Earnings Bullish 6

Stem Hits Profitability Milestone as AI Software Pivot Drives Record Margins

· 4 min read · Verified by 14 sources · By AI Intelligence Brief Editorial
Share

Key Takeaways

  • Stem (STEM) achieved its first full-year positive adjusted EBITDA in 2025, signaling a successful transition from low-margin battery hardware to high-margin AI software and services.
  • This trend of financial maturation is echoed across the AI sector, with companies like Riskified and Everspin reporting record margins and increased adoption of recurring revenue models.

Mentioned

Stem company STEM Riskified company RSKD Everspin company Arun Narayanan person Aware company SES AI company

Key Intelligence

Key Facts

  1. 1Stem (STEM) achieved its first full-year positive adjusted EBITDA of $7 million in 2025.
  2. 2Software and services revenue at Stem grew 25% year-over-year, now comprising 55% of total revenue.
  3. 3Riskified (RSKD) reported 65% annual revenue growth while reducing headcount by 3% through AI adoption.
  4. 4Everspin (MRAM) product sales increased 22% due to data center and industrial automation demand.
  5. 5Red Violet (RDVT) achieved a record 83% adjusted gross margin driven by its identity resolution AI.
  6. 6SES AI (SES) reported 10x revenue growth in 2025, reaching $21 million with a shift toward ESS products.
Metric
Full-Year Revenue $156M $344.6M $14.8M (Q4)
Revenue Growth 8% 65% 12% (Q4)
Gross Margin 38% (GAAP) 52% (Non-GAAP) 50.8% (GAAP)
Profitability Status Positive Adj. EBITDA GAAP Net Income Positive Non-GAAP Net Income Positive

Analysis

The fourth-quarter earnings cycle for the AI and machine learning sector has revealed a definitive shift in strategy: the transition from hardware-centric growth to high-margin, software-driven profitability. Leading this charge is Stem (STEM), which reported its first full-year of positive adjusted EBITDA at $7 million. This milestone was not merely a result of cost-cutting but a fundamental pivot in the company's revenue mix. Software, services, and edge hardware now account for over 55% of total revenue, a strategic move away from the volatile, low-margin battery hardware resale market. Stem’s PowerTrack software, which utilizes AI to optimize energy storage and grid interaction, saw its Annual Recurring Revenue (ARR) climb to $61 million, representing a 16% year-over-year increase. This transition allowed the company to post record GAAP gross margins of 38%, proving that the 'AI-as-a-Service' model in the energy sector is reaching a point of sustainable scale.

This theme of AI-driven margin expansion is further validated by Riskified (RSKD), which operates in the fraud prevention space. Riskified reported a staggering 65% year-over-year revenue growth, reaching $344.6 million for the full year. Perhaps more significant for the broader AI industry is Riskified’s operational efficiency; the company ended 2025 with a 3% decline in headcount, explicitly citing AI adoption as the primary driver for this productivity gain. This suggests that the 'AI dividend' is beginning to manifest not just in product offerings but in the internal cost structures of the providers themselves. Riskified’s competitive win rates, which remain above 75%, indicate that their machine learning models for transaction risk are maintaining a significant technological moat even as they scale internationally, with non-U.S. merchants now contributing 46% of their revenue.

Stem’s PowerTrack software, which utilizes AI to optimize energy storage and grid interaction, saw its Annual Recurring Revenue (ARR) climb to $61 million, representing a 16% year-over-year increase.

In the infrastructure layer, Everspin (MRAM) is capitalizing on the physical requirements of the AI revolution. The company reported a 22% increase in MRAM product sales, driven by intense demand in data centers and industrial automation. As AI workloads move toward the edge, the need for high-density, non-volatile memory like Everspin’s STT-MRAM becomes critical. The company is currently on track to launch a 256Mb monolithic device using TSMC’s 16nm FinFET process in late 2026, a move that will likely cement its position in the satellite and aerospace markets where radiation-hardened, high-performance memory is a prerequisite. Everspin’s ability to maintain a 50.8% gross margin despite the volatility of licensing revenue highlights the robust pricing power currently enjoyed by specialized AI hardware components.

What to Watch

However, the sector is not without its challenges. MicroVision (MVIS) and Wallbox (WBX) illustrate the difficulties of navigating shifting market incentives and legacy product cycles. MicroVision reported a significant revenue decline as it transitions to its next-generation Movia S and Movia L lidar sensors, incurring $29.4 million in non-cash charges. Similarly, Wallbox faced headwinds from a 40% decline in the U.S. EV market and the loss of federal incentives, leading to a 10% drop in Q4 revenue. These cases serve as a reminder that while AI software provides high margins, the underlying hardware markets remain sensitive to macroeconomic shifts and regulatory changes.

Looking ahead to 2026, the industry appears focused on 'quality of revenue.' Companies like Red Violet (RDVT) are doubling down on identity resolution AI with gross margins exceeding 83%, while SES AI is projecting a 43% to 67% revenue jump as it shifts focus toward Energy Storage Systems (ESS). The prevailing sentiment among analysts is that the market has moved past the 'experimentation' phase of AI. Investors are now demanding—and in the case of Stem and Riskified, receiving—clear evidence that AI can deliver both top-line growth and bottom-line profitability. The focus for the coming quarters will be on the successful execution of high-density hardware ramps and the continued conversion of pilot programs into long-term, recurring software contracts.

Timeline

Timeline

  1. Stem Profitability Milestone

  2. Everspin 256Mb Launch

  3. Q4 2025 Earnings Cycle

  4. DoD Contract Completion

Related Stories

Earnings

Wall Street AI Pivot: Why Clearfield and Concentrix Signal a New Value Phase

As institutional capital rotates away from overextended AI valuations, a secondary wave of infrastructure and service providers is emerging as a value play. Clearfield and Concentrix represent a strategic shift from speculative hype toward tangible hardware and integration solutions essential for the next phase of data center expansion.

Earnings

Microsoft’s Valuation Hits 10-Year Low Amid AI Disruption Fears

Despite a 23% surge in adjusted net income, Microsoft shares have retreated nearly 33% over the last five months as investors weigh the disruptive potential of AI agents against traditional enterprise software models. Bank of America has reinstated a 'Buy' rating with a $500 price target, arguing that Microsoft’s dual-threat position in cloud infrastructure and software applications makes it a rare bargain.

Earnings

Spectral AI Beats Q4 Estimates, Signals Commercial Pivot with 2026 Guidance

Spectral AI reported a significant Q4 2025 earnings beat with revenue of $3.8 million and a surprise GAAP profit, bolstered by a recent $31.7 million BARDA award. The company has introduced its first revenue guidance for 2026, marking a critical transition from R&D to commercial-scale medical AI diagnostics.

Earnings

Ping An Good Doctor 2025 Results: AI Deployment and Corporate Growth Surge

Ping An Healthcare and Technology reported robust 2025 financial results, driven by a strategic pivot toward corporate health management and the rapid integration of AI across its service ecosystem. The company is leveraging its 'Managed Care' model to enhance operational efficiency and scale its digital-to-offline healthcare offerings.

How we covered this story

Every story in our ai coverage is assembled from multiple primary sources, cross-referenced for factual consistency, and scored along three independent dimensions: sentiment, operational impact, and source-cluster confidence. Single-source rumors and unverifiable claims do not pass our editorial gate. When a story shows "Verified by N sources" with N≥2, the development is independently corroborated; when N=1, we mark it explicitly so readers can weigh the signal accordingly.

Impact scoring uses a 1-10 scale weighted toward regulatory, financial, and operational consequence rather than coverage volume. A topic that runs in every outlet but moves no real decisions ranks lower than a niche regulatory filing that reshapes how operators in the ai space have to behave. Read our full methodology for the scoring rubric, our glossary for term definitions, and our trends index for the longitudinal view across the beat.

Signal on this pageWhat it tells you
Verified by N sourcesIndependent corroboration count. N≥2 is our confidence floor; N=1 is marked explicitly.
Impact score (1-10)Regulatory + financial + operational weight. 8+ signals an experienced-operator action item.
SentimentFive-tier classification trained on labeled ai-specific corpora.
TimelineWhere applicable, the related-events sequence that contextualizes today's development.