Ideal Power (IPWR) reported a third-quarter loss of $0.18 per share. Earnings Whispers calculates that as 43.8% earnings growth from a year ago. Revenue fell 72.0% to roughly $0.01 million, or $7,000 in the release. No consensus estimate was available, so there is no beat or miss to measure. The headline per-share improvement is mostly arithmetic, not operating progress. The more important development is that the two flagship milestones management had set for the fourth quarter have been pushed back or restructured, even as the sales funnel kept growing.

The narrower per-share loss deserves scrutiny. Net loss actually widened to $3.6 million from $2.9 million a year earlier. The improvement in EPS reflects a share count that more than doubled, to about 19.9 million weighted shares from 9.2 million, after roughly $40.3 million in net equity proceeds this year. Operating expenses rose to $3.9 million from $3.0 million a year ago and $3.6 million in the second quarter. The increase came from:

• higher personnel costs;
• stock-based compensation, which more than doubled to $2.3 million year to date;
• non-cash patent impairments as the company pruned its pending portfolio.

Interest income of $336,000 on the larger cash balance offset part of that increase. Revenue remains negligible and is not yet a meaningful signal in either direction.

The commercial pipeline continued to build. Management lifted the sales funnel to about $450 million, up from roughly $400 million in the second quarter, $300 million in the first quarter and $200 million at the end of 2025. New activity in the quarter included:

• a collaboration with Power Integrations on a self-powered 800-volt DC solid-state circuit breaker, with a working prototype already demonstrated;
• two NRE-funded projects expected to start this quarter, one a 125-amp breaker program with a global industrial technology company and the other an engineering study tied to a U.S. hyperscaler;
• a large new Asia customer evaluating B-TRAN for battery energy storage;
• the 63-amp reference design kit now stocked at Richardson Electronics, with a 125-amp kit expected within weeks.

Management reaffirmed that it expects AI data center revenue to begin in the second half of 2027 and said it sees no slowdown in data center investment.

That is the central tension. The call brought the first explicit near-term slippage of this story:

• The lead Asia customer pivoted from a 63-amp to a 125-amp product, eliminating the initial low-volume 63-amp prototype orders expected in the fourth quarter.
• The hyperscaler program now begins with an engineering study before hardware, displacing the end-of-2026 prototype delivery target set in prior quarters.
• JEDEC qualification moved from "complete in Q4" to "progressing toward end of year."
• Stellantis deliverables originally guided for mid-2026 are still incomplete, with the next milestone now slated for this quarter.
• Automotive revenue is expected to remain development-only for the next couple of years.

The CEO conceded that customer programs are converting to production orders more slowly than expected. Management attributed part of the delay to evolving 800-volt DC specifications, ranging from 40 amps to more than 1,000 amps, which are forcing customers to revise product plans. That is a plausible explanation, but it also means timelines depend on standards Ideal Power does not control.

The balance sheet buys time to absorb these delays. Cash stood at $38.3 million with no long-term debt, down from $41.3 million at the end of the second quarter. Full-year cash burn guidance crept up again, to $10.4-$10.6 million from $10.3-$10.5 million, but that pace still implies a multi-year runway. The cost of that cushion was substantial dilution, which current shareholders have already absorbed. Bears have legitimate ammunition in the rising expense base, the slipping milestones and a funnel that has quadrupled in a year without converting into revenue. Bulls can point to NRE-funded work, which signals customers are willing to pay for engineering effort, and to channel validation from Power Integrations and Richardson.

Market context is limited. Shares opened at $4.76 after the August report, and the 200-day moving average sits near $4.22. Current price and investor-sentiment readings were not available, so it is not possible to say how the stock was positioned entering this report.

The bottom line is that Ideal Power delivered a quarter in which pipeline breadth improved while execution timing weakened. The narrower per-share loss is a product of dilution, and revenue remains immaterial. Investors are now being asked to accept a later path to production orders, with the second half of 2027 data center revenue target the next major test. The fourth-quarter checkpoints are the evidence to watch: the start of the NRE projects, JEDEC completion and the Stellantis milestone will show whether this quarter's delays were a one-time reset or the start of a pattern.