Which Country for TikTok? Does SEO Still Work?

A Full Cross-Border E-Commerce Scan for a Brick-and-Mortar Factory

Prologue: Three Questions from an OEM Owner

the OEM client owns a large OEM garment factory in South China. He has been at it for twenty years.

Pajamas, casual trousers, yoga pants—hundreds of thousands of units cycle through his workshops every month, shipped to every corner of the world. Yoga-pant orders from Shenzhen's big Amazon sellers, traders in North Africa, buyers in Israel, distributors in South America—the customers he serves are all the "person-to-person" kind.

Twenty years ago he worked the office towers of Guangzhou's foreign-trade district, knocking on doors floor by floor, handing out business cards one at a time. Half of the buyers he caught back then are still on his client list today.

A few years ago, Alibaba International's sales reps got into his factory—and then he showed them the door. "Homogenization, price wars, platform commissions, rule risk…" he counted them off one by one.

In the spring of 2026, a friend told him that TikTok was on fire in Spain.

So he came to me with his questions and his confusion:

• How do you launch TikTok in Europe? How do you register a European business license?
• If I do TikTok Spain, can a mainland-China identity open a store?
• North Africa, Israel—if they're allowed, can I do those too?

The three questions sound like a multiple-choice problem.

But run the whole questionnaire from end to end—six rounds of online research, cross-checked against official documentation, against post-mortems from cross-border practitioners, against the EU's new compliance rules—and you find that the OEM client isn't facing an A/B/C/D to pick from.

the OEM client isn't choosing a TikTok. He's looking for a map.

This article is about drawing that map—the 2026 cross-border e-commerce map.

I. The Real Map of TikTok Shop

The first task is a course correction.

When the OEM client heard "TikTok is on fire in Spain," the picture in his head was: Spaniards are scrolling TikTok, which means he can open a little TikTok store in Spain, ship out small sample batches, keep going if it works, walk away if it doesn't.

The first step of that reasoning is right (Spaniards are indeed scrolling TikTok). The second step is wrong.

TikTok has users in 150 countries. TikTok Shop's Seller Center is officially open for local stores in only about 16.

Note on measurement: "16 countries" is an approximate count based on the public market list. TikTok has never published a fixed number, and the exact figure shifts in small steps with policy.

Here is the approximate list of countries where TikTok Shop had opened seller onboarding (local stores via Seller Center) as of Q2 2026:

Two additional categories of "non-local-store" markets require separate explanation:

• Middle East fully-managed cross-border model: Saudi Arabia (KSA) and the UAE have long operated as "fully-managed cross-border pools"—consumers can shop, and in 2024 Saudi Arabia averaged roughly 30,000 orders per day—but the local Seller Center is not yet independently open to local companies. TikTok's MENA regional headquarters opened in Riyadh in Q1 2025, and local stores are listed as an expansion priority, but as of 2026-04 they had still not officially launched as an independent channel.
• The other 4 GCC states (Qatar, Kuwait, Bahrain, Oman), Egypt, and Israel: visible to consumers or reachable cross-border, but none have a local store-opening channel.
• India: the main TikTok app has been completely banned since 2020, and in 2025 the Indian government reaffirmed there would be no lifting—TikTok Shop does not exist in India, and cannot.
• Pakistan: TikTok is available but Shop does not support local-entity registration; Pakistani businesses typically enter obliquely via UK or Indonesian entities.

Sources: TikTok Shop official Seller Center, TikTok Newsroom's 2025-06 Japan announcement, TechBuzz China MENA interviews, DD News's 2025 statement on the India ban, and dpl.company's official list tracking.

the OEM client's three target markets—Spain, Israel, North Africa—differ sharply in real feasibility:

• Spain: ✅ Open. A mainland-China entity can open a cross-border store.
• Israel: ❌ Not open. In 2023 there were rumors of an early-2024 launch that never materialized; as of 2026-04 there had been no official activation announcement.
• North Africa (Morocco, Egypt, Algeria, Tunisia, Libya): ❌ All closed. Morocco and Egypt are TikTok-user heavyweights (Egypt has over 40 million), but TikTok Shop has no local merchant side.

The first answer to the OEM client's three-country question: Israel and North Africa have no TikTok Shop. It's not a question of whether you choose them—there is no option to choose.

Being on fire is one thing. Opening for business is another. And cross-border stores versus local stores drag in yet another set of choices.

!Figure 1 · TikTok Shop's 2026 global open-market map

Figure 1 · TikTok Shop's 2026 global open-market map

II. Decoding the EU5 One-Store-Many-Countries Mechanism

Since Europe is doable, the second task is to understand how Europe is done.

In March 2025 TikTok Shop rolled out a European expansion policy that introduced a mechanism China's cross-border circle has underrated—EU5 one-store, many-countries.

You thought you could only open one Spanish store. Under the new rules, that qualification can extend to three more countries.

The specific rules (per the official expansion channel within EU5):

Correction on measurement: this is not "one onboarding automatically activates local stores in four countries." In reality—the main site opens first, and the other three countries must be manually activated in the back end (the qualification review has passed once, but each site still requires its own application); logistics, taxes, and pricing are still handled independently per country. Inventory and orders are visible together in the back end.

One compliance foundation can extend across the four EU5 countries—but it is not one-click, four-store.

What this mechanism means—

Spain alone has a population of 46 million. Germany 84 million + France 68 million + Italy 59 million.

ES + DE + FR + IT = 256 million people.

The same compliance investment (one business license, one VAT number, one local bank account, one EU responsible person) covers a market that scales from 46 million → 256 million, ×5.6.

Limits to keep in mind:

• Sub-stores by default activate only the main site; the other three countries require manual activation (though the qualification review has already passed once)
• Taxes, GPSR compliance, language, and pricing are still handled per country
• Cross-border stores (mainland-China entities) can also enjoy the EU5 mechanism, but their organic-traffic weighting is lower than that of local stores
• The Ireland site is by targeted invitation only and is not opened automatically by default

Sources: TikTok Shop's official ES/DE/FR/IT Seller Centers, TikTok Newsroom's 2025-03 Spain expansion announcement.

One compliance foundation can, in theory, cover 260 million people—one of the core dividends of the 2025 European expansion mechanism.

Reminder: "population covered" is a measure of potential market, not "sales automatically in hand"—each country still requires independent operation, localization, and ad spend.

!Figure 2 · The EU5 one-store-many-countries radiating structure

Figure 2 · The EU5 one-store-many-countries radiating structure (2025-03 new rules)

III. The Truth About TikTok Shop's Three Markets

Chapter One sketched TikTok Shop's global silhouette—which countries are open, which are not. Chapter Two explained the EU5 one-store-many-countries mechanism.

Now zoom in on the OEM client's three target markets—Spain, Israel, North Africa—and take apart the real feasibility at the TikTok Shop layer:

To open a TikTok e-commerce operation in these regions, how much up-front financial cost is required? How many potholes 🕳️ are there to step in?

(i) Spain: the only market where the OEM client can genuinely run a TikTok Shop

Basic facts:

Sources: TikTok Shop Spain Seller Center, EU Access2Markets, Hellotax EU VAT Guide 2026, Forest Shipping.

Three-scenario cost estimate (China-entity cross-border store + adult yoga pants/sleepwear):

⚠️ Note on measurement: the table below is a range estimate based on service-provider quotes + cross-border-seller community post-mortems—not an official quote and not universally applicable. Depending on category, stocking depth, shipping method, and compliance provider, actual costs may swing 30–50% either way. Use the table for order-of-magnitude reference only; do not treat it as a budget basis.

Key cost-saving items:

• Start with adult styles (yoga pants, adult sleepwear) to sidestep the EN 14878 flame-retardancy testing for children's sleepwear (saves €500–2,000)
• Grab the new-seller 60-day 4% commission window (vs. the standard 9%)
• Use sea LCL for the first batch instead of air freight (saves $1,100)
• Clear out low-value inventory below €150 before 2026-07-01 (the EU's new €3/unit tariff takes effect)

(ii) Israel: TikTok Shop is closed, but not entirely unsolvable

Israel currently has no TikTok Shop. The typical route for Chinese sellers is:

AliExpress + a Shopify independent store + TikTok Ads (driven by Hebrew-language creatives)

Key facts:

Sources: Trade.gov Israel eCommerce Guide, VATupdate, Times of Israel, Similarweb Israel eCommerce.

(iii) The five North African countries: only Morocco is relatively viable; the rest are high-barrier or not advised

⚠️ The "feasibility score" in the table below is a subjective scoring model, based on a combined judgment of policy, logistics, payments, and sentiment—not any official or third-party metric. For directional reference only; not a basis for decisions.

Sources: Jumia Morocco official, GOEIC Decree 43 (2016) latest revised edition, trade.gov country guides, UN Security Council Resolution 2819.

The truth about the OEM client's "three target markets": Israel has no TikTok Shop. In North Africa only Morocco is viable. Spain has one, but it isn't cheap.

Before betting on any market you "heard is on fire," ask one question first—is TikTok Shop actually open there?

IV. The TikTok Factory Content Account

The TikTok cross-border business is not just the TikTok Shop route.

There is a parallel track—the factory content account—that opens no store at all, just posts everyday TikTok videos to attract B2B inquiries, moving the entire deal-closing step back offline.

This track only truly matured in 2025.

The benchmark that cracked it is LC Sign (a Guangzhou factory making light boxes and signage), whose owner Tony Zhu majored in business English, appears on camera himself, and does factory-comedy skits mimicking American/Mexican/Thai accents.

The data:

⚠️ Note on the conversion figure: this is public data from a single LC Sign account at a specific stage, not a universal industry benchmark. The vast majority of factory content accounts, in their launch phase, are closer to "1,000 views = 0 inquiries." Treat this figure as a ceiling reference from a top account that has cracked it, not a level a new account can expect.

Sources: Rest of World's 2024 feature "Chinese factory owners turn to TikTok," MIT Technology Review 2025-04, Jungle Communications' 2025 case analysis, TikTok accounts' public data.

LC Sign is not "one account gone viral" but a multi-account matrix strategy—7 TikTok accounts each cut a different category, a different angle, a different rhythm, jointly lifting traffic. Once a single account hits its ceiling, the matrix approach became the natural choice. This is the key signal of the factory content account evolving from "single star account" to "content industrialization."

Another example is a smart-TV factory in Jiangxi, @factory_in_china, with 484,000 followers and a peak of 18.2 million views on a single video, running a single-account route.

LC Sign isn't running ads. It's letting the whole world watch its factory.

The content form of these accounts is nothing like a "sell-you-goods ad":

• Sample-making workshops
• Raw-material warehouses
• Workers eating lunch
• The boss smoking
• Packing and shipping

No "Buy Now." No "Limited Time Offer." Only "who I am, where I am, what I'm making"—a form of genuinely-real, personified proof. As for why this kind of proof is especially valuable in 2026, Chapter Five will make it clear.

Some realities on the ground:

• The US region is the main battlefield: the US has TikTok's largest user base, English is the world's business lingua franca, and the apparel category has few public success stories—which means a blue-ocean opportunity.
• A real person must appear on camera: an AI-voiced factory account gets only 30–50% of the traffic of a real-voiced one. The algorithm applies an invisible weight to "human authenticity."
• Account region-binding is strict: TikTok's algorithm makes a fivefold joint determination—IP, SIM, time zone, language, App Store account country—of an account's "physical birthplace." Running a US-region account from a Chinese IP is nearly impossible to scale. Every player who has cracked it uses a US residential IP + a US phone number + a real person speaking English on camera.
• "1,000 views ≈ 1 inquiry" is a top-account ceiling, not a new-account baseline: LC Sign's numbers are widely cited in the cross-border circle, but the vast majority of accounts come nowhere near it in their first 6–12 months. If views don't rise, inquiries naturally don't come—but even if views rise, inquiries don't necessarily follow; content personification and category fit matter just as much.

Amazon sellers fight for the first page. TikTok factory owners fight for "human authenticity."

For the OEM client, the temptation and the obstacles on this road are equally concrete:

• Temptation: authentic factory content has zero material cost, and 20 years of B2B experience is a natural content goldmine.
• Obstacle: his own English level / his willingness to appear on camera / technical infrastructure like a US residential IP / the content output capacity of 3 videos a week.

!Figure 7 · The LC Sign model · The factory-TikTok conversion baseline

Figure 7 · LC Sign's 148K-follower-stage figures: 1,000 views ≈ 1 inquiry · 3–10% inquiry-to-order (single-account, specific-stage data, not an industry benchmark)

V. The Cognitive Hub: The Silent Rewrite of AI Search

Having come this far, we have already seen TikTok Shop's global map (Chapter One), the EU5 one-store-many-countries mechanism (Chapter Two), the real feasibility of the OEM client's three target markets (Chapter Three), and the LC Sign baseline of the TikTok factory content account (Chapter Four).

But behind all of it, something bigger is happening at the same time.

Something that has rewritten the underlying scoring system of four seemingly independent tracks: independent stores, SEO, TikTok content accounts, and LinkedIn.

In the past, when the cross-border circle discussed these topics, it treated them as four separate tracks. Starting in 2026, their underlying rules were rewritten by the same one thing—

Google Search became a display shelf for AI.

Let's take it apart step by step.

(i) The starting point: why Alibaba International got phased out of the old-school trade factory

the OEM client's factory took on Alibaba International sales reps a few years back. They worked together for a while, and then he showed the reps the door.

Four reasons:

• CPC is expensive: ¥20–50 per click
• Low conversion rate: 1–3%
• Cost per valid inquiry ¥500–2,000
• Homogenized price wars: the platform steers toward "whoever's cheaper wins," and the buyer's mindset is trained into that of a "wholesale reseller"

Sources: Jiuling Network, FOBShanghai forum's 2025 Alibaba International operations post-mortem reports.

This isn't Alibaba's fault. It's the structural failure of the "platform-stall model" in a B2B foreign-trade context.

The consensus in the old-school trade circle: independent-store inquiries are 3–5× better than platform inquiries—in quality, close rate, and customer caliber alike. The reason is simple—an independent store's inquiries come from a proactive action taken after multiple rounds of comparison on Google, not the stall-style "casting a wide net."

Those buyers the OEM client caught working the office towers of Guangzhou twenty years ago were "proactive action." Independent-store inquiries are "proactive action" too. They are the same thing.

(ii) The rules of the game changed: the hard data of AI Overview

But Google Search in 2026 is not Google Search in 2015.

Google AI Overview (also known as SGE, Search Generative Experience)—that AI-generated summary at the top of the results page—is rewriting the search game.

A few widely cited data sets (single-study measurements):

⚠️ Reliability note: different SEO-industry outfits (Seer / Semrush / Ahrefs / BrightEdge / ALM) differ enormously in test sample and methodology; the figures above are "single-study measurements" and have not yet formed an industry consensus. Treat them as indicators of directional trend (AIO is expanding, the organic clicks of uncited brands are taking a hit, and cited brands show a positive spillover)—do not treat the specific percentages as portable benchmarks.

Never mind the exact percentages; look at the directional meaning.

If your brand is not cited by AIO—your name isn't in that AI summary—multiple studies all point to a significant drop in organic CTR (single-study figure ~61%).

If your brand is cited by AIO—your name and link appear in the AI summary—multiple studies all point to a positive spillover in clicks (single-study figures: organic +35% / paid +91%).

The exact numbers vary with methodology; the direction of the trend is stable—AIO has reallocated search traffic to "the brands AI trusts."

SEO isn't dead. SEO became a new species.

!Figure 3 · Search click-through rates in the AI Overview era

Figure 3 · AI Overview's structural rewrite of organic and paid clicks

(iii) The essence: old game vs. new game

Before 2025, you were competing with Google for first place. Starting in 2026, you're competing with AI for citations.

Many people haven't yet made this cognitive turn.

!Figure 4 · SEO's old game vs. new game

Figure 4 · Six dimensions compared: from "keyword ranking" to "AI personified citation"

(iv) How to make AI willing to cite you

Whom AI Overview cites and whom it doesn't isn't decided by keyword density, nor by how many backlinks you have.

In 2024–2025 Google updated its internal algorithm documentation multiple times, distilling the mechanism into four letters:

E-E-A-T

The essence of these four letters is one question: AI wants to know whether you genuinely exist.

Why does AI need this?

Because in an age when large models like ChatGPT and Claude frequently "hallucinate," Google, Anthropic, and OpenAI all rank "verifiable genuine existence" as the top priority for AI-generated content. For AI Overview to cite a brand is for AI to vouch for that brand with its own reputation—it must confirm the brand isn't made up.

This is why Schema.org structured data has surged in weight since 2024:

Technical SEO in the AI era has gone from "let the search engine crawl you" to "let AI verify you."

The counterintuitive result is that AI actually prefers "small and real" accounts over "large and generic" mega-sites.

An independent store with only 5 articles—each with a real-person author headshot, a real case, and structured data—may have a higher chance of being cited by AIO than a content farm with 500 SEO articles all authored by "Admin."

AI doesn't look at how much you wrote; it looks at whether you're real.

Back to the OEM client—

His 20-year factory, his real workshops, his verifiable address, his old customers he has actually shipped to, the products he has sampled—none of this mattered to 2015's Google, but in the eyes of 2026's AI Overview it is all a goldmine.

the OEM client's "genuine existence" is itself an asset. The only question is how to digitize it.

This is also why independent-store SEO in 2026 is no longer about "writing keywords" but about "constructing an AI-verifiable chain of evidence of genuine existence"—a real-person author page, a Schema markup of the factory address, structured data for customer cases, traceable descriptions of product specs, authoritative external citations… every element answers AI's same one question: are you real?

(v) The B2B buyer's new decision chain

Now look at the buyer-side data.

Gartner and 6sense's joint research on the B2B purchase-decision chain shows:

• Before contacting sales, B2B buyers have already completed 70% of the decision
• They pass through an average of 27 touchpoints
• 80% of inquiries are buyer-initiated

Put this data back in the AIO context—

A buyer opens Google and asks "best pajama manufacturer Shantou low MOQ"—

2015: Google returns 10 blue links, and the buyer opens 3–5 to compare.

2026: Google returns an AI-generated summary that cites 3–5 brands. The buyer will only click those 3–5.

Most of the buyer's decision-funnel work is finished before he ever contacts you. The moment he does contact you, he has already seen you in AI's citation shelf.

Or he hasn't.

!Figure 5 · The B2B buyer's decision chain, 2015 vs. 2026

Figure 5 · From a linear funnel to 27 distributed touchpoints

(vi) What the hub means

Once you understand the five sub-sections above, look back—independent stores, TikTok factory content accounts, LinkedIn, even the front-end search of TikTok Shop—these four seemingly independent tracks:

• Independent store: you want it to be cited by AI. Through genuine existence, authoritative content, structured data.
• TikTok factory content account: you want it to prove you genuinely exist. The factory workshop, the boss himself, the sample-making masters, the shipping scenes—this is the deeper explanation, in the AI era, of the LC Sign phenomenon we saw in Chapter Four: why the algorithm applies an invisible weight to "human authenticity"—because AI's trust system needs this personified proof.
• LinkedIn: you want it to verify your real identity. A real headshot, a real company, a real résumé—these are the external chain of evidence that AI trusts.

The three tracks—independent store, TikTok, LinkedIn—look independent. But they now share one scoring system: does AI believe you.

This is AI search's silent rewrite—it rewrites not just SEO's rules but cross-border e-commerce's entire trust infrastructure.

!Figure 6 · AI search: the new hub of cross-border e-commerce

Figure 6 · Four tracks look independent · in reality they share one scoring system: does AI believe you

VI. Two Truths of the Gray Market

The cross-border circle has an open secret: you can buy someone else's TikTok account to sidestep real-name verification, entity qualifications, and compliance review.

China has plenty of brokers selling TikTok accounts and stores for various countries. That's a fact.

But the structure of the gray market is completely different across markets—that is the core of this chapter.

(i) The US region: a mature gray ecosystem, and an underrated price

The US gray market is the world's most mature. Public pricing (2026 Q1):

Sources: Zhihu's cross-border e-commerce board, Ebrun, Extrabux's social-account trading platform guide, TKFFF cross-border seller community.

The problem is—being able to buy one doesn't mean being able to run one.

At the end of 2024, US TikTok Shop launched a large-scale purge:

⚠️ Reliability note: the closure rate and survival rate are aggregated post-mortems from cross-border community practitioners, not statistics from any official or independent third-party audit. The specific numbers may be over- or under-estimated, but the direction—"high closure rate, low survival rate, tightening re-review"—is cross-validated by multiple independent sources. Treat them as qualitative conclusions; do not use them as precise probabilities.

Sources: Sina Finance, Jiemian News, cross-border e-commerce practitioners' post-mortems on Zhihu.

Now do a rough calculation (based on the community figures above, for order-of-magnitude reasoning only):

• One-time input ¥100–160K (store + first stocking batch + logistics)
• 6-month survival probability (community figure) <20%
• After failure: goods payment frozen + stocking sunk + store dead

If you estimate at a <20% survival rate, reaching an 80% probability of "cracking it once" requires 2–3 attempts, and cumulative input could land in the ¥300–480K range—already approaching the entry-level cost of the legitimate route.

⚠️ This is a hypothetical extrapolation based on community figures, not measured data. Individual variation is enormous—some crack it on the first try, some fail five times. The extrapolation is meant to illustrate the "expected-cost" way of thinking, not to serve as a budget basis.

The part of the logic that holds: the gray route is cheap on paper, but once you layer in the closure rate, the expected cost is not necessarily lower than the legitimate route—this is a thinking framework, not a definite conclusion.

!Figure 8 · The illusion of the gray market being "cheap"

Figure 8 · On-paper cost vs. expected cost with the failure rate factored in · US TikTok Shop

(ii) The EU: the counterintuitive truth—the open market simply doesn't exist

This is the most counterintuitive finding across the six rounds of research.

The Chinese cross-border circle (Zhihu, Telegram, cross-border forums, AMZ123, Cifnews)—turned up not a single stable public quote for a "Spain/EU TikTok Shop live store."

Completely unlike the US ecosystem of "mature pricing, account dealers everywhere, 3,000 accounts sold a month."

Why?

There are three structural reasons:

Reason one: GPSR traces liability back to a natural person

The EU's GPSR (General Product Safety Regulation), effective 2024-12-13, stipulates that a non-EU seller must designate a Responsible Person within the EU whose name and address must be publicly printed on the product detail page.

This responsible person bears administrative and criminal liability for product safety, recalls, and consumer complaints.

Not "the company takes the fall," but "a natural person takes the fall."

The EU's GPSR responsible-person name and address, publicly printed on every product detail page—this isn't a compliance requirement, it's a "notice board for assigning blame."

Reason two: EU VAT fraud is a criminal offense

The tax authorities of Spain, Germany, and France audit VAT fraud extremely strictly. Once a shell entity is investigated, the natural person acting as nominee bears joint and several criminal liability. TikTok's parent company has already been fined 500+ million euros by the EU in 2023–2025. That sensitivity has propagated to the gray market—the number of people willing to act as nominees has plummeted.

Reason three: EU5 "one-store-many-countries" is a new rule that only landed in 2025-03

The gray-market supply chain needs time to develop. From 2020 to 2024, it took the US TikTok Shop gray market four years to form a mature ecosystem. This EU rule set only fully got going in 2025-03—the gray-market product line hasn't had time to grow yet.

The result of stacking all three reasons:

The EU gray market is more likely to appear in the form of "compliant nominee holding"—a shell company in Poland / Ireland / the Netherlands + a nominee VAT + a proxy RP. Expensive, illiquid, highly concentrated—not the US "account dealer" model.

For the buyer, the type of risk is upgraded from "platform account closure" to "criminal liability for a natural person."

In the US, gray-market operations are a "might-lose-money" business. In the EU, they're a "might-go-to-prison" business.

Why doesn't the EU gray market exist?

Because the people bold enough to try got criminalized before they had the chance.

VII. The Four Sets of Numbers Behind the Entity Route

If you decide to take the legitimate route, the first question is: in whose name do I open the store?

The common entity routes on the market:

• 🇺🇸 US LLC (Delaware / Wyoming)
• 🇮🇪 Ireland Ltd
• 🇩🇪 Germany GmbH
• 🇪🇸 Spain SL

Three-year-bill comparison (calculated for a "small-to-mid-scale cross-border seller" scenario, non-audit measurement):

⚠️ Reminder: the table below is based on public service-provider quotes + cross-border community cases. Quotes vary enormously between handling agents (a 2–3× range is common). Annual cost is heavily affected by accounting, auditing, and whether compliance events are triggered; actual numbers may swing 30–50% either way. Use for horizontal order-of-magnitude comparison; do not take directly as a quote.

Sources: Stripe Atlas official, Company Bureau Ireland, OpenForest, CostaLuz Lawyers Spain, Commenda Ireland Incorporation Cost 2026, Lextax Spain SL Guide.

Once the three-year cost numbers are laid out, some counterintuitive facts appear:

Counterintuitive point one: a US LLC is the cheapest cross-border entity

The $2,500–4,000 three-year combined cost (about ¥18–28K) is only 1/3 to 1/10 of an EU entity. The reasons:

• Stripe Atlas $500 all-in-one (LLC + EIN + Mercury + Stripe)
• Delaware is non-resident-friendly, with an annual Franchise Tax of just $300
• The single-member LLC + Form 5472 compliance framework is mature

The qualification threshold for a US TikTok Shop local store—that's about all it costs.

Counterintuitive point two: an Ireland Ltd is "the long way around" for TikTok EU5

The Ireland Ltd is recommended in many cross-border articles as the "EU value-for-money king." Here's the correction—if your goal is a TikTok Shop EU5, an Ireland Ltd is actually the long way around.

Three reasons:

• The mainstream choice in China's cross-border circle for TikTok Shop EU5 is a Germany GmbH, Spain SL, or Netherlands BV (the Netherlands BV isn't EU5, but it's heavily used among EU cross-border sellers)
• TikTok Shop EU5's "one store spans five countries" has a native top-push set of DE/ES/FR/IT
• The Ireland site is by targeted invitation only; if an Irish entity wants to join TikTok Shop EU5, the main site must be DE/ES/FR/IT (not Ireland)

The Ireland Ltd's real value scenario lies elsewhere:

• ✅ Stripe Ireland: Stripe's EU clearing entity is in Dublin, so an Irish entity opens Stripe most smoothly
• ✅ Amazon EU unified entity
• ✅ 12.5% corporate tax + English-language environment
• ✅ A future UK–EU dual bridgehead

A US LLC is a universal cross-border ticket. An Ireland Ltd is an EU multi-platform pass. A Germany GmbH is the native TikTok Shop EU5 ticket. A Spain SL is a single-market Spain specialty ticket. Three tickets do three jobs—none is a universal one.

Counterintuitive point three: a company from a non-EU5 country can't open an EU5 local store

• Netherlands BV / Portugal Lda / Malta / Cyprus: non-EU5, no eligibility for a TikTok Shop local store
• Estonia e-Residency: restricted for Chinese passport holders since 2025 (three years of EEA residency required); in practice there are cases of an OÜ being rejected when opening an ES local store
• These entities still have value for cross-border SaaS, Stripe, and Amazon EU, but they are not the answer for TikTok Shop

There is no best entity. There is only the entity determined by "where you want to sell."

!Figure 9 · The four sets of numbers behind the entity route

Figure 9 · US LLC / IE Ltd / DE GmbH / ES SL three-year combined cost, compared

VIII. Account × Device × IP: A Thin, Passing Glance

This layer is the technical truth of the operational level.

Let me offer just a restrained bit of the concept:

• A TikTok account has its "physical birthplace" jointly determined by five signals: IP, App Store account country, SIM card, time zone, system language
• The regional weight of the For You Feed went from 29% in 2024 → 47% in 2026—the algorithm keeps getting stricter
• Commercial VPNs have a detection rate of >99.7%
• 1 device = 1 account = 1 IP is the compliance principle (sharing one device across multiple accounts = 100% linkage and closure)
• TikTok's region-switch detection triggers a graduated ladder of penalties: CAPTCHA → traffic throttling → shadowban → closure

The players who have cracked this track all use a US residential IP, a US phone number, and a dedicated device.

On this layer, we'll leave it there. The technology isn't complex; compliance is.

And an honest reminder:

The rules of this layer change every three months. A tutorial is out of date the moment it's written.

IX. Data-Reliability Statement: Where This Map Can Be Trusted, and Where It Can't

Having written this far, it's worth laying the report's "data undertone" on the table, face-up.

This is an industry-extrapolation report, not an audit or an academic paper. Its data comes from three kinds of sources—official documentation (high credibility), cross-border community practitioners' post-mortems (medium credibility), and stitched-together service-provider quotes (credibility depends on the specific provider).

Layered by credibility:

What this map can do:

• ✔️ Establish a cognitive framework and see clearly the several mainstream routes of 2026 cross-border e-commerce
• ✔️ Understand the trends (AIO rewriting SEO, EU5 expansion, the rising marginal cost of the gray market)
• ✔️ Find the direction worth digging into next

What this map cannot do:

• ❌ It cannot serve directly as a budget basis
• ❌ It cannot be the final judgment on market selection
• ❌ It cannot be an "investment-decision document" for any specific route

Before any real decision, every route needs a fresh "on-the-ground price check": get specific provider quotes, find post-mortems from peers who have cracked it, find the latest official version of the policy.

A map is for setting direction. Budget and decision are made once you've traveled near the destination.

Epilogue: The Map Is Still Being Drawn

the OEM client's first question was "how do I register a European business license? Should I open a Spain TikTok?"

After six rounds of research, that question still has no single "right answer."

But we can at least say this much clearly:

• TikTok Shop's Seller Center is officially open for local stores in only about 16 countries (plus the fully-managed cross-border model of Saudi Arabia/UAE)—of the OEM client's three target markets, neither Israel nor North Africa has a local store-opening channel
• EU5 qualification can extend across four countries (2025-03 new rule)—one compliance foundation can, in theory, cover a 256-million-person potential market, but beyond the main site, the others require manual activation—not one-click for four
• AI Overview has rewritten the search game—SEO isn't dead; it has become a game of "being cited by AI" (specific percentages vary by outfit, the trend direction is stable)
• The TikTok factory content account is the short-video-side embodiment of an "AI personified trust asset"—LC Sign is a top account that has cracked it; "1,000 views = 1 inquiry" is that account's specific-stage figure, not a new-account baseline
• The EU gray market barely exists—GPSR's natural-person criminal liability criminalized the bold before they even acted (a structural judgment)
• The four sets of numbers behind the entity route (orders of magnitude)—US LLC ¥18–28K / Ireland Ltd ¥78–184K / Germany GmbH ¥80–150K / Spain SL ¥100–200K, with wide variation between providers
• TikTok's account region-binding rules keep tightening—the algorithm's regional weight went from 29% → 47%, and a tutorial is out of date the moment it's written

This map is far clearer than it was before the OEM client asked, but it is far from finished.

The 2026 world of cross-border e-commerce is at a moment when multiple layers of rules are being rewritten at once—platform rules, search rules, compliance rules, payment rules, logistics rules—and no single party has a complete blueprint for the future.

What six rounds of research drew is not an answer, but a coordinate system.

The market waits for no one to finish drawing the map. But at least it's better than having no map.

【Data as of: 2026-04-17】