Top Strategic Techno 815761 NDX
Top Strategic Techno 815761 NDX
Top Strategic Techno 815761 NDX
By Analyst(s): Gene Alvarez, Tom Coshow, Jasleen Kaur Sindhu, Dan Ayoub, Mark
Horvath, Nick Jones, Soyeb Barot, Frank Buytendijk, Marty Resnick, Bill Ray, Sylvain Fabre,
Moutusi Sau, Bart Willemsen
Overview
Opportunities
■ Agentic AI brings imperatives and risks and will enable organizations to transform
the nature and efficiency of work, processes and decision making. However, it will
also drive the need for advancements in AI governance technology. The technology
created to defend organizations from the effects of disinformation will protect
people, organizations and society.
■ New frontiers of computing keep expanding the potential for benefit but also bring
threats. Quantum computing will break today’s cryptography, exposing everyone to
risk. Tiny, ultra-low-cost wireless tags and sensors will enable new business models
and ecosystems. New energy-efficient compute models will meet the demand for
more computing and sustainably. The growing numbers of computing models
provide an opportunity for integration and orchestration to optimize the use of all
models.
AI imperatives and risks abound as organizations move forward with AI agents. This,
combined with other aspects of AI, will drive a need for AI governance platforms within
organizations, enabling all to use AI responsibly and ethically.
Malicious actors using AI to accelerate the spread of disinformation can cause significant
damage to an organization, its customers, partners and employees. Enterprises will need
technologies to track the spread of information by, or about, their organization to assess
the truth of that information and create trust. Organizations must also protect themselves
from malicious actors using synthetic media to gain real-time access to their systems and
spread misinformation.
Organizations will need to meet growing compute demand while lowering their carbon
footprint. They also must integrate and orchestrate many compute models, operating
them as one in the most efficient way to meet their rising computing needs.
The increasing demand for computing and the lack of energy to support it drives the need
for new energy-efficient compute models. The optimization of the growing numbers of
new computing models working with all existing models will push organizations to focus
on integration and orchestration of computing.
Advances in the way humans and machines work together are creating a new level of
human-machine synergy. The creation of next-level interactions between physical and
virtual experiences will bring together the physical and digital world through spatial
computing.
Humans will have robots working side-by-side with them in the same environment and
even can become teammates. Robots’ ability to perform more than one function will
integrate them into humans’ daily work and home experience.
Wearables and implanted technologies, along with polyfunctional robots, will forever
change how humans and machines work together, moving us into a world where all these
technologies exist to benefit humans.
Hybrid Computing
By 2028, AI agent machine customers will replace 20% of the interactions at human-
readable digital storefronts.
By 2028, at least 15% of day-to-day work decisions will be made autonomously through
agentic AI, up from zero percent in 2024.
Agentic AI systems autonomously plan and take actions to meet user-defined goals.
Current AI assistants and large language models (LLMs) perform tasks including
generating text, summarizing content or basic use of tools, but they haven’t been able to
take action on their own “initiative.” Instead, they’ve acted on users’ prompts or followed
orchestrated processes, but agentic AI is changing that. It offers the promise of a virtual
workforce of agents that can assist, offload and augment human work or traditional
applications.
AI agency is a spectrum. It runs from traditional systems with limited agency that perform
specific tasks under narrowly defined conditions to future agentic AI systems with full
agency that learn from their environment, plan approaches, make decisions and perform
tasks independently.
Additionally, the behavior of AI agents that may formulate unique plans for each request,
goal or scenario makes testing and validating their behavior much more challenging than
both traditional software and more mainstream applications of generative AI (GenAI).
Finally, monitoring and governing the behavior of AI agents in deployment requires new
techniques and tools, since their behavior will change and adapt as they memorize both
shared and user-specific context over time. The deployment of AI agents that adapt
overtime must come with a security plan that monitors the AI agents and has forward-
looking guardrails.
Agentic AI is at the forefront of current R&D efforts by many major vendors in the AI
marketplace. The technology is on the cusp of being capable of moving from the lab to
early, lower risk, innovation-tolerant use cases (e.g., in accelerating software engineering
beyond the capabilities of existing AI code assistants).
However, the leap from the lab to production for more business-critical applications and
processes that would enable operational efficiencies and improvements at scale is not a
trivial one. It is entirely dependent on the emergence of reliable and predictable patterns,
practices and technologies for delivery, supporting and governing a virtual workforce that
can plan, act and adapt over time.
Actions:
■ Treat AI agents like Tier 1 digital co-workers that you delegate work to. Rethink
collaboration models, workflows and team strategies to maximize the benefit of AI
agents that can uncover and act on derivative events that human teammates might
not notice.
For more information, see Top Strategic Technology Trends for 2025: Agentic AI.
AI Governance Platforms
Back to top
By 2028, enterprises using AI governance platforms will achieve 30% higher customer
trust ratings and 25% better regulatory compliance scores than their competitors.
AI governance platforms also assess risks related to data quality and privacy violations,
offering continuous model monitoring and correction to ensure fairness. Additionally,
these platforms guide AI models through structured processes, track usage, monitor
performance and help organizations comply with regulations like the General Data
Protection Regulation (GDPR), California Consumer Privacy Act (CCPA) and AI regulations.
AI’s rapid integration into various sectors — especially highly regulated industries — has
highlighted the urgent need for AI governance platforms. As AI technologies become more
widespread and sophisticated, the risks of bias, lack of transparency and data privacy
concerns grow. Increasing global regulations on AI use and data privacy push
organizations to adopt robust governance methods.
There are several emerging examples of firms who have adopted AI governance practices
as a part of their AI operations. For example, HSBC has implemented its Principles for the
Ethical Use of Data and AI by establishing governance frameworks that prioritize
transparency and accountability, regularly checking for fairness and bias, protecting
customer privacy and training employees on ethical practices. 4
Actions:
■ Ensure multiple perspectives are considered when designing and evaluating ethical
and responsible AI methods, leading to more balanced and fair AI systems. Involve a
diverse range of stakeholders, including ethicists, technologists and affected
communities in the AI governance process.
■ Define clear lines of responsibility and accountability for AI-related decisions and
actions. This will enhance transparency and trust, making it easier to address issues
when they arise.
For more information, see Top Strategic Technology Trends for 2025: AI Governance
Platforms.
Disinformation Security
Back to top
By 2028, 50% of enterprises will begin adopting products, services or features designed
specifically to address disinformation security use cases, up from less than 5% today.
■ Intelligence monitoring for narratives spread through mass or social media (i.e.,
those targeting an executive leadership team, products, service or brand).
Disinformation is already a digital arms race: phishing, hacktivism, fake news and social
engineering are all being turbocharged by adversaries intent on sowing fear, spreading
havoc and committing fraud. Given the wide availability and advanced state of artificial
intelligence and machine learning tools being leveraged for nefarious purposes, the
number of disinformation incidents targeting enterprises is expected to increase over time.
Left unchecked, disinformation can cause significant and lasting damage to any
organization.
Disinformation security has come to the fore because of the following recent
developments:
■ Misinformation and disinformation were recognized as the top global threats for the
next two years by the World Economic Forum in 2024. As more organizations are
targeted with disinformation attacks, hype and concern around this category will
continue to grow.
Actions:
For more information, see Top Strategic Technology Trends for 2025: Disinformation
Security.
Criminals and state actors show an understanding of this importance as they have
recently demonstrated attack scenarios such as “harvest now, decrypt later.” In these
attacks, encrypted data is stored after exfiltration, with the expectation of being able to
decrypt these stolen secrets, data and other sensitive information in the future when
decryption by a quantum computer becomes more approachable.
The approach toward PQC involves creating a repeatable process collectively referred to
as “crypto-agility.” Crypto-agility is the capability to transparently swap out encryption
algorithms and related artifacts in an application, replacing them with newer, presumably
safer, algorithms. A consistent approach is needed, because, given past and future
developments in cryptography, this will not be the last time we have to switch encryption
methods.
Adopting PQC won’t be easy. No drop-in alternatives exist for current cryptographic
algorithms. This means discovery, categorization and reimplementation efforts will be
necessary. Additionally, new algorithms have different performance characteristics from
non-PQC ones. For example, key and ciphertext sizes are larger, and encryption and
decryption times are longer, which may impact performance. This means current
applications must be retested and, in some cases, rewritten.
Actions:
For more information, see Top Strategic Technology Trends for 2025: Postquantum
Cryptography.
Ambient invisible intelligence is enabled by ultra-low-cost, small, smart tags and sensors,
which will deliver large-scale affordable tracking and sensing and, in the long term, will
enable a deeper integration of sensing and intelligence into everyday life. Ambient
invisible intelligence is driven by three key technologies:
■ Energy harvesting, especially from ambient RF energy. This enables small battery-
free tags and sensors with effectively infinite life span and eliminates the size and
cost of battery components.
■ Low-cost, low-energy electronics, enabling chips that are efficient enough to use
harvested power to run sensors, perform simple computation and send short-range
wireless messages.
In the longer term, we expect that ultra-low-cost electronics will remain in items throughout
their life. The messaging formats involved will be standardized, and the gateways to
receive ambient messages will become an integrated feature of homes and offices.
For example, this could occur as part of Wi-Fi access points or while embedded into
consumer white goods. This will enable new ecosystems, like smart packaging
communicating with domestic refrigerators and clothing items communicating with
washers and dryers. Small intelligent tags associated with objects could also provide
unforgeable provenance and new ways for objects to report on their identity, history and
properties (e.g., to satisfy EU requirements for digital product passports).
In the short term, ambient devices will focus on communications, location and simple
environmental sensing. However, as electronics improves and ambient concepts are
incorporated into wireless standards, we expect device capabilities to grow to include
simple intelligence and, perhaps, peer-to-peer communications.
A side effect of ambient intelligence will be that much more real-time data will be
available about processes, products and logistics to be used by analytics and AI for
optimizing processes. Ambient invisible intelligence will also accelerate the trend to
invisible analytics identified by Gartner in other research.
However, privacy will be a major concern when small inconspicuous sensors and tags are
embedded in objects for their entire life. Vendors must address privacy concerns and
obtain consent for some types of data use. Some concerned consumers will likely want
ways to physically disable or destroy tags and sensors.
Actions:
■ Identify information shadows and early use cases where ambient intelligence
delivers return on investment in 2025 through 2028 and pilot the most promising
examples. Consider this as a replacement for some current or planned RFID use
cases.
■ Analyze the privacy implications of any proposed use of this technology and ensure
that users can disable it if required (e.g., by physical destruction of the tag or
verifiable deletion of keys).
For more information, see Top Strategic Technology Trends for 2025: Ambient Invisible
Intelligence.
Energy-Efficient Computing
Back to top
Organizations face growing legal, commercial and social pressure to improve their
sustainability, and IT organizations must play their part. IT impacts sustainability in many
ways. For example, water consumption by data centers, e-waste and recyclability are all
important, but in 2024, the leading consideration for most IT organizations is their carbon
footprint. Carbon is generated by developing and running applications, storing data and
networking. Compute-intensive applications are likely to be the biggest contributors, which
consume the most energy. Examples include AI training, simulation, optimization and
media rendering.
There are three levers that IT can use to control the carbon footprint of its systems:
■ Hardware efficiency. More modern processors and disks are more efficient. Special-
purpose devices, such as graphics processing units (GPUs), are more efficient than
general purpose equivalents for some applications. However, replacing hardware
incurs an embodied carbon cost, which may outweigh the efficiency improvements.
Short-term tactics include using greener energy, retiring inefficient hardware, improving
utilization or shifting jobs to greener cloud regions. In the medium term, good practices
are emerging for efficient coding, architecture and algorithms. Some algorithms may be
ported to more efficient hardware such as GPUs.
Demand for IT, however, is inexorably rising — especially in areas with the largest carbon
footprints, such as AI and optimization. In the long term, improvements of one to two
orders of magnitude will be required, which is more than current technology can deliver.
More radical approaches will be required, but in 2024, many are still academic research
topics or, at best, early prototypes. Examples include neuromorphic systems, which can
execute some types of AI and certain other tasks very efficiently.
Starting in the late 2020s, we expect to see a number of optical computing systems
emerge for special-purpose tasks, such as AI and optimization. These will perform some
operations using dramatically less energy than their silicon counterparts (e.g., offering
100x or better improvements). In the long term, technologies such as DNA storage,
ceramic storage and quantum computing also promise significant sustainability benefits.
■ New hardware, cloud services, skills, tools, algorithms and applications will be
required. The enterprise hardware landscape will become more complex and diverse
involving multiple new computing architectures and paradigms.
■ Energy prices will likely rise in the short to medium term, especially for green energy
where demand will outstrip supply in some regions.
Actions:
■ Identify current and planned systems that will have large carbon footprints. Consider
rearchitecting and migrating some or all onto significantly more efficient hardware,
such as GPUs.
For more information, see Top Strategic Technology Trends for 2025: Energy-Efficient
Computing.
Hybrid Computing
Back to top
It’s often said that the future of computing is “quantum.” It’s not. The future of computing
is hybrid. Hybrid computing combines different compute, storage and network
mechanisms to solve computational problems. For example, it can combine
neuromorphic, quantum, photonic and, eventually, bio and carbon computing
technologies. It involves creating a hybrid environment built on an orchestration
framework that uses the respective strengths and capabilities of the various mechanisms.
New use cases across multiple industries (i.e., manufacturing and logistics, financial
services, life sciences, materials and drug discovery) will focus mainly on two areas. First,
higher levels of automation, working toward running complete autonomous businesses.
Second, augmenting human capability, offering real-time personalization at scale and
ultimately using the human body as a computing platform itself.
Organizations adopting hybrid cloud will have to rearchitect systems and applications to
better integrate and interface across compute mechanisms. Application and system
architectures will become more complex as they leverage more diverse hybrid compute
mechanisms, amplifying the need for robust modularity and interface design.
Hybrid compute environments will require a robust software, storage and network
orchestration layer alongside associated services. They’ll need these elements to enable
organizations to use multiple compute mechanisms for given portions of the application
scope.
Introducing hybrid computing poses significant security risks, as each module part of the
system will operate autonomously on computation and decision making. This means
each module will generate data elements and pass them to other modules as part of the
application workflow. The governance and security of data pipelines and overall systems
will thus require an overhaul.
Actions:
■ Make sure that hybrid computing works by ensuring that distributed data
management is in order. First, rethink data retention — decide what information is to
be stored, where, how and for how long. Second, create a universal data fabric to
generate and maintain all the metadata between the different data domains. Third,
implement DataOps to manage data observability throughout the various data
pipelines.
■ Counter the computing cost increase over the coming years with hybrid computing.
Do so to gain the opportunity to scale and achieve benefits that far outweigh the
cost of optimizing existing infrastructures.
■ Create opportunity maps and rapidly scale performance by building efficient and
resilient autonomous computing environments. Do so because innovating while
addressing climate sustainability will be vital as energy costs increase and compute
energy consumption grows, largely driven by GenAI.
For more information, see Top Strategic Technology Trends for 2025: Hybrid Computing.
Human-Machine Synergy
Spatial Computing
Back to top
By 2026, 30% of manufacturing processes will use spatial computing to streamline and
improve efficiencies by being more dimensionally accurate.
By 2028, 20% of people will have an immersive experience with persistently anchored,
geoposed content once a week, up from less than 1% in 2023.
By 2033, spatial computing will grow to $1.7 trillion, up from $110 billion in 2023.
Spatial computing combines physical and digital objects in a shared frame of reference
beyond screen-based displays. This involves spatial mapping (see Glossary key at the end
of this document) and identification of people, places and things within the physical world
as a foundation for anchoring digital content that intersects with the physical world’s
spatially anchored, indexed and organized content.
Spatial computing digitally enhances the physical world with technologies such as
augmented reality and virtual reality. It does so seamlessly, creating the next level of
interaction between physical and virtual experiences. This expands the potential of
physical and digital objects, and their monetization possibilities.
Consumer demand is growing for immersive and interactive experiences in sectors such
as gaming, manufacturing, education, financial services and e-commerce. Additionally, the
need for sophisticated visualization tools for better decision making and increased
efficiency drives the uptake of spatial technology in industries such as healthcare, retail
and manufacturing.
Hype relating to spatial computing technologies and applications is increasing. This has
been amplified by the introduction of new head-mounted displays (HMDs) that enable
immersive digital experiences in the physical world. These devices include XREAL’s Air 2
and Beam Pro, 6 Apple Vision Pro 7 and Meta Quest 3. 8 Such devices have created the
potential for new business models and opportunities for monetizing physical-digital
interactions.
In the next five to seven years, organizations’ use of spatial computing will increase their
operational effectiveness through streamlined workflows and enhanced collaboration. It’ll
offer real-time contextual information to users for improved decision making, especially in
logistics and manufacturing.
■ The overcoming of some accessibility issues — especially for those with visual
impairment — through adaptive interfaces that can adjust text size, color contrast or
enable screen readers to access content more comfortably
■ Cost: HMDs are expensive, as is digitizing assets. Although phones and other mobile
devices offer spatial computing experiences, functionality may be limited.
■ Issues with HMDs: HMDs are heavy, making them uncomfortable. They also
consume much battery power and require frequent charging. HMDs tend to isolate
users from those around them. Concerns also exist about the potential for accidents
when wearing HMDs.
■ Lack of a killer app in the consumer market: There is no application or use case
driving long-term adoption. Today, there is more short-term interest with little
stickiness.
■ Device fragmentation: This remains an issue for developing and deploying spatial
computing content.
Actions:
■ Invest in the necessary infrastructure for spatial computing, such as internet high-
speeds, wireless, ubiquitous coverage, low latency and reliability, along with
compatible devices. Assess where spatial computing could expand the utility and
reach of your organization’s products and services.
■ Identify specific use cases where spatial computing can add value for your
organization. Evaluate, for example, the potential to enhance customer experiences,
improve operational efficiency or facilitate remote collaboration.
For more information, see Top Strategic Technology Trends for 2025: Spatial Computing.
Polyfunctional Robots
Back to top
By 2030, 80% of humans will engage with smart robots on a daily basis, up from less
than 10% today.
Task-specific robots, custom designed to repetitively perform a single function, are being
replaced with polyfunctional machines capable of doing just about anything. These next-
generation robots can take on a multitude of tasks, and seamlessly switch between them
as required, improving efficiency and providing a faster return on investment — thus
accelerating adoption across industries and use cases.
However, one can envision a single, polyfunctional, robot providing elder (and child) care,
cooking meals and cleaning the house. This trend is timed between three and 10 years,
reflecting that we’re still a long way from that ideal, but polyfunctional robots provide
other benefits which will see them adopted much more quickly in other environments.
Today’s robots often present a single point of failure in a business process, and one which
requires significant investment to deploy and maintain. Business processes are often
rearchitected to ensure that an expensive robot is fully occupied for as much time as
possible, while accepting that 100% utilization is rarely possible.
This rearchitecting further contributes to the cost (and risk) of deployment, while creating
a production bottleneck that puts efficiency savings at risk if the robot does not perform
as hoped. This can even threaten production itself if the robot fails in any way. These
risks have slowed robot adoption outside of specialist industries with volume
requirements large enough to allow for significant redundancy.
Tomorrow’s polyfunctional robots are a different breed. These robots can take on different
tasks, but are also designed to fit into a human-shaped world, thus removing the need for
architectural changes to the process (or bolt-down infrastructure in the working
environment), making for fast deployment, low risk and easy scalability. Many are
humanoid, or partly humanoid, in form factor. They can thus substitute for (or be
substituted by) a human worker, which increases flexibility, as humans and robots can
collaborate within the same working space, to the benefit of both.
This market is still developing, with new products being developed by existing players
such as Huandi-owned Boston Dynamics, ABB and new entrants such as Tesla and
Unitree. The range of products is wide: a robot dog might cost between $75,000 (Boston
Dynamics’ SPOT) and $1,600 (Unitree’s Go2). Similar disparities exist in humanoid and
torso form factors. Though the industry is still deciding what constitutes the minimum
functionality required, and thus what price is acceptable, the development of hardware is
being accompanied by huge leaps in software.
Actions:
For more information, see Top Strategic Technology Trends for 2025: Polyfunctional
Robots.
Neurological Enhancement
Back to top
By 2030, 30% of knowledge workers will be enhanced by, and dependent on, technologies
such as BBMIs (both employer- and self-funded) to stay relevant with the rise of AI in the
workplace, up from less than 1% in 2024.
By 2034, 50% of Fortune 500 companies will collect bioinformation from employee-used
devices, prompting a revamp of privacy policies.
■ The extraction of information from the brain (e.g., thoughts and emotions)
Much work is being done in this area. For example, Neuralink obtained FDA approval for
human trials of implantable chips in May 2023, 10 a business-critical development, with
the first patient operated on in January 2024. 11 Additionally, the sums being invested in
neurological enhancement are significant. In April 2024, brain-chip maker Blackrock
Neurotech received a $200-million investment from crypto firm Tether. 12
■ Medical research
■ Mindfulness
■ Workplace monitoring
■ Personalized education
However, implementation and adoption challenges exist. They fall into several categories:
■ Cost and technology issues: These include the high cost of early products, limited
battery life, limited mobility and wireless connectivity options and the complexity of
integrating different data systems.
■ Social acceptance: More advanced functionality means more invasive and risky
solutions. People will be resistant to implants because of the risk of initial surgery.
This will be compounded by the need for repeat surgery because first-generation
devices will have a short life span. Social acceptance, especially for the more
conspicuous form factors, may be a long way off.
■ Security: UBMIs and BBMIs interface directly with the human brain. This creates
security challenges and new vulnerabilities to individuals and companies.
■ Ethics and privacy: The use of UBMIs and BBMIs raises serious ethical concerns,
including issues such as altering users’ perceptions of reality, memories or even their
personalities. The right to think freely must be revisited, as must the concept of
privacy. 13
Actions:
■ Set up proofs of concept of existing solutions with acceptable form factors (e.g.,
headphones) where you have applicable use cases (i.e., accelerated training). Do
this for solutions that are already providing high returns and have high levels of
acceptability.
■ Keep customers and your business safe by implementing data anonymity and
privacy for the collection and management of data from brain-wearables.
For more information, see Top Strategic Technology Trends for 2025: Neurological
Enhancement.
■ AI-augmented development
■ Democratized generative AI
■ Intelligent applications
■ Machine customers
■ Platform engineering
■ Sustainable technology
2
Article 27: Fundamental Rights Impact Assessment for High-Risk AI Systems, EU
Artificial Intelligence Act.
3
Growing Public Concern About the Role of Artificial Intelligence in Daily Life, Pew
Research Center.
4
HSBC’s Principles for the Ethical Use of Data and AI, HSBC.
5
AI Ethics at Unilever: From Policy to Process, MIT Sloan Management Review.
6
Millions of Apps in AR Glasses, XREAL.
7
A Guided Tour of Apple Vision Pro, Apple.
8
Do What You Love in Entirely New Ways, Meta.
9
Quick Answer: What Is Industrial Metaverse?
10
Elon Musk’s Neuralink Says It Has FDA Approval for Human Trials: What to Know,
Washington Post.
11
‘ People Think It’s Like the Matrix’: Neuralink’s First Patient on Having a Brain Chip,
Euronews.
12
What $200 Million in Crypto Cash Means for Blackrock Neurotech, Forbes.
13
Your Brain May Not Be Private Much Longer, Vox.
Hybrid Computing
Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.
Alternative Proxies: