What the hell is happening in China?

I had my first trip to China at the end of April 2026. This is a set of observations and insights from that trip in no particular order.

1. China has temufied biotech

It is not necessarily news that there will always be a tradeoff between doing the novel risky thing – and dying by science failure; and doing the validated incremental bet – and dying by competition. However, with acquisition demand being more or less constant for a particular target-modality pair (there are only so many pharmas in the world!) and the supply curve shifting far to the right by China printing more early-stage assets for known targets with known modalities – I think the next few years will be the first time we see Western biotech companies unable to exit despite having promising data on me-too or me-better. I suspect exits for GalNAc siRNAs (including multitargeting siRNAs) and antibodies (including bispecifics and trispecifics) will become a bloodbath. In other words, if something is incremental or obvious, it is no longer worth doing – and the competition penalty is much higher than before.

2. Being the China middleman for pharma will die out

Summit Therapeutics is arguably one of the first times a lot of people realized what is happening with Chinese biotechs. Around 2022, the legend himself identified a PD-1 x VEGF bispecific antibody in China and flipped the asset by taking it through later-stage trials in lung cancer, eventually making the stock go vertical. In subsequent years, a number of VC firms and early-stage biotechs rushed to China to repeat the playbook, some successfully (Aiolos Bio, Kailera). That held a lot of value as pharma was establishing its own presence in the region, but I think in 2026 this playbook is quickly becoming outdated since pharma now has a front-row seat to all the same assets and would compete for them. Essentially, today, anyone acquiring assets from China is betting on them being overlooked by bigger players. While hoping that those same big players would want the asset once it has gone through the next stage of trials.

David Li had a great point that the lack of late stage / dev sophistication in Chinese biotechs reduces their likelihood of being able to partner with pharma, so marketing and packaging might still be the unique angle that western VCs and biotechs bring to the table. That being said, this is not the forte of everyone trying to be the middleman either.

3. I mentioned that Western biotech companies will struggle to exit with very straightforward bets, but I don't think the same will necessarily apply to Chinese biotechs

Chinese biotechs are asset-centric, but they are basically never single-asset. You can find early-stage Chinese biotechs with 10+ development candidates. First, it is obviously much cheaper to get a drug to DC in China than in the US. Someone also mentioned to me that this has been further amplified because Chinese biotechs in the past lacked access to later-stage development resources (people and capital), and so were always encouraged to do a breadth-first strategy. With the right portfolio construction, this approach amplifies your probability of finding the right partner or acquirer.

4. You don't need to manufacture drugs in China to run trials in China

This has been by far the most common misconception I hear from people. You do need to meet Chinese QC standards (whether for IIT or IND), but it is not written anywhere that you have to manufacture with a Chinese CDMO. People choose to do this because it's much cheaper and faster (and because often the same Chinese CDMOs can manufacture for your US/Europe IND).

5. Not manufacturing in China doesn't protect your IP

One of the most common reasons I hear for why someone isn't pursuing a clinical trial in China and goes to Australia or New Zealand is IP protection. To be completely blunt, I think it is one of those answers that has been templated by VCs and memorized by their portfolio companies without realizing that the priors are outright wrong. All biotech protection is built on IP and a mix of know-how. The longer a particular therapeutic direction has evolved, the less protection you get from know-how. It is not a secret to anyone how to design antibodies, siRNAs, and even small molecules. Even for more novel modalities, I think people generally overestimate how much manufacturing secret sauce their internal kitchen possesses.

So the only protection you'd truly ever have is in the patent. The patents in most cases become public before you even finalize your GLP-tox. To think that Chinese biotechs haven't already scraped them and haven't already worked around your 80% sequence similarity threshold is naive. So, to me, the calculus becomes - you can run your trial in China, be fast and cheap to get human data, and have your tech be scraped by Chinese biotechs; or you can run a trial elsewhere, take a few years and spend $20M++ more to get human data, and have your tech be scraped by Chinese biotechs nevertheless.

The only exception I can think of is novel target discovery, but in most cases, unless there is a lot of public validation, you won't see target crowding on genes that look biologically obscure.

6. Chinese biotechs can't exist without US markets

China is a single-payer system, with the CCP holding a lot of pricing power. One side effect of this is that the rare disease market is largely non-existent (rare disease drugs have to be expensive to recover the costs, and no one likes paying $2M for Zolgensma). This also explains why the Chinese market is orders of magnitude smaller than the US market despite being much larger by patient volume. All of this means that Chinese-born drugs will continue to target US markets for exits and approvals.

7. The bar for novel is higher than you think

In vivo CAR-Ts are not novel (despite being new). Gene editing and epigenetic editing are no longer novel. So whatever you are doing – you have to push the envelope with your tech and therapeutic applications much further than before. I understand that there is a reason people hate doing truly novel science as biotechs – too much novelty can be a recipe for failure. But staying within the realm of the known will almost certainly be.

8. Investigator-initiated trials are still the Wild West

Order No. 818 (came into place on May 1st, 2026) added some rules, with CMC and tox requirements going up, but none of these fundamentally changed that the IIT is still the fastest and cheapest way to get first-in-human data for complex modalities. The major change is that hospitals now have to belong to Tier 3A, so IITs will be limited to only top-tier hospitals in the country. Which would increase competition for investigators. Manufacturing and preclinical package requirements are largely the same as before (i.e., ambiguous and largely unspecified).

9. There are emerging precedents for the US accepting IIT human data as part of IND filing

Internally, we've been operating under the assumption that IIT data will not be part of our IND package and that geopolitical tensions are likely to block us from using it altogether. Even after assuming this, IIT seems clearly worth doing since having human data is the inflection point recognized by both pharma and investors. So the fact that the FDA can recognize this data as legible is a cherry on top! Goes without saying that this requires early FDA communication.