Analysis

CCP General Secretary Xi Jinping–once again–torched China’s military leadership and purged two of the People’s Liberation Army’s (PLA) most senior leaders. While headline-grabbing, what does such a move signal for Chinese elite politics, China’s economy, and US-China relations? 

First, while purges can be signals from the CCP’s central leadership about the party’s intended policies and views, their meaning is often not clear, especially in the immediacy of their application, despite the confidence in which analysts declare significance. The reality is that we really know very little about the internal machinations of the party elite, but dramatic news about a purge is often touted as crucial evidence in lieu of other signals from an otherwise opaque system. Analysis that points to what this means for the PLA’s combat readiness against Taiwan or Xi’s attempts to stop a suspected coup should therefore be qualified. 

Second, purges are a feature, rather than a bug, of the CCP’s institutional structure. They are one of the main mechanisms by which the party leadership maintains discipline, internal cohesion, and loyalty. Recent purges, for example, come after a purge just last year of other senior military leaders, which similarly rocked Chinese politics. This is the CCP doing what it has done for over a century, not necessarily a sign of a new development. Moreover, China is hardly unique. This is what autocratic systems do.

Third, if we take Xi’s purging of those close to him within the military leadership as a sign of his personal interest in more closely controlling China’s military, then one possible inference is that Xi is continuing to centralize his power over the party and to minimize potential threats. After all, having the military on side is vital to autocratic survival. With a caveat that we know little about China’s elite politics, one potential interpretation could be that Xi is pushing for personal rule over institutional rules. That is not new. Xi has pushed for centralized personal rule since he became General Secretary. US businesses engaged in the PRC should therefore not be surprised at Xi’s continued purging of those he deems a threat to either himself or the party. Rather, it is business as usual in the PRC. US businesses operating in the PRC should expect more of the same in the future. 

Collectively, this means that elite politics can be a useful tool for determining broad trends in Chinese politics, but individual news items should not drive business decisions. Rather, longer term signals–from how China is dealing with involution, driving AI innovation, and collapsing distinctions between the party and society–reveal more useful information that US companies can  incorporate in their planning than any instance of palace intrigue. A PLA purge therefore does not necessarily mean a change in how Xi manages the party, or by extension China’s economy.

Analysis

Following recent calls between Trump and Xi, Chinese state media is promoting an image of warming US-China relations. At the same, however, the US government continues to hedge its bets. President Trump revealed his newest initiative to bolster US self-reliance: “Project Vault” to be jointly financed by $1.67B in private funds and a $10B loan from the US Export-Import Bank. The project intends to create a US Strategic Critical Minerals Reserve that will contain both the rare earths and critical minerals necessary for US civil needs (a military reserve already exists). Unspoken in the announcement, but clear in its subtext, is the US government's desire to insulate US supply chains from the volatilities of a China-controlled rare earth market. 

The new project is a large step forward in President Trump’s on-going efforts to bolster US technological  independence from China. In December of 2025, the State Department launched its Pax Silica initiative, aimed at creating an alliance of friendly nations to ensure AI and supply chain security. Current signatories include Australia, Japan, South Korea, Singapore, and the UK, with Taiwan recently inking its own Pax Silica declaration. This arrangement was founded in tandem with a more loosely organized “critical minerals club,” which comprised Australia, Japan, South Korea, Malaysia, and Thailand at the end of 2025, but that was formalized and expanded last week at another event hosted by the State Department: the 2026 Critical Minerals Ministerial. Attended by representatives from more than fifty countries, it resulted in eleven additional memorandums of understanding and, if successful, will open up additional mining resources globally. While Project Vault was touted as a central achievement at the Ministerial, the Departments of Energy and Defense, along with the US International Development Finance Corporation, additionally committed an additional $10B in funds to related projects. 

The matter of creating a critical minerals reserve has strong, bipartisan support in Congress, although the devil is always in the details: How much should be held in reserve, who should hold it, and where it should be held remain open-ended questions. Empirical evidence also suggests that we should temper expectations. When Japan got into a dispute with China back in 2010, China withheld rare earth shipments and caused Japanese rare-earth prices to jump ten-fold. In the ensuing years, Japan has worked to decrease its reliance on China and now get approximately 60% of its rare earths from China, relative to a previous 90%. While this is a significant reduction, it also gives a sense of how slowly this process is likely to proceed. Consequently, while we may see a long-term market reorientation after Congress finalizes project specifics, US companies should expect Chinese dominance of rare earth supply chains to continue in the short term and expect recent volatility to remain.

Analysis

Washington is accelerating efforts to reduce China-linked exposure in strategic infrastructure across the Western Hemisphere, raising the geopolitical sensitivity surrounding one of the world’s most critical maritime channels–and chokepoints. This push extends beyond Panama to include heightened scrutiny of port concessions, telecommunications backbones, energy grids, and logistics corridors across Latin America, where US policy makers increasingly view Chinese commercial presence as a potential vector for political leverage, intelligence collection, and long-term supply chain dependence. 

Following a 2025 visit by Secretary of State Marco Rubio, Panama withdrew from China’s Belt and Road Initiative, and increased scrutiny of foreign concession agreements. Danish operator Maersk assumed transitional management of the terminals, and US investment firm BlackRock remains positioned for a potential long-term role amid regulatory uncertainty. Beijing has condemned the ruling and responded by suspending new infrastructure discussions and tightening customs inspections of Panamanian imports into China, signaling economic retaliation without directly disrupting canal transit. These responses suggest that while China may avoid overt escalation, it is prepared to exploit trade dependencies to signal dissatisfaction and impose incremental costs.

In late January 2026, Panama’s Supreme Court declared the long-standing port concessions of Hong Kong–based CK Hutchison unconstitutional, ending its decades-long management of the Balboa and Cristobal terminals at either end of the canal. Although framed domestically as a response to audit findings alleging $1.2 billion in lost revenue, the ruling coincides with broader US efforts to limit Chinese commercial footholds in high-leverage logistics nodes throughout the region. Moreover, the ruling marks a return of national political agency over critical logistics nodes that had long functioned as independent commercially driven entities.

The canal remains governed by neutrality guarantees, and Chinese vessels retain full access. What is changing is the strategic framing of infrastructure control. US policymakers are placing greater emphasis on supply-chain resilience and the security of both physical and digital systems within the canal zone, including Chinese-manufactured port equipment. Even absent formal restrictions, heightened great-power competition is elevating the political, regulatory, and compliance risk profile of what was previously treated as routine commercial infrastructure.

For US businesses, the immediate impact is not operational bifurcation, but increased geopolitical exposure. A more Western-aligned port management structure may reduce long-term coercion and counterintelligence risk, yet it also raises the likelihood of episodic regulatory friction or retaliatory measures from Beijing. Firms with concentrated exposure to canal transit, Latin American agricultural trade, or China-linked supply chains should anticipate the increased likelihood of greater volatility in shipping insurance, customs processing times, and investment approvals. The canal is unlikely to fragment structurally, but it is no longer insulated from strategic rivalry. Infrastructure control in the Americas is becoming an explicit variable in corporate risk calculations rather than a background assumption of global trade stability.

Analysis

China is moving away from Western-style "frontier" AI breakthroughs that rely on massive hardware scale, pivoting instead toward a swarm-based strategy rooted in software coordination. By deploying teams of smaller, specialized AI agents, PRC innovators  are building a technology system that depends less on the most advanced semiconductors and is more resilient to external pressure. This shift became highly visible in January 2026 with the release of Kimi K2.5 by Beijing-based startup Moonshot AI. This model uses an architecture where a primary system orchestrates up to 100 specialized sub-agents to complete complex tasks in parallel. The focus is no longer on creating one massive "brain," but on distributed execution across many smaller, coordinated components.

This design reflects a broader transition toward system-level optimization. Instead of needing the world's most powerful computer chips to make AI better, developers are now connecting specialized units that each handle a small piece of a larger task. This allows the system to become more capable through organization rather than just raw power, which helps get around current chip restrictions A similar logic appears in the military sector, as recently demonstrated by the National University of Defense Technology where formations of over 200 autonomous drones were managed by a single controller, operating independently even when communication links were disrupted.

However, this path comes with significant technical trade-offs. While the swarm approach offers resilience, it introduces a heavy orchestration tax, where the sheer amount of communication among 100 different agents can slow response times and spike energy use. Furthermore, while these swarms excel at specific tasks, they often lack the "general reasoning" found in larger Western models. There is also the danger of emergent chaos; small errors or hallucinations in one unit can ripple through the network, creating feedback loops that amplify mistakes. Most importantly, while these agents are small, training them still requires a massive "compute floor" of high-end hardware, meaning the chip bottleneck is only bypassed during deployment, not development.

Ultimately, Chinese AI development is being defined by a new metric where success is measured by how many intelligent pieces can work together under pressure. This strategy prioritizes synergy, survivability and flexibility over immediate commercial returns. As these architectures mature, the result is a technology sector built to operate autonomously in the face of foreign competition, ensuring that technical knowledge accumulates across the system even when individual firms face sanctions.

For US companies, the competition is shifting from a race for intelligence to a race for cost. While American firms lead in building the "smartest" models, Chinese firms are winning on the "cheapest" execution, with models like Kimi K2.5 often running at a fraction of the cost of their US counterparts. If US businesses only focus on chasing the next massive breakthrough while ignoring the efficiency of coordinated swarms, they risk being priced out of the global market. The danger isn't just smarter AI from China—it’s cheaper, more modular AI that becomes the world’s default infrastructure while Western models remain expensive luxuries.

Analysis

Most espionage against US companies is undetected and the damage only realized after it is too late. Even less is caught or reported. A federal jury’s conviction of former Google engineer Linwei Ding for engaging in IP theft is therefore an important example for how to spot and stop economic espionage. While this is the first successful case of the Department of Justice prosecuting Chinese espionage targeting US AI innovation, Linwei Ding is only the latest example of corporate espionage. So what can US companies learn about the risk of IP theft–and how to prevent it–from Ding’s case?

Ding’s background contained multiple risk indicators consistent with similar cases of IP theft and illegal technology transfer to China. Ding’s signals, if properly monitored, should have alerted Google sooner as to his malintent. In December 2022, for example, he publicly disclosed his role as a CTO at a PRC AI company, concurrent with his role at Google. In service of this role, he copied source material from Google using Apple Notes, asked a colleague to badge him in while he was in the PRC, and used a series of public engagements in the PRC to fundraise using his Google credentials. Despite this, his unusual behavior, erratic travel, undisclosed external affiliations, and physical security anomalies were initially undetected. 

Yet Ding, while likely not an uncommon example, is nonetheless an extreme and blatant case of economic espionage. The far more damaging–and more challenging to detect–cases are those that either do not involve deliberate malintent, or that are more subtle than Ding’s public announcements of his parallel life. Ding’s patterns of behavior are strong signals for the types of indicators and detection opportunities that companies should employ, but they are not fail-safe measures. Instead, preventing economic espionage entails continuous monitoring of multiple risk vectors, and supporting a culture of openness with employees so that they feel comfortable reporting if they are approached to hand over sensitive information. 

Ding’s conviction may be the first in terms of the specifics of the case, but it will not be the last. As the CCP mounts pressure for PRC state-owned and private companies to engage in more widespread IP theft against US companies, the problem will be with us for a long time. US companies need to plan carefully and sensitively for how best to defend their employees, detect potential infiltration, and deter attacks against their IP. Failure to do so will be catastrophic for individual US firms, entire sectors, and US economic vibrancy and national security.