Innventure, Inc. Q2 FY2026 earnings call

Leadership Transition

• Bill Haskell will retire as CEO of InVenture on October 1 after six years leading the company; this is a planned succession.
• Dr. Bill Grieco, founding CEO of Raffinity and former InVenture CTO, will take over as CEO. He has deep internal knowledge of all InVenture businesses and is prepared to lead the company's next phase.

Core Business Updates

• Excelsius: The market now accepts two-phase direct-to-chip cooling will be adopted, the only open question is when. Smaller early adopters currently face structural constraints (GPU allocation limits, power access shortages, site development barriers), but this has strengthened the long-term opportunity by pushing Excelsius to focus on relationships with large, dominant hyperscalers. A major independent third-party benchmark test validated Excelsius' Nucool technology: it cooled NVIDIA B200 GPUs 9-14°C cooler than standard single-phase cooling, held 9°C of headroom below the GPU's throttle temperature at 50°C facility water (where single-phase exceeds the throttle limit), and enables 5% more GPUs per fixed data center power envelope. Excelsius updated its commercialization strategy to focus exclusively on four core customer segments: chip manufacturers, server OEMs, server ODMs, and hyperscalers. The company will now track progress against four leading milestone metrics instead of revenue: chipmaker engagement for reference design inclusion, OEM/ODM co-development partnerships, executed SOWs with leading hyperscalers, and continued third-party benchmark validation.
• Aeroflex: Commercial pipeline has grown to ~$35 million, up 9% quarter-over-quarter, with expanded global partnerships in Latin America and Europe. Co-manufacturing equipment is installed and operational at its Italian facility, with product qualification underway. The business is moving toward independent capital raising to limit parent company funding needs.
• Raffinity: Engineering design for its 10-kiloton demonstration plant remains on track for completion by the end of 2026, and extended duration process scaling milestones have been met on schedule. Raffinity is also positioned to independently raise capital going forward.

Strategy and Communication

• InVenture will continue its core company-building model: spinning out operating companies around breakthrough technologies, providing initial capital and back-office support to enable focus on commercialization.
• Management will shift from providing quarterly revenue targets (which are unreliable for pre-adoption innovative technology) to reporting progress against key market adoption milestones, building credibility through delivered results instead of forward projections.
• Capital allocation policy remains unchanged: capital above a required parent reserve will be distributed to shareholders, and management prioritizes minimizing dilution for existing InVenture shareholders by financing Aeroflex and Raffinity at the operating company level where possible.

Consolidated Q2 2026 revenue was $1.0 million, compared to $0.5 million in Q2 2025 and $1.4 million in Q1 2026. Excelsius contributed $0.9 million to total revenue, equal to 90% of consolidated revenue. No separate financial performance figures were provided for Aeroflex or Raffinity, only operational updates.

• Management has suspended all formal revenue targets for Excelsius and consolidated InVenture, and will only reinstate targets once structural adoption constraints ease or core adoption KPIs are achieved.
• The timeline for Excelsius to reach break-even has been extended beyond 2026, driven by market dynamics rather than changes to technology or market opportunity conviction.
• The previous target for consolidated InVenture to achieve positive cash flow in 2028 has been withdrawn; management will revisit the timeline once there is greater visibility into the pace of Excelsius adoption.
• Excelsius cannot provide visibility into 2027 commercial revenue at this time, and will only update the market on progress when core milestones are completed, not on a fixed quarterly schedule.

• Near-term order bookings and revenue will remain lumpy and highly unpredictable until broad two-phase cooling adoption is established, due to structural constraints on smaller early adopters and long decision timelines for large hyperscaler customers.
• The delayed path to revenue and positive cash flow for Excelsius means InVenture will require additional capital in the future, creating potential dilution for existing shareholders.
• Data center development is facing growing regulatory and community pushback over water and power use, which can delay or block projects even for large hyperscalers.
• The DarkNX deployment was removed from the 2026 forecast after DarkNX lost access to its planned original site due to power allocation issues, and the timing of a potential redeployment at an alternate site is uncertain.

Q: The prior $50 million of Excelsius bookings included the DarkNX project – is the DarkNX order still intact, what caused the original site to fall through, and will it transfer to a new site? / A: Management does not disclose specific dollar amounts for individual customer orders, but expects the DarkNX order will transfer to an alternate site. The original site was lost due to issues securing a sufficient power envelope for the project. The order was removed from the 2026 forecast because the timing to identify and develop a new site is currently uncertain. (188 characters)

Q: What is the typical timeline from securing an executed hyperscaler SOW to commercial deployment and revenue, and is there visibility into meaningful 2027 commercialization? / A: After securing an SOW, the full cycle from proof of concept to reference design inclusion to volume commercial deployment typically takes 2-3 years for large hyperscalers, and Excelsius is in varying stages of this process with multiple large customers. There is no visibility into specific 2027 revenue at this time, as management is focused on hitting its four core adoption milestones first. (234 characters)

Q: What is the third-party validation study background, and what is the Jacobs reference design mentioned in results? / A: The third-party testing for the recent benchmark took approximately 60 days and was completed in late July 2026, right before the earnings call. The Jacobs reference design is a separate independent engineering study completed 1-1.5 years prior that analyzed data center-level power and infrastructure benefits of two-phase cooling, which found two-phase enables 5% more GPUs within a fixed power envelope. (240 characters)

Q: What is the current cash position and plan for future capital raising, and are there strategic options for Aeroflex? / A: InVenture has $41.5 million in cash on hand as of Q2 end, with access to additional capital via its existing standby equity purchase agreement, so cash is not an immediate issue. Management will raise additional capital opportunistically when needed, prioritizing minimizing shareholder dilution. Both Aeroflex and Raffinity are pursuing independent operating company-level capital raising and are expected to be self-funding after 2026, with Aeroflex seeing growing commercial pipeline momentum. (311 characters)

Q: When did InVenture shift from debating if two-phase cooling would be adopted to debating when it will be adopted, and what was the catalyst? / A: InVenture was internally convinced two-phase would be adopted early on, after proving the technology outperformed existing air and single-phase solutions in initial testing. The catalyst for the broader market shifting to a 'when not if' frame is growing industry consensus that increasing GPU heat densities will eventually require two-phase, reflected in 2030 two-phase market projections rising to $9 billion, up from prior lower estimates. (267 characters)