Treasury's COINS Rules Spare China Licensing, Target Joint Ventures

Treasury's proposed outbound-investment rules never mention pure licensing deals with Chinese biotech companies — and Treasury confirmed that omission on a press call to Endpoints News reporter Max Bayer. That silence is the one thread optimists are holding onto. It is thinner than it looks, because almost everything built around licensing got caught.

The proposed COINS regulations explicitly prohibit joint ventures. They also name "pharma compounds" as a prohibited technology. Together, those two clauses could capture much of the dealmaking that has defined the past two years — "the M&A story in biopharma for the last 2 years," as Bayer put it on the Post-Hoc Live podcast the morning after the rules dropped, in conversation with Endpoints executive editor Drew Armstrong.

Bayer had broken the May 2026 story that biotech would appear in the rules at all. His October 21 follow-up, built on five background sources close to Treasury, reported the outcome would not be as favorable as the industry hoped.

What does the rule actually say?

Three prongs matter to anyone doing China deals. Only one favors the industry.

• Licensing: Not mentioned. Treasury confirmed the omission. Pure "write a check, take the asset" contracts appear preserved.
• Joint ventures: Explicitly prohibited. Treasury declines to define the term and points to the "plain English meaning" — capital and/or assets contributed by two parties, with profits and losses shared. That's broad enough that co-development, co-commercialization and equity-linked structures may fall inside.
• Prohibited technology: "Pharma compounds" — enumerated as small molecules, cell and gene therapies, monoclonal antibodies and protein therapeutic recombinants. Bayer reads the list as sweeping in bi-specifics, tri-specifics and the wave of in vivo CAR-T deals sourced from China.

The joint-venture clause is where the legal ambiguity lives. Bayer described Treasury's approach this way: "The regs go far as to say, 'We're not going to define this. We're going to refer to the plain English meaning of the term,' which they describe as the contribution of capital and or assets by two parties and the sharing of profits and losses. Extremely broad."

The technology list is no narrow carve-out either. "Within these joint ventures, what is described as a prohibited technology is pharma compounds, effectively anything you might want to do a deal with," Bayer said. "Small molecules, cell and gene therapies... Monoclonal antibodies and also protein therapeutic recombinants."

Armstrong added the structural reality the drafters may have missed: few China deals are pure licensing. Co-development pacts, royalty streams and knowledge transfer are routine — and each raises the question of whether equity or asset contribution has effectively created a joint venture.

Why the final text surprised the sector

A week before publication, the prevailing assumption was that biotech would get a mention and escape largely unscathed. Then Bayer started receiving cryptic messages that the national-security wing had gained the upper hand. By Friday, October 16, he was told pharma was in the regs. He and Armstrong had earlier come close to reporting that licensing was included, then stepped back on separate sourcing.

The final text surprised people on breadth. "This is not going to be as sweet as people thought it was going to be in that they were going to sort of smoothly escape this tunnel of regulations with no mention at all to the deal flow and the kinds of technologies and compounds that people have been making a ton of money off of," Bayer said.

One detail he flags as a signal of intent: the rule effectively asks industry whether it is too harsh. He reads that as evidence the drafters felt constrained by pressure elsewhere in the administration and left room to narrow the rule after the comment period.

Armstrong's read is that the document targets equity transfers, but real deal construction is messier than that test. He cites a Gilead deal announced overnight — a China transaction with a billion-dollar-plus upfront plus milestones — as a narrowly written structure that looks squarely permissible. He doesn't claim the timing was deliberate.

Which structures get hit first?

The NewCo model is the most exposed. The playbook — source a molecule in China, spin it into a US company, develop it, sell it to pharma — typically involves equity stakes and often looks like a joint venture, exactly what the rule prohibits. China-focused funds raising from US limited partners face the same problem.

Armstrong relays an argument from a trusted source: clean licensing preserves late-stage access to Chinese innovation, but the US loses early venture-stage participation, where much of China's most advanced work happens. The result is an ecosystem that buys finished drugs rather than one that participates across global innovation. China hawks would take that trade.

Bayer raises a counter-current: the NewCo model may already have been losing favor as large pharma built search-and-evaluation teams inside China. He cites colleague Kyle's reporting on roughly five or six Roche deals in China since August 2026, and singles out Roche as the most interesting case — Swiss-domiciled, with Genentech as a major US R&D hub.

Who wrote the rule — and who won

Two factions fought inside the administration. Companies that met Treasury officials came away believing the department did not want its thumb on the market's scale. The national-security wing wanted leverage against China, including in unrelated negotiations — Bayer notes the run-up to President Xi's September 2026 visit to Trump as backdrop.

The concept is simple: COINS mirrors an older tool. "CFIUS has been an existing Treasury policy that monitored, reviewed, and restricted inbound investment into the US from countries of concern," Bayer said. "COINS was like the outbound equivalent of that."

Per Bayer's sourcing, several concerns pushed Treasury toward restraint: fear of handing European pharma first and only access to China-originated medicines; worry about being framed as limiting Americans' access to new cancer drugs; and midterm-year politics. Armstrong pushes back on the "Europe will launder the deals" argument — multinationals have major US operations and government relationships, so re-domiciling in London, Switzerland or France is harder than it sounds.

China, meanwhile, is likely to escalate rather than negotiate. It has already discussed restricting technology exports; there are reports of limits on knowledge workers leaving; and its latest five-year life-sciences plan, covered by Endpoints in early October 2026, is described as extremely ambitious. Armstrong's expectation, based on tariff experience, is escalation first.

What happens next?

The 45-day comment window, plus what Armstrong believes is a grandfather clause, creates an incentive to close pending China deals fast — before December 7. Bayer expects every affected party to file Federal Register comments in precise legal language, and the loudest hand-wringing to come from venture capital.

The strategic question is whether industry attacks the rule broadly or picks a target: the "pharmaceutical compounds" designation if lawyers find the JV definition survivable, or the JV definition itself if they conclude deals can't be struck under it. Expect aggressive direct lobbying too. Chris Pilkington, the assistant secretary of Treasury leading the investment security division, has been the main meeting target all summer — and Bayer wonders whether the campaign escalates to CEO-level calls, naming Albert Bourla and Dave Ricks as likely holders of Scott Gottlieb's and the president's phone numbers.

For anyone holding shares in a US biotech sourcing molecules from China, the next 45 days are the risk window.