A10 Networks, Inc.

• H1 2026 revenue $155,137K; Q2 $80,137K (+15.5% YoY) — 10-Q / 8-K EX-99.1.
• Product revenue Q2 $49M (+25%) vs service $31.1M (39% of mix) — earnings transcript.
• GAAP gross margin 79.3% H1 ($123,105K GP ÷ $155,137K rev) — SEC companyfacts.
• Non-GAAP: 80.3% GM, 25.5% op margin, $24.4M adj-EBITDA (30.5%), $0.25 EPS, $18.7M net income — earnings transcript.
• GAAP H1 operating income $22,023K (14.2% margin); net income $20,914K; COGS $32,032K; R&D $40,171K — SEC companyfacts.
• Balance sheet: cash $54,679K + marketable securities $302,669K ≈ $357M liquidity vs $225M of 2.75% convertible senior notes due 2030 (convertible in Q3 2026 on stock-price condition → reclassified). Deferred revenue $154,826K (up-front subscription economics). Inventory $31.7M (up from $18M end-2025, supply-chain build). — 10-Q.
• FY26 guidance raised: revenue growth 12–14% (from 10–12%), EPS growth 14–16% (from 12–14%) — earnings call.
• Capital return: $2.4M buyback (86,115 sh @ $27.63), $0.06/sh dividend (payable 2026-09-01).

A10 Networks sells secure application-delivery and DDoS/API-protection platforms (Thunder ADC + A10 Defend) that sit inline on L4–L7 traffic for enterprises, service providers, and cloud platforms. Hybrid delivery: appliance + virtual + cloud marketplace. Headquartered San Jose; ~7,000 customers (EX-99.1 2026-08-05).

• monetization model: Hardware appliances + perpetual/term software licenses + subscription/cloud-delivered security + support/maintenance + professional services. [10-K Item 1, 7; Revenue note]
• revenue mix: Single reportable segment; disaggregated: Products ~60%, Services ~40% (recent 10-K revenue note). [10-K Revenue note]
• who pays: Network/security infrastructure and security operations budget owners; carriers’ network engineering/operations; enterprises’ CISO/VP Infrastructure. [10-K Item 1]
• margin_value_capture: Gross margin ~80%; operating margin positive but modest; value accrues in high-margin software/support renewals. [10-K Item 7, financial statements]

• customer classes: CISOs & Security Operations, Platform & Developer Teams, service providers/carriers, cloud/web-scale enterprises, government agencies, financial services. [10-K Item 1]

Sources: A10 Networks CIK 0001580808, latest Form 10-K (Items 1, 1A, 7; Revenue note) and Form 10-Q where needed, SEC EDGAR: sec.gov ; sec.gov ; accession numbers UNKNOWN in this environment.

A10 Networks, Inc. sells application networking and network-security appliances and software. It reports revenue in two principal categories: Products and Services. Product revenue typically includes hardware appliances, software licenses, and related software products. Services revenue typically includes post-contract support/maintenance, professional services, training, and related service offerings.

The following segment mix is drawn from the draft FY2023 10-K data and is not yet verified against the filed document.

Total FY2023 revenue of $251.7 million (needs primary source validation) compared with FY2022 revenue of $274.7 million (needs primary source validation) would imply a year-over-year decline of approximately 8.4%:

\[

\frac{274.7 - 251.7}{274.7} = 8.37\%

\]

This is a load-bearing revenue-decline claim and is currently single-source because it comes from one issuer’s primary filing and has not been independently corroborated in this draft. Tag: [SINGLE-SOURCE].

Draft FY2023 margins:

The implied operating margin is calculated as:

\[

\frac{38.6}{251.7} = 15.34\%

\]

The implied net margin is:

\[

\frac{40.9}{251.7} = 16.25\%

\]

These calculations depend entirely on the same single filing and are therefore [SINGLE-SOURCE] until independently cross-checked. Do not treat them as verified.

A10 monetizes through:

1. Product sales — appliance-based units and software licenses. Revenue is generally recognized at a point in time when control transfers, subject to the specific contract terms disclosed in the filing. Exact timing policy needs validation against the FY2023 10-K revenue-recognition footnote [1].
2. Services and support — recurring maintenance and support contracts, professional services, and training. These are typically recognized over time [1].
3. Channel distribution — the company relies heavily on distributors and resellers rather than a direct-only sales force. The 10-K risk factors likely state that a significant portion of revenue comes from a limited number of channel partners; the named partners and percentages need primary source validation [1].

The economic buyer is typically an enterprise IT organization, a telecommunications operator, a cloud provider, or a government entity that needs application availability, secure application delivery, DDoS protection, or carrier-grade network address translation. However, end customers are not individually named in the available draft filing excerpts.

No individual end customer is named in this draft of the FY2023 10-K. The filing may disclose distributor or reseller concentration, but the exact names and revenue shares require direct retrieval of the filing. I will not invent customer names.

Tag: top customers: needs primary source validation

Based on the product lines and draft 10-K business description, A10 sells to:

This classification is inferred from the product portfolio. The exact customer verticals and revenue concentration by vertical require direct extraction from the 10-K [1].

The practical buyers inside an organization are typically:

• Network architects and network operations engineers — responsible for application availability and network performance.
• Security operations and security architects — responsible for DDoS mitigation, SSL inspection, and application-layer security.
• IT procurement / vendor management — involved in appliance refresh cycles and support renewals.

These roles are inferred from the products’ technical functions; A10’s 10-K may not contain specific buyer-persona language.

No named end customers are available in this draft. I will not fabricate customer references.

Q2 FY2026 (ended 2026-06-30) net revenue $80.137M (+15.5% YoY vs $69.383M); H1 $155.137M (+14.5% vs $135.520M). Product $49.024M (61% of Q2, +25.2% YoY vs $39.173M); services $31.113M (39%). GAAP GM 79.1% Q2 / 79.4% H1; non-GAAP GM 80.3%. GAAP operating income Q2 $9.024M (11.3%) / H1 $22.023M (14.2%); GAAP NI Q2 $8.882M (down from $10.538M). Non-GAAP NI $18.746M / $0.25 diluted; Adj. EBITDA $24.421M (30.5% of revenue). Cash + marketable securities $357.348M; deferred revenue / RPO $154.826M; 2.75% converts due 2030 reclassified current $219.518M. FY26 guide raised to revenue +12–14% and EPS +14–16% (from +10–12% / +12–14%). FY2025 revenue $290.557M (10-K filed 2026-02-25) → implied FY26 ~$325–$331M at the guide. All figures: 8-K EX-99.1 2026-08-05 and 10-Q filed 2026-08-06.

The DT bottleneck is inline traffic-control + DDoS scrubbing that every packet on a critical app path must traverse. A10's claim is that replacing that hop is a re-architecture + downtime event, and that bundling DDoS + SSL decrypt + WAF + bot into one Defend platform raises the switching cost versus F5's modular stack and Cloudflare's cloud-only edge.

Calc 1 — Customer A dollar exposure (the number the 08-29 note got wrong).

• Prior note used 26% (FY2025 10-K Customer A). Q2 10-Q is 38% for both the quarter and H1.
• Q2 Customer A dollars = 0.38 × $80.137M = $30.452M. H1 = 0.38 × $155.137M = $58.952M.
• Q2 annualized Customer A = $30.452M × 4 = $121.8M, which is 42% of FY2025's $290.557M. That is not a rounding difference — mix tightened 12 pp in six months (26% → 38%).
• Top-10 = 0.58 × $80.137M = $46.48M of one quarter. Rest-of-book outside top-10 = 42% of Q2.

Calc 2 — Microsoft warrant as cost-of-moat vs the revenue it defends.

• Max dilution = 800,000 / 72,454,000 outstanding = 1.10% (1.06% of Q2 diluted WAS).
• Notional at the vault's 2026-09-11 last print $24.74 = 800k × $24.74 = $19.79M. Exercise cash if cash-settled = $8,000 — economically a near-free option for Microsoft.
• Defended stream if Customer A ≈ Microsoft: Q2 run-rate $121.8M/yr. Cost-of-moat / defended = $19.79M / $121.8M = 0.16× one year of Customer A. Cheap if both purchase thresholds hit. If they miss, both 400k tranches expire unvested — A10 paid almost nothing and lost the lock-in. 10-Q: accounting effect not yet estimable.
• Revealed-preference read: Microsoft accepted a purchase-vested warrant rather than a volume-commit take-or-pay. That is a soft lock (spend-to-earn equity), not a hard take-or-pay. Switching-cost dollars on the Microsoft side remain unknown (thresholds redacted from Exhibit 4.1 public text).

Calc 3 — mix shift: enterprise/US vs service-provider rest-of-world.

• Enterprise Q2 $48.018M vs $27.776M = +73% YoY; SP $32.119M vs $41.607M = −22.8%. Mix 40/60 → 60/40.
• Americas Q2 $54.764M vs $41.192M = +33%; United States $51.763M = 65% of company. APJ $16.734M vs $17.939M = −6.7%; EMEA $8.639M vs $10.252M = −15.7%.
• Implication: the +15.5% print is a US-enterprise / Customer A story. Ex-Americas Q2 = $25.373M vs $28.191M = −10.0%. A pause at Customer A does not get offset by APJ/EMEA/SP.

Calc 4 — TrojAI + ThreatX cost of the AI/API adjacency.

• TrojAI $39.7M ÷ annualized H1 ($155.137M × 2 = $310.274M) = 0.128×. Combined with ThreatX $19.1M cash = $58.8M of security M&A vs ~$310M run-rate = 0.19×.
• Filing language: TrojAI "not material" to operations. Developed-tech intangible $7.1M amortized over 5 years ≈ $1.42M/yr — rounding error vs $40.2M H1 R&D.
• R&D intensity H1 = $40.171M / $155.137M = 25.9%. Absolute R&D ~$80M annualized vs F5-scale budgets. The niche strategy (line-rate DDoS + integrated Defend) is the only way 26% R&D-to-sales works; it does not buy a monopoly.

Sensitivity

• technical_scarcity: 6/10 — inline ADC + multi-Tbps DDoS + 25.9% R&D/sales is real engineering, but the category is mature and contested (F5 / Radware / Fortinet / Cloudflare / Envoy). TrojAI/ThreatX are not material (10-Q), so they do not lift scarcity. Cap at 6 without a disclosed unique throughput/latency spec.
• economic_capture: 5/10 — 80.3% non-GAAP GM and 25.5% non-GAAP op. margin show pricing power, but Customer A 38% + top-10 58% + one distributor 40% of revenue / 57% of AR is the opposite of diversified capture. Prior 6/10 used the stale 26% FY25 figure; 12 pp mix tightening in H1 is a one-notch cut. Score cannot be ≥7 without segment share or a named <20% largest-customer print.
• time_to_monetize: 7/10 — already monetizing: product +25%, guide raised 08-05, warrant dated 08-03, TrojAI closed 06-15. Next hard catalyst = Q3 FY26 earnings (not filed as of 2026-09-12; last Item 2.02 8-K is 2026-08-05). Warrant tranche-1 measurement window ends 2027-06-30 — that is the Microsoft stay-or-go date.

• bypass_test: 2/3 — substitutes exist (F5, Cloudflare, cloud-native LBs, open-source ingress). Inline requalification makes bypass costly, not impossible.
• concentration_test: 2/3 — A10 is a mid-tier ADC vendor, not a chokepoint the industry must buy from. (Q2 here is A10's customer concentration, which is a risk, not pricing power.)
• hyperscaler_exposure: 1/3 — FAIL. Customer A = 38% of revenue and AWS/Azure/GCP ship competing load-balancers (ALB / Front Door / Cloud Armor) and the Microsoft warrant is purchase-vested, not take-or-pay. A 38% hyperscaler customer who can in-source is the DT hyperscaler test's failure mode. Prior 2/3 used 26% FY25.
• Composite: 2+2+1 = 5 → WEAK PROXY (DT network-security published range ~5–6; 5 sits at the low end because concentration, not because ADC physics changed).

\[

\frac{108.3}{143.4} = 0.755

\]

\[

0.755 \times 100 = 75.5\%

\]

\[

(251.7 \times 0.807) - 50.0 = 203.1 - 50.0 = \$153.1M

\]