A10 Networks Raised Guidance and the Stock Fell Anyway — Here's Why the Math Doesn't Work at This Price

A10 Networks beat Q2 estimates and raised full-year guidance, then watched its stock fall more than 9% the next day and sit roughly 28% lower over the prior month. Now at $26.84, the stock has given back a substantial chunk of its 52% year-to-date run.

That price action is the real story. The earnings were solid. The guidance was higher. What the market is telling you is that the underlying structure of this growth doesn't match the valuation.

The Microsoft dependency is the entire thesis, and everyone already knows it

Microsoft accounted for approximately 37% of A10's Q2 revenue. Management clarified on the call that the recently expanded multi-year partnership with Microsoft is AI-relevant and not driven by one-off prebuys. That reassurance did not calm investors because the structural risk hasn't changed — it has been reinforced.

When one customer accounts for more than a third of your revenue, your growth story is that customer's growth story. A10 is telling you it is a leader in AI infrastructure networking and security. What the numbers say is that A10 is a supplier whose fate is tied to a single hyperscaler's deployment rhythm.

The company has been pushing the narrative that enterprise now represents 60% of Q2 revenue, up from roughly half on a trailing twelve-month basis. That sounds like diversification progress. It also sounds like a company that knows it has a problem. Enterprise sales cycles are six to nine months, according to management. The diversification they're counting on is a pipeline, not a closed book.

The GAAP numbers tell a different story from the adjusted ones

Non-GAAP diluted EPS came in at $0.25, up 19% year-over-year, beating the $0.24 estimate. GAAP diluted EPS was $0.12, down from $0.15 in the prior year quarter. GAAP net income was $8.9 million, down from $10.5 million a year ago.

Revenue grew 15.5% year-over-year to $80.1 million. GAAP earnings shrank. The non-GAAP adjustments bridge a gap that is wider than the headline beat suggests. The non-GAAP gross margin held at 80.3%, which is genuinely impressive for a hardware-adjacent business. The adjusted EBITDA margin — earnings before interest, taxes, depreciation, and amortization, a rough cash-earnings proxy — was 30.5%. On that basis, A10 looks like a profitable, cash-generative operator. On a GAAP basis, it looks like a company where the additions back — stock compensation, depreciation, acquisition-related costs — are eating real economic value.

Free cash flow is declining, and the debt is rising

Trailing twelve-month free cash flow stands at $60.9 million, down 13.7% from the prior year. Management guided that full-year 2026 free cash flow will exceed the roughly $65 million generated in 2025. That is a recovery thesis, not a growth thesis.

Meanwhile, total debt sits at $440 million against $238 million in total equity, for a debt-to-equity ratio of 92%. The company has $357 million in cash and marketable securities, which nets out to negative net debt of roughly $138 million. That's clean enough. But ROIC — return on invested capital, the return the business generates on the money deployed in it — is 8.8%.

An 8.8% ROIC on a stock that trades at roughly 46 times trailing earnings is not a high-return business. It's a moderate-return business priced as if it were exceptional. The market is paying for a trajectory it hasn't seen yet.

The TrojAI acquisition is a roadmap promise, not revenue

A10 acquired TrojAI in June, a company that provides AI red-teaming and runtime protection capabilities. Management described the combination as pairing A10's "hardware-based AI firewall" with TrojAI's "software-based red teaming and runtime protection." The acquisition is expected to have no material financial impact in fiscal 2026.

Management indicated that native AI security solutions derived from this technology could reach market in one to two years. The acquisition is strategically coherent — it adds capabilities around the Model Context Protocol for securing agentic AI workflows, something management said TrojAI was "meaningfully ahead" of A10's internal roadmap.

It is a coherent bet for 2028, not a justification for today's stock price.

What the stock price is actually saying

The stock fell 9.2% in the sessions following the earnings release, and was down roughly 28% over the prior 20 days. Year-to-date, it had gained roughly 52% before this pullback. The current price of $26.84 is well below the 52-week high of $38.49.

That sell-side movement after a raised outlook is not a sign that the business is collapsing. It is a sign that the market is resetting the relationship between what A10 has delivered and what the stock has been priced to deliver. A stock that has doubled year-to-date cannot absorb the same amount of skepticism as one that hasn't moved. The math works against it.

The cross-currents

The cross-currents here are:

• Microsoft deployment pace. If Microsoft's AI infrastructure build stays on pace, A10 benefits as a proven supplier with an expanded multi-year agreement. If the pace slows or Microsoft diversifies its vendor base, 37% of revenue becomes a 37% drag. Directionally: the multi-year nature of the agreement provides near-term visibility, but hyperscaler vendor consolidation is a constant risk.
• Enterprise diversification. Management is confident in pipeline quality and improving visibility. The sales cycle is six to nine months, so results flow later in 2026 and into 2027. Directionally: this is the most important variable for long-term valuation but the least visible today.
• Margin sustainability. Memory pricing and component constraints are expected to persist for several quarters. Management says it maintained margins by not absorbing all cost pressures, which implies passing costs to customers or trimming lower-margin deals. Directionally: gross margins near 80% are structurally high but not immune to component cost spikes.
• AI security productization. The TrojAI integration is real but distant. Native solutions are one to two years out. Directionally: this is a genuine strategic asset that doesn't appear on an earnings sheet this year.

The investor implication

A10 Networks is not a bad company. It has real products, a strong balance sheet, 80% gross margins, and a foothold in a customer whose AI spending trajectory is the most visible in the industry. The Q2 beat and guidance raise were legitimate.

The problem is the price. A stock that trades at 46 times trailing earnings, with 8.8% ROIC, declining free cash flow, and a single-customer concentration ratio of 37%, is not being paid for what it has delivered. It is being paid for what it might deliver if Microsoft keeps buying, enterprise diversifies, and the AI security roadmap materializes on schedule.

All three of those conditions have to hold. That is not a thesis. That is a wish list with a 46x multiple attached.