#459: Tariff Turmoil: Will It Mark The End Of A Three-Year Rolling Recession?, & More

While many observers fear that the Trump tariff policy is a recipe for economic and geopolitical disaster, we believe that what looked at first glance like the largest and most regressive tax increase in US history could turn out to be quite the opposite. Now that President Trump has asked Treasury Secretary Bessent to take the lead from Peter Navarro and Howard Lutnick in negotiating with our allies, what once seemed like a chaotic situation based on incomprehensible “reciprocity” calculations could have been a setup—premeditated or otherwise—for serious negotiations that will lead to lower tariffs and non-tariff barriers, neither of which would have been possible without the shock therapy that President Trump administered. Still influential in the Trump Administration, Elon Musk has been a strong advocate for this solution to the tariff and non-tariff trade barriers that have evolved over the last 50 years.

During the past week of extreme volatility in the stock and bond markets, our working assumption was that President Trump has been aiming for robust economic growth and a strong stock market during the second half of this year into 2026 ahead of the midterm elections. How many times this week did he mention that the economy and stock market are going to boom!

Even before the tariff controversy, we had been expecting strong growth to begin sometime in the second half, because we do believe that the last leg of a three-year rolling recession will result in negative Gross Domestic Product (GDP) growth for the first and second quarters. During the past three years, as one cohort of the economy after another capitulated to the interest rate shock that started in 2022, high-end consumers and the government propped GDP up. Now, both are giving way, with the government entering its first recession in 30 years. As a result, the Administration and the Federal Reserve will have more degrees of freedom to stimulate than most investors have been expecting. Now that much of the economy has seized up in response to the fear of tariffs, the drop in activity is likely to be more severe than otherwise would have been the case, a clarion call for tax cuts, deregulation, and lower interest rates. 

Last week, the U.S. Food and Drug Administration (FDA) announced a groundbreaking plan to phase out its longstanding requirement for animal testing in the development of monoclonal antibodies and other drugs.¹ Instead, the FDA will encourage the use of “more effective, human-relevant” testing methods, including AI models and lab-grown human cell systems, to evaluate drug safety and efficacy.

Since 1938, Federal law has mandated animal trials in drug development.² Now, rapid advancements in AI, multi-omics, and microphysiological systems have made possible—and preferable—the pursuit of more predictive pathways, as shown below in Gen 2 and Gen 3 methods. The FDA’s initiative should lower the time and money necessary to develop efficacious drugs for human trials, with positive implications for AI-forward biotech companies and the patients they seek to serve. 

Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Bunne et al. 2024 and Rood et al. 2024³ as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.

The FDA’s initiative represents a transformative opportunity for biotech firms already leveraging AI in drug discovery and development. They are already using these sorts of techniques in earlier stages of the pre-clinical process and might be able to use some of those data to fulfill the new requirements. In addition, because these firms have streamlined preclinical processes, the removal of mandatory animal testing should accelerate their progression to clinical trials relative to conventional drug developers, as shown below.

Note: These calculations anticipate that AI drug development efforts can reduce target-to-lead, hit-to-lead and lead optimization stages of development to less than 18 months in total, consistent with company commentary. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, as of April 11, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Forecasts are inherently limited and cannot be relied upon.

The FDA’s directive signals its commitment to innovative, predictive drug development methodologies. By reducing reliance on lengthy and costly animal studies, AI-powered and human-relevant approaches could lower drug development costs, accelerate time-to-market, and increase therapeutic effectiveness. ARK’s analysis, shown below, suggests that AI-developed therapeutics could boost returns on investment to more than 20%, a 5x improvement relative to the pharmaceutical industry's current average research and development (R&D) returns of ~4%. Ultimately, the integration of AI throughout the drug pipeline process could reduce overall development costs by four-fold.⁴

Note: 10% discount rate. *Companies pursuing this strategy for drug development, though this does not imply that every asset the company has at that particular stage is worth that amount or that company return on R&D dollars will meet the modeled value. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.

While AI-centric drug developers like AbSci, AbCellera, and Recursion Pharmaceuticals stand to benefit most immediately, the policy's implications extend into the multiomics tools and programmatic biological space as well. Companies providing sequencing, single-cell molecular analysis, spatial transcriptomics, and other advanced multiomics tools should benefit as biopharma firms adapt workflows to accelerate and enhance AI-driven drug discovery. 

The benefits should extend beyond biotech to the healthcare system at large. Currently, only ~5% of rare diseases can be treated by approved therapies.⁵ The reason is clear: high costs focus drug developers only on markets with large addressable patient populations. As AI increases the efficacy of drug development, a higher percentage of rare diseases should become addressable. ARK estimates that AI could reduce the time to market by 50% and the odds of failure by another 50% at each stage of the development process, potentially enabling biotech companies to address 75% of the rare disease patient population profitably. As a result, the quality of patient lives should increase as the need for costly hospitalizations and outpatient healthcare declines. 

Note: The percent of rare disease patients addressable assumes a required 15% return on R&D spend. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of April 13, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

In summary, the FDA's announcement this week heralds not only a regulatory shift but also the dawn of a new era for molecular medicine, one characterized by faster, more cost-effective, and more accurate drug development benefiting innovators, patients, and the healthcare system.

Last week, at a Walmart Supercenter in Dallas, Zipline launched a drone delivery service⁶ that will transport 65,000 items to households in under two minutes. ARK’s research estimates that drones could slash delivery costs by 94%, from $5.40 to $0.35 per trip, compared to traditional delivery services during the next five to ten years. For more on our research on autonomous delivery, check out ARK’s Big Ideas 2025.⁷  

Source: Zipline 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

Amazon is no longer just a marketplace. Its new "Buy for Me" button⁸ should transform its service into an AI-powered purchasing agent functioning as a lead generator across not only its own inventory but also third-party websites, completing transactions with its one-click checkout. With this capability, Amazon is integrating product discovery and checkout, collapsing the e-commerce funnel into a single interaction. 

This breakthrough extends the historical arc of marketplace innovation that has lowered the friction associated with shopping, as shown below. In early economies, bartering gave way to a standardized currency that streamlined the exchange of value. Department stores introduced fixed pricing and centralized inventory control, eliminating negotiation. The internet removed geographic limitations, and then mobile platforms introduced real-time, on-demand commerce. Each wave of innovation compressed the path to purchase by making it faster and easier to complete transactions. 

Note: Data points from earlier eras are estimates based on anecdotal accounts. Source: ARK Investment Management LLC, 2025, based on a range of data as of April 11, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.

AI is ushering in the next phase: marketplaces as autonomous agents. Instead of facilitating transactions, platforms like Amazon are beginning to take over the entire purchasing process: consumer participation is minimal, their intent inferred, their preferences learned, and their transactions automatic. Marketplaces are shifting from reactive to proactive systems.  

As AI advances, marketplaces could evolve further into dynamic interfaces that orchestrate supply, optimize price, and personalize fulfillment—all without direct user input. This new phase of marketplaces will be defined by how seamlessly they blend intelligence with infrastructure. Marketplaces using AI to optimize price, inventory, and speed are likely to shape the future of economic activity.