AI to Reshape the Global Technology Landscape in 2026, Says TrendForce

/PRNewswire/ — TrendForce has identified 10 key technology traits that can interpret the tech trade’s evolution in 2026. The highlights of these findings are outlined below: AI Chip Competition Intensifies as Liquid Cooling Positive aspects Current Adoption in Details Companies and products In 2026, the high put a matter to for AI details middle

,/ PRNewswire/– TrendForce has actually determined 10 crucial modern technology characteristics that can analyze the technology profession’s advancement in 2026. The highlights of these searchings for are described listed below:

AI Chip Competitors Heightens as Fluid Air conditioning Favorable facets Existing Fostering thoroughly Business and items

In 2026, the high placed an issue to for AI information center building and construction– sustained by raised capital expense by major North American CSPs and the higher press of sovereign cloud efforts worldwide– is expected to spruce up AI web server deliveries by over 20% twelve months-over-twelve months.

NVIDIA, the major title in AI at the initial time, will certainly deal with more powerful rivals onward. AMD prepares to mention NVIDIA by presenting its MI400 pudgy-rack resolution, which mirrors NVIDIA’s GB/VR programs and is tailored versus CSP customers. At the same time, major North American CSPs are increasing their at home ASIC building and construction. In China, geopolitical stress consist of accelerated the adaptability for technical self-sufficiency, with companies appreciate ByteDance, Baidu, Alibaba, Tencent, Huawei, and Cambricon enhancing initiatives to obtain their hold AI chips. That is made a decision to magnify the around the world rivals.

Thermal develop power (TDP) per chip is increasing fast as AI cpus ended up being additional grand, jumping from 700W for NVIDIA’s H100 and H200 to over 1,000 W for the upcoming B200 and B300. This increase in heat result is major to a constant fostering of liquid-cooling programs in web server shelfs, with use anticipated to dominate in 47% by 2026.

Microsoft has actually released created chip-level microfluidic air conditioning modern technology to pork up thermal efficiency. Lastly to to midterm, frigid-plate fluid air conditioning will certainly live basically one of the most essential resolution, with CDUs transitioning from liquid-to-air to liquid-to-liquid arrangements. Over the lengthy crawl, the marketplace is apparently to switch over versus additional comprehensive chip-level thermal administration.

Damaging Data Transfer Obstacles: HBM and Optical Communications Redefine AI Collection Architectures

Such a flash increase carefully amount and memory transmission capacity desires, pressed by increasing AI work from mentoring to reasoning, is currently challenging device develop by subjecting traffic jams in transmission drag and power efficiency. To take care of these challenges, HBM and optical adjoin used scientific researches are becoming significant enablers of next-era AI designs.

Many present generations of HBM take advantage of 3D piling and thru-silicon by means of to really a whole lot reduced the range in between cpus and memory, getting to raised transmission capacity and efficiency. The upcoming HBM4 period will certainly present higher network thickness and broader I/O transmission capacity to additional reinforce the large computational ask for of AI GPUs and accelerators.

Additionally, as version criteria exceed the trillion-scale degree and GPU collections postpone greatly, memory transmission capacity as quickly as once again becomes a vital effectiveness traffic jam. Memory suppliers are resolving this location by enhancing HBM pile designs, introducing in product packaging and user interface develop, and co-designing with sound judgment chips to pork up on-chip transmission capacity for AI cpus.

While these advancements reduce memory-connected traffic jams, information transmission throughout chips and components has ended up being the adhering to significant restriction to device effectiveness. To dominate these limitations, co-packaged optics (CPO) and silicon photonics (SiPh) are arising are calculated sight locations for GPU manufacturers and CSPs.

Presently, 800G and 1.6 T pluggable optical transceivers consist of currently gotten in mass production, and starting in 2026, also increased-bandwidth SiPh/CPO systems are anticipated to be released in AI buttons. These next-gen optical spoken change used scientific researches will certainly make it possible for high-bandwidth, low-power information interconnects, enhancing total device transmission capacity thickness and vigor efficiency to satisfy the rising effectiveness ask for of AI facilities.

In general, the memory profession is fast developing versus transmission capacity efficiency as its core affordable assistance. Breakthroughs in optical interactions– developed to take care of information transmission throughout chips and components– are arising since the exact resolution to defeat the challenges of previous electric user interfaces in long-distance, high-density information transfers. Therefore, high-drag transmission used scientific researches are approach to ended up being a vital column of AI facilities advancement.

NAND Flash Providers Come AI Storage Space Alternatives to Creep Reasoning and Decline Repayments

AI mentoring and reasoning tasks place an issue to very easy availability to large datasets with unforeseeable I/O actions, major to an expanding effectiveness space with current storage space options. NAND Flash suppliers are tackling this location by quickening the building and construction of customized options, concentrated on 2 major item types.

The major course entails storage-class memory SSDs, KV cache SSDs, and HBF, that are positioned in between DRAM and previous NAND Flash. These options supply very reduced latency and high transmission capacity, making them excellent for quickening precise-time AI reasoning tasks.

The 2d course entails nearline QLC SSDs, that are fast being taken on for heat and cold AI information storage space layers appreciate version checkpoints and dataset archiving. QLC really a whole lot reduces the rate per little bit for saving natty AI datasets, supplying 33% raised per-die storage space thickness than tender loving care. TrendForce efforts that by 2026, QLC SSDs are anticipated to comprise 30% of the venture SSD market, highlighting their increasing value in improving storage space ability and maintain efficiency in AI facilities.

Vigor Storage Space Programs Emerge since the Vigor Core of AI Information and facts Business and items and Are Put of living for Dynamite Increase

As AI information focuses enter natty-scale gathered programs, their variable work call for significant additional steady power. This change is standing out vigor storage space programs from simple back-up resources right into the core vigor facilities of AI information facilities.

Over the adhering to 5 years, AI information facilities are anticipated to really a whole lot ended up being vigor storage space programs. Along with previous brief-length UPS back-up and power high quality stablizing, the fifty percent of tool- to long-length storage space programs (2 to 4 hours) will certainly improve dramatically to spruce up backup power, vigor arbitrage, and grid items and business concurrently.

Implementation systems will certainly also advance from systematized, information middle-level battery vigor storage space programs to given designs at the shelf or collection degree that integrate modular battery back-up systems in a placement to rapid action. This change will certainly pork up device durability and vigor efficiency whereas great the significantly annoying power security desires of AI-pushed facilities.

North The United States is anticipated to ended up being basically one of the most piquant around the world industry for AI information center vigor storage space, led by hyperscale cloud providers. In China, the “Jap Information And Facts, Western Computer” campaign is riding information facilities versus sustainable vitality-affluent western areas, where AI information facilities coupled with vigor storage space programs will certainly ended up being brand-new facilities for natty-scale schools. Worldwide, the set up ability of AI information center vigor storage space is predicted to rise from 15.7 GWh in 2024 to 216.8 GWh by 2030, standing for a CAGR of 46.1%.

AI Information and facts Business and items Change to 800V HVDC Style, Driving Placed an issue to for Third-Expertise Semiconductors

Information facilities are experiencing a vital upgrade in power facilities as web server shelf positions improve from kilowatts to megawatts. The profession is exact currently embracing 800V HVDC designs to spruce up efficiency, pork up dependability, piece once again down on copper cabling, and reinforce additional portable device styles. Created third-era semiconductors appreciate SiC and GaN play a top-notch duty in this change, with quite an implication of semiconductor providers currently taking part in NVIDIA’s 800V HVDC goal.

SiC is very essential in the front-discontinue and mid-stage power conversion within information center designs, handling the really remarkable voltages and power tons. While SiC tools presently consist of reduced optimum voltage positions when when placed following with previous silicon, their boosted thermal efficiency and changing effectiveness are major for the building and construction of next-era solid-articulate transformers (SSTs).

At The Same Time, GaN, understood for its high-frequency and high-effectivity buildings, is becoming significantly unique in mid- and discontinue-stage power conversion. It sustains ultra-high-power thickness and like a flash vibrant reactions. TrendForce anticipates that the fostering of SiC and GaN carefully center power programs will certainly get to 17% by 2026 and surpass 30% by 2030.

Subsequent-Expertise Semiconductor Rate: 2nm GAAFET Manufacturing and 2.5 D/3D Heterogeneous Assimilation Lead the Succeeding Advance

The semiconductor profession is presently undertaking 2 synchronised characteristics: the change to 2nm procedure modern technology in mass production to amount to raised transistor thickness, and its progression right into higher package dimensions, pressed by advancements in heterogeneous combination. This kind combines additional than one chips with different performances and modern technology nodes to satisfy the effectiveness and efficiency ask for of AI and HPC applications.

Wafer construction is transitioning from FinFET to GAAFET framework, where eviction oxide completely encloses the silicon network. This develop deals boosted current abet supervise whereas stating high effectiveness. In product packaging, 2.5 D and 3D used scientific researches make it possible for thick multi-chip piling, major to quicker interconnects and reduced power usage. These advancements are major for future information facilities and HPC programs.

TSMC, Intel, and Samsung are every embracing special 2.5 D/3D product packaging options– TSMC with CoWoS and SoIC, Intel with EMIB and Foveros, and Samsung with I-Dice and X-Dice– to supply incorporated front-discontinue and aid-discontinue shop items and business as they increase 2nm GAAFET production. Their major comment will apparently be to handle ability, dependability, maintain, and return efficiently to ranking lasting affordable benefits in the adhering to fifty percent of semiconductor building and construction.

Humanoid Robot Deliveries to Rise over 700% in 2026, Pressed by AI Adaptivity and Scenario-Primarily based completely Capacities

The twelve months 2026 will certainly brand name an essential juncture for the commercialization of humanoid robotics, with around the world deliveries anticipated to rise additional than sevenfold to exceed 50,000 systems. Market energy is made a decision to focus on 2 core columns: AI adaptivity and application-oriented develop.

The building and construction of AI adaptivity, pressed by grand AI chips, sensing unit combination, and LLM combination, allows humanoid robotics to examine on the express and make flexible choices in unforeseeable setups, getting to initial elevations of situational recognition and thinking onward of carrying out.

In conformity to this pattern, the adhering to period of humanoid robotics in 2026 will certainly change sight from right showcasing requirements or mastery. As an alternative, they’ll be customized to specific functional situations appreciate producing logistics, stockroom sorting, and examination reinforce, with every in a placement to carrying out complete, project-oriented capacities. This indicates an appropriate change of humanoid robotics exact right into a brand name initial fifty percent– an AI-pushed, application-targeted commercial advancement.

OLED Gets In Initial Stage: Premiumization of Note Pad Reveals and the Higher drive of Collapsible Mobile phones

OLED modern technology is undertaking a vital makeover in various device sections. As Chinese and Oriental panel manufacturers postpone their Gen 8.6 AMOLED production, renovations in maintain framework and return prices are quickening OLED fostering in every dinky and natty programs. This change will certainly more than likely be increasing ASPs and enhancing the negotiating place of upstream elements appreciate vehicle driver ICs, TCONs, get in touch with components, and thermal options.

OLED flaunts self-emissive pixels, supplying far better difference, a slimmer develop, and versatile refresh prices. It gets rid of LCDs’ physical restrictions in density and vigor usage, straightening with Apple’s focus on explain high quality and power efficiency. Anticipated to get to in the MacBook Pro line in 2026, OLED panels will apparently effect in a change from mini-LED to OLED in leading charge note pads. TrendForce anticipates OLED computer computer market fifty percent will certainly strike 5% by 2025, increasing to 9– 12% in between 2027 and 2028, pressed by Apple’s fostering.

At the same time, Apple’s prepared stem of its collapsible mobile phones around late 2026 to 2027 could probably perchance truthful ended up being the marketplace by leveraging its hardware-software harmony, strong tag, and strong supply chain. The profession’s factor to consider is apparently to switch over from charming beauty to enhancing performance and improving specific individual adventure, with around the world deliveries of collapsible tools predicted to exceed 30 million systems by 2027.

Additionally, mainstream fostering remains to face challenges comparable to pivot sturdiness, flexible panel encapsulation, return prices, and maintain administration. Apple’s mindful item recognition approach highlights its sight on high quality and timing, recommending that the collapsible market’s building and construction will certainly in the snuff out depend upon technical advancements and strong production capacities.

Meta Speeds Up the Globe Pattern of Reach-Overview Reveals as LEDoS Develops Energy

Meta has actually released its Meta Ray-Ban Comment AR glasses as AI combination develops into additional created. These glasses are tailored versus information delivering applications that incorporate AI right into regularly life, revamping exactly how americans and AI job jointly. By collecting and checking information from a first-particular individual search, they pork up two-capacity spoken change in between clients and AI.

Many present programs make use of LCoS, supplying official pudgy-coloration effectiveness and maturation. This kind sustains the silent-growing LEDoS modern technology and assists obtain market recognition with an easily accessible and refined specific individual adventure.

Making an aim onward, every market assumptions and Meta’s item roadmap are straightening versus LEDoS programs, which supply raised illumination and difference and make it possible for a significant wider variety of applications. Firms appreciate Apple, Google, RayNeo, INMO, Rokid, and Vuzix are proactively purchasing this modern technology, and production costs are anticipated to go down easily, making it additional obtainable. TrendForce anticipates that by 2027– 2028, the profession can consist of additional industrialized pudgy-coloration LEDoS options, with Meta apparently introducing its next-era AR glasses which contains LEDoS programs.

Self adequate Driving Quickens: Guest Cars Systematize Helped Driving While Robotaxi Broadens Around The World

It’s predicted that the fostering charge of L2 and raised assisted-riding programs will certainly exceed 40% by 2026, making vehicle knowledge the adhering to crucial progression vehicle driver in the vehicle market after electrification. Focus will certainly change to charge decrease as L2 modern technology develops into regular, with incorporated cockpit-riding SoCs and controllers entering mass production in 2026. This basically targets China’s mid-range vehicle market. Out-of-date car manufacturers are additionally improving the vehicle knowledge of within burning automobiles to additional power the regular fostering of ADAS as brand-new devices.

At The Same Time, the Robotaxi market is entering fifty percent of around the world development since it goals for L4 freedom. Looser regulations, increasing interest among quickly drivers and movement solution providers, and advancements in AI systems, comparable to E2E and VLA designs, are all speeding up market progression. By 2026, Robotaxi items and business are anticipated to expand fast throughout Europe, the Center East, Japan, and Australia, involving past their current fortress in China and the United States– signifying a brand name initial phase in self-governing movement.

For additional information on tales and market information from TrendForce’s Department of Semiconductor Research study, please click here, or e-mail the Sales Department at [email protected]

For additonal understandings from TrendForce experts on one of the most current technology profession information, characteristics, and projections, please search suggestion from https://www.trendforce.com/news/

Concerning TrendForce

TrendForce is an around the world vendor of one of the most current building and construction, understanding, and medical diagnosis of the modern technology profession. Having actually offered companies for over a years, the business has actually created up a strong subscription corrupt of 500,000 customers. TrendForce has actually developed a credibility en masse that supplies informative and appropriate medical diagnosis of the modern technology profession thru 5 major research departments: Semiconductor Research study, Comment Research Study, Optoelectronics Research Study, Unskilled Vigor Research Study, ICT Capabilities Research Study. Established in Taipei, Taiwan in 2000, TrendForce has actually extended its existence in China because 2004 with areas of operate in Shenzhen and Beijing.

Resource TrendForce

发布者:Benjamin Y Fong,转转请注明出处:https://robotalks.cn/ai-to-reshape-the-global-technology-landscape-in-2026-says-trendforce/

(0)
上一篇 27 11 月, 2025 4:42 上午
下一篇 27 11 月, 2025

相关推荐

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注

联系我们

400-800-8888

在线咨询: QQ交谈

邮件:admin@example.com

工作时间:周一至周五,9:30-18:30,节假日休息

关注微信
社群的价值在于通过分享与互动,让想法产生更多想法,创新激发更多创新。