SES AI (SES) Q2 2026 earnings review
First Broad Commercial Validation, But Steep H2 Ramp Required
SES AI generated $5.1M in Q2 revenue, down sequentially from $6.7M but validating its strategic pivot by recording sales across all four segments (ESS, drones, materials, and Molecular Universe) for the first time. Gross margin is stabilizing, improving to 22.6% from 18.1% in Q1. However, execution risk remains high: the company reaffirmed FY26 revenue guidance of $30-35M, which implies a steep accelerating revenue ramp in the second half of the year (requiring roughly $18-23M in H2 vs $11.8M achieved in H1). Net loss reversed its improving trend, widening sequentially to $17.9M due to lower QoQ revenue and a bad-debt provision on a legacy EV contract.
🐂 Bull Case
For the first time, all four business lines generated revenue. The Sol-Ark certification for UZ Energy's ESS systems unlocks immediate US market potential, while the first 'Search-in-a-Box' multi-year subscription validates the Molecular Universe AI platform's commercial viability.
Despite the sequential uptick from a one-time bad debt provision, operating expenses are down 26% YoY. With $163M in liquidity and a CapEx-light model, the company has sufficient runway to fund operations without near-term dilution.
🐻 Bear Case
To hit the midpoint of FY26 guidance ($32.5M), SES must generate $20.7M in H2—an average of over $10M per quarter. This requires flawless execution in ramping the Korea drone cell facility and scaling US ESS deployments.
A bad-debt provision related to a legacy EV service contract increased OpEx sequentially, contradicting management's aggressive cost-reduction narrative and highlighting lingering risks from the abandoned OEM commercialization path.
⚖️ Verdict: ⚪
Neutral. The commercial milestones are tangible and validate the pivot away from passenger EVs. However, the required H2 revenue acceleration leaves little room for error, and the margin mix shift continues to weigh on the long-term profitability profile.
Key Themes
Sol-Ark Certification Unlocks US Market (Macro Tailwind)
UZ Energy's low-voltage residential batteries received closed-loop compatibility certification from Sol-Ark, a top US hybrid inverter producer. Management noted this is particularly timely given recent US FCC restrictions on foreign-produced inverters, creating a macro tailwind for certified players and a direct catalyst to accelerate UZ's revenue growth in North America.
NDAA Drone Capacity Approaching Full Scale
The Chungju, South Korea facility is one month away from scaling up to 1 million NDAA-compliant cells per year, with full capacity expected starting in Q4. Management reports hosting line audits for major American and allied drone makers, with strong demand driving line of sight for orders well into 2028. This segment is expected to contribute meaningfully to revenue starting in Q4.
Molecular Universe (MU-3.0) Commercial Validation
The AI4Materials platform generated its first direct revenue via a multi-year 'Search-in-a-Box' subscription commitment with a major global battery manufacturer. The release of MU-3.0 (an agentic workflow platform integrating autonomous labs on-premises) is a major technological innovation. MU-4.0, which will generate new molecules based on desired properties, is slated for release later this year.
Legacy EV Bad-Debt Contradicts Cost Reduction Narrative
Despite management's stated goal of reducing YoY OpEx by more than 20%, Q2 GAAP operating expenses reversed course, increasing sequentially to $20.3M from $19.1M in Q1. This was directly attributed to a bad-debt provision related to a legacy EV service contract. This contradicts the otherwise positive narrative of flawless operational efficiency and raises questions about other lingering legacy receivables.
Gross Margin Mix Shift
While gross margin showed a stable sequential improvement (from 18.1% in Q1 to 22.6% in Q2), it remains drastically lower than the 73.7% margin achieved in Q2 2025. This reflects the structural mix shift from high-margin automotive OEM service contracts to lower-margin ESS hardware sales.
Implied H2 Revenue Ramp
First half revenue totaled $11.8M. To meet the reaffirmed $30-35M FY26 guidance, SES must generate $18.2-$23.2M in the second half. This implies an accelerating trajectory that is highly dependent on flawless execution of the Chungju drone cell ramp and swift conversion of the MU and Materials sales pipelines.
Other KPIs
Materials contributed $1.2M in Q2, proving that the discovery pipeline generated by the Molecular Universe platform can be monetized. Roughly half a dozen customers are in second-phase testing for these materials, keeping the company on track to secure its first commercial supply agreement in H2.
Cash, cash equivalents, and short-term investments stood at ~$163M, down from $178M in Q1. The CapEx-light model is intact, and management stated this runway is sufficient to fund operations and execute 2026 growth initiatives, while maintaining flexibility for inorganic growth (M&A).
Guidance
Reaffirmed. This implies accelerating growth in the second half of the year, driven by the drone cell capacity ramp-up in Q4, expanded ESS Edge Box distribution in the US, and closing material supply agreements.
Reaffirmed. Decelerating/Reversing from the 22.6% achieved in Q2, indicating management expects a heavy influx of lower-margin ESS hardware or drone cell sales in the second half of the year to dilute the blended average.
Key Questions
Legacy EV Bad Debt Exposure
The sequential increase in OpEx was driven by a bad-debt provision for a legacy EV service contract. Are there any remaining legacy receivables on the balance sheet that carry similar risk, and what is the total exposure?
H2 Revenue Mix Dependency
To achieve the midpoint of your $30-35M guidance, you need roughly $20.7M in H2. How much of this required acceleration is derisked through existing backlog versus pipeline conversions in drones and materials?
Molecular Universe Economics
You announced your first multi-year subscription for 'Search-in-a-Box'. Can you provide color on the pricing structure or average contract value for these enterprise AI subscriptions?