2025 was quite a year for artificial intelligence. We asked for predictions from the people working every day to take AI into the future where they see AI headed in 2026. The response was so great, we broke it down in two parts. Click here if you missed Part One.
Happy New Year…. and now check out AI Insider’s Expert Predictions on Artificial Intelligence: Part Two.
Crystal Foote, Founder and Head of Partnerships at Digital Culture Group
Diversity and Cultural Intelligence Will BecomeAI’s Strategic Edge
In 2026, the most effective AI-driven campaigns will come not from those with the largest datasets, but from brands that embed cultural intelligence into the core of their tech infrastructure. AI is only as effective as its operator; without diverse talent guiding its development and deployment, campaigns risk automating bias instead of unlocking insight. Authentic cultural resonance requires more than data—it demands teams with lived experience and cultural fluency. Agencies that eliminated DEI roles missed a critical opportunity to evolve those positions into AI and innovation leadership. The expertise required to detect and mitigate bias, understand nuance, and build inclusive strategies already existed within those roles. To succeed in the next wave of AI, brands must reimagine their organizational structures—infusing diversity not as a layer of review, but as a strategic function at the foundation of how AI is built, trained, and applied.
Voice Will Become the Next Frontier for Contextual Targeting
As consumers increasingly interact with technology through voice, advertising is poised to follow. Next year, voice-activated ad targeting will become a breakout channel, driven by a shift in behavior: people are speaking to their devices more than ever, replacing typed searches with spoken commands. This behavioral shift signals a major opportunity for brands to rethink how, where, and when they show up in consumers’ lives.
Asking Alexa for a recipe, cueing up a podcast on Google Home, or checking the weather through Siri are no longer just utility moments—they’re emerging touchpoints for real-time, hyper-relevant advertising. If brands and platforms don’t yet have infrastructure in place to retarget based on voice commands, they’re behind. Voice offers the immediacy and intimacy of conversation, paired with the potential for contextual targeting that reflects a consumer’s exact need in that moment. In a world where convenience, relevance, and personalization are paramount, voice isn’t just a media channel—it’s a strategic unlock.
Not Every Trend Is Yours to Chase: Brands will Re-evaluate When to Join a Cultural Movement
Too often, brands jump into cultural conversations without understanding the audience they’re speaking to. The result? Cringe-worthy content, outdated references, and missed opportunities. The pressure to stay culturally relevant will push brands to chase emerging trends, especially those tied to Gen Z and Gen Alpha.Howver the smartest marketers will ask themselves is this really our audience? It’s no longer about being “in culture” for the sake of optics—it’s about being in alignment with the audience you’re trying to reach. Cultural cachet doesn’t guarantee purchase intent – resonance does.
In the year ahead, brands will need to pivot, reassessing whether their values, offerings, and internal identity genuinely align with the generation they’re pursuing. When marketers treat culture as a static playbook or jump into conversations they don’t fully understand, the result is dissonance. Relevance in 2026 won’t be about looking youthful – it will be about using real-time audience signals, emotional context, and resonance tools to ensure they’re not just speaking, but connecting.
M&A Activity Will be Driven by Cultural Intelligence, Not Just Data
Startups are outperforming legacy platforms due to their agility and cultural fluency. This shift will drive a new wave of M&A activity in 2026. Just as holding companies once acquired influencer agencies to buy credibility, they’ll now seek out AI startups that offer clarity—tools that decode what people feel before they act. The value won’t be in the data alone, but in the cultural intelligence behind it.
The Smartest Startups Will Seek Collaboration, not Acquisition
In 2026, the most meaningful AI innovation in advertising won’t come from big tech—it will come from startups. Agile and unencumbered by red tape, these companies will lead by building culturally intelligent, partnership-driven solutions that reflect how people actually live and choose. They’ll use AI not just to automate but to anticipate—decoding emotional, behavioral and cultural signals in real time. The smartest founders won’t wait for acquisition offers; they’ll build power through collaboration. By joining forces with other nimble innovators, they’ll form ecosystems of shared data, creativity and cultural fluency that big tech can’t replicate. Together, these partnerships will create value through clarity, resonance and measurable outcomes shaping the future of advertising on their own terms.
Octavian Tanase, CPO at Hitachi Vantara
Agentic AI Will Dominate Enterprise Innovation
Agentic AI will see a tremendous surge in enterprise adoption in 2026. People will be able to create autonomous modules seamlessly embedded in business workflows that can make decisions on their behalf. While enterprises have already begun to experiment with generative AI, these agentic systems will go further by enabling self-directed automation and decision-making at scale, making real impact on businesses while underscoring the need for governance, security and trust in AI systems.
Data Emerges as the Core Competitive Asset
How companies control, manage and use proprietary and high-quality data will create] competitive advantage. Organizations are deluged with an ever-increasing volume of data, which they are feeding into AI models, but as compute and algorithms become increasingly commoditized, proprietary and high-quality data will define competitive advantage. So, companies will focus even more on securing high-quality proprietary data and activating it across edge, core and cloud to refine, fine-tune, correlate or augment their AI and agentic models as a differentiator. Organizations that build trusted data foundations will lead in innovation and resilience.
Sovereign Clouds Become Strategic Infrastructure
Many governments now view AI as a matter of national security due to the sensitive personal, defense and financial data involved and therefore seek to retain that data within their jurisdictions. They also see AI infrastructure as critical to control, including security and regulatory assurances. So sovereign cloud ecosystems will accelerate growth, particularly for countries focusing on digital sovereignty. Secure, compliant and energy-efficient data centers will become essential for trusted AI development and regulatory assurance.
AI Workloads Will Redefine Data Infrastructure
Data infrastructure is evolving to handle AI pipeline demands, including AI preparation, model training and inference. This infrastructure will emphasize ultra-low latency, scalability and elasticity across data lakes and block storage systems.
AI Drives Hardware Bottlenecks to the Compute Layer
AI is driving the bottleneck to the compute layer, away from networking and storage. Compute will move closer to data and storage, with technologies such as GPU Direct, allowing for faster insights and greater efficiency across data-intensive environments.
Hybrid Cloud and Kubernetes Will Anchor Future IT Models
Enterprises increasingly need the flexibility to deploy, scale and rebalance workloads seamlessly across environments based on performance, security or cost. As a result, hybrid cloud strategies will move towards Kubernetes to maximize workload mobility, automation and cost optimization. This will enable the blending of the agility of a public cloud with the control and compliance of private infrastructure.
Power and Density will Drive Data Center Modernization
AI and data-heavy workloads demand exponentially more compute and storage, which is straining energy use and cooling capacity. Power and energy efficiency is an increasing challenge for AI infrastructure and data centers. Flash storage offers better performance with lower power consumption and a smaller footprint so that organizations can consolidate applications, reduce operational costs and free up power for AI innovation. Companies will look to adopt systems that enable them to consolidate applications through high-density or flash-based technology versus traditional hard disk environments. With sustainability needs increasing and regulations growing, enterprises will modernize around compact, energy-efficient architectures that deliver performance, business and sustainability returns.
Workforce Reskilling Becomes Essential in an AI World
Autonomous systems will grow to dominate business processes over time, but success still depends on people. This means that enterprises need to invest in reskilling people to use AI. The barrier to entry is no longer high, and the use cases of AI and agentic AI will continue to expand across verticals, including product development, customer sentiment analysis, human resources or finance. Much of the next wave of innovation in agentic AI will come not only from computer scientists but from functional industry experts.
Soumendra Mohanty, CSO of Tredence
As agentic AI shifts from copilots to autonomous workflows, leaders must adopt three critical factors:
- The first is algorithmic empathy — the ability to understand how AI works, how it generates outputs, and the risks associated with it.
- Second is human–AI collaboration — with clear clarity on when to rely on AI and when human judgment must step in.
- Lastly, as agentic AI workflows become increasingly autonomous — organizations will need strong controls, observability, and governance frameworks to mitigate risks and accelerate performance.”
Don Boxley, CEO and Co-Founder of DH2i
AI Outages Become the New “Ransomware Moment”
In 2026, the biggest wake-up call for enterprises will be unexpected AI outages. As more organizations rely on AI systems for customer service, fraud detection, claims processing, supply chain routing, and decision automation, even a few minutes of downtime will create real-world business disruption. We’re moving into an era where AI is fully embedded into workflows, which means the databases, pipelines, and connections behind those AI systems must be architected for continuous availability. The companies that treat AI like a traditional app are going to run into the same wall we saw with ransomware years ago: you don’t realize how fragile the architecture is until it breaks.
What I’m seeing going into 2026 is a shift from ‘How do we deploy AI?’ to ‘How do we keep AI running, resilient, and trustworthy every second?’ The winners will be the companies that build durable foundations – resilient failover, airtight DR strategies, and secure, persistent connections between every environment where the data and compute live. AI will only be as reliable as the infrastructure supporting it. Businesses have to treat availability and security as non-negotiable if they want AI to successfully transform outcomes.
Multi-Cloud Fragmentation Becomes a Crisis
Whether they planned it or not… by 2026, nearly every enterprise will be operating in a patchwork of public cloud, private cloud, containers, and edge environments. When apps need to talk to each other securely, or when data must move quickly and reliably to support analytics and AI, that fragmentation will become a real liability. Teams are already discovering that traditional networking and legacy failover approaches simply don’t work at multi-cloud scale. The complexity isn’t slowing down – so the resiliency architecture and network connectivity has to evolve to match the world we’re deploying into.
What I expect to see in 2026 is a massive shift toward secure, lightweight, point-to-point connectivity models built on zero-trust principles. Companies need a way to ensure constant uptime, fast recovery, and secure movement of data across clouds without wrestling with brittle tunnels or static network overlays. High availability isn’t just about servers anymore — it’s about the entire distributed fabric staying resilient. Businesses will choose solutions that let them seamlessly failover across clouds, maintain jurisdictional control, and securely reach any resource from anywhere. That’s the only way to operate confidently in a multi-cloud world.
Disaster Recovery Moves From “Backup Plan” to “Active Architecture”
For years, disaster recovery has been the fire extinguisher in the hallway — something everyone pays for but hopes they’ll never have to touch. That thinking won’t make it through 2026. Regulators are tightening the screws in finance, healthcare, and government. Cloud regions are going dark without warning. Geopolitical tensions and climate disasters are taking entire data centers offline. The idea that a single cloud or region can keep you safe is becoming a dangerous illusion. Disruption isn’t the exception anymore. It’s the operating environment.
The companies that don’t get caught flat-footed will treat resilience as a living, breathing part of their architecture — not an afterthought. Cross-region and cross-cloud failover will shift from “nice to have” to the only sane way to run a business. And whether critical apps come back online fast enough will depend on secure, low-latency connections that don’t crumble under pressure. In 2026, resilience becomes a board-level concern. The organizations that invest in it now will be the ones still delivering uninterrupted services when everyone else is scrambling to recover.
Richard Copeland, CEO of Leaseweb USA
Trusted Execution Environment Technology Will Reshape Distributed Compute and Multi-Cloud Architecture
In 2026, Trusted Execution Environment (TEE) technologies will finally move from ‘interesting concept’ to real-world game changer. We’re going to see organizations secure memory and hardware in a way that simply wasn’t practical before, which opens the door for decentralized compute in a very big way. Companies will be able to safely split compute across multiple clouds, regional providers, and even on-prem environments, instead of keeping all their workloads under one hyperscaler’s roof. This will bring a level of flexibility and resilience that hasn’t been possible until now.
What is interesting to note here is that the shift isn’t driven by budgets or hype, but by behavior. When you can secure workloads at the hardware level, you’re suddenly free to architect systems around business needs instead of who owns the data center. It unlocks more creative architectures for blockchain, AI, and high-performance computing, and gives organizations confidence that they can spread their risk without compromising security.”
AI Becomes Truly Agentic – Replacing Tasks, Not People – and Drives a New Phase of Cloud Repatriation
“AI is no longer just a tool for optimization. In 2026, agentic AI starts replacing full workflows, and that shift will separate companies that understand how to use AI from those that fight it. The real impact isn’t that AI replaces jobs, but that it replaces the tasks people shouldn’t be doing in the first place – the repetitive, time-sucking operations that drain teams. Organizations that lean into agentic AI will run faster, make decisions earlier, and redirect people into work that actually moves the business.
As AI becomes more embedded in day-to-day operations, more companies will realize that complexity and cost are pushing them away from the hyperscalers. They’re seeing outages, noisy-neighbor issues, unpredictable billing, and environments so complex that one failure cascades through the whole stack. AI workloads, especially GPU-heavy ones, run better, and more cost-effectively, when the infrastructure is simpler, more transparent, and built for their exact workloads. That’s why 2026 will be a major year for cloud repatriation back to regional providers and bare-metal platforms built for performance.”
GPU Optimization and AI-Driven Attacks Will Push Companies Toward Regional Cloud Providers for Security and Stability
“GPU optimization becomes a headline topic in 2026. Today, most companies only use about 60 percent of the GPU power for which they are paying. Next-gen optimization software is going to flip that on its head, giving organizations the ability to squeeze full value out of their infrastructure. That matters not just for cost control, but for AI reliability. When your model performance becomes a competitive advantage, you can’t afford wasted compute, unpredictable throttling, or hardware carved into fractional units you can’t see. This is where optimized IaaS and regional GPU clouds start to shine.
At the same time, attackers are getting smarter, and they’re starting to use AI too. The largest, most complex cloud environments become the biggest targets – when bad actors can spin up their own LLMs. Hyperscalers have hundreds of thousands of tenants, which means hundreds of thousands of potential attack surfaces (and pockets to pick). Regional providers have tighter vetting, cleaner environments, and fewer noisy neighbors. In 2026, security-conscious organizations will realize that the safest place to run AI and high-value workloads often isn’t the biggest cloud, it’s the one that actually keeps out the wrong people.”
Chief Technology Office at Booz Allen
- Solidifying U.S. Telecom Leadership Through AI-RAN (from Chris Christou, 5G/NextG Technologies Leader): AI-RAN will become the cornerstone of AI-native 6G, where networks are continuously learning, adapting, and optimizing in real time, enabling more secure, resilient connectivity for defense, industry, and society at large.
- The Role of Physical AI in Tech Development (from Bill Vass, CTO): As AI becomes embedded across mission systems in 2026, organizations will increasingly need to close the loop between field operations and synthetic training – creating continuous cycles for training, testing, deploying, and monitoring models to ensure they are safe and ready for any battlefield environment or scenario.
- The Agentic AI Evolution (from Sahil Sanghvi, Emerging Tech Leader): Agentic AI marks the next leap in computing, and over the next year, the focus will shift to scaling multi-agent architectures responsibly while building the governance and safety frameworks to manage autonomous decision-making.
Yehuda Niv, CEO of Spines
AI isn’t changing how books are written. It’s changing who gets to publish them. Book publishing in 2026 will be defined by accessibility, where creative voices once sidelined by cost or complexity finally get a seat at the table. AI will democratize opportunity by turning publishing into a truly open, global stage for human creativity.
Jason Hardy, CTO of AI at Hitachi Vantara
- Sovereign AI becomes the defining infrastructure battleground of 2026. Governments worldwide will treat AI capability as critical national infrastructure, similar to energy or telecommunications. This creates a dual economy: hyperscaler-dependent regions versus those building government-sponsored AI platforms for domestic innovation. The gray area between cloud and on-prem will sharpen into regulatory boundaries, fundamentally reshaping where and how enterprises can deploy AI. Hitachi Vantara is uniquely positioned here. Our heritage working with governments and critical infrastructure, plus our focus on hybrid, aligns perfectly with sovereign AI requirements.
- ROI means having a plan, not hope. CIOs face pressure to ‘do AI,’ but successful ones in 2026 will resist the headlong rush and instead focus on specific, measurable outcomes. ROI combines direct cost recovery with learning value: understanding what infrastructure improvements enable AI success while ensuring the AI investment itself becomes self-sustaining. Hitachi Vantara helps CIOs build that plan. We’re not selling a headlong rush into AI. We’re coaching enterprises on establishing clear outcomes, identifying the right workloads, and building infrastructure that actually supports sustainable AI economics.
- Everyone obsesses over GPU costs while ignoring that GPUs represent maybe 40% of total AI infrastructure spending. Storage, networking, power systems, cooling, these ‘peripheral’ costs add up faster than the compute itself. 2026 forces enterprises to think holistically about ‘AI factories’ rather than GPU shopping, creating demand for turnkey, optimized solutions. This is exactly why Hitachi Vantara builds turnkey AI infrastructure. We optimize across all components, balancing GPU, storage, network, and power costs to deliver the best total economics. Buying GPUs in isolation is like buying an engine without the rest of the car.
Nathan Xu, CEO and Founder of Plaud AI.
AI & Productivity
As we head into 2026, what’s the biggest gap between what AI companies are building and what professionals actually need?
The most significant gap is context and taste. Most AI tools excel at processing data, butstruggle to understand human intent or tone, and they lack the nuance and refinement thatmake a significant difference in the experience.
People don’t need another chatbot or dashboard in their day job. They need tools that listen, understand, and amplify what truly matters in their conversations, without adding complexity or risk. At Plaud, we believe in the power of being present, and want to help people focus fully in the moment while our intelligence handles the rest, quietly and responsibly.
Most AI tools focus on what happens on screens: virtual calls, documents, emails. What’s being left behind, and why does that matter for 2026?
What’s being left behind are the real conversations — the impromptu chats, coffee catch-ups, in-person brainstorms, client calls — the spaces where real ingenuity happens. AI has been obsessed with the digital layer of work, but not so much with people. In 2026, I believe the companies that will thrive will be those that bridge human elements across every medium, not just what’s typed or shared online. That’s why we designed Plaud to work in all scenarios: in-person, over the phone, or on screen. It’s how creativity naturally happens.
Future of Work
The return-to-office debate is still raging. How will conversation intelligence tools change the equation in 2026, for better or worse?
Conversation intelligence will make work location irrelevant, but connection essential. Whether you’re in the office, remote, or hybrid, what matters most is how knowledge is utilized. Next year, AI note-taking won’t just transcribe meetings; it’ll connect insights in every scenario across teams and time zones. The future of work isabout how well we listen, understand, and remember together.
Of course, with this comes responsibility, and keeping trust front and center is something I think about a lot. We don’t want this technology to risk becoming surveillance instead of empowerment, and ethics, such as consent, transparency, and security, are core product designs in our Plaud infrastructure.
What’s one workplace trend everyone’s talking about that won’t matter in 2026, and one that no one’s talking about that will?
That’s easy. “AI-native.” It’s a buzzword for many companies right now, but it will becomeirrelevant very soon because each product, industry, and company will be powered by AI, much like the evolution of the internet.
The trend that will matter? The rise of human-AI co-presence- AI that captures and learns from your conversations without interrupting them. Plaud is built around this, ensuring the context and nuance of your discussion are aligned and not just ingested by AI.
Enterprise AI Adoption
Looking at 2026, what industries will be transformed by conversation intelligence, and which ones are overlooking it?
Industries built on trust and robust dialogue, such as law, consulting, healthcare, and education, are already transforming. I see more adoption in SaaS and Finance coming, as knowledge workers and even Sales teams look to distill hours of business discussions into clear, actionable insights without compromising privacy or interrupting their natural train of thought.
What’s the line between helpful and violation of privacy when companies record workplace conversations in 2026? How do businesses navigate this without losing trust?
The line is consent and clarity, not just in legal terms, but in people’s expectations about being listened to. Plaud is built for trust, and we design our tools so users decide what’s recorded, where it lives, and when it’s gone. Nothing is used to train our AI, and no one listens in.
Before capturing a conversation, it’s good practice to give others a quick heads-up. If you’re unsure how to do that or what your local laws require, a quick Google search helps you stay confident.
In 2026, the companies winning with AI won’t be the ones collecting the most data; they’ll be the ones earning the most trust. The future belongs to products that work responsibly, not just innovatively.
Technology & Hardware
Software-only AI tools have dominated. Why will 2026 be different for hardware + AI integration?
Because real intelligence comes from where humans naturally communicate, not just where data is stored. Software is powerful, but it’s reactive. Hardware gives AI senses, the ability to perceive voice, emotion, and context in real time, securely at the edge. By combining the two, we can create experiences that feel effortless and nuanced.
Our philosophy at Plaud is to push the boundary of human-AI collaboration through a combination of natural physical sensors and software. That is why we built Plaud as a portable device—slim, elegant, attached to the back of your phone or worn on your body—to act as the most natural sensor for your offline context.
Workplace Intelligence
If you could fix one broken thing about how professionals work today before 2026, what would it be?
The fragmentation of conversation.
A significant portion of our work is conducted through conversation. Conversation is the most human, powerful form of expression.
It seems like every idea, every expression, and every insight is scattered across documents, calls, emails, and apps. Too often, the intelligence sparked in our minds never gets captured, understood, or turned into action.
If we can unify things, we can unlock the real potential of human-AI partnership. That’s what Plaud exists to do: to amplify natural human intelligence and bring together conversations and insights across every medium.
Five years from now, what will we look back on from 2025 and think was obviously wrongabout AI?
“General-purpose AI (one model to rule all tasks) is the near-future.”
The real value turns out to be in verticalised, fine-tuned, workflow-embedded models, or human-AI systems tightly stitched into domain & process. 2025’s AI race was about scale: more models, faster launches, more tokens, more noise.
However, the future belongs to more specialized AI that understands context and tightly integrates into industry workflows and knowledge.
The next great leap in AI won’t be artificial at all, I think it’ll be deeply, fundamentally human.
