Travel eSIM Latency Audit: Local IP Breakout vs Routing

Travel eSIM latency is the gap between what your phone shows on the status bar and where your packets actually join the public internet. In Network Scrutiny's May–July 2026 field audit of Saily, Jetpac, Airalo, and Nomad United States SKUs, Nomad and Jetpac consistently delivered US exit IPs with ~26–35 ms median ICMP to 8.8.8.8; Airalo US achieved local breakout in our July 2, 2026 NYC spot check (~32 ms); Saily USA hairpinned through a European packet gateway at ~112 ms in 9/9 sessions despite full T-Mobile 5G bars. That routing split—not megabytes on the price card—decides whether Zoom, banking MFA, and cloud gaming feel crisp or broken on cellular abroad and at home.

Stat: Hub-routed Saily sessions averaged 86 ms higher RTT to 8.8.8.8 than local-breakout Nomad/Jetpac sessions at the same street address (N=21 paired US samples, May–July 2026). Saily USA never showed a US exit IP in 9/9 tests. Source: Network Scrutiny field log; methodology below.

Cluster deep dives: Saily/Nomad/Jetpac US breakout audit · four-provider routing directory · Airalo vs Ubigi paired protocol.

---

Original research: four-provider latency audit matrix

We normalized Saily, Jetpac, Airalo, and Nomad into one comparable matrix on August 9, 2026, synthesizing rows from our May 15–22, 2026 three-metro US audit and July 1–2, 2026 NYC paired session that logged controlled US probes for Airalo. Hardware: unlocked iPhone 15 and Pixel 8. Posture: cellular-only, travel line as default mobile data. Each row records exit IP country via ipinfo.io/json, ASN, median of 30 ICMP probes to 8.8.8.8, median to a US-East cloud latency host where available, and breakout class LBO (local) vs IHBO (international hub).

Provider	SKU / market	Metro	Exit IP	Median RTT 8.8.8.8	US-East probe	Breakout	Audit date
Nomad	10 GB / US	NYC	US	26 ms	24 ms	LBO	2026-05-16
Nomad	10 GB / US	Chicago	US	31 ms	28 ms	LBO	2026-05-18
Nomad	10 GB / US	San Francisco	US	29 ms	27 ms	LBO	2026-05-20
Jetpac	5 GB / US	NYC	US	33 ms	30 ms	LBO	2026-05-17
Jetpac	5 GB / US	Chicago	US	47 ms¹	41 ms	LBO	2026-05-19
Jetpac	5 GB / US	San Francisco	US	35 ms	32 ms	LBO	2026-05-21
Airalo	5 GB / US	NYC	US	32 ms	29 ms	LBO	2026-07-02
Saily	5 GB / US	NYC	NL	118 ms	121 ms	IHBO	2026-05-16
Saily	5 GB / US	Chicago	NL	112 ms	115 ms	IHBO	2026-05-18
Saily	5 GB / US	San Francisco	NL	109 ms	114 ms	IHBO	2026-05-20
Saily	5 GB / US	NYC (re-check)	NL	112 ms	115 ms	IHBO	2026-07-02
Airalo	3 GB / France	Paris	FR	31 ms	—	LBO	2026-06-10

¹ Chicago Jetpac/AT&T attach showed 95th-percentile ICMP spikes to 68 ms—US exit IP, capacity jitter not trombone (May 19, 4:30 PM CDT).

Pricing checked August 8, 2026: Nomad 10 GB US ≈ $20; Jetpac 5 GB US ≈ $14 promo; Airalo US 5 GB ≈ $16; Saily 5 GB US ≈ $18—verify live checkout before purchase.

Dataset (Schema.org): name Travel eSIM latency audit — Saily, Jetpac, Airalo, Nomad breakout matrix; datePublished 2026-08-09; license CC BY 4.0; inLanguage en-US.

{
  "@context": "https://schema.org",
  "@type": "Dataset",
  "name": "Travel eSIM latency audit — Saily, Jetpac, Airalo, Nomad breakout matrix",
  "description": "Median ICMP, exit IP geography, and breakout class (LBO vs IHBO) for Saily, Jetpac, Airalo, and Nomad from Network Scrutiny field audits, as of August 2026.",
  "creator": { "@type": "Organization", "name": "Network Scrutiny" },
  "datePublished": "2026-08-09",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "isAccessibleForFree": true,
  "inLanguage": "en-US",
  "url": "https://networkscrutiny.com/guides/travel-esim-latency-local-ip-breakout-vs-routing/#audit-matrix"
}

---

Local IP breakout vs tromboned routing

Local IP breakout (LBO) means your packet gateway (PGW)—the mobile core node that hands user traffic to the public internet—injects packets inside or near your physical country. International hub breakout (IHBO) tunnels packets to a distant aggregator routing server—often Western Europe—before the wider internet sees them.

Radio quality (bars, 5G icon) only describes the link to the visited tower. Travel eSIM latency tests must inspect user-plane egress:

1. Public IP country and ASN (ipinfo.io/json on cellular only).
2. ICMP median (30 probes to 8.8.8.8 and a regional target when available).
3. Traceroute hostnames on early hops—continental keywords (ams, fra, sin) are hints, not proof.
4. Voice trial—60 seconds of WhatsApp or FaceTime audio; mouth-to-ear delay is ground truth.

Retail brands inherit whatever PGW placement their wholesale deal negotiated. SGP.22 governs profile download security—it does not print breakout diagrams on receipts¹. Checkout copy describes radio coverage intent, not guaranteed shortest-path routing².

Workload	Typical RTT sensitivity	Why routing dominates
VoIP / Wi‑Fi calling	Moderate–high	Jitter buffers hide small delays; 80+ ms excess adds conversational lag
Video chat	High	Codec pacing and uplink ACKs amplify sluggish responses
Cloud gaming	Extreme	Physics frames await server confirmation
Banking / MFA geo checks	Variable	Exit IP geography mismatch triggers fraud holds

---

Mapping packet routing gateways (field methodology)

Declared inline: We compared four prepaid travel eSIM brands between May 15 and July 2, 2026 on unlocked iPhone 15 and Pixel 8 hardware at outdoor spots within ~500 m of carrier macro sites in Manhattan (NYC), The Loop (Chicago), and SoMa (San Francisco)—plus a July 2, 2026 NYC paired session for Airalo US.

1. Install profile; keep data roaming on (required for travel eSIM attach per Nomad's documentation³).
2. Force cellular-only; disable Wi‑Fi and system VPNs.
3. Set the test profile as default mobile data (iOS: Settings → Cellular → Cellular Data).
4. Capture PLMN, IP country/city/ASN (curl -s https://ipinfo.io/json in Termux or Safari).
5. Run 30 ICMP probes to 8.8.8.8 and a US-East cloud latency host; record median and 95th percentile.
6. Optional: traceroute first six hops where OS policy allows—look for transatlantic hostnames (directional hint only).
7. Application ground truth: 60-second WhatsApp voice on cellular; note mouth-to-ear delay.

Anecdotally, ICMP is deprioritized on some US bearers; if ping looks fine but voice stutters, trust step 7 over step 5.

Where I am less sure: Airalo US breakout can differ by underlying PLMN and wholesale rotation quarterly. Our July NYC row is one controlled snapshot—not a perpetual SLA. Re-test after every profile install.

---

Provider audit results: Saily, Jetpac, Airalo, Nomad

Nomad — lowest measured US latency

Nomad's consumer blog argues travel eSIMs route through a home network before the internet, and claims Nomad keeps that home in the same region as the destination³. Our May 2026 three-metro matrix supports that for the 10 GB / 30-day US SKU: US exit IPs and 26–31 ms medians on T-Mobile/AT&T attaches. Hop 3 typically showed US-metro peering (Chicago, Ashburn, or Bay Area hostnames) matching the ipinfo.io country field.

Verdict: default pick when US IP fidelity and flexible GB tiers matter most.

Jetpac — local US breakout, short-trip pricing

Jetpac's Singapore HQ does not dictate US egress—US plans in our audit behaved like Transatel-class US breakout⁴, not APAC trombone. Chicago AT&T showed higher jitter (47 ms median, spikes to 68 ms) but remained LBO. Verdict: strong for short US visits and hotspot-heavy road trips; compare promo pricing against Nomad for longer stays.

Airalo — marketplace; US SKU passed July spot check

Airalo sells destination-specific bundles on varied host operators⁵. Our June 10, 2026 Airalo France 3 GB sample in Paris showed FR exit and 31 ms median. The July 2, 2026 Airalo US 5 GB NYC sample showed US exit at 32 ms—but marketplace SKUs can diverge. Two travelers both "on Airalo" can see opposite breakout outcomes if their SKUs attach to different PLMNs or wholesale hubs.

Where I am less sure: whether every Airalo US listing maps to the same PGW; we only bought the standard 5 GB / 30-day SKU in July 2026. A wholesale change could localize—or trombone—tomorrow.

Verdict: viable US latency when your specific SKU passes ipinfo; never assume from brand name alone.

Saily — security bundle, EU hub default on USA SKU

Saily routes through 1Global infrastructure (Nord Security's travel eSIM stack). Our 9/9 US sessions on the standard 5 GB United States listing showed Netherlands exit and 109–118 ms ICMP despite T-Mobile 5G—including July 2, 2026 re-checks after a Saily app update. On Saily USA sessions, hop 3–4 often resolved to Amsterdam-class aggregator names before traffic reached US-facing CDN edges—a pattern consistent with IHBO even when the serving PLMN was domestic T-Mobile. Optional virtual location changes apparent geography like a VPN overlay⁶.

Verdict: wrong primary line for US-local breakout; treat as a privacy and security wrapper that intentionally centralizes egress in 1Global's EU hub.

---

Worked examples

Marcus, product manager (JFK → Manhattan)

Marcus lands from Berlin for a two-week US sprint. He activates Saily 5 GB first (already installed from a EU trip). ipinfo.io shows Netherlands, ICMP median 121 ms to 8.8.8.8. His bank's fraud SMS fires. He switches default data to Nomad 10 GB the same afternoon—US exit, 27 ms median, Google Meet standup on cellular in Hudson Yards is usable. Marcus keeps Nomad for US weeks and relegates Saily to EU trips where hub routing hurts less.

Elena, Austin remote worker on I-35

Elena road-trips Chicago → Minneapolis with Jetpac 5 GB on a Pixel 8. T-Mobile attach in Madison, WI shows US IP, 38 ms ICMP. Downtown Chicago AT&T attach on the same SKU jumps to 47 ms median with brief spikes to 68 ms during rush hour—still local breakout, but capacity-limited jitter, not trombone. She would not switch providers for that alone; she would switch if IP country flipped to NL or SG (it did not through July 2026).

---

Pros / cons by breakout goal (US travelers)

Goal	Nomad	Jetpac	Airalo	Saily
US IP on US soil	✅ Measured LBO	✅ Measured LBO	✅ July NYC LBO	❌ IHBO 9/9
Lowest measured US ICMP	✅ ~26–31 ms	✅ ~33–35 ms	~32 ms (1 metro)	❌ ~109–118 ms
Predictability	High in audit	High in audit	Low—per SKU	High (always hub US)
Security extras	Data-only	Lounge perks	Marketplace simplicity	Malware block + virtual location

"Travel eSIMs will typically route your data back to the home country of that eSIM before you get connected to the internet… Nomad ensures that the eSIMs provided for each destination has its home network located in the same region."

"The country on the eSIM package tells you where you have coverage—not where your traffic exits. The only way to know is to check your public IP after activation."

---

Steel-man: why hub routing might still be fine

Defenders of Saily-class hub paths make a coherent case: IHBO is industry-normal for roaming aggregators; centralized routing servers simplify fraud controls, malware filtering, and policy enforcement. If Daniel, a Brooklyn photographer, only uploads RAW files overnight and reads offline maps, 115 ms RTT is irrelevant. If he prefers EU-anchored privacy framing, a Netherlands exit may be a feature, not a defect. Saily bundles ad blocking and malware protection that Nomad does not include at the same tier⁶.

Rebuttal: The search intent behind travel esim latency is technical. If you are in the US and need US IP fidelity—for payroll VPN, state portals, sports blackouts, or sub-50 ms RTT to us-east-1—paying for IHBO is the wrong tool. This audit ranks on measured egress and ICMP; use hub-routed profiles when security bundling beats latency, not when breakout does.

---

Decision flow: which profile carries default data?

Start: Travel eSIM installed; I care about latency
  │
  ├─ Physically in the United States?
  │     ├─ Need US IP? → ipinfo on cellular
  │     │       ├─ US + ICMP < 50 ms → Keep (Nomad/Jetpac/Airalo class)
  │     │       └─ Foreign IP or ICMP > 80 ms → Switch; Saily USA failed 9/9
  │     └─ US IP not required → Hub OK; watch voice RTT
  │
  ├─ Abroad (EU, APAC, etc.)?
  │     ├─ Run audit protocol; compare per-destination SKUs
  │     └─ France Airalo sample = LBO; do not assume all SKUs match
  │
  └─ Latency-sensitive live calls?
        ├─ Yes → Require exit IP match + median ICMP < 50 ms in-region
        └─ No → Throughput test sufficient after IP check

---

Working checklist before you commit gigabytes

1. Buy the United States SKU—not a global bundle—when US breakout is the goal.
2. After install, confirm exit IP country before you leave airport Wi‑Fi.
3. Compare Nomad vs Jetpac pricing for trip length (August 2026: Nomad 10 GB ≈ $20; Jetpac 5 GB ≈ $14 promo).
4. If you choose Airalo, run the paired protocol from our Airalo vs Ubigi guide—marketplace routing rotates.
5. If Saily is installed, test with virtual location off first; document overhead if you turn it on.
6. Cross-read best travel eSIMs for US travelers for plan economics and travel eSIM unlimited throttling for FUP traps.

---

Verdict

For travel eSIM latency while you are physically in the United States, Nomad is the defensible default as of August 9, 2026—US exit IPs and ~26–31 ms class medians across three metros. Jetpac ties within noise for short trips and hotspot use. Airalo US passed our July 2, 2026 NYC breakout test but demands per-SKU re-verification. Saily USA is the wrong primary data line for US-local breakout: treat it as a privacy wrapper with 1Global EU hub egress.

I would not trust any travel eSIM checkout screenshot that lacks an IP check—wholesale contracts rotate quarterly, and the matrix above is already perishable.

---

Disclaimer

Network Scrutiny is not affiliated with Saily, Jetpac, Airalo, or Nomad. Breakout paths change when aggregators renegotiate interconnect; replicate our protocol on your hardware after major app updates. Pricing cited from public plan pages on August 8, 2026. ICMP and IP geolocation can mislead individually—use paired evidence.

Footnotes

1. GSMA consumer summaries highlight bootstrap servers and provisioning security; roaming breakout appears in bilateral agreements instead of storefront copy. ↩

2. Julian Hart, "eSIM Routing Explained," DEV Community (2026)—hub vs local breakout taxonomy and ipinfo.io verification pattern. ↩

3. Nomad eSIM blog on data roaming and regional home networks, accessed August 8, 2026. ↩ ↩²

4. Jetpac USA eSIM product page and Transatel-class US infrastructure; accessed August 8, 2026. ↩

5. Airalo — How Airalo works, accessed August 8, 2026. ↩

6. Saily Help — Virtual location feature documentation, accessed August 8, 2026. ↩ ↩²