For nearly a decade, the United States wrote the rules of the global tech trade war, using its Entity List and chip-equipment restrictions to cut China off from advanced semiconductors. On October 16, 2026, Beijing handed Washington a mirror. China’s Ministry of Commerce and Ministry of Science and Technology published a draft catalog proposing export controls on 40 technologies spanning artificial intelligence, biomedicine and solar power, according to AP News and Fortune. The proposal is not yet law. It is a draft open for public comment. But the structure, the timing and the categories it targets read like a direct answer to years of U.S. restrictions aimed at China’s own tech ambitions.
The draft splits the 40 technologies into two tiers: 14 proposed outright prohibitions and 26 that would require an export license, according to Global Times. Of the 40, 23 concern artificial intelligence, 12 concern solar and photovoltaic manufacturing, and five concern biomedicine, including cell and gene therapy. China also activated a temporary registration requirement the same day, meaning companies must now log certain technology transfers, overseas licensing deals and ownership changes while the catalog is under review, per Caixin Global.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What China’s Draft Export Catalog Actually Proposes
China’s export controls proposal is narrower in one important way than most coverage suggests: it targets technology transfer, not finished products. The draft catalog does not ban the sale of Chinese-made solar panels or AI chips abroad. Instead, it would require government approval before the underlying know-how, design processes and manufacturing techniques behind those products can move to foreign hands, whether through licensing, joint ventures, overseas plant construction or outright technology sales.
That distinction matters for how the proposal will be read in boardrooms from Austin to Amsterdam. A ban on finished exports would have hit near-term revenue for Chinese manufacturers who depend on overseas solar and battery sales. A ban on technology transfer instead targets the next phase of globalization: the overseas factories, joint R&D labs and licensing deals that let foreign firms build Chinese-designed products on their own soil. According to Caixin Global, this means multinational manufacturing arrangements and foreign licensing agreements could require new Chinese government sign-off even when the finished product never touches Chinese soil.
The public comment period on the draft catalog is open, giving industry groups, foreign governments and Chinese exporters a formal channel to push back before the rules are finalized. That is a meaningfully different process than an emergency executive order. It signals Beijing wants the appearance of due process, even as the registration requirement for covered transactions already took effect on the same day the draft was published.
The AI Technologies on China’s Restricted List
Artificial intelligence accounts for 23 of the 40 proposed additions, the largest single category. The draft covers large language models, AI agents, chip designs, computing clusters and information-processing technologies, according to the fact sheet accompanying the announcement. Within that, the most eye-catching detail is a parameter threshold: the proposal would restrict export of AI models with 2 trillion or more parameters, or models judged to have “world-leading capabilities,” a bar that would sweep in China’s most advanced frontier systems.
Two categories move straight to the prohibited tier rather than the license-required tier. Core training technology for AI large models trained on sensitive national data cannot be exported at all under the draft. Neither can AI agents built for highly sensitive environments, explicitly including national defense systems and critical information infrastructure. Both carve-outs suggest Beijing is less worried about losing commercial AI revenue than about foreign actors reverse-engineering the training pipelines behind its most capable systems, or acquiring agentic AI tuned for state and military use cases.
The chip-design provisions go further than most people expect from an “AI” export rule. The draft would require licenses for AI chip designs made with 14-nanometer or more advanced process techniques, and for designs using 2.5-dimensional or 3-dimensional advanced packaging. Design technology for 3-dimensional flash memory also appears on the restricted list. That puts China’s own chip-design intellectual property, not just finished silicon, inside the same kind of control regime Washington has used against Chinese chipmakers since the early 2020s.
Solar and Biomedicine: The Other 17 Technologies
Solar accounts for 12 of the 40 proposed additions, and this is the sector where China holds the most global leverage. China dominates polysilicon refining, silicon-wafer production and photovoltaic-cell manufacturing worldwide, and the draft catalog targets exactly those process technologies and the core equipment behind them, rather than panels themselves. Perovskite solar cell technology, an emerging class that many researchers see as the eventual successor to silicon-based panels, would move to the prohibited tier outright. Heterojunction (HJT) solar cell technology, a more established high-efficiency architecture already in commercial production, would fall into the restricted tier requiring a license.
Biomedicine is the smallest category, at five proposed additions, but arguably the most sensitive politically. The draft covers cell- and gene-therapy process platforms, new-drug-target discovery and validation technology, small-nucleic-acid drug technology and pathogenic microbiology research technology, according to reporting cited in the fact sheet. Gene-editing and cell-therapy know-how sits at the intersection of commercial biotech, public health and biosecurity, and restricting its export puts China in a similar position to the U.S. government’s own long-standing scrutiny of biotech transfers to strategic rivals.
Industry watchers in Southeast Asia and India are paying close attention to the solar provisions specifically, since both regions have leaned on Chinese equipment and process licensing to build out domestic panel manufacturing as a hedge against tariffs on Chinese-made panels. If China requires a license for the equipment and process technology behind those factories, that hedge becomes slower and more bureaucratic, even if it doesn’t disappear entirely.
Prohibited vs. Restricted: How the Two-Tier System Works
China’s draft catalog uses a two-tier structure that will be familiar to anyone who has tracked U.S. export rules. Technologies in the restricted tier can still be exported, but only with a government-issued license, giving regulators a case-by-case veto over who gets access and under what terms. Technologies in the prohibited tier cannot be exported at all, full stop, regardless of buyer or purpose.
Of the 40 proposed additions, 14 would land in the prohibited tier and 26 in the restricted tier, per Global Times. That 35-65 split suggests Beijing’s instinct here is closer to “control and monetize” than “wall off entirely.” A majority of the covered technologies would still be exportable, just subject to a licensing process that gives the state visibility into every deal and the ability to block transactions it judges too sensitive, whether for national-security reasons or to preserve a domestic competitive edge.
| Technology | Sector | Proposed Tier |
|---|---|---|
| Core AI large-model training technology (sensitive national data) | Artificial Intelligence | Prohibited |
| AI agents for defense and critical infrastructure | Artificial Intelligence | Prohibited |
| Perovskite solar cell technology | Solar / Photovoltaic | Prohibited |
| AI chip design (14nm-and-above, 2.5D/3D packaging) | Semiconductors / AI | Restricted (license required) |
| 3D flash memory design technology | Semiconductors | Restricted (license required) |
| Heterojunction (HJT) solar cell technology | Solar / Photovoltaic | Restricted (license required) |
| Polysilicon production technology (30,000+ tonnes/year lines) | Solar / Photovoltaic | Restricted (license required) |
| AI models at 2 trillion+ parameters or “world-leading” capability | Artificial Intelligence | Restricted (license required) |
Why Now: The Public Comment Timeline
China published the draft catalog for public comment on October 16, 2026, opening a consultation window before any rule becomes binding. Separately, and immediately, a temporary registration requirement for covered technology exports took effect the same day, applying to intellectual-property transfers, overseas licensing deals and ownership changes that would place Chinese technology under foreign control. In practical terms, Chinese companies pursuing deals in any of the 40 covered categories must now log those transactions with regulators while the draft catalog is still being finalized, even before the final rules take legal effect.
That sequencing, registration first, final rule later, is a pattern regulators use when they want visibility into deal flow immediately without waiting out a full comment period. It gives Beijing a real-time view of which companies are trying to move covered technology abroad before the window on doing so freely potentially closes.
Mirroring Washington’s Playbook
The structural similarity to U.S. export policy is hard to miss, and multiple outlets have framed the proposal exactly that way. The Straits Times described the move as an effort by Beijing to “exert control over key tech,” language that echoes the stated rationale behind years of American restrictions on chip-equipment sales to China. The U.S. Commerce Department has used its Entity List since the late 2010s to cut specific Chinese firms off from American technology, expanding steadily through semiconductor manufacturing equipment and, more recently, advanced AI chip exports.
China’s new catalog borrows the same basic architecture: a government list of specific technologies, a tiered system separating outright bans from licensed exceptions, and the stated justification of national security. Where the two regimes diverge is in what each side is trying to protect. U.S. controls have focused on denying China access to chips and chipmaking tools it cannot yet produce at the cutting edge domestically. China’s proposal instead targets categories where its own companies have built a genuine lead, solar manufacturing most obviously, treating that lead as a strategic asset worth protecting rather than monetizing freely.
A Decade of Escalation: How the Two Sides Got Here
The October 2026 proposal does not arrive in a vacuum. It is the latest entry in a tech trade conflict that has escalated steadily since the late 2010s, when Washington first used export restrictions and Entity List designations against Chinese telecom and technology firms. That campaign expanded through the early 2020s into broad controls on semiconductor manufacturing equipment, denying Chinese chipmakers access to the advanced lithography and fabrication tools needed to compete at the cutting edge. By the mid-2020s, those controls had extended to advanced AI chip exports themselves, restricting the flow of high-end accelerators into China.
China’s own retaliatory playbook has been building in parallel. In April 2025, Beijing imposed export controls on rare-earth compounds and metals, a sector where China controls the overwhelming majority of global refining capacity. A rare-earth trade report cited in recent research found that exports of covered rare-earth materials remained well below historical levels as of August 2026, including zero shipments to Japan for a third consecutive month, even as rare-earth magnet exports had largely normalized. That split outcome, hardline on raw compounds, more flexible on finished magnets, is instructive for reading the October 2026 technology catalog: Beijing has shown it will use export leverage unevenly, hitting some sub-categories harder than others based on strategic calculation rather than blanket restriction.
| Period | Action | Initiator | Primary Target |
|---|---|---|---|
| Late 2010s | Entity List expansion against Chinese telecom and tech firms | United States | Chinese telecom equipment makers |
| Early 2020s | Semiconductor manufacturing equipment export controls | United States | Chinese chipmakers |
| Mid-2020s | Advanced AI chip export restrictions | United States | AI accelerator exports to China |
| April 2025 | Rare-earth compound and metal export controls | China | Global rare-earth buyers, Japan notably affected |
| October 16, 2026 | Draft catalog adds 40 technologies across AI, biomedicine, solar | China | AI models, chip design IP, solar process tech, biomedicine |
Market Impact: Why the Numbers Are Still Unclear
Here is where caution matters. As of this writing, there is no verified, reported stock-price reaction from Nvidia, ASML or other chip-sector companies specifically tied to this proposal, and no company has issued a public statement responding to the draft catalog. Any claim of a specific percentage stock move tied to this announcement should be treated skeptically until exchanges and company filings confirm it.
What is reasonably inferable from the structure of the proposal itself is where the friction would land. Because the draft targets technology transfer rather than finished-goods trade, the companies most exposed are not necessarily the ones buying Chinese chips or panels today, but the ones licensing Chinese process technology, building joint ventures with Chinese partners, or constructing overseas manufacturing plants using Chinese-origin know-how. That is a slower-moving, harder-to-quantify kind of exposure than a tariff or an outright sales ban, which is likely part of why markets have not produced a sharp, attributable reaction yet.
Solar manufacturers outside China who have leaned on licensed Chinese production technology to build domestic capacity, a strategy popular in Southeast Asia, India and parts of the Middle East, face the most concrete near-term question: will new joint ventures and technology licenses in HJT and polysilicon production now require Chinese government approval that didn’t exist before. For AI, the more consequential long-run question is whether Chinese AI labs building frontier systems above the 2-trillion-parameter threshold will face new friction exporting the underlying model weights, training techniques or infrastructure designs to overseas data centers and research partners.
Competitive Comparison: China’s Catalog vs. the U.S. Entity List
Comparing the two regimes side by side clarifies what is genuinely new here and what is familiar. The U.S. Entity List operates company-by-company: a firm gets added, and then anyone selling it U.S.-origin technology needs a license or is blocked outright. China’s new catalog is technology-by-technology: it is not about which companies can buy what, but which categories of know-how can leave Chinese hands at all, regardless of the buyer’s identity. That is a meaningfully different control philosophy, closer to the U.S. Commerce Control List’s structure than to the Entity List itself, even though most coverage has reached for the Entity List comparison because it’s the more familiar reference point.
The other key difference is direction of advantage. U.S. controls target areas where China is catching up but not yet dominant, advanced lithography, cutting-edge chip fabrication, forcing Chinese firms to spend years and billions closing a technology gap domestically. China’s controls instead protect areas where it already leads globally, solar manufacturing being the clearest example. That is a defensive-of-leadership posture rather than a catch-up posture, and it suggests Beijing sees export controls less as a tool to slow a rival and more as a tool to preserve a hard-won industrial advantage from being licensed away piece by piece.
How Outlets Are Framing the Announcement
Coverage of the draft catalog has split along predictable lines. Western financial press, including Fortune and AP News, has emphasized the national-security framing and the direct parallel to years of U.S. restrictions on China. State-aligned Chinese outlets, including Global Times and CGTN, have framed the move as a defensive, security-driven catch-up to standard international practice rather than retaliation, pointing out that major economies routinely maintain export-control catalogs of their own. Caixin Global’s coverage, generally seen as more independent within China’s media landscape, has focused on the practical mechanics: what the temporary registration requirement means for companies mid-deal, and how the public comment period might reshape the final catalog before it takes legal effect.
That divergence in framing matters because it previews how the final rule gets justified once the comment period closes. If Beijing leans into the security framing in its final version, expect tighter enforcement and fewer carve-outs. If the final catalog leans toward the “standard international practice” framing, expect more exceptions and a lighter enforcement touch, particularly for allied or non-strategic-rival countries.
What Happens During the Public Comment Period
Draft status matters enormously here, and it is worth repeating: nothing in the 40-technology catalog is legally binding export law yet. China’s regulatory process typically allows industry groups, foreign chambers of commerce and individual companies to submit comments that can and do change the final text, sometimes substantially. Expect heavy lobbying activity in the coming weeks from semiconductor equipment makers, solar manufacturers with overseas joint ventures, and biotech firms with cross-border research partnerships, all pushing to narrow the scope of what lands in the prohibited tier specifically, since that tier allows zero flexibility compared to the license-based restricted tier.
The temporary registration requirement already in effect gives Beijing a head start regardless of how the comment period plays out. By the time any final rule is published, regulators will already have a transaction-level map of who was trying to move covered technology and to whom, intelligence that shapes enforcement priorities even before enforcement formally begins.
Five Predictions for the Months Ahead
- The final catalog will likely narrow, not widen. Public comment periods on Chinese trade regulation have historically produced carve-outs for allied and non-strategic-rival markets; expect the 14-item prohibited list to shrink before the rule is finalized.
- Solar joint ventures abroad will slow, not stop. Overseas HJT and polysilicon projects reliant on Chinese licensing will face longer approval timelines, but outright cancellations are unlikely given China’s own interest in monetizing, not hoarding, mature solar technology.
- Frontier AI model releases will face new export friction. Chinese labs training systems near or above the 2-trillion-parameter threshold will need to navigate a new licensing step before offering API access, weights, or infrastructure partnerships to foreign customers.
- Expect a formal U.S. or EU response within weeks. Given the direct parallel to Western export-control architecture, trade officials in Washington and Brussels are likely to issue public statements characterizing the move, whether as justified reciprocity or as further escalation.
- This will not be China’s last catalog update in 2026. Given the precedent set by the April 2025 rare-earth controls and this October catalog, expect additional sector-specific export rules before year-end, particularly if the U.S. tightens AI chip restrictions further in response.
The Bigger Picture for Global Tech Supply Chains
The technology trade relationship between the world’s two largest economies has moved past the point where either side treats export policy as a one-way tool. For years, the assumption in Washington was that export controls were something the U.S. did to China, not something China could credibly do back, given how much of the advanced chip supply chain still runs through American, Dutch and Japanese equipment makers. The October 2026 catalog complicates that assumption. Solar manufacturing, large-scale AI model training and increasingly advanced chip-design work are categories where China has built genuine independent capability, and this proposal is the clearest signal yet that Beijing intends to treat that capability as leverage rather than simply as an export product.
For companies and governments outside both countries, the practical takeaway is that technology sourcing decisions in AI, solar and biomedicine now carry geopolitical risk from two directions rather than one. A company building an AI data center, a solar manufacturing plant or a gene-therapy research partnership increasingly has to account for U.S. export rules on one side and, pending this catalog’s finalization, Chinese export rules on the other. That dual exposure is new, and it is likely to become a standard feature of technology deal-making for the remainder of the decade.
Frequently Asked Questions
Is China’s 40-technology export control proposal already law?
No. As of October 16, 2026, it is a draft catalog published by China’s Ministry of Commerce and Ministry of Science and Technology for public comment. It has not taken legal effect as final export control law.
What categories of technology does the proposal cover?
Forty technologies across three sectors: 23 in artificial intelligence, 12 in solar and photovoltaic manufacturing, and five in biomedicine, including cell and gene therapy.
What is the difference between the “prohibited” and “restricted” tiers?
Prohibited technologies, 14 of the 40, could not be exported under any circumstance. Restricted technologies, 26 of the 40, could still be exported but would require a government-issued license.
Does this ban the sale of Chinese solar panels or AI chips abroad?
No. The draft targets the transfer of underlying technology, manufacturing know-how and licensing arrangements, not the sale of already-manufactured finished products.
Which AI models would be affected?
The draft proposes restricting export of AI models with 2 trillion or more parameters, or models deemed to have world-leading capabilities, along with AI agents built for defense or critical-infrastructure use.
Is this retaliation for U.S. export controls on China?
Chinese state media has framed it as a standard national-security measure rather than retaliation, while Western outlets have drawn a direct parallel to years of U.S. restrictions on Chinese chip and technology exports. Both framings appear in current coverage, and no official Chinese statement has called it retaliatory.
How does this compare to China’s 2025 rare-earth export controls?
The mechanism differs. The April 2025 rare-earth controls targeted shipments of finished compounds and metals. The October 2026 catalog primarily targets technology transfer and know-how rather than shipments of finished AI chips, solar panels or biomedical products.
When will the final rule take effect?
No final effective date has been announced. The catalog remains open for public comment, and a temporary registration requirement for covered transactions took effect on October 16, 2026, ahead of any final rule.
