International Patents Are a Time Game, Not a Money Game

The first time a founder hears the phrase "international patent," they usually picture a six-figure bill. Lawyers in five countries. Translations. Filings in languages they don't read. A wall of fees stretching out for two decades.

The reality is different in a way that took me a while to see. There is no such thing as an "international patent." Patents are country-specific. What WIPO and the international system actually offer isn't a global patent — it's a way to defer the expensive country-by-country decision until you've had a chance to figure out which countries are worth the spend.

That changes the entire shape of the problem. International patent strategy isn't a money game where you decide how much to invest up front. It's a time game where you use two specific windows to keep your options open for as long as the system allows. Get this wrong on day one and you've burned $30K+ buying protection in countries you'll never sell in; get it right and the same money buys you roughly 30 months of optionality before you have to commit to anything irreversible.

This post is the fourth in a five-post series on US patents for builders. The primer covered what a patent actually is. The second post argued that filing should be the exception, not the default. The third argued that reading patents is one of the highest-leverage moves an engineer can make. This one is about what to do once you've decided a filing is worth it — and how to do it across borders without burning $30,000 on day one. The fifth and final post picks up after the filing decision, when someone actually fights over the patent.

There is no global patent

The single most useful thing to internalize about international patents is the part of the system most marketing copy hides: every patent is granted by, and enforceable in, exactly one country.

If someone copies your invention in Japan and you only have a US patent, you have no claim. The US patent doesn't reach there. To get protection in Japan, you have to actually file in Japan. The same is true for every other country.

What WIPO (the World Intellectual Property Organization) does is run the plumbing that makes filing in many countries less painful. It does not run a global patent registry. There is no single place to file once and be done.

The two pieces of plumbing worth knowing are the Paris Convention and the Patent Cooperation Treaty (PCT). They're often described together, but they serve different roles, and the timing windows are the heart of why the system is cheaper than it looks.

The 12-month Paris Convention window

The Paris Convention solves the global queue problem with one simple rule, codified in Article 4 of the convention: file once, and you have 12 months to file anywhere else without losing your place in line. It's signed by 180+ countries, including the US, every major EU country, China, Japan, Vietnam, and almost every place a software company would want protection.

Here's how the rule actually works. The day you file your first patent application in any signatory country, you start a 12-month clock. Within that window, any other Paris Convention country has to treat a follow-up application from you as if you had filed it on your original date. That date is called your priority date .

Why does the priority date matter? Because patents are awarded to whoever files first. If you wait too long after going public, someone else might publish or file something similar, and that becomes prior art against you. The priority date freezes your position in the global queue on day zero. You then have a year to decide where else to file without losing that position.

This is the first, simplest, cheapest move in international patent strategy: file once in your home country, then take the full 12 months before doing anything else. You haven't committed any money beyond the first filing. You've bought a year of optionality.

The only thing that can ruin this is publishing your invention before you file. Once it's public, the priority clock has already started in some countries' counting (or your novelty has been destroyed entirely). The first filing has to come before any public disclosure if you want clean priority everywhere.

The 30-month PCT window

A year goes fast, and the Paris window often closes before a startup actually knows where its market is. The PCT is the system's answer to that gap: Article 22 of the PCT stretches the international decision window from 12 months out to roughly 30 months from your original priority date, in exchange for one additional filing fee. (A handful of countries still use the older 20-month deadline; for any country you actually care about you'll want to double-check.)

The PCT works by letting you file a single "international application" with WIPO that's recognized by ~150 countries. It doesn't grant a patent. It buys you time and produces a useful search report. At month 30, you have to enter the national phase in each country you actually want — meaning you file there, pay their fees, and start their examination process.

So the full timeline of windows looks like this:

What this gives you is roughly 2.5 years of validation time before you have to commit real money to specific countries. For a software startup, that's enormous. Two and a half years is the difference between "we have an idea and a prototype" and "we have customers in three countries and know exactly where the legal protection matters."

The PCT filing itself isn't free — typical cost runs $1,500 to $3,000 in fees plus legal work. But it's much cheaper than filing in each country individually on day one, and it preserves your priority date everywhere.

A cost playbook for builders

Putting the windows together, three patterns cover most situations a founder will face.

Pattern 1: US-first. You're aiming primarily at the US market. File a non-provisional in the US first — $8,000 to $20,000 with an attorney. At month 12, file a PCT application — $2,000 to $3,000 — to keep international options alive. At month 30, enter the national phase only in countries where you have actual customers or strategic reasons. Per-country national phase entry typically runs $3,000 to $10,000 depending on translation requirements.

Pattern 2: Home-country-first. You're based outside the US and want to start cheap. File a provisional or full application in your home country — often well under $1,000 in fees plus modest legal work. Use the Paris window to validate. At month 12, decide whether to file a US non-provisional, a PCT, or both. This is the slowest path but the cheapest day-one commitment.

Pattern 3: Defensive publication. If you're not sure the invention is worth patenting at all, publish it publicly — a blog post, a conference talk, a published spec. Once it's public prior art, no one else can patent it either. You've protected your freedom to operate without paying the patent system anything. The catch is that you've also given up your own ability to patent it (in most countries — the US has a 1-year grace period, but other countries don't).

The pattern most product-thinking founders end up using, in my experience, is some mix of patterns 2 and 3. They publish most of what they build, and they patent only the small set of things that are both technically novel and commercially central.

What this looks like for a Vietnamese founder

A specific question I keep running into from founders here: how does this actually work if you're based in Vietnam and aiming for both Vietnamese and US markets?

The Vietnamese system follows the same Paris and PCT plumbing as everyone else, with a few wrinkles worth knowing.

Vietnamese founders have a famous cautionary tale on the cost of skipping the priority window. In July 2000 a US distributor, Rice Field Corp, quietly registered the TRUNG NGUYEN trademark at the USPTO before the actual coffee company did — founder Đặng Lê Nguyên Vũ discovered it almost a year later and had to negotiate the registration's withdrawal before Trung Nguyên could re-file in May 2002. The dispute was trademarks not patents, but the principle is identical: whoever files first owns the priority date, and the priority date owns the country.

The patent-side equivalent shows up regularly. Vietnam Patent & Trademark Office's case write-up describes a Thai applicant who filed a series of 38 patent applications in Vietnam without claiming priority from earlier Indonesian filings; several of the Vietnamese applications then got novelty rejections — based on the applicant's own earlier Indonesian patents, which had become prior art once they published. The applicant beat themselves. That's the failure mode in miniature: the priority date isn't optional paperwork, it's the entire reason the international system works.

For a Vietnamese founder, the non-obvious move is that this works in your favor. A $1,500 home-country filing in Vietnam gives you a priority date that any of the 180+ Paris Convention countries — including the US — has to treat as if you'd filed there on the same day. The Vietnamese priority date is one of the cheapest globally enforceable priority anchors a software founder can buy.

A few specifics worth knowing about filing in Vietnam:

• Non-residents must use a local agent. You can't file directly. The agent fees are usually $500 to $1,000 on top of the government fees.
• Filing costs are low compared to the US. Filing, search, and examination fees together typically run $700 to $1,500. Maintenance fees are similarly modest — often $50 to $200 a year.
• Total Vietnam filing cost for a non-resident is roughly $1,500 to $3,000 — meaningfully cheaper than the US filing alone.
• Examination takes 18 to 24 months after substantive examination begins, with a total timeline to grant of roughly 3 to 4 years.
• Enforcement is weaker than the US. Patent litigation in Vietnam is rare, and damages are smaller. The patent is more useful for licensing, fundraising, and acquisition value than for active enforcement.

For a Vietnamese founder building something they think will travel internationally, the playbook I'd consider:

The math works out roughly the same as the math for a US-based founder, with the home-country filing acting as the cheapest possible entry point. A few thousand dollars at month zero buys you a priority date and 12 months of optionality. Another few thousand at month 12 buys you another 18 months. By the time you're spending serious money on country-by-country national phases, you should know exactly which countries actually matter to your business.

Counterargument: when the time-game playbook is wrong

A reasonable reader at this point might push back. "You're recommending the cheapest, slowest path — home country first, PCT at month 12, national phase only where you have customers. But IBM doesn't operate that way. Qualcomm doesn't operate that way. The companies with the strongest patent portfolios file in dozens of countries on day one. Why is the slow path right if the people with the most experience pick the fast one?"

The objection is real, and it has a named voice behind it. Manny Schecter, IBM's former Chief Patent Counsel, has been explicit about why IBM files the way it does: "Having a large portfolio enables us to work from a position of strength." IBM holds roughly 50,000 active patents worldwide and uses that density for cross-licensing leverage, freedom-to-operate clearance, and defensive bargaining against peers with comparably-sized portfolios. At that scale, day-one broad international filing is rational — the marginal cost of one more country is small compared to the strategic value of dense, overlapping international protection.

The thing the objection misses is that this is the exception path, calibrated for a specific situation that almost no individual founder is in. The conditions that make IBM's strategy correct are:

If you can answer yes to any of those three questions, the time-game playbook is too slow for you and broad day-one filing makes sense. If you can't, the playbook is the correct default.

A few more situations where the windows don't help and the answer is just "don't file internationally":

• You have no plan to sell or operate outside your home country. A patent in a country you don't sell in mostly costs you maintenance fees forever.
• Your invention is easy to design around. International protection of an easily-redesigned mechanism is just expensive paperwork.
• You can't afford the eventual national phase costs. If the $3,000 to $10,000 per country at month 30 would break the bank, you're better off filing in fewer countries up front rather than getting locked into PCT timelines you can't follow through on. WIPO's PCT Yearly Review 2025 tracked roughly 698,500 national phase entries worldwide in 2023 — but a large share of PCT applications never make it that far, because the month-30 bill arrives whether or not the business is ready for it.
• The invention is software-only, and the country you'd want to file in (Europe, India) has narrow software-patent eligibility. The cost-benefit gets much harder to justify.

The pattern across all the failure modes is the same: international patents earn their cost when there's a real market on the other side of the border and you're not in IBM's strategic situation. Without one or the other, the windows don't help — they just delay the moment you have to admit the patent isn't pulling its weight.

Pulling this all back together

What kept getting clearer to me as I worked through these four posts is that the patent system is much more product-shaped than it looks at first. The legal language is intimidating, the fees are confusing, and the international layer feels like another tax on builders. But underneath, the system has a few useful properties for anyone thinking about IP as a product decision:

• The default setting (don't file anything yet) is usually correct.
• The cheap moves (publish defensively, file at home, search before building) cover most of what an engineer actually needs.
• The expensive moves (full international filings, multi-year prosecution, NPE-resistant portfolios) are rare, and the system gives you a long runway to figure out whether you need them.

It's the same lesson I keep running into across product thinking: the question isn't usually "should we do the most ambitious version of this?" It's "what's the smallest commitment that keeps our future options open?" The patent system is unusually good at letting you make that smallest commitment first.

What the time game leaves you with

• There is no global patent — every patent is country-specific, and international strategy is really about how to file in many countries efficiently
• The Paris Convention gives you 12 months from your first filing to add other countries while keeping your original priority date
• The PCT extends that decision window to roughly 30 months from the first filing — at the cost of a $2,000 to $3,000 PCT fee
• Together, these windows give you about 2.5 years of validation time before you have to commit real money to specific countries
• For most founders, the right pattern is "file at home cheaply, then PCT, then enter only the countries that earn it" — not "file everywhere on day one"
• Defensive publication is a free alternative — give up your own patentability in exchange for blocking everyone else's
• For a Vietnam-based founder, total filing cost is unusually low (~$1,500 to $3,000 for a Vietnam patent), making the home-country-first pattern especially efficient
• Vietnam enforcement is weaker than the US, so VN patents are most valuable for licensing, fundraising, and acquisition rather than litigation
• The international system is designed around deferring expensive decisions, not making them — almost every fee can be pushed several months later than founders assume
• The system stops helping when there's no real market on the other side of a border — international patents only earn their cost when international demand is real

The part I want to carry out of this series is the bottom item above. Patents at every level — domestic, international, defensive — work the same way. They earn their keep when there's something real on the land they fence in. The market is always the harder problem. The patent system, surprisingly, gives you years to figure out if you have one.

Treaties, fees, and country guides

• Paris Convention for the Protection of Industrial Property — WIPO text (Article 4) — Used for: the 12-month priority window and the principle of national treatment across signatory countries.
• PCT Article 22 and WIPO time-limits page — Used for: the 30-month national-phase deadline and the handful of countries still on the 20-month version.
• WIPO PCT Yearly Review 2025 — Used for: the 698,500 national phase entries figure and the conversion-rate context in the counterargument.
• WIPO PCT Applicant's Guide — Used for: the PCT timeline (search, publication, optional examination) and country coverage.
• USPTO Fee Schedule — Used for: US non-provisional filing costs and PCT-related US fees.
• Manny Schecter interview on IBM's patent strategy (IPWatchdog, 2013) — Used for: the "position of strength" framing in the counterargument and IBM's 50,000-patent portfolio scale.
• Vietnam Patent & Trademark — case study on lost priority rights — Used for: the Thai-applicant 38-application case where prior Indonesian filings destroyed novelty.
• Viet An Law, "Register a Patent in Vietnam Using a PCT Application" — Used for: Vietnam national phase entry process and local agent requirements.
• IP-Coster, "PCT Vietnam Guide" — Used for: Vietnam government fee ranges and timeline-to-grant figures.
• Vietnam IP Law (Law on Intellectual Property, current consolidation) — Used for: 20-year patent term, maintenance fee structure, and enforcement context.