Merck & Co is deepening its bet on artificial intelligence with a long-term partnership with Google Cloud, committing up to $1 billion over several years to expand its AI infrastructure, talent base, and access to advanced models, including the Gemini Enterprise platform.

The agreement, unveiled at Google Cloud Next, formalizes an existing collaboration and signals a shift from experimentation to large-scale deployment across the pharmaceutical value chain. Executives from both companies said the partnership will embed Google engineers within Merck’s operations, effectively integrating external AI capabilities into core research, regulatory, and commercial workflows.

“I easily see us investing a billion over the next several years in this, in those capabilities,” said Dave Williams. “We’re not just buying tokens. It is really the tool set” Google Cloud offers, he added, pointing to the combination of Gemini Enterprise, engineering support, and underlying infrastructure.

The scope of the collaboration is unusually broad. Rather than focusing on a single use case, the companies plan to deploy AI across drug discovery, clinical development, regulatory submissions, manufacturing optimization, and commercial operations. That breadth reflects a growing industry view that AI’s impact in pharmaceuticals will be cumulative, driven by incremental efficiency gains at each stage rather than a single breakthrough application.

Williams indicated the partnership is likely to run for at least a decade, underscoring both the scale of the investment and the complexity of integrating AI into highly regulated scientific workflows. The absence of a fixed timeline suggests the companies are prioritizing iterative deployment over rigid milestones, a common approach in large-scale digital transformation programmes.

The deal strengthens Google Cloud’s position in a high-value vertical where data intensity, regulatory complexity, and compute demands create barriers to entry. Alphabet Inc. has been pushing aggressively to position its AI stack, particularly Gemini Enterprise, as enterprise-grade infrastructure capable of handling sensitive workloads in sectors such as healthcare and finance.

“We’ve always said we wanted AI to play a positive role in society. One of the ways is to help people find cures to illnesses,” said Thomas Kurian. “They have the domain knowledge. We’re bringing the AI tools and platform and cyber capability to help them build using these tools.”

At the operational level, Merck is targeting measurable efficiency gains. Williams said AI will be used to run computerized simulations of laboratory experiments, reducing reliance on time-consuming physical trials in early-stage research. The company also plans to expand its use of AI in regulatory processes, an area where documentation requirements are extensive and often repetitive.

Merck has already been applying AI to prepare sections of clinical study reports for about two years, with what Williams described as positive results. The next phase involves scaling that capability across a wider set of documents and jurisdictions, turning isolated productivity gains into system-wide efficiencies.

“We feel there’s a tremendous opportunity there, and it’s a huge information challenge,” he said.

One of the more immediate impacts has been on market access. Williams said the company has used Google’s technology to cut by half the time and cost required to compile regulatory dossiers — detailed submissions needed in many countries to secure reimbursement approval for new medicines.

“This isn’t a pilot,” he said. “We’re submitting dossiers in markets using this new capability, and we’re now scaling it globally.”

That transition from pilot to production is remarkable. Many pharmaceutical companies have experimented with AI in limited settings, but fewer have moved to full-scale deployment tied to core business outcomes such as regulatory approval timelines and revenue realization.

The logic is that drug development remains one of the most expensive and time-intensive processes in modern industry, often taking more than a decade and billions of dollars to bring a single therapy to market. Even marginal reductions in time-to-approval can translate into substantial financial gains, while also accelerating patient access to treatments.

But the partnership is seen as a reflection of a broader shift in how pharmaceutical companies are approaching technology. Firms are increasingly forming deep alliances with cloud providers, effectively outsourcing parts of their digital infrastructure while retaining control over proprietary data and scientific expertise.

The Merck–Google Cloud collaboration sits within a wider competitive landscape, where major technology firms are racing to secure long-term contracts with healthcare companies as AI adoption accelerates. These partnerships are becoming foundational, shaping not only operational efficiency but also the pace and direction of innovation in drug discovery.

What distinguishes this deal is its scale and integration. By committing up to $1 billion and embedding external engineers within its teams, Merck is treating AI not as an auxiliary tool but as core infrastructure — a shift that could redefine how pharmaceutical research and development is conducted over the next decade.

The model, if successful, is expected to compress development timelines, streamline regulatory pathways, and lower operational costs. But it also raises execution risks, particularly around data governance, regulatory compliance, and the challenge of aligning advanced AI systems with the rigorous standards required in clinical science.

Palantir Technologies has secured a $300 million contract with the United States Department of Agriculture, deepening its reach inside the federal government as Washington turns to artificial intelligence and data integration to manage agricultural risks increasingly shaped by geopolitics.

The agreement builds on existing collaboration and reflects a broader policy shift: farmland, crop output, and agricultural logistics are now being treated as components of national security infrastructure. The USDA is expected to deploy Palantir’s platforms to consolidate fragmented datasets across land ownership, crop production, and supply chains into a centralized system designed to improve oversight and decision-making.

This pivot comes at a time when U.S. agriculture is navigating overlapping shocks. Farmers are facing elevated input costs, unstable export demand, and growing uncertainty tied to geopolitical tensions. The trade dispute with China, a major buyer of U.S. soybeans, has already demonstrated how quickly demand can collapse. Late last year, Chinese pullbacks disrupted pricing and left producers with excess supply, forcing many to reconsider planting strategies.

The pressure has intensified with the Middle East conflict. Higher energy prices, linked to shipping disruptions and instability in key oil transit routes, have fed directly into fertilizer costs, a critical input for large-scale farming. Because fertilizer production relies heavily on natural gas and global transport networks, even marginal disruptions can cascade into sharp cost increases. For farmers, that translates into tighter margins and difficult decisions over crop selection, acreage, and investment.

The Trump administration has attempted to cushion the blow. In December, Donald Trump announced a $12 billion bailout for farmers affected by the trade war. But industry analysts say such measures offer only temporary relief, particularly when global factors beyond domestic policy control are driving cost pressures.

Within this context, the USDA’s partnership with Palantir is aimed at improving situational awareness. By integrating real-time data on inputs, yields, logistics, and market conditions, policymakers hope to better anticipate disruptions and respond more effectively. The system could also support scenario modelling, allowing officials to assess how shocks—such as export restrictions or fuel price spikes—would ripple through supply chains.

Another driver behind the deal is growing concern over foreign ownership of U.S. farmland. Lawmakers and policy analysts have warned that acquisitions linked to Chinese entities could carry implications, particularly if they provide visibility into or influence over food production. A report from the Foundation for Defense of Democracies recommended reforms to the Agricultural Foreign Investment Disclosure Act, urging tighter reporting rules “to prevent China and other adversarial countries from exploiting commercial land transactions to gain a strategic edge over the United States.”

Palantir’s tools are expected to play a role in closing those gaps by improving transparency around land transactions and ownership structures. This aligns with a broader effort in Washington to strengthen oversight of critical assets, from semiconductors to energy infrastructure, as geopolitical competition intensifies.

However, the deal means that Palantir will continue expansion beyond traditional defense work. Founded in the aftermath of the September 11 attacks, the company built its reputation on intelligence and counterterrorism applications. It has since evolved into a key provider of AI-driven analytics for both military and civilian agencies.

Its Maven Smart System, used by U.S. forces in Iran, highlights the convergence between battlefield and data-driven decision-making. Chief executive Alex Karp has framed this shift in stark terms, telling CNBC: “The fact that you can now target more precisely … has shifted the way in which war is fought.”

That same analytical capability is now being applied to agriculture, where precision, whether in targeting threats or optimizing production, is increasingly valuable.

But the company’s growing influence has not come without controversy. Palantir has faced sustained criticism over its work with U.S. Immigration and Customs Enforcement and the Department of Homeland Security, amid reports that its platforms have been used for surveillance. Civil liberties advocates argue that the expansion of such technologies into domestic sectors raises questions about data privacy and government overreach.

Those concerns are likely to follow the USDA deployment, particularly as it involves large-scale aggregation of sensitive data related to land ownership and agricultural activity. How that data is governed, who has access, and how it is used will be closely scrutinized.

On the market side, Palantir’s trajectory underlines both enthusiasm and skepticism around AI-driven business models. The company’s stock surged more than 25-fold between 2022 and the end of 2025, propelled by strong demand for its platforms. This year, shares have declined about 18%, as investors reassess valuations and growth expectations.

Short sellers remain vocal. Michael Burry has described the stock as “wildly overvalued,” a view that underscores broader concerns about whether current AI-driven gains can be sustained. Karp has responded forcefully to such criticism, saying: “I do think this behavior is egregious and I’m going to be dancing around when it’s proven wrong.”

The USDA contract adds a new dimension to that debate. It positions Palantir not just as a technology vendor, but as a partner in managing one of the most fundamental components of economic stability: food supply. It also illustrates how the boundaries between defense, economic policy, and domestic infrastructure are becoming increasingly blurred. In effect, Washington is applying a national security framework to agriculture, using advanced analytics to monitor, predict, and respond to risks that extend well beyond farm boundaries.