The Shell Game
TI-2026-050C ยท Letter 3 of 5 ยท Page 299Follow the IP addresses and you find VPS providers. Follow the VPS providers and you find shell companies. Follow the shell companies and you find the Seychelles.
The Go_SSH fleet's 320 IPs are not randomly distributed. They arrange themselves into three distinct tiers โ each serving a different operational purpose, each occupying a different risk position, each existing in a different jurisdictional reality. What initially appeared to be a swarm of independent European VPS nodes is, on closer inspection, a structured network of offshore entities operating through bullet-proof hosting providers.
I. The Map
320 IPs. 43 countries. But the distribution is not random.
| Region | Go_SSH IPs | % | libssh IPs | % | Contrast |
|---|---|---|---|---|---|
| Western Europe | 112 | 35.0% | 85 | 3.6% | 9.7ร over-represented |
| North America | 71 | 22.2% | 125 | 5.3% | 4.2ร over-represented |
| Southeast Asia | 49 | 15.3% | 250 | 10.7% | ~comparable |
| East Asia | 24 | 7.5% | 540 | 23.1% | 3.1ร under-represented |
| South Asia | 21 | 6.6% | 100 | 4.3% | ~comparable |
| South America | 7 | 2.2% | 50 | 2.1% | โ |
The contrast is stark. The Go_SSH fleet is 9.7 times more concentrated in Western Europe than the libssh swarm. Western Europe โ France, Germany, UK, Spain, Finland, Italy, Netherlands โ hosts 35% of Go_SSH IPs but only 3.6% of libssh nodes. Conversely, East Asia (China, Korea, Japan, Hong Kong) hosts 23.1% of libssh but only 7.5% of Go_SSH.
Two operations. Two completely different geography. Two different procurement ecosystems.
The Country Leaders
| Rank | Country | Go_SSH IPs | % | Notable |
|---|---|---|---|---|
| 1 | ๐บ๐ธ United States | 68 | 21.3% | DigitalOcean dominant. Multiple cloud providers. |
| 2 | ๐ซ๐ท France | 43 | 13.4% | Absent from libssh. OVH/Scaleway ecosystem. |
| 3 | ๐ฉ๐ช Germany | 28 | 8.8% | Contabo, Hetzner, IONOS. Budget VPS capital. |
| 4 | ๐ฌ๐ง United Kingdom | 23 | 7.2% | w1n ltd (AS42237). Small hosting operators. |
| 5 | ๐ฎ๐ณ India | 21 | 6.6% | Broadband + cloud mix. Also in libssh. |
| 6 | ๐ธ๐ฌ Singapore | 18 | 5.6% | Cloud hub presence. Regional node. |
| 7 | ๐ป๐ณ Vietnam | 16 | 5.0% | Growing hosting market. Cheap VPS. |
| 8 | ๐ฐ๐ท South Korea | 13 | 4.1% | Korea Telecom. Residential + hosting. |
| 9 | ๐ฎ๐ฉ Indonesia | 8 | 2.5% | High total_hits (382). Persistent scanning. |
| 10 | ๐ง๐ท Brazil | 7 | 2.2% | Regional presence. Growing cloud market. |
France at #2 is the headline. Forty-three IPs โ 13.4% of the fleet โ on French hosting infrastructure. France is essentially absent from the libssh swarm. This is not random variation. This is a procurement preference. The Go_SSH operator has access to French hosting markets that the libssh operator does not โ or chooses not to โ use.
Why France?
France offers a specific combination: GDPR privacy protections that slow abuse responses, major hosting providers (OVH is Europe's largest, Scaleway is growing), competitive pricing (โฌ3-5/month for basic VPS), and excellent network connectivity to Europe and Africa. OVH alone operates 400,000+ servers globally. For a scanning operation that needs cheap, disposable European IPs, France is the optimal market. The Go_SSH operator knows this market well enough to exploit it at scale.
II. The Three Layers
The geographic data reveals structure. The Go_SSH fleet is not a flat collection of equivalent nodes. It has tiers โ three distinct layers of infrastructure, each with different characteristics, different risk profiles, and different operational roles.
Three shell companies registered in offshore jurisdictions. They provide the persistent infrastructure โ the nodes that survive when scanning IPs get burned. These entities generate 47.5% of all observed sessions from just 8 IPs.
109 sessions ยท 4 IPs ยท Recon role ๐๏ธ Omegatech LTD ยท AS202412 ยท ๐ธ๐จ Seychelles
28 sessions ยท 3 IPs ยท 100% success ๐๏ธ Private Layer ยท AS51852 ยท ๐ต๐ฆ Panama
53 sessions ยท 1 IP ยท 100% success
Small companies and individuals providing dedicated scanning resources. Near-100% success rates suggest they target only pre-validated hosts. They sit between the disposable fleet and the offshore core.
37 sessions ยท 1 IP ยท 94.6% success ๐ค ISAEV Igor ยท AS200730 ยท ๐ฐ๐ฟ Kazakhstan โ ๐ต๐ฑ Poland
35 sessions ยท 1 IP ยท 100% success
Major cloud providers. High IP count, low sessions per IP. This is the scanning tier โ 280+ IPs on DigitalOcean, Contabo, Hetzner, IONOS, Google Cloud, Amazon, and others. Each IP scans a few targets, gets flagged, gets replaced. Cost: โฌ5-10/month per node.
47 IPs ยท Few sessions each โ๏ธ Contabo ยท AS51167 ยท ๐ฉ๐ช
33 IPs ยท Budget VPS โ๏ธ Hetzner ยท AS24940 ยท ๐ฉ๐ช
18 IPs ยท Cloud instances โ๏ธ IONOS ยท AS8560 ยท ๐ฉ๐ช
9 IPs ยท German hosting
The Economics of Each Layer
Layer 3 costs roughly โฌ1,500-3,000/month (300 VPS ร โฌ5-10). This is the cost of anonymity โ disposable probes that absorb blocklisting and abuse reports. Layer 2 costs less in hosting but more in trust โ these dedicated operators must be reliable, consistent, and discreet. Layer 1 costs the most to establish (registering shell companies, obtaining ASNs, setting up RDAP records) but is the most durable โ offshore entities can operate for years without takedown.
III. The Seychelles Connection
Two companies. Same country. Same IP space. Different names.
ASN: 197170 (TECHTIES-AS)
Registration: Seychelles (SC)
RDAP: techties-mnt ยท NL
Abuse: abuse@tech-ties.net
IP Space: 91.92.40.0/24, others
Role: Recon (mega-command execution)
Omegatech LTD
ASN: 202412 (OMEGATECH-AS)
Registration: Seychelles (SC)
RDAP: NL (Netherlands)
IP Space: 91.92.x.x, 94.154.x.x
Role: Validation (100% success)
Both Seychelles-registered. Both with Dutch RDAP records. Both operating in the 91.92.x.x address space. The IP allocation overlap is the strongest indicator: IP address blocks are assigned by Regional Internet Registries (RIPE, in this case) to specific organizations. Two different companies operating in the same /16 prefix (91.92.0.0/16) suggests either a shared upstream provider or โ more likely โ related registrations that obtained adjacent allocations.
The Matryoshka Pattern
In offshore finance, it's common to create multiple shell companies in the same jurisdiction to layer ownership and compartmentalize risk. If Company A is investigated, Company B continues operating with its separate ASN, separate RDAP, and separate abuse contact. The Go_SSH operator appears to follow this model: multiple Seychelles shells providing redundancy and operational continuity. If TechTies is taken down, Omegatech continues. If both Seychelles entities are disrupted, Private Layer in Panama provides geographic redundancy.
Why the Seychelles?
The Republic of Seychelles โ 115 islands in the Indian Ocean, population 100,000, GDP $1.7 billion โ has become a preferred jurisdiction for bullet-proof hosting companies. Why:
โข Minimal corporate transparency. The Seychelles International Business Companies Act allows registration with minimal disclosure. Beneficial ownership is not publicly recorded. Directors can be nominees.
โข No data retention requirements. Unlike EU countries (which require ISPs to retain data for 6-12 months), the Seychelles has no mandatory data retention. There is nothing to seize.
โข Limited cooperation with EU/US law enforcement. While the Seychelles has mutual legal assistance treaties (MLATs) with some countries, enforcement is slow and under-resourced. A takedown request that would be processed in days by a German court takes months โ if it's processed at all.
โข Low registration cost. An International Business Company can be registered for under $1,000. An ASN can be obtained through RIPE for a few hundred euros. The total cost to establish a bullet-proof hosting shell in the Seychelles is under $5,000.
The Pattern Repeats
TechTies appeared in the libssh swarm (TI-2026-049F) with 35 IPs. Now it appears in Go_SSH with recon duties. Omegatech appears in Go_SSH with validation duties. Both share IP space and jurisdiction. This is not two isolated hosting decisions โ this is an infrastructure strategy built on Seychelles corporate law and European IP allocation.
IV. The Panamanian Anchor
Private Layer INC deserves separate attention. Panama-registered, operating from Switzerland, 53 sessions from a single IP โ all successful. A 100% success rate from a single node means this IP never attempts a connection it won't win. It has advance knowledge of which hosts accept which credentials.
Private Layer is a known entity in the bullet-proof hosting ecosystem. Unlike TechTies and Omegatech (which appear to be purpose-built shells), Private Layer is a commercial service โ a hosting provider that markets itself on privacy and resilience. Its Swiss infrastructure provides EU-adjacent performance with Swiss data protection law.
The Jurisdictional Triangle
Seychelles for corporate registration (opaque ownership, no disclosure). Netherlands for RDAP/IP allocation (RIPE membership, EU peering). Panama for operational redundancy (different legal system, different continent). Switzerland for infrastructure hosting (neutral jurisdiction, strong data protection). Four countries, four legal systems, four sets of courts that would need to coordinate to take the operation down simultaneously. This is not ad hoc โ this is legal engineering.
V. The Individuals
Layer 2 contains two entities that don't fit the corporate shell model: w1n ltd (UK) and ISAEV Igor (Kazakhstan).
w1n ltd โ The UK Small Operator
UK-registered company. AS42237. A single IP in Sweden generates 37 sessions with 94.6% success. The UK registration provides banking access in pounds sterling โ useful for European payment processing. The Swedish hosting provides good connectivity and Swedish privacy law (one of the strongest in the EU). This entity bridges UK financial infrastructure with Nordic hosting.
ISAEV Igor โ The Personal ASN
An individual named Igor Isaev. Kazakhstan. AS200730. A single IP in Poland. 35 sessions, 100% success. This is the most anomalous entity in the fleet: a personal ASN registered to an individual, not a company. Personal ASN registrations require providing real identity to the Regional Internet Registry (RIPE). Either this is a real person whose ASN has been co-opted, or it's a straw man registration โ a name used for ASN procurement by someone else.
The Kazakhstan-to-Poland routing is notable: the ASN is registered in Kazakhstan but the infrastructure is hosted in Poland. This geographic split between registration and hosting mirrors the Seychelles-to-Netherlands pattern of TechTies and Omegatech.
Pattern: Registration โ Hosting
TechTies: registered Seychelles โ hosted Netherlands
Omegatech: registered Seychelles โ hosted Netherlands/Germany
Private Layer: registered Panama โ hosted Switzerland
ISAEV: registered Kazakhstan โ hosted Poland
w1n: registered UK โ hosted Sweden
Every entity separates its corporate registration from its physical hosting. This is deliberate: it forces any legal action to proceed in two jurisdictions simultaneously โ one for the corporate records, one for the server seizure. Neither jurisdiction alone can take down the operation.
VI. The Inverted Pyramid
The session data reveals something counterintuitive about the Go_SSH fleet's structure:
Layer 2 (2 small operators, 2 IPs): 72 sessions (18.0%)
Layer 3 (280+ cloud IPs): 138 sessions (34.5%)
The fleet has 280+ cloud IPs but they generate FEWER sessions than 8 offshore IPs.
Layer 3 is high-IP, low-activity. Layer 1 is low-IP, high-activity.
The pyramid is inverted: fewer nodes do more work.
This inversion makes sense through the two-phase architecture lens. Layer 3 performs Phase 1: try a credential, disconnect, move on. Each IP touches each target once โ low session count per IP, distributed across 280 nodes. Layer 1 performs Phase 2: revisit validated hosts with the mega-command. Each IP returns to many hosts โ high session count, concentrated on a few IPs.
The disposable fleet does the dangerous work (trying passwords that might fail, getting blocklisted). The offshore core does the valuable work (profiling hosts for exploitation). Risk and value flow in opposite directions.
VII. The Absence of China
The libssh swarm has 228 Chinese IPs โ 9.7% of its fleet. The Go_SSH fleet has three. 0.9%.
This absence is as informative as any presence. Chinese cloud infrastructure โ Alibaba, Tencent, Huawei Cloud, UCloud, ByteDance โ is abundant, cheap, and used extensively by the libssh operator. The Go_SSH operator avoids it entirely. Three IPs from a fleet of 320 is statistical noise, not a procurement strategy.
Possible explanations:
1. Different operator nationality. Chinese cloud platforms often require Chinese phone numbers, Chinese payment methods, or Chinese identity verification. A European operator cannot easily procure Chinese cloud VPS. The Go_SSH operator's European focus suggests they operate from โ or at least bank in โ Europe.
2. Deliberate avoidance. Chinese cloud providers are increasingly cooperative with international law enforcement (especially for cybercrime involving non-Chinese targets). An operator aware of this might deliberately avoid Chinese infrastructure to reduce exposure.
3. Different supply chain. The Go_SSH operator's procurement chain runs through European markets (OVH, Contabo, Hetzner, DigitalOcean) and offshore shells (Seychelles, Panama). This is a self-contained ecosystem that doesn't need Chinese cloud.
The Geographic Negative Space
What a threat actor does NOT use is as revealing as what they use. The Go_SSH operator avoids: Chinese cloud, Russian hosting, Middle Eastern infrastructure. They concentrate in: Western Europe, North America, Southeast Asia, and offshore jurisdictions. This profile โ European VPS, Seychelles/Panama shells, absence of Chinese/Russian infrastructure โ is consistent with a professional operation based in Western or Central Europe with banking access in euros.
VIII. The Thread Forward
The shell game reveals itself. Three offshore entities. A jurisdictional triangle spanning Seychelles, Panama, Netherlands, and Switzerland. A three-layer infrastructure that separates risk from value. France as an unexpected node in the geographic map. And the absence of China as a negative indicator of the operator's identity.
The next letter steps back from the Go_SSH fleet to examine what happens when its validated credentials are used. 333 hosts profiled. Some with GPUs. Some abandoned for months. Some are honeypots (including ours). The mega-command outputs went somewhere โ to a triage system that sorts hosts by value and queues them for exploitation. What does that exploitation look like? What payloads follow? What does the operator do with a GPU server that no one is watching?
The shell game shows us who operates the fleet. The next letter shows us why.
TI-2026-050C ยท The Shell Game ยท Series: The Toolmakers (3/5)
Classification: TLP:GREEN ยท Confidence: HIGH
Sources: Cowrie honeypot sessions, RIPE RDAP records, ASN allocation data, campaign geographic analysis
Cross-references: TI-2026-050B, TI-2026-049F, TI-2026-049C
Published: June 30, 2026 ยท shuffle-on.com