Top Strategic Techno 815761 NDX

Download as pdf or txt
Download as pdf or txt
You are on page 1of 31

Top Strategic Technology Trends for 2025

21 October 2024 - ID G00815761 - 40 min read

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

Initiatives:Technology Innovation and Strategy

We’ve identified the 10 strategic technology trends that will have


the most impact in the next five years and beyond — trends that
span AI imperatives and risks, new frontiers of computing and
human-machine synergy. Tracking these will help IT leaders shape
the future with responsible innovation.

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.

■ Human-machine synergy is increasing, with the creation of next-level interactions


between physical and virtual experiences. Robots that perform more than one
function will integrate into humans’ daily lives. Technology will bring the ability to
directly access and improve thoughts and emotions for enhancing human cognition
and performance, and bringing about new ways to help people thrive.

Gartner, Inc. | G00815761 Page 1 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Recommendations
■ Identify opportunities to add agentic AI to workflows where significant demand for
scale and efficiency exists and adaptability is required. Ensure fairer AI systems by
considering multiple perspectives when designing and evaluating AI methods.
Include deepfake detection in systems such as identity verification.

■ Develop policies to ease the transition to new cryptographic algorithms. Identify


information blind spots and early opportunities to collect data from your physical
environment using ambient intelligence. Use computing more efficiently by
switching to greener cloud providers. Manage the complexity of using diverse
computing models building a robust and scalable orchestration layer for
provisioning and managing resources.

■ Invest in the necessary infrastructure for spatial computing. Adopt a policy of


“polyfunctionality by default” for all robot deployments. Set up proofs of concept for
neurological enhancement solutions where you have applicable use cases.

What You Need to Know


Shape the Future With Responsible Innovation
Organizations face the challenge of continually having to innovate to meet business
challenges and disruptions. As new technologies arise, they present many opportunities
but increasingly bring ethical challenges and considerations. Organizations must act
responsibly to balance innovation while retaining the trust of their customers, employees
and partners. This research will help you shape the future for your organization with
responsible and ethical innovation.

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.

Gartner, Inc. | G00815761 Page 2 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


New frontiers of computing are being created, requiring organizations to look differently
at how they compute. As new security measures will be needed, information in the
shadows today must become visible in the future.

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.

In these new frontiers of computing, quantum computing threatens to break today’s


cryptology, exposing everyone to risk. A new cryptology is needed to protect organizations
and society. Tiny, ultra-low-cost wireless tags and sensors will make possible real-time,
large-scale tagging, tracking and sensing — enabling new business models and
ecosystems.

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.

Humans will become integrated with machines through wearable or implanted


technologies, and neurological enhancement will give us the ability to directly access and
improve thoughts and emotions. This will enhance human cognition and performance,
bringing about new ways to help humans.

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.

Gartner, Inc. | G00815761 Page 3 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Figure 1: Top Strategic Technology Trends for 2025

Trend Profiles. Click links to jump to profiles

AI Imperatives and Risks New Frontiers of Human-Machine Synergy


Computing

Agentic AI Postquantum Cryptography Spatial Computing

AI Governance Ambient Invisible Polyfunctional Robots


Platforms Intelligence

Disinformation Security Energy-Efficient Computing Neurological Enhancement

Hybrid Computing

AI Imperatives and Risks


Agentic AI
Back to top

Analysis by Tom Coshow, Gary Olliffe

Strategic Planning Assumptions:

Gartner, Inc. | G00815761 Page 4 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


By 2028, 33% of enterprise software applications will include agentic AI, up from less
than 1% in 2024.

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.

The goal-driven planning capabilities of agentic AI also promises to deliver more


adaptable software systems, capable of completing a wide variety of “undefined” tasks
within a domain, rather than only those designed into the software.

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.

Emerging and existing AI features, such as agent and multiagent frameworks, AI


guardrails and function-calling, enable systems that use LLMs to have agency. Agentic
systems using these capabilities can plan, act and adapt based on their context to meet
goals in complex environments, dramatically increasing AI’s potential. For example,
agentic AI can examine data, do research, compile a list of tasks to complete and then
take those actions in the digital world or physical world via APIs or robotic systems.
Agentic AI, with its ability to take action autonomously or semiautonomously, has the
potential to realize CIOs’ desire to increase productivity across the organization. 1 This
motivation is driving both enterprises and vendors to explore, innovate and establish the
technology and practices needed to deliver this agency in a robust, secure and trustworthy
way.

Gartner, Inc. | G00815761 Page 5 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


The key challenges facing organizations that are building early AI agents lie in the need to
establish high levels of trust and confidence. Agents that can autonomously select and
use tools (via features APIs and function calling) must be constrained by robust
guardrails that ensure their actions are aligned to the provider’s intentions and properly
reflect user intent. Frameworks for enforcing AI guardrails are growing in maturity, but
commonly also rely on LLMs as part of their processing, leaving a margin for error.

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:

■ Identify opportunities to add agentic AI to workflows where significant demand for


scale and efficiency exists and adaptability is required. Rethink entire workflows
across silos from an automation-only perspective and add humans back into new
workflows at strategic points. Start small on use cases where high-quality data is
accessible and behavior is verifiable.

■ 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.

Gartner, Inc. | G00815761 Page 6 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Put in place guardrails to ensure agentic AI is constrained to a defined role and set of
capabilities. Do so to prevent it from taking incorrect actions that cause damage.

For more information, see Top Strategic Technology Trends for 2025: Agentic AI.

AI Governance Platforms
Back to top

Analysis by Jasleen Kaur Sindhu, Moutusi Sau, Svetlana Sicular

Strategic Planning Assumptions:

By 2028, enterprises using AI governance platforms will achieve 30% higher customer
trust ratings and 25% better regulatory compliance scores than their competitors.

By 2028, organizations that implement comprehensive AI governance platforms will


experience 40% fewer AI-related ethical incidents compared to those without such
systems.

AI governance platforms are technology solutions that enable organizations to manage


the legal, ethical and operational performance of their AI systems. These platforms, part
of Gartner’s evolving AI Trust, Risk and Security Management (AI TRiSM) framework, help
enforce enterprise AI policies. They offer key features such as creating, managing and
enforcing policies for responsible AI use, explaining how AI systems work, model life cycle
management and providing transparency to build trust and accountability.

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.

Gartner, Inc. | G00815761 Page 7 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


For example, the EU’s AI Act 2 requires high-risk AI systems to respect human rights,
accelerating the demand for comprehensive AI governance platforms. Public concern
about AI harms is also growing, with a 2023 Pew Research Center 3 survey finding that
52% of respondents are more concerned than excited about AI. High-profile cases of AI
bias and disinformation have further highlighted the need for better oversight and
governance.

Despite the benefits, implementing AI governance platforms comes with challenges. AI


governance practices are still emerging, with varying guidelines across regions and
industries. The fast pace of AI development makes it hard to establish consistent and
future-proof governance practices. Cultural barriers and resistance to change can also
impede adoption. However, there are significant opportunities.

Technological advancements in explainable AI, automated model monitoring and scalable


AI governance tools can help organizations implement robust governance practices.
Adopting AI governance platforms can help organizations stay compliant with evolving
regulations, reduce legal risks and build trust. Transparent and fair AI systems can
improve consumer trust and brand reputation, providing a competitive edge.

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

Similarly, Unilever has an AI governance framework that includes an AI assurance


compliance process, involving internal reviews, third-party assessments and continuous
monitoring to ensure AI systems align with ethical standards and operational goals. 5

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.

Gartner, Inc. | G00815761 Page 8 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Perform regular audits of AI systems to identify and mitigate risks to ensure
continuous alignment with governing standards and regulatory requirements.

■ Allocate resources to develop in-house responsible and ethical AI expertise.


Dedicated AI governance teams are more likely to scale AI initiatives successfully
and responsibly.

For more information, see Top Strategic Technology Trends for 2025: AI Governance
Platforms.

Disinformation Security
Back to top

Analysis by Dan Ayoub, Akif Khan

Strategic Planning Assumption:

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.

Disinformation security is an emerging category of technologies aimed at systematically


discerning trust. Its objective is to provide methodological systems for ensuring integrity,
assessing authenticity, preventing impersonation and tracking the spread of harmful
information.

Primary use cases for enterprises include:

■ Detecting use of synthetic media in unauthorized contexts (i.e., identity verification,


real-time communications or claims validation).

■ Intelligence monitoring for narratives spread through mass or social media (i.e.,
those targeting an executive leadership team, products, service or brand).

■ Preventing the impersonation of individuals doing business with an organization


(i.e., employees, contractors, suppliers and customers).

Disinformation security expands traditional enterprise defenses through use of emerging


technologies by:

Gartner, Inc. | G00815761 Page 9 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Strengthening controls for validating identity and ensuring integrity of real-time
communications within workforce and consumer applications to prevent fraud.
Deepfake detection can be integrated into workflows for identity, authentication,
claims validations, messaging applications and phone systems.

■ Mitigating account takeover through continuous risk scoring with contextual


awareness across session and event data within an online user journey.
Impersonation prevention extends beyond authentication by leveraging a continuous
adaptive trust model for risk assessment.

■ Protecting brand reputations by proactively identifying harmful narratives, how they


are spreading and their source for strategic incident response communications.
Reputation protection includes the ability to proactively identify harmful narratives
targeting a brand, product or executives.

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.

■ Threat actors attempt to manipulate public opinion at critical times of importance


through social media influence operations and fake news websites. Elections are
being conducted around the globe throughout 2024, with disinformation security
playing a central theme.

■ Deepfake attacks are increasing rapidly, with as many as one in 10 executives


already reporting their businesses being targeted. Biometric authentication was
thought to be safe, accurate and reliable, but deepfakes show that no single
technology layer is impenetrable.

Gartner, Inc. | G00815761 Page 10 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Harmful narratives are being amplified with AI to quickly spread online, damaging
reputations and manipulating audiences before organizations are prepared to
respond. Enterprises are now adopting narrative intelligence solutions pioneered for
government and defense use cases.

Actions:

■ Ensure deepfake detection is covered as a feature capability within broader systems


such as identity verification, biometric authentication and claims validation. Look for
dedicated solutions to authenticate third-party content or monitor integrity of real-
time workforce communications systems.

■ Move beyond authentication to incorporate impersonation prevention through


continuous adaptive trust models that continuously assess risk across the entire
user journey. Look for context-aware solutions that incorporate behavioral analytics
within probabilistic orchestration models.

■ Incorporate digital risk protection services, narrative intelligence and media


monitoring as part of a layered brand safeguarding strategy. Leverage tools to
proactively identify information harmful to the organization and cross-functional
teams to develop response strategies.

For more information, see Top Strategic Technology Trends for 2025: Disinformation
Security.

New Frontiers of Computing


Postquantum Cryptography
Back to top

Analysis by Mark Horvath, Bart Willemsen

Strategic Planning Assumption:

By 2029, advances in quantum computing will make most conventional asymmetric


cryptography unsafe to use.

Gartner, Inc. | G00815761 Page 11 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


In 2019’s Top Strategic Technology Trends, we pointed at a long-term trend to prepare for:
quantum computing (QC). Ever since, QC developments have progressed steadily and
expectations of its availability before the end of 2020 proved on point. QC is expected to
mean the end of several types of conventional cryptography used widely in billions of
devices and over 80% of communications over the global internet.

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.

Postquantum cryptography (PQC) provides data protection that is resistant to QC


decryption risks. However, switching cryptography methods in existing architectures is not
an easy task, as the new algorithms are not drop-in replacements for existing asymmetric
algorithms. As such, organizations must prepare a longer lead time to ready themselves
for robust protection of anything sensitive or confidential. Owners of sensitive assets will
not know if their data has been harvested and later decrypted until it’s too late. These are
not just risks to the business but to any individual or group of individuals whom that data
may be about as well.

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:

■ Develop policies to ease the transition to new algorithms. Do so because adopting a


policy-based program will reduce confusion and arbitrary choices, and increase
manageability.

Gartner, Inc. | G00815761 Page 12 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Build a cryptographic metadata database of all in-use cryptographic algorithms. Use
it to identify the expected end-of-life targets in the short-, mid- and long-term time
scales. Create a key life cycle policy to reflect the risks to asymmetric keys.

■ Implement crypto-agile application development and move to production after


extensive testing. Vet and test new PQC algorithms to understand their
characteristics, uses and performance. Upgrade or replace hardware where
necessary.

For more information, see Top Strategic Technology Trends for 2025: Postquantum
Cryptography.

Ambient Invisible Intelligence


Back to top

Analysis by Nick Jones

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:

■ Low-power wireless, low-energy Bluetooth will dominate in 2025, but technologies


such as Wi-Fi and 5G are also exploring extensions to support ambient wireless.
Innovations, such as backscatter wireless, will likely also be used in the future.

■ 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.

Gartner, Inc. | G00815761 Page 13 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Through 2028, early examples of ambient invisible intelligence will focus on solving
immediate problems by enabling low-cost, real-time tracking and sensing of items to
improve visibility and efficiency. Such examples will be limited to single organizations or
tightly coupled supply chains because of current systems’ need for infrastructure.
Examples of processes that could benefit include retail stock checking, perishable goods
logistics and organizations managing large numbers of packages, such as couriers and
postal services.

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.

Gartner, Inc. | G00815761 Page 14 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Look for new AI and analytics opportunities enabled by real-time information (e.g.,
related to how objects are used, and how and where they are stored).

■ 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).

■ Collaborate with partners and industry-standards organizations to look for longer-


term ecosystem opportunities based on lifelong knowledge of an object’s identity
and properties.

For more information, see Top Strategic Technology Trends for 2025: Ambient Invisible
Intelligence.

Energy-Efficient Computing
Back to top

Analysis by Nick Jones

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:

■ Architecture, code and algorithms. Develop applications with more efficient


algorithms, architectures and data structures.

■ 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.

Gartner, Inc. | G00815761 Page 15 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Greener power. Power runs IT systems and cools data centers. Carbon intensity is
the amount of carbon emitted for each kilowatt hour of electricity generated. It will
be almost zero for renewable sources and very high for fossil-fuel sources. Some IT
organizations will choose when and where to run jobs to exploit regional differences
in the supply carbon intensity.

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.

However, the path to more sustainable IT will be disruptive, because:

■ 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.

■ Migrating to new computing platforms and architectures may be complex and


expensive.

■ Energy prices will likely rise in the short to medium term, especially for green energy
where demand will outstrip supply in some regions.

Gartner, Inc. | G00815761 Page 16 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Some organizations will face a rearchitecting sustainability chasm, where improved
efficiency is required, but new technologies aren’t sufficiently mature.

■ If IT leaders fail to improve sustainability, the performance of their business could be


impacted. Sustainability constraints could limit their ability to deploy advanced
solutions on current platforms.

Actions:

■ Adopt short-term tactical solutions, such as improving utilization, switching to


greener cloud providers, shift workloads to greener cloud regions, run systems at
times and locations where the local supply has a lower-carbon intensity.

■ 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.

■ Monitor emerging technologies such as optical and neuromorphic computing and


pilots when available.

For more information, see Top Strategic Technology Trends for 2025: Energy-Efficient
Computing.

Hybrid Computing
Back to top

Analysis by Soyeb Barot, Frank Buytendijk

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.

Hybrid computing helps organizations explore and solve problems by combining


specialized compute, storage and network mechanisms — combined. This helps
technologies, such as AI, perform beyond current technological limits.

Gartner, Inc. | G00815761 Page 17 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Hybrid computing will be used to create highly efficient, high-speed transformative
innovation environments. These environments will perform more effectively than
conventional environments, leveraging the strengths and capabilities of each individual
computation mechanism. They’ll be able to deal with high-dimensional optimization
problems scalable for large, complex problems — using significantly less energy.

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.

However, hybrid computing requires exploring highly experimental technologies and


scaling successful initiatives. Many of the involved technologies are not only nascent but
highly complex, requiring specialized skills. Organizations seeking to create value from
using hybrid computing must have a high tolerance for cost and challenges with
complexity, security and trust.

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:

Gartner, Inc. | G00815761 Page 18 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Manage the complex nature of a hybrid computing architecture that may easily
become “messy” by building a robust and scalable orchestration layer. This may
involve establishing delivery platforms that support distributed data management
fabric, multiple software architectures and deployment of applications to diverse
hybrid computing environments.

■ 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.

■ Evaluate each computing mechanism for relevance in your environment and


establish reference architectures to guide the adoption and integration of each —
spanning data, infrastructure and application environments.

■ 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

Analysis by Marty Resnick

Strategic Planning Assumptions:

By 2026, 30% of manufacturing processes will use spatial computing to streamline and
improve efficiencies by being more dimensionally accurate.

Gartner, Inc. | G00815761 Page 19 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


By 2027, over 40% of large organizations worldwide will use a combination of Web3,
spatial computing and digital twins in metaverse-based projects aimed at increasing
revenue.

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.

For example, spatial computing will enable:

Gartner, Inc. | G00815761 Page 20 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ New monetization models, such as the monetization of physical assets (e.g., users
can click to buy event tickets, lease parking spaces or rent vehicles)

■ Virtual collaboration for R&D, product design and training

■ 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

■ Immersive experiences across product, sales and market funnels

■ Interaction with digital twins in a manufacturing environment as part of industrial


metaverse initiatives 9

■ Visual search for real-time information overlaid on physical products

■ New consumer experiences that’ll increase market opportunities in consumer


electronics, gaming and brand marketing

■ More inclusive hiring because devices will enable employees to collaborate in a


shared space, irrespective of their physical location

However, several challenges exist to the uptake of spatial computing:

■ 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.

■ Complicated user interfaces: Many HMDs have more of an “onboarding process”


with tutorials needed to fully understand actions that may be less than intuitive.

■ 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.

Gartner, Inc. | G00815761 Page 21 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Data privacy and security: Spatial computing uses cameras and sensors to collect
data about the user’s environment, actions and interactions. Devices can store vast
amounts of this personal data, and organizations must ensure this data’s safety. The
lack of device management capabilities could also result in security vulnerabilities,
but the sensitivity of stored data also poses ethical and legal issues.

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.

■ Adopt a platform to bring multiple spatial computing experiences together. Ensure it


can optimize the streaming of content in areas with low bandwidth.

For more information, see Top Strategic Technology Trends for 2025: Spatial Computing.

Polyfunctional Robots
Back to top

Analysis by Bill Ray, Nick Jones

Strategic Planning Assumption:

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.

Gartner, Inc. | G00815761 Page 22 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Not only will this accelerate adoption in manufacturing and warehouses, it also creates
new opportunities for single-function robots that won’t make economic sense. Most
consumer homes, for example, lack the space for multiple robots accomplishing different
tasks, and it wouldn’t be economical to have such a fleet of electronic servants.

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.

Gartner, Inc. | G00815761 Page 23 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Training the robot to a new task now consists of showing the robot what to do. However,
the polyfunctional robot isn’t just a mimic — it interprets the task and intended outcome,
enabling it to modify the instructions based on the desired objective, rather than just
copying what it has seen. This kind of training — combined with better motors, batteries,
sensors and actuators — makes polyfunctional robots one of the most revolutionary
trends on this list.

Actions:

■ Adopt a policy of “consider polyfunctionality first” for all robot deployments, to


maintain flexibility and anticipate future demands by considering how the robot
might need to retain value as demand changes.

■ Start drawing up essential company policies for robot coexistence (robotology),


consider how your staff will be expected to interact with robots and how they might
react.

For more information, see Top Strategic Technology Trends for 2025: Polyfunctional
Robots.

Neurological Enhancement
Back to top

Analysis by Sylvain Fabre, Frank Buytendijk

Strategic Planning Assumptions:

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.

Neurological enhancement improves human cognitive abilities using technologies that


read and decode brain activity. It “reads” a person’s brain to provide “brain transparency”
by using unidirectional brain-machine interfaces (UBMIs) or bidirectional brain-machine
interfaces (BBMIs; see Glossary for definitions) and a range of other approaches. Soon,
neurological enhancement will also be able to “write” to the brain, enhancing its function.

Gartner, Inc. | G00815761 Page 24 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


The UBMIs and BBMIs at the heart of neurological enhancement measure electrical
activity in the user’s brain and also monitor the user’s mental state. BBMIs enable some
modification of the brain’s state based on analytics and insights. This modification occurs
in two ways: via noninvasive electrostimulation through a head-mounted wearable, or an
invasive implant.

Neurological enhancement’s brain-communicating capability sets it apart from other


technologies that also employ wearables and implants. It’ll enable organizations to
monetize:

■ The addition of capability to the brain (e.g., in information processing, memory,


learning and gaming)

■ 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

Neurological enhancement has huge potential in three main areas:

■ Human upskilling: Enhances cognitive abilities such as memory, attention, learning


and problem solving. This is vital to help people keep up with AI and for workers to
stay relevant — being an “enhanced” human may become a condition of
employment.

■ Next-generation marketing: Enables brands to know, in real time, what consumers


are thinking and feeling as they interact with the world.

■ Performance: Enhances human neural capabilities to optimize outcomes in areas


ranging from the prevention of industrial accidents, to learning, to healthy aging. For
example, it could prevent drivers from falling asleep, personalize education for each
learner and enable an aging population to stay working for longer.

Use cases will evolve in this sequence:

■ Medical research

Gartner, Inc. | G00815761 Page 25 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Medical applications

■ Mindfulness

■ Workplace monitoring

■ Personalized education

■ Neurological enhancements for business professionals

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.

■ Sensing: Electrical activity in the brain is 3D and in depth, so measuring activity on


the surface can’t capture the full “digital twin” of the brain’s electrical activity. Neither
can an embedded implant.

■ 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.

Gartner, Inc. | G00815761 Page 26 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


■ Protect your organization from the issues of workers using their own consumer
neurological enhancement devices by preparing now. Involve legal counsel early on
to establish policies for unauthorized implantables.

■ 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.

Changes Since Last Year


For 2024, Gartner identified 10 strategic technology trends (see Top Strategic Technology
Trends for 2024):

■ AI-augmented development

■ AI trust, risk and security management

■ Augmented connected workforce

■ Continuous threat exposure management

■ Democratized generative AI

■ Industry cloud platforms

■ Intelligent applications

■ Machine customers

■ Platform engineering

■ Sustainable technology

Gartner, Inc. | G00815761 Page 27 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Evidence
1
2024 Gartner CIO Generative AI Survey. This survey was conducted online from 30
January through 12 February 2024 to examine how CIOs are thinking about generative AI
and the current role of the CIO in generative AI initiatives. This serves as an update to the
2023 Gartner CIO Generative AI Survey. In total, 83 CIOs who were members of Gartner’s
Research Circle participated. Members from North America (n = 42), EMEA (n = 29),
Asia/Pacific (n = 7) and Latin America (n = 5) responded. (Gartner’s CIO Research Circle
members include enterprise-level CIOs/CTOs, divisional CIOs/CTOs and heads of the
office of the CIO, representing a mix of industries and organization sizes.) Disclaimer:
Results of this survey do not represent global findings or the market as a whole, but reflect
the sentiments of the respondents and companies surveyed.

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.

Gartner, Inc. | G00815761 Page 28 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


Acronym Key and Glossary Terms
Bidirectional A neural interface that enables two-way communication between
Brain- a human brain and a computer or machine. It can measure the
Machine user’s mental state and also influence it.
Interface
(BBMI)

Spatial The creation of 3D (x, y, z) datasets, also called “point clouds,” to


Mapping produce a 3D visualization of a (static) indoor or outdoor
environment.

Unidirectional A neural interface that enables one-way communication between


Brain- a human brain and a computer or machine. It can measure the
Machine user’s mental state and enable the user to control computers or
Interface machines via thought.
(UBMI)

Document Revision History


Top Strategic Technology Trends for 2024 - 16 October 2023

Top Strategic Technology Trends for 2023 - 17 October 2022

Top Strategic Technology Trends for 2022 - 18 October 2021

Top Strategic Technology Trends for 2021 - 19 October 2020

Gartner, Inc. | G00815761 Page 29 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


© 2024 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered trademark of
Gartner, Inc. and its affiliates. This publication may not be reproduced or distributed in any form
without Gartner's prior written permission. It consists of the opinions of Gartner's research
organization, which should not be construed as statements of fact. While the information contained in
this publication has been obtained from sources believed to be reliable, Gartner disclaims all warranties
as to the accuracy, completeness or adequacy of such information. Although Gartner research may
address legal and financial issues, Gartner does not provide legal or investment advice and its research
should not be construed or used as such. Your access and use of this publication are governed by
Gartner's Usage Policy. Gartner prides itself on its reputation for independence and objectivity. Its
research is produced independently by its research organization without input or influence from any
third party. For further information, see "Guiding Principles on Independence and Objectivity." Gartner
research may not be used as input into or for the training or development of generative artificial
intelligence, machine learning, algorithms, software, or related technologies.

Gartner, Inc. | G00815761 Page 30 of 30

This research note is restricted to the personal use of fatih.balyeli@exaion.com.


AI Imperatives and Risks New Frontiers of Computing Human-Machine Synergy

Agentic AI Postquantum Cryptography Spatial Computing

AI Governance Platforms Ambient Invisible Intelligence Polyfunctional Robots

Disinformation Security Energy-Efficient Computing Neurological Enhancement

Hybrid Computing

Gartner, Inc. | G00815761 Page 1A of 1A

This research note is restricted to the personal use of fatih.balyeli@exaion.com.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

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:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy