Many clients have been asking us what our expectations are for artificial intelligence (AI) and how it will ultimately shape the economy, capital markets, and the overall job landscape. While it’s still early to draw firm conclusions, the signs increasingly point to the beginning of a meaningful societal shift—one that appears to be unfolding at a pace and scale that could surpass the adoption speed of any prior technology cycle. If the first few months of this year are any indication, 2026 could mark the beginning of a more accelerated phase of this development.

What makes this powerful transition somewhat easy to overlook at the moment is that it is occurring alongside a period of elevated geopolitical uncertainty. Much of the market’s recent attention has been focused on issues in Iran and the potential implications for global energy and commodity supply. While these risks are important, they have also diverted attention from the dominant secular trend of our time: the steady and rapid advancement of AI. Beneath the headlines, this technology is already beginning to influence productivity, corporate decision-making, and capital allocation in tangible ways.

To frame the current moment, it is helpful to consider how quickly AI has entered the mainstream. When OpenAI introduced ChatGPT just over three years ago, it marked a clear inflection point. The platform reached approximately 100 million users within just two months—the fastest adoption of any consumer application to date. In response, large-scale technology companies including Alphabet, Microsoft, Amazon, Meta, and Oracle committed substantial capital toward building AI infrastructure. By the end of this year, their combined investment over this four-year period is expected to approach $1.5 trillion.

For perspective, the buildout of the internet’s infrastructure in the late 1990s required roughly $500 billion of telecom investment. While significant at the time, today’s AI investment cycle is both larger in magnitude and unfolding at a much faster pace.

Not surprisingly, this rapid acceleration led many investors last year to seriously question whether AI enthusiasm had moved ahead of fundamentals. While adoption was in its initial stages, measurable productivity gains and clear economic returns were slower to materialize than initially anticipated.

That narrative began to shift earlier this year with the emergence of what is often referred to as an “agent layer” (for example, systems like OpenClaw). Unlike earlier AI tools that primarily respond to prompts, these systems are designed to take action on a user’s behalf.

In practical terms, this could mean beginning the day with a fully organized agenda. An agent reviews your inbox, prioritizes key messages, schedules meetings, and prepares summaries with suggested responses. For individuals managing overflowing email inboxes and high volumes of information, this has the potential to meaningfully enhance productivity and free up time for higher-value activities. In these cases, AI may augment rather than replace roles. However, positions that are highly repetitive or process-driven could face increasing pressure. In a recent projection, Goldman Sachs estimated that about 7% of jobs could be replaced by AI, while 63% could be augmented and the remaining 30% would probably not be impacted. There will also be new jobs created, but given the rapid pace of change, it may take time for skills to catch up with need. From today’s vantage point, individuals and organizations that understand how to effectively leverage AI tools are likely to have a distinct advantage.

Financial markets have already begun to reflect these dynamics. Public software companies experienced notable volatility as investors reassessed the durability of certain business models. The possibility that some functionality could be replicated at lower cost has led to greater scrutiny of valuation multiples, particularly for companies reliant on recurring revenue streams. While we believe many established software providers will remain relevant, the competitive landscape is clearly evolving. Companies will generally need to adapt—both in terms of product offering and pricing—and those engaged in more commoditized solutions may be at greater risk.

Another important consideration is that many of the most influential AI companies remain private, limiting visibility into the true pace of adoption and underlying economics. That could change if companies such as SpaceX, Anthropic, or OpenAI ultimately enter the public markets. Given their scale, any such listings could meaningfully expand the investable universe and provide valuable insight into how AI is being implemented and monetized.

One recent data point is particularly notable: in early April, Anthropic reported that its revenue run rate had reached $30 billion, up from $9 billion at the end of last year. This level of growth suggests that enterprises are moving beyond experimentation and are increasingly willing to allocate meaningful dollars toward AI solutions. In our view, this reinforces the idea that AI is transitioning from a thematic concept to a tangible driver of economic activity.

It is also worth noting that Anthropic has recently introduced its latest model, Mythos, on a highly limited basis. In early testing, the company reported that the model identified thousands of high-severity cybersecurity vulnerabilities. Notably, these capabilities were not explicitly pursued or designed but rather appear to have emerged as the model’s overall reasoning and problem-solving abilities advanced. Given the sensitivity of these capabilities, Anthropic has restricted access to a small group of partner organizations using the model internally to identify and address vulnerabilities. This cautious approach reflects a broader concern: that if deployed more widely, similar tools could potentially be misused by bad actors to discover and exploit weaknesses at scale. In our view, this example highlights one of the core dilemmas in AI: tools like Mythos can meaningfully strengthen an organization’s defenses, but if misused, could also create significant risks.

Given the pace of development so far this year, 2026 may be the first period in which AI begins to have a measurable impact on productivity. While there are still many open questions—and, as recent developments highlight, meaningful advances are often accompanied by equally difficult concerns—we expect the near-term effects to be constructive. Increased efficiency, improved decision-making, and broader adoption across industries have the potential to support both economic growth and corporate earnings in a more tangible way.