Trends 2025

Zero Trust is a cybersecurity strategy that assumes that users and devices should not be trusted by default, whereas traditional security approaches assume that anything inside the corporate network can be trusted. The reality is that this assumption no longer holds true. The zero-trust model of security prompts you to question your assumptions of trust at every step. This approach is now befitting of society as a significant decline in public trust towards institutions is occurring both in the U.S. and abroad.

Technology can be both an instigator and a solution. We have written about this erosion of trust in previous trends (Fake News, Trends 2017; DeTrust, Trends 2022) that undermine truth and question identity. The ability of AI to generate realistic fake content (Deep Fakes, Trends 2019) only exacerbates concerns about the spread of misinformation and manipulation of public opinion. However, we have also ruminated on the promises of blockchain technology (Blockchain, Trends 2016, 2018) which requires building a trust layer above trustless technology. Crypto/blockchain technology is seen by many as essential in realizing the productivity gains of decentralized autonomy (Web3, Trends 2022) and Agentic AI (Trends 2025), but this too is at risk from advances in Quantum Computing (Trends 2018), which could compromise its underlying encryption protections.

As Generative AI and other technologies continue to become more deeply ingrained in our society, there is a growing concern about our ability to separate fact from fiction. A lack of trust in institutions can lead to societal divisions and polarization, especially when AI is used in sensitive areas like politics and healthcare. The stakes are high. Low-trust societies devolve to tribalism (smaller “circles of trust”) and are associated with impaired economies, higher crime and corruption, and ill-defined norms. High-trust societies tend to have stronger democracies, richer economies, better health, and fewer social ills.

Despite the hope of an Energy Transition (Trends 2024), consumption of all sources of energy is increasing. Whether it be coal, oil, natural gas or biomass (wood), though the share of these fossil fuels as a percentage of energy consumed is going down, the actual amount is increasing. In 2024, the United States reduced coal usage by 55 million tons versus 2023 - and the United Kingdom went to basically zero with the closure of its last coal-fired power plant in October; yet, according to the Energy Institute’s Statistical Review of World Energy, global production of coal exceeded 9 billion tons, the highest ever recorded and more than double than at the start of the millennium. Why? China’s coal production increased 83 million tons during the first 11 months of this year while India’s grew by 66 million tons. The offsets of reduced fossil fuel usage in the (mainly) developed West (U.S., U.K., Germany, etc.) has been more than offset by the increasing needs in emerging markets, particularly in Asia.

Carbon-neutral forms of energy such as nuclear, wind and solar are indeed growing at a faster pace; that said, for the foreseeable future, the world will use these newer sources of energy to add to – rather than replace – existing sources. Case in point, wood, as measured by percent of energy consumed, is at all-time lows, though global energy production from wood reached an all-time high in 2022, more than what was produced in either 1800 or 1900. With the advent of Artificial Intelligence and other powerhungry technologies, societies that fail to accept this fact do so at their own peril, for not only is affordable energy essential to progress, it is also essential to national security.

Agentic AI represents a significant evolution in artificial intelligence, moving beyond passive information processing to proactive, autonomous action. In previous years we’ve highlighted how AI has continued to evolve from machine learning to Generative AI to multimodal AI (see CAIS Trend 2020, ChatGPT Trend 2023, Multimodal Trend 2024). Unlike traditional Generative AI models, which typically require human prompts to initiate tasks, Agentic AI systems possess builtin decision-making frameworks. This allows them to independently plan, strategize, and execute actions within predefined constraints. This trend has profound implications across various sectors, transforming how businesses, governments, and individuals interact with technology (Productivity Boom, Trends 2024).

The emergence of autonomous agents is fueling the rise of “digital labor,” providing an always-on, 24/7 workforce. This trend is exemplified by Salesforce, where over 80% of customer service interactions are now resolved through their Agentforce platform. Similar productivity gains are being observed among Salesforce customers who are deploying agents at scale within their own organizations.

While currently confined to the digital realm, Agentic AI is increasingly manifesting in physical forms through robotics. As reasoning and decision-making capabilities within these systems advance, robots will embody the proactive, autonomous nature of Agentic AI. These systems will dynamically adapt their actions based on real-time feedback and learn from past experiences to optimize future performance.

The benefits of Agentic AI are substantial, offering increased efficiency, scalability, and enhanced problem-solving capabilities. However, the deployment of such systems also necessitates careful consideration of ethical and operational challenges, issues related to accountability, trust (Zero Trust, Trends 2025), and the potential for unintended consequences. Despite these challenges, the transformative potential of Agentic AI in redefining operational paradigms and enhancing decision-making at scale, positions it as one of the most impactful trends in the AI landscape.

Birth rates in the U.S. have been declining to below replacement rate and are at their lowest level in over 40 years. Contributing factors stem from trends we have written about in the past (Sex Recession, Trends 2019 and Family 2.0, Trends 2016), involving cultural/societal, scientific/technological and economical dynamics.

The cultural changes include an evolving definition of a nuclear family. Both men and women are marrying late as they place a higher importance on their careers. In some cases, they don’t see the need to get married and have babies; they can be married and just have pets as opposed to children or not get married at all. Gender roles are also changing; women used to be primarily homemakers, which led to formation of traditional families. As more women pursue professional careers and independence, they see less need to compromise and get married if they don’t find the right partner. There are burgeoning feminist movements such as 4B going on in South Korea that would seem to contribute to this trend of lower birth rates. 4B is an absolutist feminism movement wherein women adhere to the “4 Nos” (Korean language term “bi” for “no”). These are do not date men, do not marry men, do not have sex with men, and do not have children with men. Not coincidentally, South Korea has the lowest fertility/birth rates globally at 0.7 births per woman in 2023; almost one third of the replacement rate of 2.1 births per woman needed to maintain a stable population.

Scientific reasons for lower birth rates are lower sperm counts and lower fertility rates. Contributing factors include obesity and other lifestyle factors such as excessive alcohol consumption and unhealthy diets. In addition, certain medications are now being studied to see if their use is impacting fertility rates. Further, overexposure to more processed foods as well as environmental toxins (such as plastics) are also believed to be impacting fertility. Now more than ever, environmental toxins (generally called endocrine disrupters) are present in everything from plastic packaging and toys to sofa covers and cosmetics. These endocrine disrupters mimic or block the hormones responsible for many of the body’s essential functions, including reproduction. Economics, especially in the U.S., can also be a limiting factor. More young couples see having a baby or babies as a luxury that only the affluent can entertain. Childcare costs alone remain an almost prohibitive expense for most Americans. According to a U.S. Department of Labor study, U.S. families can spend up to 16% of their median annual income on full-day care for just one child. Further, affordability related to housing and other amenities is beyond the middle-class to some extent. To take advantage of economies of scale, there are also examples of multi-generational households, where families are living together and saving money by sharing expenses, but at the same time making it more challenging to have babies.

The Trump administration’s formation of the Department of Government Efficiency (DOGE) has the stated objective of reducing government spending and burdensome regulation – a frustration of many Americans. President Trump has tasked Elon Musk to lead the effort working with/under the White House Office of Management and Budget. Cutting government spending has been very difficult historically given the political constraints and promises elected officials make on the campaign trail. Unlike the private sector, there are obstacles to head-count reductions such as civil-service protections to consider. That said, long-term changes are needed to address the U.S. federal deficit. Annual deficit spending is approaching $2 trillion, and the two most obvious ways to reduce this number would be to increase tax receipts or cut spending. The success of either approach could be very important to improve the country’s fiscal position, and through announcing the DOGE initiative, President Trump is signaling his clear preference to lead with spending reductions. However, the political will to address non-discretionary spending (Medicare/Medicaid, Social Security, and interest on federal debt) is likely a requirement for real progress, as nondiscretionary spend is trending toward two-thirds of all outlays. One thing we know is that Trump and Musk enjoy challenging the status quo, and therefore, investors should pay attention to any progress on this effort.

Economic warfare continues to dominate global policy, with tariffs, sanctions, and regulatory actions shaping the economic and geopolitical landscape. The Trump administration’s aggressive trade policies, including a proposed 25% tariff on imports from Mexico and Canada, aim to address trade deficits, illegal immigration, and drug trafficking. These tariffs threaten deeply integrated North American supply chains, particularly in sectors like automotive manufacturing, where components often cross borders multiple times. Such disruptions could drive up production costs, ultimately increasing prices for consumers and straining U.S. manufacturing competitiveness. The tariffs on China, covering $370 billion worth of goods at their peak, sought to counter intellectual property theft and unfair trade practices. While these measures were intended to bolster U.S. industries, they triggered retaliatory tariffs from China, disrupted global supply chains, and contributed to rising inflation domestically.

Meanwhile, the European Union has taken a leading role in challenging the dominance of multinational tech corporations, particularly through its high-profile antitrust cases against Google. These actions reflect the EU’s broader strategy to curtail monopolistic behaviors and protect competition within its markets. Adding another layer to this dynamic is the push for a global minimum tax, spearheaded by the Organisation for Economic Co-operation and Development (OECD) and supported by the U.S. and EU. Designed to curb tax avoidance by multinational corporations, this policy represents a coordinated effort to level the playing field and ensure that corporations pay their fair share regardless of jurisdiction. These developments highlight the increasing role of economic policies in reshaping industries and markets, demanding that businesses adapt to a landscape of growing regulation and geopolitical tension.

The socio-economic climate today reflects a “vibecession,” a pervasive sense of public dissatisfaction with institutions and the establishment despite strong economic indicators such as low unemployment. This discontent stems from rising inflation, income inequality, and a growing disconnect between policymakers and the realities faced by everyday citizens. While traditional economic metrics suggest stability, the prevailing sentiment is one of frustration and distrust in established systems.

The recent targeted killing of United Healthcare CEO Brian Thompson has brought these antiestablishment sentiments into sharp focus. While many mourn Thompson’s death, others view the act as a grim expression of growing anger against perceived exploitation by large corporations, particularly in the healthcare sector. This tragic event underscores the widening chasm between corporations and the public they serve. The healthcare industry, often criticized for high costs and lack of transparency, has become a symbol of systemic inequities, intensifying public dissatisfaction. Social media amplifies these grievances, providing a platform for anger and solidarity that can further escalate tensions. The incident also highlights the dangerous consequences of leaving these sentiments unaddressed, as extreme actions can emerge from unresolved systemic frustrations. The vibecession reflects a broader cultural and economic shift, where traditional institutions are increasingly seen as failing to address structural inequalities. Leaders are under mounting pressure to not only acknowledge these grievances but to act decisively in implementing reforms that rebuild public trust.

For the last 18 months, housing affordability for first time homebuyers has been worse than at any time before in the United States, according to an index from the National Association of Realtors. While the homeownership rate is above the 60-year average, many experts still say the U.S. is short somewhere between 3.8 million and 6.8 million homes, according to Vox. This has given rise to the Yes In My Backyard, or YIMBY, movement whose aim is to encourage and incentivize more housing mostly through the reduction or elimination of zoning barriers. Another key aspect of the movement is to provide representation for a wider demographic of people, many of whom are not homeowners, at local planning meetings, a forum often dominated by homeowners. More scalable solutions are also being explored including litigation and pushing certain zoning decisions to a state level from a local level. Via their persistence, the YIMBY movement has given political cover to local officials to loosen or rewrite local zoning laws in many jurisdictions across the country. This has led to many creative solutions such as Accessory Dwelling Units (ADUs) and repurposing older commercial buildings such as malls and their anchor stores as condos and apartments.

2024 will be considered the year that Autonomous Vehicle (AV) services finally found traction in the U.S. economy. Companies such as Tesla have promised some level of Full Self Driving (FSD) since as early as 2018; and even though Tesla still failed to completely deliver what they promised even in 2024, they did achieve some progress, and companies such as Waymo (a subsidiary of The Alphabet Company) made considerable progress. We shared Elon’s enthusiasm in 2018 for FSD as we wrote about it in a piece wherein we espoused how the trend of a Passenger Economy (Trends 2018) will have profound implications for industries such as auto, insurance, technology. We were a bit early, but we finally saw that happen in 2024.

There are two flavors of Robotaxis, or AV, that are coming to the fore. One is a driverless fleet service that companies such as Waymo are offering, featuring vehicles armed with an expansive sensor suite with 29 cameras (in case of Waymo), LiDAR’s, and radar that take passengers around. LiDAR, in particular, adds depth perception that a camera lacks. Currently, Waymo is the only viable option, as Cruise was initially banned in San Francisco due to an accident. Ultimately, their funding was pulled by General Motors. There are others such as Zoox by Amazon, but they are notably behind. Waymo’s vehicles are estimated to cost anywhere from $100k to $200k currently, but those costs are expected to come down over time. Waymo operates primarily in the cities of San Francisco and Phoenix currently, and will be launching in Austin and Atlanta in partnership with Uber. This is a Level 4 AV approach in which the car is truly driverless, but is expected to be driven under limited conditions and not to be operated unless all conditions are met. The second approach is a FSD one by Tesla in which currently the driver is passive and not engaged with the steering wheel but must be available to drive when the system requests. This is currently a Level 3 FSD approach. Tesla is not using expensive hardware such as LiDAR or radar,but is relying mainly on cameras and a machine learning/Artificial Intelligence model to ultimately achieve Level 5 FSD which involves a driverless car and ability to drive. This is a much cheaper option as Tesla cars are supposed to cost only $30,000-$40,000 and have a large installed base of 5-6 million cars. But it carries considerable risk as relying on cameras alone has proven to cause quite a few accidents and Tesla’s disengagement rates are an order of magnitude inferior to their peers. Ultimately, Tesla plans to allow all existing Tesla owners to have their cars participate in the AV fleet and earn passive income.

Going forward there will be lots of puts and takes as systems scale and we get more data. Waymo has already taken 22% of share in the city of San Francisco (per Yipit data) after starting from close to zero a year back. That is identical to the share of Lyft, trailing Uber at 55%. For Robotaxis, hardware and operating costs will eventually come down but profitability will be sensitive to utilization rates and minimizing dead time, which again has to be balanced by the fleet size to match peak loads. There are also considerations of what kind of business model Waymo might employ; will they partner with the likes of Uber and just license their software or take on the cost of building and maintaining the infrastructure and fleet operating costs? Over time, as the costs of the AV system goes down with scale, the costs of the ride sharing networks are expected to go up due to labor as well as insurance costs rising. In terms of Tesla, if they succeed with their camera-only and AI-model approach and improve their safety data, they could become the low-cost provider and end up with the majority of market share. Regulation will be an area that they will need to overcome, as providers such as Waymo have been working with regulators for north of 10 years. It will be interesting to see how the Robotaxi industry and eco-system evolves over the next 5 years.