From Chip Controls to DNA Synthesis Screening: New Supply Chain Security in the AI Era
When the CEOs of the world's leading AI companies call for mandatory DNA and RNA synthesis screening, it reflects a deeper shift: the core of tech governance is moving to supply chain and infrastructure management.
3 Key Takeaways
- Shift in Governance Core: The core of tech governance is shifting from algorithms and models to physical supply chains and critical infrastructure nodes.
- Expanding Regulations: Supply chain regulations in the AI era are expanding beyond semiconductors to include life sciences and DNA synthesis.
- Converging Security & Biotech: Biomanufacturing and genomic data face new compliance and security standards as national security and biotechnology increasingly converge.
When the CEOs of four world-leading AI companies—OpenAI, Anthropic, Google DeepMind, and Microsoft AI—jointly addressed the US Congress to call for a mandatory screening system for DNA and RNA synthesis, many focused on the risk of bioweapons. However, looking at this event within the context of recent global tech competition, it reflects a deeper shift: the core of tech governance is moving from algorithms and model capabilities to supply chain, infrastructure, and crucial node management.
Over the past five years, fierce global competition has centered around semiconductors. The US has used export controls to restrict China's access to high-end chips and advanced manufacturing equipment, while Japan and the Netherlands have strengthened controls on the export of key materials and equipment. A common logic lies behind these policies: the key determining national competitiveness is often not the end product, but the core capabilities and supply chain nodes that are difficult to replace. Now, this same thinking is gradually extending to the life sciences sector.
The rapid advancement of AI is changing the way R&D is conducted in life sciences, moving biotechnology from medical and research topics into the realm of national security and industrial strategy. If semiconductors are the computing infrastructure of the AI era, then DNA synthesis capabilities, biological databases, automated experimental platforms, and biomanufacturing systems may become equally important strategic assets in the future.
The same logic moving down: from models to chips to DNA synthesis
- Algorithms and model capabilityThe focus of the past five years
Evaluations, red-teaming, and capability thresholds — the first layer of AI governance to take shape.
- Semiconductors and fab equipmentControls already exist
US export controls limit China’s access to advanced chips and fab tools; Japan and the Netherlands tightened materials and equipment. The shared logic: competitiveness turns less on the end product than on supply-chain nodes that cannot be substituted.
- DNA and RNA synthesis2026 · four AI CEOs write to Congress
The chief executives of OpenAI, Anthropic, Google DeepMind, and Microsoft AI jointly urged Congress to mandate DNA and RNA synthesis screening — the first time the supply-chain-node logic lands on life sciences.
- Biological databases
Where sequence, structure, and experimental data accumulate. After AlphaFold, the strategic value of the data itself rose sharply.
- Automated laboratory platforms
What turns a model’s suggestion into an executed experiment — the layer that sets the distance between conceiving something and producing it.
- Biomanufacturing systems
The end point of scale production. The semiconductor experience suggests the least substitutable nodes usually sit here too.
Mature control regime in placeNow being proposedNot yet governed
AI is Reshaping the Life Sciences Industry
For decades, breakthroughs in life sciences mainly came from upgrades in research equipment, accumulated researcher experience, and lengthy experimental verification. This model did not change the nature of scientific research, but it meant that developing new drugs, protein engineering, and vaccine design often required massive investments of time and cost. In recent years, the addition of AI has begun to alter this rhythm.
AlphaFold, developed by Google DeepMind, is one of the most representative examples. Protein structure prediction had long been considered a major challenge in biology, often requiring research teams months or even years to complete partial analysis. The emergence of AlphaFold significantly boosted research efficiency, allowing scientists globally to understand protein structures and functions at unprecedented speeds. This technology is not just a single breakthrough, but a symbol that AI has begun to serve as an essential research tool in life sciences.
In addition to protein research, AI is also being integrated into drug discovery, antibody design, materials science, and genetic engineering. Many international pharmaceutical companies have established AI-assisted R&D pipelines, hoping to shorten drug development cycles and reduce the cost of failure. Tech companies like NVIDIA, Google, and Microsoft also continue to increase their investments in biotechnology-related platforms and models, reflecting the market's high expectations for AI's scientific research capabilities.
However, increased technical capability also means a lower barrier to entry. A recent special analysis by *Nature* on AI and biosecurity pointed out that the scientific community's focus has shifted from "whether AI can participate in biological design" to "to what extent AI can help design biologically active molecules, and whether existing governance frameworks are sufficient to cope with the rapid development of such capabilities." Such technologies have typical dual-use characteristics; the same tools can assist researchers in developing cancer treatments, but could also be used to design potentially harmful biological agents. The boundary between innovation and risk is blurring more than ever before.
Tech Governance is Shifting from Models to Supply Chains
It is easy to view this open letter as merely a warning about risk from AI companies, but the direction of governance it proposes is what truly deserves attention. The four companies' common stance is not to restrict the AI models themselves, but to demand a more comprehensive screening system for DNA and RNA synthesis. This reflects a growing consensus among policymakers and the industry: as AI capabilities rapidly proliferate, what usually needs to be managed is not the knowledge itself, but the critical nodes where knowledge enters the physical world.
In fact, the US government has already begun constructing an applicable framework. In 2024, the White House Office of Science and Technology Policy released the *Framework for Nucleic Acid Synthesis Screening*, requiring DNA and RNA synthesis providers to implement mechanisms for sequence alignment, customer verification, risk assessment, and anomaly reporting. In the same year, the US government also updated policies concerning Dual Use Research of Concern (DURC) and Pathogens with Enhanced Pandemic Potential (PEPP), strengthening oversight of high-risk life sciences research. These measures indicate that the US is gradually extending biosecurity governance from the research end to the supply chain end.
This thinking is strikingly similar to the semiconductor industry. Over the past few years, the focus of the US's high-end chip export controls on China was not on restricting mathematical knowledge or algorithm research, but on controlling high-performance GPUs, advanced manufacturing equipment, and critical manufacturing capabilities. Policymakers target the most difficult-to-replace nodes in the industrial chain. Similarly, in the biotech sector, even if someone uses AI to design a new DNA sequence, they still require a DNA synthesis platform to achieve physical production. From a governance perspective, managing these critical nodes is far more actionable than restricting the technology itself.
This also signifies that global tech governance is entering a new phase. In the past, nations focused on those who possessed the technology; in the future, they will pay more attention to those who control capabilities. DNA synthesis platforms, automated labs, biological databases, and high-throughput research equipment could all become new strategic nodes. As national security becomes increasingly intertwined with life sciences infrastructure, the importance of supply chain governance will rise rapidly.
The same logic applied twice — only one row differs
| Governance question | Semiconductors | Life sciences |
|---|---|---|
| Is the knowledge itself controlled? | No. Mathematics, algorithmic research, and publication all sit outside the controls. | No. The four companies do not propose restricting the models but the step where a sequence enters the physical world. |
| Which node is actually controlled | High-performance GPUs, advanced process equipment, and key manufacturing capacity. | DNA and RNA synthesis platforms — the single doorway through which a sequence becomes a molecule. |
| Why that node | It is hard to substitute, and it is physical: inspectable and countable. An algorithm is not. | Even a sequence designed with AI still needs a synthesis platform to be made. Controlling that step is more enforceable than restricting the technique. |
| Where the rules stand | Export controls, allied coordination, and an enforcement record already exist; Japan and the Netherlands have also tightened materials and equipment controls. | The 2024 OSTP Framework for Nucleic Acid Synthesis Screening calls for sequence comparison, customer verification, risk assessment, and anomaly reporting. What the four chief executives asked Congress for is to make it mandatory — that is, it currently rests largely on voluntary adherence. |
| Who actually carries compliance | Equipment makers and foundries. Customer screening, end-use judgement, and licensing all land on them. | Synthesis service providers — an industry far smaller in scale and legal capacity than the equipment makers carrying the equivalent duty. |
| Where Taiwan currently sits | An established critical node, and a real participant in rule coordination. | A complete health system, national insurance data environment, biotech research capability, and a semiconductor base — the ingredients are present, while the role in standards, regulation, and industry governance is not yet defined. |
Compared row by row, the governance logic is nearly identical: neither controls knowledge, both target the node where knowledge becomes physical, and for the same reason — that node is hard to substitute and it can be seen and counted.
The fourth row is where they part. Semiconductors already have export controls, allied coordination, and an enforcement record; nucleic acid synthesis rests largely on voluntary adherence, which is what the four chief executives asked Congress to change. The fifth row marks a gap easily overlooked: the industries carrying compliance differ greatly in size, and placing an equivalent duty on far smaller providers breaks the regime at the point of enforcement.
The last row is a note for Taiwan. The position in the left-hand column took two decades to grow, and it came not only from capacity but from a record of taking part in writing the rules.
Source: Impactful Creative, compiled from semiconductor export-control practice, the 2024 White House Framework for Nucleic Acid Synthesis Screening, and the open letter from four AI companies to the US Congress, as described in this articleThe Next Strategic High Ground: The Intersection of AI and Life Sciences
From an industrial development perspective, the significance of this shift goes far beyond biosecurity. Over the past three years, global AI competition has mainly centered around large language models, data centers, and high-performance computing power. Over the next decade, a grander transformation is likely to stem from AI's reshaping of the scientific research system.
As AI begins to participate in drug development, new material design, bioengineering, and energy technology research, its sphere of influence will expand from the information industry into the entire real economy. By then, nations will not only be competing on model capabilities, but also on biological data, experimental platforms, clinical resources, manufacturing capacities, and regulatory systems. Whoever can build a complete innovation ecosystem will stand a better chance of seizing the commanding heights in the next wave of industrial development.
For Taiwan, this trend is especially worthy of attention. Over the past twenty years, Taiwan has become a critical node in the global supply chain by relying on its semiconductor industry. In the future, if it can further integrate AI, healthcare, biotechnology, and digital infrastructure, there will be opportunities to establish new advantages in the next wave of tech competition. Taiwan possesses a comprehensive healthcare system, a valuable national health insurance data environment, biotech R&D capabilities, and a world-leading semiconductor industry foundation. In the context of AI and life sciences gradually merging, these conditions could form a unique competitive edge.
However, new opportunities come with new responsibilities. Going forward, the international community will not only focus on innovation capability, but also on data governance, research ethics, biosecurity, and industrial management competence. As the world begins to establish AI biosecurity norms and supply chain management systems, Taiwan also needs to think early about its role within international standards, regulatory frameworks, and industrial governance.
Semiconductor competition has made the world re-understand the importance of supply chain nodes. The convergence of AI and life sciences has allowed this governing logic to expand into new territories. As the world starts discussing DNA synthesis screening, biosecurity, and AI scientific research, the issues genuinely emerging are no longer just bioweapon risks, but evaluating which capabilities will become the next generation of critical infrastructure, and who will hold the power to set related rules and standards.
Great power competition in the 20th century revolved around energy and industrial capacity, while the first two decades of the 21st century focused on semiconductors and digital platforms. Looking ahead, the intersection of life sciences and AI is highly likely to become a new strategic high ground. By then, nations will be vying not just for technological leadership, but for control over critical capabilities, supply chain nodes, and industrial order. From chip controls to DNA synthesis screening, global tech governance is broadcasting one consistent signal: the core of the next round of competition will be built upon those critical infrastructures that connect innovation capabilities to the real world.
Related Articles
The Statement That Never Came: A Chinese Missile Test Exposes the Deeper Contest in Pacific Diplomacy
After a Chinese submarine test-launched a strategic missile into the Pacific, the Forum's foreign ministers could not agree on a statement. A failure to agree on wording repays more study than a strongly worded communiqué: the same consensus rule that let small states constrain nuclear powers in 1985, and blocked Beijing's regional security package in 2022, has now stopped the Forum's own document. The institution has not changed — the structure of interests has.
Saudi Arabia, Turkey, and Pakistan Align: The Middle East Builds a Second Security Layer Beyond the United States
The Mecca Joint Defence Agreement treats an armed attack on any one of the three as an attack on all — language that drew immediate comparison to NATO's Article 5. But reading it as another NATO misses the larger shift: regional powers are no longer only choosing among architectures designed elsewhere. Seventy years after the Baghdad Pact traced nearly the same geographic axis, the architects have changed.
Papua New Guinea in the New South Pacific Order: Security Alliance via the Pukpuk Treaty, Multi-Market Hedging, and Compartmentalized Diplomacy
Papua New Guinea is advancing military alignment with Australia under the Pukpuk Treaty and granting US defense access, while simultaneously expanding trade ties and praising major investments with Beijing. This 'compartmentalized alignment' strategy seeks multi-market hedging, but as security, energy, and port infrastructure blur, the cost of policy ambiguity is rising rapidly.