Innventure, Inc.

Consolidated revenue in Q2 2026 was $1 million, versus $0.5 million in Q2 2025 and $1.4 million in Q1 2026. Accelsius contributed $0.9 million, or 96% of revenue. Net loss was $34.9 million, compared with $27.8 million in Q1. Adjusted EBITDA was a loss of $22.6 million versus $18.4 million in Q1. G&A expense was $14.5 million, down 22% versus Q2 2025. Cash and restricted cash ended at $46.5 million, versus $60.4 million at the end of Q1. Management said it had previously expected Accelsius to exit 2026 near cash-flow breakeven at about a $100 million annualized revenue run rate, but now expects breakeven to extend beyond this year. It also said it is suspending revenue targets, removing DarkNX from the 2026 forecast, and no longer targeting consolidated positive cash flow for Innventure in 2028.

Bill Haskell framed the quarter around a major strategic shift in how Innventure communicates progress, putting more emphasis on milestones than quarterly revenue targets. He said the market debate has moved from whether two-phase cooling will be adopted to when, and argued that the same constraints hurting smaller early adopters are improving the long-term case for Accelsius. He also highlighted a planned leadership transition, describing Bill Grieco as well positioned to run the company and saying he is more bullish than ever on Accelsius and two-phase cooling.

Dave Yablunosky focused on the financial impact of slower-than-expected adoption at Accelsius. He cited $1 million in consolidated revenue, a $34.9 million net loss, a $22.6 million adjusted EBITDA loss, and $46.5 million of cash and restricted cash at quarter-end. He said the company raised about $13 million via standby equity draws at an average price of $6.21, fully repaid convertible debentures earlier this year, and now expects Accelsius breakeven timing to extend beyond 2026 because GPU access, power availability, and site allocation remain constrained for smaller customers. He also said Innventure will aim to fund AeroFlexx and Refinity increasingly at the operating-company level and raise capital thoughtfully to reduce dilution.

Analysts focused on the DarkNX order, the path from hyperscaler discussions to deployment, and whether 2027 remains a realistic commercialization year for Accelsius. Management said the DarkNX order should transfer to a new site, but timing is uncertain because the original site was lost due to power-envelope issues. On hyperscalers, John Hewitt said conversations are active, but a typical path from statement of work to proof of concept to deployment can take two to three years, so it is too early to give specific 2027 revenue visibility. Management also said the new technical validation was completed around the end of July, and clarified that the Jacobs study was a separate, earlier data-center-level analysis rather than the third-party integrator test.