RADCOM's $111 Million Cash Pile Finally Has a Job — Just as the Orders Slip Again
Revenue fell 24% year-over-year, but the real news is a $20M buyback replacing a two-year "M&A first" promise, and a server-cost squeeze that both delays RADCOM's deployments and sells its product
RDCM · Earnings Call · 2026-11-11
Cash-rich, revenue-light
RADCOM is a $195 million-market-cap telecom assurance vendor sitting on $111.4 million of cash, cash equivalents and short-term bank deposits, and no debt. That is roughly 57% of its own market value in the bank. On Tuesday it reported third-quarter revenue of $14.04 million, up 19.4% sequentially from $11.8 million — and down 23.6% year-over-year from $18.4 million. Non-GAAP operating income of $0.7 million (4.9% of revenue) replaced a $2.2 million operating loss in Q2, but compares with $3.8 million and a 20.9% margin a year ago. GAAP net income was $441,000 against $3.5 million. Strip out the quarter-over-quarter optics, which flatter a business lapping its own bad quarter, and the year-over-year line is the honest one: this is a company recovering from a self-inflicted air pocket, not one compounding. Management reaffirmed — the operative word, since this outlook was revised in July — a full-year 2026 revenue range of $57 million to $63 million, midpoint $60 million, and promised a return to double-digit growth in 2027.The paradox: server inflation is both the disease and the cure
Here is the genuinely interesting piece. RADCOM attributes its revenue timing to deployment delays caused by server infrastructure costs at its carrier customers. “In customer deployments, our platform has delivered up to 70% lower TCO than other solutions.” — Benny Eppstein, CEO · 2026-11-11 Because RADCOM is pure software, it can run on servers operators already own — including those of the incumbent vendor it is replacing. So the exact cost pressure throttling its revenue is the argument winning its deals. The evidence is in the two logos closed this quarter, both competitive replacements of the incumbent assurance provider. CETIN Networks in Slovakia, and VNPT — a Vietnamese Tier-1 with more than 30 million subscribers, landing RADCOM in roaming across multiple cloud environments. “It's basically modernization of the legacy platform that was supported with competitors up to that point.” — Benny Eppstein, CEO · 2026-11-11 These are new logos of a specific kind: displaced incumbents, not greenfield wins. That is a better signal than a spreadsheet of bookings implies, and it explains why Tier-1 operator displacement has become the core of the native assurance pitch. The macro cross-read is unusually clean. Know Trend's global tape shows server-side and data-center capex themes being sold hard across every window: "traditional server business" (HPE, DELL, AMD) is down roughly 29% over 30 days, and "AI data center" — a 63-name cluster — is a decliner over 30, 90 and 360 days. RADCOM's revenue hiccup is the small-cap, telco-shaped shadow of that same squeeze. Its customers are not refusing to buy assurance; they are refusing to buy servers.What management quietly stopped saying
For a company that spent over a year telling investors M&A was job one, the capital-allocation paragraph has been rewritten. In August 2025: “We're looking into potential M&A as the first priority. We are progressing with a few candidates.” — Benjamin Eppstein, Chief Executive Officer · 2025-08-13 In February 2026, asked point-blank about allocating $199 million: “Our first priority remains to look into M&A. And it's our first priority, this is what we are trying to accelerate.” — Benny Eppstein, Chief Executive Officer · 2026-02-11 Today, that sentence is gone. In its place: a share repurchase program of up to $20 million over up to 18 months — about 10% of the market cap — with the Israeli creditors' 30-day objection window closed and “no objections” — Hod Cohen, CFO · 2026-11-11. Acquisitions survive only as "we also remain open to targeted acquisitions and partnerships." A demotion, not a departure, but a demotion. The second silence is more striking. ServiceNow and the ecosystem-partnership story — the one that got its own analyst question in August 2025 (“The partnership is actually going very well... a few of the connectors are already in place.” — Benjamin Eppstein, Chief Executive Officer · 2025-08-13) — is entirely absent from this call. No AWS. No ServiceNow. No channel leverage. What replaced it is a narrower, more credible story: own the assurance layer, monetise AI agents (RADCOM Neura, the newly shipped ADM module) inside the accounts you already have. Fewer partners, more product. That telco buyer is under pressure, and it is consensus this week: among the telecom reporters in this window, cost reduction and churn dominate the conversation. RADCOM's pitch to AT&T and Rakuten — service quality as a retention weapon — is aimed precisely at that nerve.The credibility file
One thing did not change: the shape of the delay. In May 2026 Benny Eppstein expected the new customers to land in Q4: “Q4 definitely will reflect some of the general new customer that currently we are engaging with.” — Benny Eppstein, Chief Executive Officer · 2026-05-19 On this call, the same cohort was dated differently: “Those delayed upgrades will come at early Q1 2027.” — Benny Eppstein, CEO · 2026-11-11 And a year-plus slip, stated as progress: “We already received a couple of orders that planned for early 2027. So those orders that we expected to get early 2026, and we said that we pushed to 2027 is happening.” — Benny Eppstein, CEO · 2026-11-11 The Q2 shock was severe enough that an investor used the following call to complain about disclosure itself:That is the real question mark over the 2027 double-digit growth promise. It is a promise made from a base that has been cut once, from a pipeline management describes as "higher-single digits" POCs and field trials across North America and Europe, with a stated 6-to-12-month lag from contract to revenue. The math works — but only if "early 2027" means early 2027.The bad results of second quarter, they came quite surprising, and we had no clue after the Q1 call... is there more visibility for third and fourth quarter? Or might that happen again.