Three months ago RADCOM told investors that its best customers had not left, they had simply hit pause. The upcoming third-quarter report, due before the open on November 11, is the first real chance to test that claim. After four straight strong quarters, the network-assurance specialist saw revenue drop 33% in the second quarter. It cut its full-year outlook from 8% to 12% growth down to an actual decline. Management blamed a multiple-fold jump in server infrastructure costs, which led Tier 1 operators to defer private-cloud and on-premise expansion phases. The company described the shortfall as deferred rather than lost, but it also conceded the warning came with no visible signal in the prior quarter. That makes visibility, not just the numbers, the real issue this time.

The Street expects non-GAAP earnings of $0.06 per share on revenue of about $13.1 million. That would be a modest step up from the $11.8 million posted last quarter, but still down nearly 29% from the $18.4 million a year earlier, so this remains a contraction story. The Earnings Whisper sits at $0.08, a couple of pennies above consensus. That hints that some investors expect cost control to do more of the work than the official estimates assume. The earnings bar deserves attention. Last quarter's non-GAAP loss was roughly $1.5 million on $11.8 million of sales. A swing to about $1 million of profit on only $1.3 million of extra revenue cannot come from gross margin alone, even at RADCOM's steady 76% level. It requires real restraint on operating expenses, so the cost line may be as telling as the top line.

Guidance frames the second half neatly. With roughly $30.4 million booked in the first half, the reaffirmed $57 million to $63 million range implies $26.6 million to $32.6 million over the final two quarters. Consensus for the year sits at $57.1 million, essentially the floor of that range. That positioning is telling: analysts have taken management at its word on direction but are not giving it credit for any recovery before 2027. A third quarter at or above $13.1 million, paired with a reaffirmed range, would keep the low end intact. Anything that pushes the midpoint of $60 million into play would suggest the deferrals are thawing faster than feared.

The qualitative checklist matters just as much. Management said normalization might arrive in the fourth quarter but more likely in the first quarter of 2027. Any change in that timing will move the stock more than a penny of EPS. The post-quarter wins also need follow-through. These included the Satin Networks displacement in Slovakia, a gateway into a group serving more than 12 million subscribers, plus a new Tier 1 customer in Asia-Pacific and a European renewal. Evidence that these deals are converting to revenue, or that the AI-oriented Neura platform and newly launched ADM module are pulling pipeline forward, would support the claim that competitive position is improving even as spending stalls. Progress on the $20 million to $25 million buyback, funded from a debt-free $109.7 million cash pile, plus confirmation of positive second-half free cash flow, would reinforce the balance-sheet floor.

Sentiment has warmed noticeably, flipping from solidly bearish before the last report to mildly bullish now. The stock, however, has barely budged, slipping 2.5% since the guidance cut, roughly in line with a softer S&P 500. At about $10, shares trade well below the 200-day moving average near $12.16. They sit only a few percent above the post-earnings low of $9.73, in a range far narrower than the previous quarter's violent $9.40 to $15.78 swing. The market appears to have priced in the reset and is waiting for a reason to move.

Ultimately this report hinges on one question: whether the server-cost shock is a timing problem or a structural one. Stable revenue, a return to non-GAAP profit and an unchanged 2027 recovery timeline would validate management's narrative. Another sequential stumble, or a slide in that timeline, would suggest the air pocket is deeper than anyone has modeled.