it’s not just cancer that tech tools can detect early.
Scientists are leveraging nanotechnology that fits inside a smartphone camera in order to diagnose a wide variety of diseases.
The idea is that this nanotechnology makes phase imaging, an advanced way of viewing cells, available to patients at home.
If this technology makes it through trials, patients will be able to take an image of their saliva or a pinprick of blood at home using phase imaging on their camera and send it directly to their medical provider for analysis.
One at-home medical test that uses smartphone technology that’s already cleared for use is the Minuteful Kidney test from Healthy.io.
The test is able to measure a patient’s albumin-to-creatinine ratio (ACR) in order to diagnose chronic kidney disease.
Patients first receive a test kit in the mail and the Minuteful Kidney app provides instructions for collecting a urine sample and dipping a test strip in the urine. After that, the patient takes a photo of the test strip with their smartphone and the app analyzes it.
The app utilizes AI technology and computer vision in order to assess albumin levels. The results are available almost immediately.
The technology has been available in Europe for more than 18 months and more than 540,000 people have already enrolled.
3. 早期疾病检测技术的精进
新兴技术方案正聚焦于更早期地发现疾病。
这一创新有望拯救无数生命。
举例来说，如果卵巢癌能在仍局限于局部阶段时被诊断，其五年存活率可达到93%。
对于黑色素瘤患者来说，类似的趋势同样显著：当癌症仍然局限于最初部位时，五年存活率可高达99%；然而，一旦癌症扩散，这一生存率则骤降至32%。
某些癌症类型——如胰腺癌、肺癌和卵巢癌——以其早期诊断的高难度而闻名。
为了提高存活率，研究人员与科技初创公司正加大资源投入，致力于研发前沿技术，从而在更早的阶段检测到癌细胞。
以胰腺癌为例：
一家名为Biological Dynamics的企业研发出一款“芯片实验室”（lab-on-a-chip）测试，可通过分析特定的生物标记物来识别胰腺癌迹象。目前，该测试正进入人体临床试验阶段。
该公司迄今已募集了1.25亿美元的资金，并希望在胰腺癌测试的成功基础上，进一步推出针对肺癌和卵巢癌的检测方案。
利用AI进行早期诊断是医疗科技领域的另一趋势。
麻省理工学院（MIT）的研究人员正利用人工智能模型来评估患者罹患肺癌的风险。
这一工具利用来自美国和台湾患者长达6年的肺部扫描数据进行训练，从中识别和分类各类影像特征。
目前，该工具最擅长预测一年内可能发生的肺癌，同时也能在最早提前6年预估疾病风险。
更重要的是，科技工具的早期检测能力并不限于癌症。
科学家们正利用适用于智能手机相机的纳米技术进行各种疾病的早期诊断。
这个构想是借助相位成像（phase imaging）这类先进的细胞观察手段，使患者在家中就能获取所需图像。
一旦该技术通过测试，患者便可在家中拍摄唾液样本或少量血液的图像，并直接发送给医疗服务提供者进行分析。
事实上，一款已获许可、利用智能手机进行家庭检测的设备已投入使用，那就是Healthy.io的Minuteful Kidney肾脏检测工具。
该测试可测量患者的白蛋白与肌酐比值（ACR），从而诊断慢性肾病。
患者会先收到邮寄的检测套件，然后按照Minuteful Kidney应用程序的指引采集尿液样本，并将试纸浸入其中。接着，患者只需用智能手机拍摄试纸图像，应用程序即可对其进行分析。
该应用程序借助人工智能和计算机视觉技术评估白蛋白水平，并几乎能即时生成结果。
这项技术在欧洲已应用超过18个月，并已有超过54万人参与测试。
4. Quantum Computing Moves Toward Real-World Application
Quantum computing has been a topic of discussion since the 1980s.
Fast forward to 2024 and the world is finally close to achieving real-world applications for this type of computing.
Traditional computers we know today operate on binary code (either 0 or 1). Quantum computers use qubits, which allows a piece of data to exist in two states at the same time (both 0 and 1).
All of this technology boils down to an increased speed of computation.
Complex calculations done by today’s computers might take millions of years, but they can be solved in minutes with quantum computing.
Technological and financial issues have plagued the industry, but momentum has been building in recent months.
In fact, $1.7B was invested in the quantum computing startups last year.
So far, the leader in the race to quantum computing appears to be IBM.
Osprey, IBM’s quantum computer, was unveiled in November 2022 and boasts 433 qubits.
The company says they’ll have a computer with 4,000 qubits by 2025.
However, tech experts say quantum computers need millions of qubits to be fully functional. Many expect that to come to fruition by 2027.
Alphabet has been operating a quantum computing division for the past six years and, in March 2022, announced it would become a standalone company called Sandbox AQ.
They received “nine figure” funding in 2022 and had raised another $500 million by February 2023.
Although projects from IBM and Alphabet show promise, other lesser-known companies are creating competition and raising large amounts of funding in the space.
For example, Origin Quantum, a Chinese company, raised $148.2 million in 2022, the largest amount of the quantum computing industry for the year.
Q-CTRL was launched as Australia’s first VC-backed quantum tech company in 2017.
Company leaders are focused on overcoming the industry’s challenges associated with hardware error and instability through quantum control infrastructure software.
The company closed a $27.4 million Series B funding round in early 2023.
Q-CTRL already has contracts with more than 8,000 users. That includes the US government and Australian defense agencies, as well as companies like IBM, IonQ, and others.
As companies race to develop this new technology, leaders in several industries are anticipating the potential impact of quantum computing. That impact could be huge for a number of sectors.
A report from McKinsey says quantum computing will provide the highest value in the life sciences and financial sectors.
Up to $700 billion in value may be realized by 2035, according to McKinsey.
In the life sciences industry, quantum computing could be valuable in simulating chemical processes, optimizing pharmaceutical designs, and advancing the development of personalized medical treatment through genomics.
In the financial sector, quantum computing could drastically reduce market risks, improve fraud detection, and speed up customer onboarding.
Quantum computing could have various consumer-facing impacts, as well.
Take electric vehicle charging, for instance.
It takes an average of 10 hours to fully recharge an electric car at home. Even the fastest charging speeds still take 20 minutes.
With quantum technology, charging time could potentially be cut down to just three minutes at home and a few seconds at high-speed charging stations.
4. 量子计算迈向真实应用场景
自上世纪80年代以来，量子计算一直是备受关注的话题。
迈入2024年，世界终于正接近将这种计算方式落地于真实应用中。
与当下以0或1为基础的二进制计算机不同，量子计算机使用“量子位”（qubit），允许同一数据同时存在于0和1两种状态。
这一技术特性最终带来了运算速度的惊人跃升。
当今计算机可能需要数百万年才能完成的复杂运算，通过量子计算却有望在数分钟内解决。
尽管技术和资金难题曾长期困扰该行业，但近几个月来，量子计算领域已明显呈现积蓄力量的趋势。
事实上，去年就有约17亿美元的投资流入量子计算初创公司。
他们于2022年11月发布的量子计算机“Osprey”拥有433个量子位（qubits）。
IBM还声称，将在2025年推出拥有4000个量子位的量子计算机。
然而，科技专家指出，要实现量子计算的全面性能发挥，仍需达到数百万个量子位的量级。许多人预计，这样的里程碑有望在2027年之前实现。
Alphabet（谷歌母公司）在过去六年中一直运营着自己的量子计算部门，并于2022年3月宣布将其独立为一家名为“Sandbox AQ”的公司。
该团队在2022年获得了九位数级别的资金注入，并在2023年2月前又追加融资5亿美元。
2024年12月9日，谷歌宣布研发出了一款运算能力超强、适用量子计算机的芯片，宣称这种芯片只用5分钟即可完成现有运行速度最快的计算机要10尧（10的25次方）年才能完成的任务。
谷歌当天在网站发布消息说，除了高速运算能力，这款名为“威洛”的芯片还有突出的纠错能力，为研制“实用的大规模量子计算机铺平了道路”。
尽管IBM和Alphabet的项目前景可观，但一些相对不为人知的企业也在为这一领域注入活力并吸引了大量资金。
例如，中国企业本源量子（Origin Quantum）在2022年成功募得约1.482亿美元，成为当年量子计算行业融资额最高的公司。
另外，Q-CTRL于2017年在澳大利亚成立，成为首家由风投支持的澳大利亚量子技术企业。
该公司的领导层正专注于通过量子控制基础设施软件，解决行业在硬件误差与不稳定性方面面临的挑战。
该公司在2023年初完成了2740万美元的B轮融资。
目前，Q-CTRL已与超过8000位用户达成合作关系，其中包括美国政府及澳大利亚国防机构，以及IBM、IonQ等知名企业。
随着各家公司竞相研发这项全新技术，不同行业的领导者正密切关注量子计算可能带来的深远影响。对许多领域来说，这种影响或将巨大无比。
麦肯锡的报告指出，量子计算将在生命科学和金融领域创造最高价值。
根据麦肯锡的分析，到2035年，量子计算所创造的价值有望高达7000亿美元。
在生命科学领域，量子计算可用于模拟化学反应过程，优化药物设计，并通过基因组学推动个性化医疗的发展。
在金融业，量子计算有望大幅降低市场风险、提升诈骗侦测能力，并加快客户入驻流程。
量子计算技术同样可能对消费者生活产生直接影响。
以电动汽车充电为例：
目前，在家为电动车充满电平均需要10个小时，即使是最快速的充电方式也需要20分钟左右。
借助量子技术，这一过程有望被大幅缩短，或许在家中仅需3分钟即可充满电，而在高速充电站可能仅需数秒。
5. Cyber Threats Grow More Advanced
From attacks on random consumers to government-sponsored cyber warfare, cybercrime is a persistent and growing threat.
A 2022 global survey from Hiscox showed 43% of companies reported a cyber attack.
The most alarming statistic from that report was the fact that 20% of attacked organizations said the cost of the damage threatened their solvency.
Some predictions show that a single data breach was expected to cost an average of $5 million in 2023.
IBM’s research says it’s even higher: $9.44 million.
In fact, virtual crime has become so prevalent the $155 billion cybersecurity industry is expected to swell to $376 billion by 2029.
Deepfake attacks are one of the most sophisticated ways hackers are gaining access to businesses.
In a recent survey from VMware, 66% of participating IT leaders said they’d experienced a deepfake-related attack within the past 12 months. That’s an increase of 13% since 2021.
Deepfake technology utilizes AI/deep learning to create convincing videos, images, and audio of fake events and people.
The technology has been around for a few years, and it’s getting better and better for hackers.
A type of machine learning technology called generative adversarial networks is making the deepfake models nearly impossible to detect.
Plus, the emergence of 5G networks is making it easier to manipulate videos in real time.
Deepfakes are especially useful to cyber criminals who commit BEC (business email compromise) scams.
Manipulating face-to-face verification methods with deepfakes is another possibility for cyber criminals.
Identifying deepfakes and other incoming threats is very much a defensive game for organizations. Security professionals are always one step behind attackers.
However, cybersecurity pros are using AI and other advanced technology solutions in order to detect and stop attacks as early as possible.
A recent report from IBM found that organizations that use AI cybersecurity tools and automation extensively save $2.22 million compared to those who don’t use these solutions.
Not only can AI tools recognize attacks before human operators, but they can also be configured to stop the attack and alert IT personnel before the breach gets out of control.
Deep Instinct is a cybersecurity company powered by deep learning.
The solution can scan millions of files per day and detect threats in less than 20 milliseconds.
Since its founding in 2015, the company has raised more than $321 million in funding from firms like BlackRock and Chrysalis Investments.
5. 网络威胁日益高级化
从对普通消费者的随机攻击到国家层面有组织的网络战，网络犯罪始终是一个持续蔓延的威胁。
2022年，Hiscox进行的一项全球调查显示，有43%的企业曾遭遇网络攻击。
更令人警醒的是，该报告指出，被攻击企业中有20%表示，所受损失已严重影响其偿付能力。
有数据显示，2023年一次数据泄露的平均成本可高达500万美元。
IBM的研究指出，这一数字或可攀升至944万美元。
事实上，虚拟犯罪的泛滥已使价值1550亿美元的网络安全产业预计在2029年前膨胀至3760亿美元。
深度伪造（Deepfake）攻击正成为黑客渗透企业防线的最高明手段之一。
VMware近期的一项调查显示，参与调查的IT领导者中有66%在过去12个月内遭遇过深度伪造相关的攻击，比2023年增加了13%。
深度伪造技术借助AI与深度学习，可生成栩栩如生的视频、图像和音频，呈现出从未真实发生的人物和事件。
该技术已问世数年，并在黑客手中不断精进。
其中，“生成式对抗网络”(GAN)技术的应用，使深度伪造模型几乎难以被识别。
同时，5G网络的兴起亦让视频实时篡改更加轻而易举。
深度伪造对从事BEC（商业邮件诈骗）的网络罪犯来说尤为有利。
利用深度伪造干扰面对面身份验证，也将成为网络犯罪分子的一大潜在手段。
对于企业而言，识别深度伪造及其他潜在威胁更像是一场永不止息的防守战——安全专家往往步步紧跟，却总慢攻击者半拍。
不过，网络安全专业人士正借助AI与其他先进技术解决方案，尽可能及早地发现并阻止攻击的发生。
IBM近期的一份报告显示，大量使用AI网络安全工具与自动化方案的组织，较未使用这些技术的对手可节省约222万美元的成本。
AI工具不仅能在人工干预前识别攻击，还可在事态失控之前自动中断攻击并及时提醒IT人员。
Deep Instinct是一家借助深度学习驱动的网络安全公司。
其系统每天可扫描数百万个文件，并在不到20毫秒的时间内检测出威胁。
自2015年成立以来，该公司已从包括贝莱德（BlackRock）与Chrysalis Investments在内的多家投资机构筹集了超过3.21亿美元的资金。
6. Enterprise and Public Use of IoT Expands
Despite ongoing chip shortages, the number of global IoT connections is expected to grow to 27 billion by 2025.
With the advancement of 5G networks and relief on the horizon for semiconductor manufacturing, some estimate the enterprise IoT market could grow to $483 billion by 2027.
Business leaders from a variety of industries are adopting and expanding their use of IoT technology in order to improve connectivity and data collection.
In a 2022 survey from IEEE, IoT was ranked as one of the top five most impactful technologies of 2023.
Businesses are seeing relatively quick ROI on their IoT projects. More than 60% say they’re reaching financial payback in just three years.
Manufacturing is one industry that’s investing in IoT (nicknamed the “industrial internet of things” or IIoT) as part of an effort to create smart factories.
In the United States, 35% of manufacturers are already collecting and using data through IoT devices.
For instance, sensors on machinery gather data from the machines, store it wirelessly, and a machine-learning platform runs through the data to analyze it and assess if human intervention is needed.
This process can be used to improve efficiency, reduce waste, and predict maintenance needs.
IoT devices are also powering smart cities.
Sensors collect data from things like electric meters, trash cans, traffic signals, and more. That data is then utilized to make the cities more efficient and make the citizens safer.
When the University of Idaho partnered with Nordsense, a smart trash can company, they outfitted 58 concrete trash cans with IoT sensors.
Because these sensors monitored trash levels and altered the waste management team when a bin needed to be emptied, the team cut their working hours and fuel consumption by 50%.
6. 企业与公共领域的物联网应用不断扩张
尽管芯片短缺仍在持续，但预计全球物联网（IoT）连接数量将在2025年达到270亿。
随着5G网络的推进以及半导体制造瓶颈的缓解，有分析预测，到2027年，企业级物联网市场规模可能增长至4830亿美元。
来自各行各业的商业领袖正积极采用并扩展物联网技术，以提升互联能力与数据采集效率。
在IEEE于2022年进行的一项调查中，物联网被评为2023年最具影响力的前五大技术之一。
许多企业在物联网项目上看到较快的投资回报率，超过60%的企业表示，在三年内就能收回成本。
制造业正在投资物联网（也被称为“工业物联网”或IIoT），以推动智能工厂的建设。
在美国，已有约35%的制造企业利用物联网设备来采集并应用数据。
比如安装在机器上的传感器可实时收集设备数据并通过无线方式储存，然后由机器学习平台对这些数据进行分析，从而判断是否需要人为干预。
这一过程有助于提高生产效率、减少浪费，并预测维护需求。
物联网设备同样助力打造智慧城市。
传感器可从电表、垃圾桶、交通信号灯等众多来源收集数据，这些数据随后被用于提升城市的运行效率和市民安全感。
爱达荷大学（University of Idaho）与一家名为Nordsense的智慧垃圾桶公司合作，为58个混凝土垃圾桶配备了物联网传感器。
由于这些传感器可监测垃圾量并在需要清空时自动提醒环卫团队，团队的工作时长和燃料消耗减少了50%。
7. Ambient Computing Enables Nearly Invisible Tech
Ambient computing is a concept that builds on IoT and promises a future of nearly invisible technology.
Here’s why: with ambient computing, an AI-driven network of devices and software runs in the background (all around us) with little to no human intervention required.
Ambient computing uses both AI and machine learning to interpret data gathered from physical devices, like smart thermometers and smart watches, and make decisions on their own.
All of this technology comes together to create devices that can interact with both individuals and other devices.
With the potential to transform how we interact with everything from coffee makers to freight trucks, it comes as no surprise the ambient intelligence industry is expected to grow at an impressive 32% CAGR through 2028 to reach a total value of $225 billion.
Ambient computing is still a young technology, but use cases are already being seen in both consumer-focused and business-focused solutions.
Voice assistants and smartphone-controlled thermostats are great examples of ambient computing, but this technology has the potential to become even more invisible.
For example, a person getting off a flight could be automatically alerted that their bags are ready at a specific carousel. After their bags were picked up, they’d receive another notification that their rideshare was ready at a specific location. While they were in the car, a coffee would be automatically ordered from the nearest Starbucks and they’d be automatically checked into their hotel.
In another possibility, homeowners may no longer need garage door openers. Instead, the owner’s smartphone will communicate the location to a device in the home, which will open the garage door for them as they approach.
In factories, ambient computing is being used to monitor and schedule maintenance on machines.
When the IoT device recognizes the machine needs servicing, it links up with the software that schedules maintenance and inputs that machine on the schedule.
In retail, sensors on shelves could automatically order new inventory when stock is running low.
However, as this technology progresses, privacy and security concerns will be top of mind.
7. 环境计算：实现“隐形”技术体验
环境计算（Ambient Computing）是建立在物联网基础上的概念，为未来带来几乎“隐形”的技术体验。
其核心在于：利用AI驱动的设备和软件网络在我们周围悄然运行，几乎无需人工干预。
环境计算通过AI和机器学习来解释智能温度计、智能手表等物理设备所采集的数据，并能自行决策。
这些技术相互配合，使设备不仅能与人类互动，也能与其他设备沟通协作。
考虑到环境智能产业有望以约32%的年复合增长率在2028年达到2250亿美元的市场价值，这种潜力不足为奇。
目前，环境计算还处于起步阶段，但其应用案例已在面向消费者与企业的场景中初露端倪。
语音助手和可用智能手机控制的恒温器就是环境计算的典型示例，但未来这项技术可能会更加“隐形”。
举个例子，当一名旅客下机后，他的智能系统可自动提醒他行李在特定的行李转盘已准备就绪；取到行李后，又会收到下一条通知，告诉他约好的网约车已在指定地点等待。同时，在乘车途中，系统可自动在附近的星巴克下单咖啡，并在旅客抵达前为他完成酒店入住登记。
在另一种场景中，屋主可能再也不需要车库门遥控器。
屋主的智能手机会与家中的设备通信，当屋主接近家门口时，车库门会自动开启。
在工厂里，环境计算正用于监控并安排设备的维护保养。当物联网设备识别出机器需要维护时，它会直接与计划维护的系统相连，将该机器的维护需求自动纳入日程表中。
在零售业，当货架上的库存即将不足，传感器则可自动下单采购新的库存。
然而，随着此类技术的不断发展，隐私与安全问题将成为行业关注的焦点。
8. Adoption of Robotic Process Automation Continues to Grow
8. 机器人流程自动化（RPA）持续增长
随着人工智能与机器学习之间的界限日益模糊，各企业正在不断探索将自动化融入业务流程的方法。
在众多引发业界高层兴奋的创新技术中，机器人流程自动化（RPA）正脱颖而出。
Search demand for “robotic process automation” is up 4,700% over 10 years.
RPA的核心在于训练软件程序执行那些枯燥重复的任务。
有预测显示，到2030年RPA市场规模有望增长至250亿美元，年复合增长率接近36%。
2022年，RPA软件营收已接近100亿美元，几乎是2021年的两倍。
调查显示，目前已有五分之一的企业在使用RPA。
企业之所以青睐RPA，其中一个主要原因在于当前劳动力市场紧张，这使企业更需最大化员工的效率与生产力。
数据显示，在美国拥有500名员工的平均企业，每年因重复性任务浪费的时间成本超过140万美元。
Losing time to repetitive tasks costs businesses more than a million dollars annually.
RPA解决方案能在处理低价值、重复性任务时节省时间和资金。
例如，当RPA负责从文档中复制粘贴信息至数据库时，这一流程不仅更快，而且较人工操作更易确保准确性。
有估算指出，RPA可使工作产能提升多达50%，同时让高价值的人才转而专注于更具创造性的工作。
RPA还可用于从网站中提取数据、预约安排、收集客户信息、监督合规性以及为新员工办理入职手续等多种场景。
在RPA供应商中，UiPath明显领跑市场。
该公司的年度经常性收入（ARR）已超过10亿美元，目前在全球范围内为10,500多家客户提供服务。
9. FemTech Looks to Reform Traditional Women’s Healthcare
The FemTech sector — defined as tech related to fertility, menstruation, pregnancy, and other women’s health topics — is poised for huge growth in the coming years.
Predictions show its value could shoot up from $51 billion in 2021 to $103 billion by 2030.
An analysis by McKinsey points out several ways surging investment and growth in FemTech is trickling down to consumer products.
Virtual clinics are providing telemedicine solutions that allow women across economic classes to access quality care. Trackers and wearables are providing women with the data they need to better assess their health. Startups are bringing more attention to cultural populations (LGBTQ+ and Black women, for example) that are often marginalized in traditional care settings.
The sector saw its first-ever unicorn in 2021: Maven Clinic.
The company created a digital health platform to provide medical care specifically to women and families. They’ve treated more than 15 million patients so far.
In November 2022, they closed a $90 million Series E round to bring their total funding to nearly $300 million.
One of the most popular science-backed FemTech trackers today is breathe ilo.
Here’s how it works.
A woman breathes into the breathe ilo device for 60 seconds and the device measures the CO2 levels in her breath. The data is sent to a smartphone app, which utilizes a self-learning algorithm to predict the woman’s ovulation. (CO2 levels are highest in the first phase of a woman’s menstrual cycle)
Other tech companies, like Apple and Samsung, are looking to capitalize on FemTech without creating stand-alone devices.
For instance, the Apple Watch Series 8 tracks body temperature via two wrist sensors. The Health app displays the results and ovulation predictions.
Samsung partnered with Natural Cycles to add ovulation tracking to its Galaxy Watch 5. The temperature tracking tech from Samsung will be combined with an ovulation-predicting algorithm from Natural Cycles.
The feature will roll out to 32 markets in mid-2023.
FemTech领域——聚焦于生育、经期、妊娠和其他女性健康议题的科技领域——在未来数年内或将迎来巨大发展。
Search interest in “FemTech” has increased more than 700% in the past 5 years.
有预测显示，其市场价值将从2021年的510亿美元跃升至2030年的1030亿美元。
麦肯锡的分析指出，不断飙升的投资和增长趋势正逐渐传导至面向消费者的产品领域。
虚拟诊所正提供远程医疗方案，让不同经济层次的女性都能获得优质护理；追踪器和可穿戴设备为女性提供更全面的数据，助其更好地评估自身健康状况；初创公司则将更多关注投向此前常被忽视的群体（如LGBTQ+与黑人女性），让她们在传统医疗环境之外获得更公平的关注。
有一家企业专门为女性和家庭提供数字健康平台，目前已为超过1500万名患者提供医疗服务。
2022年11月，该企业完成9000万美元的E轮融资，总融资额已接近3亿美元。
当下最受认可且有科学依据的FemTech追踪器之一是breathe ilo。
其原理是：女性对breathe ilo设备吹气60秒，设备会测量呼气中的二氧化碳浓度，并将数据发送到智能手机App中。
App运用自学习算法预测女性的排卵期。（在女性月经周期的初始阶段，CO₂浓度最高）
其他科技公司，如苹果和三星，则希望在不推出独立设备的情况下也能分享FemTech的红利。
例如，苹果手表Series 8通过两个手腕传感器监测体温，健康App会展示结果并预测排卵期。
三星则与Natural Cycles合作，将排卵追踪功能整合进Galaxy Watch 5中，将利用三星的体温追踪技术配合Natural Cycles的排卵预测算法。
10. Wearable Devices Get Sleeker & Smarter
The market for wearables continues to trend upward.
Up until recently, many wearables were focused on sleep tracking and step counting.
In recent years, however, we’ve seen a variety of devices launched with features that go far beyond these basic functions.
A few tech companies offer smart rings with built-in near-field communication technology.
These rings enable wearers to make payments, unlock smart-key doors, and share data with the swipe of their fingers.
For instance, McLear’s RingPay acts as a contactless payment device, allowing users to swipe their ring in the same way they swipe a contactless card. There’s no need for any kind of interface or app.
Another example comes from the smart ring brand Cnick. They’ve created the Tesla key ring.
Owners can lock, unlock, and start their vehicle by simply taping their knuckle.
The ring can also be used for contactless payments.
There’s also increasing consumer demand for tech wearables that are embedded in clothing, referred to as e-textiles.
In one example, the Nadi X yoga pants use sensors and stretch bands to analyze and provide feedback on the owner’s posture.
For long-distance runners, Sensoria socks could potentially help prevent injuries.
These smart socks have textile sensors running through them that not only record basic measurements like step count and distance, but also track cadence, foot landing, and the impact generated with each step.
When paired with the mobile app, runners get real-time audio cues to correct problems that may lead to injuries.
And even babies are wearing smart clothes.
The Owlet dream sock fits snuggly on a baby’s foot and monitors heart rate, oxygen level, movement, and wakings.
10. 可穿戴设备更时尚、更智能
可穿戴设备市场持续升温。
过去许多可穿戴设备主要聚焦睡眠监测与计步功能。
但近年来，我们已看到各种超越基本功能的设备问世。
一些科技公司推出内置近场通信技术（NFC）的智能戒指。
佩戴者可用手指轻轻一扫便完成支付、解锁智能门锁或共享数据，无需任何接口或应用。
例如，McLear的RingPay就能像无接触式卡片一样一挥即付。
另一例子是Cnick品牌推出的“特斯拉钥匙戒指”，车主仅需用指关节轻触，即可锁车、解锁和启动汽车，同时这款戒指也支持无接触支付。
Search volume for “Tesla ring” has climbed more than 240% in the past 5 years.
Cnick’s rings are waterproof and never need to be charged.
消费者对嵌入于服装中的可穿戴技术（即电子纺织品）的需求也在增加。
如Nadi X瑜伽裤利用传感器和弹性织带来分析并指导用户调整姿势。
针对长跑爱好者，Sensoria智能袜可帮助减少运动损伤。
Sensoria’s smart socks track a variety of data points for walkers and runners.
这些智能袜子内嵌纤维传感器，不仅能记录步数和距离，还可追踪步频、足部着地方式以及每一步的冲击力。
通过与移动应用程序配合，跑者可在实时音频提示下纠正可能导致伤害的问题。
甚至婴儿也在穿“智能衣物”。
11. Extended Reality Expands Beyond Entertainment
As the lines blur between mixed reality, augmented reality, and virtual reality, the term “extended reality”’ has become an umbrella term designed to encompass all of the above and more.
In the next few years, extended reality has potential applications in everything from the metaverse to virtual concerts.
For starters, there are multiple B2B use cases where extended reality could drive serious innovation.
One, in particular, is the medical field.
In just one application, extended reality is being used to train both new and existing surgeons.
This technology could prove highly useful in helping surgeons understand the intricacies of new, innovative surgeries they may not have learned about in school or previous positions.
A study in Clinical Orthopaedics and Related Research found that VR training significantly increased the accuracy of surgery and the rate of completion of surgery.
Another study showed VR training improved surgical performance by up to 230%.
The automotive and manufacturing industries are two other sectors in which extended reality is expected to make an impact in the coming years.
The technicians in these industries are required to perform highly intricate processes and machine work.
Because of that, extended reality technology can be used to provide an up-close, first-person view of those processes in a safe environment where mistakes can easily be remedied.
When technicians are able to practice assembly processes before stepping onto the actual line, human error and injuries can be prevented.
Extended reality is even being used to train automotive students.
Maryland’s Vehicles for Change program is a virtual-first training program that aims to ease the mechanic labor shortage and, at the same time, provide job options for former prisoners.
The program is expected to expand to 20 new sites by 2028.
The United States military is also using extended reality for training purposes.
With this technology, military leaders can set up training operations that are too dangerous or costly to set up in the real world. It also has the benefit of reducing wear and tear on critical pieces of equipment like aircraft.
In April 2022, the Army reiterated the military’s interest in extended reality technology when officials requested new technologies from public industries as part of the Technologies for Mission Rehearsal and Training project.
It seems that current extended reality technology isn't quite as advanced as the military would like.
The Army announced a $21.88 billion contract with Microsoft to acquire 120,000 HoloLens-like headsets, but the timeline was pushed back in late 2021 and the headsets are currently going through another series of redesigns.
11. 拓展现实技术超越娱乐应用
随着混合现实（MR）、增强现实（AR）与虚拟现实（VR）之间的界限愈加模糊，“拓展现实（Extended Reality，XR）”已成为一个包罗万象的术语，涵盖上述技术及更多应用。
Search volume for “extended reality” is up more than 150% since 2019.
在未来几年，拓展现实有望应用于从元宇宙到虚拟音乐会等广泛场景。
首当其冲的是多个面向企业（B2B）的应用案例，特别是在医疗领域。
在医疗培训中，拓展现实技术正被用于为新晋与在职外科医生提供模拟训练。
这项技术可帮助他们深入理解新型创新手术的复杂细节，即便这些手术未曾在传统教育或过往工作中涉及。
《临床骨科与相关研究》（Clinical Orthopaedics and Related Research）的一项研究发现，VR训练显著提高了手术的准确性与完成率；另一项研究则显示，VR训练可将手术表现提升高达230%。
汽车业和制造业也是拓展现实有望大展身手的领域。
这些行业的技术人员常需执行精密繁琐的工序，有了拓展现实技术，他们可在安全的虚拟环境中以第一视角近距离观摩和练习。
通过在上生产线前先行模拟操作，企业可减少人为失误与工作伤害。
此外，拓展现实还正用于培训汽车专业学生。
例如，美国马里兰州的“Vehicles for Change”项目即通过虚拟化的培训形式，既缓解技工劳动力短缺，又为曾服刑人员提供就业机会。该计划预计在2028年前扩展至20个新场点。
美国军方也在利用拓展现实来进行军事训练。
借助该技术，军方可在虚拟环境中搭建现实中代价高昂或风险极大的训练场景，并减少对关键设备（如航空器材）的磨损。
2022年4月，美国陆军重申了对拓展现实技术的兴趣，在“任务预演与训练技术（Technologies for Mission Rehearsal and Training）”项目中向社会公开征求新技术。
尽管目前的拓展现实技术尚未完全达到军方预期，陆军曾与微软签订了约218.8亿美元的合约采购12万套类似HoloLens的头显设备，但进度在2021年底延迟，产品目前仍在重新设计中。
12. Edge Computing Transforms How Enterprises Use Data
In 2018, global data centers processed approximately 155 exabytes of data per month. In 2023, that number had skyrocketed to 374 exabytes.
As the world is becoming increasingly digital — and 5G technology enables larger data transfers at faster speeds — IT infrastructure is requiring more processing power than ever.
In particular, growth in data-intensive technologies like personal and enterprise use of IoT, remote health monitoring, and remote work is expected to continue pushing global data processing requirements to record highs.
To aid in the processing of all that data, companies are pouring more of their budgets into edge computing.
In 2023, it’s estimated that more than half of all new enterprise IT infrastructure will utilize edge computing.
With a focus on speed and network distribution, edge computing is designed to improve response times and save bandwidth by moving processing power physically closer to the source of data.
Cloud computing has been one popular way businesses have attempted to deal with massive amounts of data and increasing technology needs.
However, cloud computing is both expensive and resource-intensive. This is especially true for companies dealing with large amounts of data because many cloud storage services base their fees on usage.
Because of this, corporations are looking at ways to reduce their dependence on cloud computing by moving to edge computing instead.
By leaning more heavily on the edge, and in particular its ability to reduce bandwidth transmissions, companies will be able to reduce their cloud computing bill while providing a faster user experience to end-users.
Edge computing also decreases latency because data isn’t being sent back and forth between the cloud and devices. This provides enterprises with the opportunity to analyze data faster and make quicker decisions.
In fact, 45% of IT professionals say this is the biggest advantage of running workloads on the edge.
Growing reliance on AI is also fueling the growth of edge computing.
To reach its full potential AI/ML systems will need to become more data-driven. And that means more analysis and computation needs to happen at the edge.
Self-driving cars are a great example of edge AI in the real world.
The car’s sensors collect data and take action in the blink of an eye without ever sending that data to the cloud.
In another example, edge AI enables security cameras to process data and run real-time algorithms to detect suspicious activity and take action immediately.
12. 边缘计算重塑企业数据使用方式
2018年，全球数据中心每月处理的数据量约为155 EB（艾字节），而到2023年这一数字已飙升至374 EB。
随着世界日渐数字化，以及5G技术加快海量数据的高速传输，IT基础设施对算力的需求前所未有地提升。
个人与企业级物联网的扩张、远程医疗监测和远程办公的普及，将持续推高全球数据处理需求。
为应对庞大数据处理带来的挑战，企业正将更多预算投入边缘计算。
2023年起，超过半数新增的企业IT基础设施将采用边缘计算。
边缘计算的设计初衷是将算力更靠近数据源头，提升响应速度并节省带宽。
云计算曾是企业应对数据洪流与技术需求增长的热门选择，但云计算既昂贵又资源密集，尤其是当企业的数据量庞大、且云存储费用按用量计算时。
因此，许多企业正寻求降低对云计算的依赖，通过转向边缘计算减少带宽传输量，从而降低云端成本并为终端用户提供更快速的体验。
边缘计算还能减少延迟，因为数据无需频繁在云端与设备间往返。这为企业创造了更快分析数据与做出决策的契机。
事实上，约有45%的IT专业人士认为，这种快速响应能力是在边缘运行工作负载的最大优势。
对AI的日益依赖也在推动边缘计算的发展。
要充分发挥AI和机器学习（ML）的潜能，就需让其更依赖数据驱动，这意味着更多分析与计算将发生在边缘。
自动驾驶汽车就是边缘AI的绝佳实例：
汽车传感器采集数据并在瞬间做出决策，无需将数据传回云端。
在另一个场景中，边缘AI可让安防摄像头实时处理数据并运行算法，一旦发现可疑行为能立刻采取行动。
13. More Developers and Non-Developers Use Low/No Code Tools
Given the constant release of new and exciting technologies, it’s easy to assume tech companies are rolling out new products and services as fast as they can.
In reality, however, there’s a serious shortage of developers.
IDC predicts the shortage will reach nearly 4 million unfilled positions in 2025.
In a 2023 survey from Infragistics, more than 37% of respondents said they expect to continue having trouble hiring developers in 2023.
With a persistent and growing shortage of talent, creating solutions that allow developers to work more efficiently has become critical for companies that want to move initiatives forward and outpace their competition.
Fortunately, that’s precisely what low-code and no-code (LCNC) software programs can do.
By combining visual models with AI-powered tools, software developers who use LCNC are able to skip, or at least expedite, the time-consuming process of writing thousands of lines of code from scratch.
This dramatically streamlines the coding process and speeds up the development of software.
And it’s not just full-time software developers that benefit from these new tools.
Because LCNC solutions are meant to be user-friendly by design, an increasing number of non-developers are able to build software programs too.
These non-IT employees are being called “citizen developers.” Gartner predicts that large enterprises will soon have four times the amount of citizen developers as compared to professional developers.
The Infragistics survey showed more than 76% of organizations are already utilizing LCNC.
Further proof that adoption is soaring is the growth of the low-code market. It was expected to jump to $27 billion in 2023, a 20% increase over 2022.
At the core of LCNC solutions are what’s known as low-code application platforms (LCAPs).
These platforms are component-based and offer pre-defined templates.
Mendix was the leader in the 2023 Gartner Magic Quadrant analysis for low-code platforms.
Their platform is used by more than 4,000 enterprises worldwide.
Since being acquired by Siemens in 2018, Mendix’s ARR has grown more than 300%.
With an increasing shortage of software developers and an increasing number of people successfully building applications using LCNC, it's safe to assume these platforms and their use among enterprises will continue to advance at a blistering pace.
As evidence of this, Gartner predicted that LCAPs would account for $10 billion of the LCNC market in 2023 and $12.3 million in 2024.
13. 更多开发者与非开发者使用低/无代码工具
在新技术层出不穷的背景下，许多人会以为科技公司正疯狂推出产品与服务。
但现实中，专业开发者却严重短缺。IDC预测，2025年这一缺口将接近400万个空缺岗位。
一项2023年的Infragistics调查显示，超过37%的受访者在2023年仍难以招聘到足够的开发者。
在人才短缺日益加剧的情况下，让开发者更高效地工作已成为企业在激烈竞争中保持领先的关键。
幸运的是，低代码/无代码（LCNC）软件正是为此而生。
通过将可视化模型与AI驱动的工具相结合，使用LCNC的开发者无需从零编写数千行代码，大幅缩短开发周期，加快软件问世时间。
不仅是全职开发者受益于LCNC工具——这些友好易用的方案让越来越多的非开发者（即“公民开发者”）也能构建软件。
Gartner预测，大型企业中公民开发者的数量将很快达到专业开发者的四倍。
Infragistics调查显示，超过76%的组织已在利用LCNC。
低代码市场的迅猛增长则为此提供了佐证：2024年该市场规模预计达到370亿美元，较2023年增长20%。
LCNC解决方案的核心是低代码应用平台（LCAP），它们通常以组件化形式提供预定义模板。
在2023年Gartner低代码平台魔力象限分析中，Mendix居于领先地位。
该平台已被全球4000多家企业采用。
自2018年被西门子收购后，Mendix的年经常性收入增长超过300%。
随着软件开发者短缺加剧、越来越多的非专业人士成功利用LCNC构建应用，这类平台及其在企业中的应用将继续以惊人速度发展。
结论
至此，我们已盘点了当前影响企业与消费者的13大科技趋势。纵观全文，这13大科技趋势如同交织的脉络，正推动人类社会迈向更智能、更高效、更可持续的未来。从人工智能的全方位渗透到量子计算的潜能释放，从绿色能源的崛起到无所不在的物联网，这场科技浪潮已不再停留于愿景，而正以惊人的速度重塑各行各业。未来已来，每一项创新都是时代前进的信号，每一种技术都是我们下一步探索的起点。
本文作者：Josh Howarth，译：调研君
“华夏气候”除了原创，其他都是转载其他公号/媒体的，来源在开头都明显标示，“阅读原文”附有原文链接！
CCTC®
关注更多气候（狭义/泛义）资讯
请长按识别下方二维码
关注 华夏气候
广告/投稿/业务etc
微信/邮箱：995201502@qq.com
想深入了解碳排放（碳达峰碳中和）
关注更多碳（CO2）资讯
请长按识别下方二维码
关注 3060
