By Kritika Lamba
July 31 (Reuters) - SiriusXM's SIRI.O second-quarter profit declined more than expected on Thursday, taking the shine off a smaller-than-expected loss in subscriber numbers, sending its shares down about 7%.
The company warned about persistent weakness in the advertising market as economic uncertainty, cautious consumer sentiment and concerns over potential tariffs prompted advertisers to scale back spending.
SiriusXM also faces mounting pricing pressure in its streaming business, citing an oversupply of connected TV (CTV) ad inventory and intensifying competition in the audio sector.
Second-quarter profit fell 23% to 57 cents per share, down from 74 cents a year earlier, while analysts expected 75 cents.
Operating costs climbed about 4%, driven by increased investment in exclusive content,
"The management is still working on capital spending plans and going through the new tax act - there could be a revision to the guide after next quarter," said Stuart Novick, analyst at GimmeCredit.
The company lost 68,000 self-pay subscribers during the quarter, a marked improvement from the 303,000 subscribers lost in the previous quarter, beating Visible Alpha estimates for a decline of 176,840.
"There may be a little concern that the encouraging subscriber net adds don’t seem to be reflected in the outlook for the year," Novick added.
SiriusXM's podcast segment was a bright spot, with revenue surging 50% from a year ago, helping offset some of the broader market challenges.
Its quarterly net income stood at $205 million, compared with $354 million a year ago.
The company also signed a new agreement during the quarter with television host Trevor Noah to launch an original podcast spanning comedy, culture and current events.
Revenue from its Pandora and Off-platform segments fell 3% during the quarter, mainly due to a smaller subscriber base and reduced demand for music streaming.
SiriusXM posted second-quarter revenue of $2.14 billion, largely in line with analysts' average estimate of $2.13 billion, according to data compiled by LSEG.