The AI transformation is just beginning
The 2026 RBC Capital Markets Technology Private Company Conference brought together over 30 disruptive private tech companies to discuss the future of technology in the AI era. The overarching theme emerging from these conversations challenges conventional thinking: the current AI build-out is not a typical technology cycle measured in quarters or years, but rather a multi-decade transformation potentially lasting 20 years, fundamentally reshaping how investors should evaluate AI-exposed companies and their competitive positioning.
The bottleneck to AI adoption is not technological capability, but enterprise readiness. Organizations are still in early stages of getting data estates in order, building governance frameworks needed to deploy agents at scale, and rearchitecting workflows around AI-native processes. Simultaneously, they face the harder challenge of training workforces to adopt an AI-first mentality. The compounding nature of these dynamics means each successive wave of adoption — from individuals using AI tools, to teams building AI-powered workflows, to fully autonomous software organizations — builds on the last. While near-term valuation debates will dominate headlines, the more important question is positioning for a prolonged infrastructure and software spending cycle that remains closer to the beginning than the end.
"While near-term valuation debates will dominate headlines, the more important question is positioning for a prolonged infrastructure and software spending cycle that is still closer to the beginning than the end."
Matthew Hedberg, Head of Global TIMT Research, RBC Capital Markets
Data as the defining competitive moat
As AI capabilities continue to evolve but eventually commoditize over time, a consistent message emerged: data with context will be a sustainable competitive advantage. Management teams emphasized that the context, quality, breadth, and governance of a vendor or customer's data estate will be the primary source of durable competitive differentiation. Several executives were explicit that organizations unable to get their data estates in order will be structurally unable to leverage AI effectively.
Strong ARR growth and NRR will directly reflect how deeply embedded data platforms become in customer operations. High-volume data processors increasingly view their data graphs and life cycles as core differentiators, while real-time analytics platforms are positioning themselves as back-end infrastructure for next-generation companies. One founder added important nuance: code was never the true moat, and as AI-generated code becomes ubiquitous, the companies that will win are those innovating quickly and staying on the frontier as true partners to customers through the AI transition.
Many frontier AI models lack proprietary data context, creating potential value for tech vendors that can partner with AI labs for data access. For investors, platforms with proprietary, high-volume, and well-governed data assets with context are best positioned to widen leads and fend off competition. Conversely, those that cannot get their data estates in order will find themselves structurally disadvantaged in a world where data is the primary differentiator.
The engineering organization reimagined
Perhaps the most striking data point from the conference concerned productivity gains from autonomous software development. One company reported shipping 7-8x more code compared to seven months prior with only 10% additional engineering headcount, and anticipates that within five years, 99.9% of code will be AI-generated. Broader panel discussions echoed this trajectory, with companies building toward self-healing software, autonomous IT delivery, and agent-based deployment becoming mainstream.
What is emerging is a vision of the engineering organization that looks fundamentally different from today. AI writes code and alerts humans only when assistance is needed, while autonomous agents handle the bulk of vulnerability patching and infrastructure deployment. The value of a human engineer shifts from lines of code produced to the quality of judgment applied to increasingly complex decisions. However, executives emphasized that speed alone is insufficient. The full development lifecycle (including code review, monitoring, deployment testing, and security) must each accelerate in parallel for organizations to realize full productivity benefits. Enterprises that build secure and adaptive harnesses around their models will move far faster than those constrained by point solutions or legacy tooling.
For investors evaluating tech companies, the question is no longer whether AI will change how software is built, but whether management teams have the vision, tooling, and organizational culture to execute that transition at the pace markets are demanding. Those that hesitate will find the gap between leaders and laggards compounding quickly in the emerging K-shaped economy.
"Organizations that invest proactively in reskilling, that redefine job functions around oversight and judgment rather than execution, and that build cultures capable of continuous adaptation will have a structural advantage."
Matthew Hedberg, Head of Global TIMT Research, RBC Capital Markets
Platform consolidation meets point-solution innovation
A recurring tension emerged between point solutions and platforms, with every participant expressing ambitions toward platform consolidation. Near-term reality suggests point solutions will continue finding fertile ground, particularly in fast-moving areas like security, data, and infrastructure where innovation speed and problem focus allow new entrants to mature quickly. However, the longer-term gravitational pull is toward consolidation as enterprises grow fatigued by managing dozens of specialized vendors and increasingly demand integrated solutions.
Legacy platforms have an opportunity to consolidate spend, but face structural disadvantages due to slower innovation cycles. New capabilities tend to get built around legacy platforms rather than by them, creating openings for next-generation platform challengers. Executives highlighted the ability to offer access to all AI frontier models natively, provide secure and broad data estates, and enable rapid agent development within a single architecture as meaningful differentiators. For investors, this consolidation dynamic cuts both ways: it creates significant long-term opportunity for platforms with the right architecture and distribution to expand revenue per customer, while simultaneously compressing exit and standalone growth prospects for point solutions without credible paths to platform status.
Security and legacy software: Urgent transformation required
The security landscape is being fundamentally redrawn by agentic AI deployment. Executives framed emerging threats as urgent and present challenges, not future considerations. Over 50% of internet traffic is now machine-generated, and the majority of databases are being deployed by agents rather than humans. The security infrastructure built for human actors is already obsolete. The internet has been redesigned around machine behavior without corresponding redesign of governance controls, and this gap is widening rapidly.
"Several executives noted that agents are already here and already operating at scale within enterprises, while the broader market continues to treat agents as an emerging concept."
Matthew Hedberg, Head of Global TIMT Research, RBC Capital Markets
Enterprise security priorities are converging on three critical needs: knowing what agents are deployed within organizations, understanding what access and permissions those agents hold, and establishing accountability frameworks when agents malfunction. Identity security is evolving accordingly. The relevant question is no longer whether an identity is human or machine, but whether behavior is consistent with granted rights. Adding urgency to the defender side, open-weight models are lowering attack costs, shifting asymmetry between offense and defense in ways requiring defenders to spend more, move faster, and innovate continuously.
Despite disruption narratives, legacy enterprise software remains more resilient than bear cases suggest. Real ERP systems with decades of embedded trust, auditability, and governance are not replaceable by startups overnight. However, companies clinging to traditional models without meaningful adaptation will struggle. Seat-based pricing models are under accelerating attack as frontier model adoption grows and customers demand flexible, outcome-oriented commercial structures.
The forward-looking view is bifurcation: established platforms authentically embedding AI natively, rearchitecting pricing toward consumption and outcomes, and evolving security proactively will extend relevance. Those treating AI as feature bolt-ons or marketing narratives rather than genuine business model transformations will increasingly face displacement.
Internet and defense tech: complementary growth stories
In the internet sector, an emerging economic principle is reshaping cost structures: tokens are cheaper than people. AI Proptech companies reported accelerating growth with virtually no headcount expansion, while back-office outsourcing costs fell 40% in 12 months. This dynamic is forming a core CFO consideration: token cost per employee and revenue as key determinants for company headcount levels.
Additionally, companies are maintaining LLM vendor optionality. One panelist noted optimizing from 13 different LLMs down to 5, all swappable same-day depending on content, performance, and budget requirements, suggesting LLMs possess relative lack of pricing power given easy interchangeability.
For defense technology, global conflicts are supporting increasing demand for disruptive innovations. Broad consensus among defense tech firms indicates the outlook for AI, unmanned systems, and emerging technologies remains robust. Lessons from recent conflicts in Ukraine and Iran highlight the importance of new technologies, particularly unmanned systems across all domains. The Department of Defense is working to accelerate adoption pace and expand the defense industrial base by attracting non-traditional investors. While contract timing and appropriations have experienced delays, the pace of change at the DoD will continue accelerating, with ability to attract non-traditional investors remaining a core leadership focus.
Matthew Hedberg authored "2026 RBC Private Tech Conference Takeaways," published May 14, 2026. For more information on the full report, please contact your RBC Capital Markets representative.
