Aeva Technologies Inc

1. Define the business precisely

Aeva designs frequency-modulated continuous-wave LiDAR-on-chip and perception software. Its sensors measure depth, reflectivity and instantaneous velocity for automated driving, industrial automation, robotics, consumer devices and security. Customers are automotive and industrial OEMs and system developers; most 2025 units were prototypes used for evaluation, pilot or development rather than mass production. Revenue also includes non-recurring engineering.

The company designs photonics, sensing and software, works with external manufacturers, supplies prototype hardware and engineering, and seeks production nominations. OEMs and their end customers ultimately pay for reliable perception that improves automation and safety. Net revenue was only $18.1 million in 2025, up from $9.1 million, while gross loss was $0.7 million. Aeva is therefore a funded technology-development business, not yet a proven component franchise.

2. Understand the customer

OEMs can choose camera, radar, time-of-flight LiDAR, other FMCW vendors, sensor fusion or lower automation. Purchase criteria include range, velocity accuracy, interference resistance, size, power, safety, reliability, unit cost, manufacturability, software integration and supplier solvency. Automotive qualification can take years and creates switching cost after a platform is designed, but OEMs can cancel, delay or resource before production.

Industrial customers value precise velocity and distance for automation and metrology, but established machine-vision and sensor vendors offer substitutes. Aeva's brand has little independent economic value today; technical validation and production reliability matter. Customer concentration is high: three customers represented 72% of receivables, increasing bargaining and cancellation risk.

3. Explain how profits could be created

Revenue would scale with production units, price, engineering milestones and software content. Gross profit must cover photonic components, contract manufacturing, calibration, warranty and customer support. Fixed R&D, sales and public-company cost then create large operating leverage if production volume arrives.

2025 revenue of $18.1 million was below $18.7 million cost of revenue. R&D was $85.4 million, G&A $34.8 million and sales and marketing $6.7 million. Operating loss was $127.6 million and net loss $145.4 million. Stock compensation was $21.8 million. Cash used in operations was $115.1 million.

Working capital is immature: inventories rose as prototypes and programs expanded, while customer advances and engineering arrangements can create lumpy timing. A $3.8 million anticipated loss on a joint-development agreement shows that strategic contracts can consume more cost than allocated consideration.

Incremental returns depend on design wins converting to profitable, durable production rather than more engineering expense. Volume without positive gross margin destroys value. Suppliers and OEMs may capture most economics through manufacturing dependence and cost-down demands.

4. Analyze the industry and capital cycle

LiDAR has high technical barriers but intense entry and substitution. Venture and public capital funded many vendors before vehicle volumes existed. OEM delays, consolidation and failures then removed capacity. Strong program announcements attract new investment, which can reduce component prices before suppliers recover R&D.

Automotive OEMs are concentrated and powerful. Qualification, functional-safety standards and long vehicle cycles create entry barriers; the same cycles delay cash and enable re-sourcing. Contract manufacturers and semiconductor partners hold specialized process knowledge. Cameras and radar improve continuously and can reduce the required LiDAR content.

Industrial markets diversify use cases but require channel, application support and incumbent displacement. Exit from a failed platform leaves specialized R&D with limited recovery. Competition benefits OEMs through lower prices while making shareholder returns uncertain.

5. Identify the source of competitive advantage

Aeva claims differentiation from FMCW integration, direct velocity per pixel, long range and immunity to sunlight and other LiDAR interference. Silicon-photonics integration could lower size, power and cost at scale. If independently validated in production, those mechanisms can improve perception performance and simplify sensor fusion.

The advantage is not established by patents or prototypes alone. Competitors can develop alternative FMCW architectures; time-of-flight systems can improve; OEM software can combine radar and cameras; manufacturing yield can erase theoretical cost benefits. Strategic work with LG Innotek may improve manufacturing and market access, while also sharing economics and creating dependence.

Disconfirming evidence would include repeated production delays, negative gross margin at higher volume, poor yield, lost design wins, OEM re-sourcing, patents designed around, or performance that does not change customer system cost.

6. Examine the operating system

Aeva links photonic design, chips, optics, firmware, perception software, calibration, outsourced manufacturing, OEM validation and application support. Product and process must mature together: a high-performing lab sensor has little value if suppliers cannot manufacture it reliably at automotive cost and quality.

Outsourcing limits factory capital but surrenders process control and bargaining power. Customer-specific engineering aids qualification but fragments resources. Automotive programs offer scale and long life but impose cost reductions and cancellation risk; industrial programs diversify revenue but require different sales and support.

Inventory must precede prototype demand, while receivables are concentrated. The strategic trade-off is funding multiple markets versus concentrating cash on the most probable production path. The operating system remains unproved until repeatable manufacturing and positive gross margin coexist.

7. Assess financial resilience

Year-end cash and marketable securities were $121.9 million: $72.3 million cash equivalents and $49.6 million securities. Current liabilities were $35.9 million. Cash burn was $115.1 million in 2025, so liquid assets approximated only one recent year of operating use before new financing.

Aeva issued $100 million of 4.375% convertible notes due November 2032, with $96.7 million net carrying value and 4.95% effective rate. Interest may be paid in cash, shares or both. The initial conversion rate equals 6.303 million shares at about $15.864 per share; holders can demand cash at a fundamental change. This extends maturity but embeds dilution and potentially share-settled interest.

Lease liabilities were $5.7 million, with undiscounted payments of $1.9 million in 2026, $1.9 million in 2027, $1.9 million in 2028 and $0.8 million in 2029. No revolver was disclosed. Management cited cash, securities and a standby-equity facility as sufficient for at least 12 months; standby equity is access at an uncertain share price, not committed non-dilutive liquidity.

A severe scenario combines a two-year production delay, no positive gross margin, $110 million annual burn and closed technology capital markets. Cash would be consumed rapidly; R&D and programs would be cut or equity issued under weak conditions. Long note maturity prevents immediate refinancing, but interest, warrants and dilution transfer the stress to common holders. Resilience depends on funding, not operations.

8. Evaluate capital allocation and shareholder outcomes

Capital allocation is R&D prioritization and financing. Aeva spent $85.4 million on R&D and only $4.6 million on property, consistent with outsourced manufacturing. It raised $100 million of convertible principal and $32.5 million from LG Innotek for 3.510 million shares at $9.26. No dividend or repurchase occurred.

Shares outstanding rose from 54.140 million to 61.580 million. The weighted basic and diluted denominator rose to 57.023 million, while 21.984 million potential shares were excluded as anti-dilutive: 2.316 million options, 7.137 million RSUs, 0.735 million performance units, 5.492 million warrants and 6.303 million note conversion shares. Accounting anti-dilution in a loss year does not remove the economic claim.

Stock compensation of $21.8 million exceeded annual revenue. LG funding and convertible debt extend the runway but transfer future value through shares, warrants or interest. Per-share value increases only if the funded engineering produces production cash worth more than burn and the enlarged fully diluted denominator.

No material business acquisition was disclosed in 2025 or the reviewed annual history; the company built technology internally, so acquisition consideration did not create a separate goodwill or share claim. Nor was debt reduced: Aeva moved from no material funded debt to $100 million of convertible principal. The long maturity improves near-term liquidity, but interest and the 6.303 million conversion-share claim make this financing, rather than repayment, the relevant common-share outcome.

9. Define legal and regulatory exposure

Automotive safety and product liability — medium probability, very high severity, multi-year, only partly reversible. Sensor error can contribute to injury, recalls, indemnity claims and loss of qualification. Redesign cannot recover all trust or program time.

Intellectual property and export controls — medium probability, high severity, long duration, partly reversible. Patent disputes, trade restrictions or semiconductor controls can block markets or sourcing. Licensing and redesign may help at material cost.

Cybersecurity and software integrity — medium probability, high severity, persistent, partly reversible. Compromise of sensing software or development systems can disrupt customers and expose confidential designs. Restored systems, patched code and stronger controls can cure operations, but leaked intellectual property, lost qualification time and damaged OEM trust cannot be fully recovered.

OEM contracts, disclosure and governance — recurring medium probability, moderate-to-high severity, potentially multi-year, partly reversible. Program cancellation and optimistic order-book interpretation can create litigation or capital-market consequences. A cancelled program can be replaced only after a long qualification cycle, so the lost revenue window is not recoverable; disclosure controls, settlements and board remedies can cure process failures, but financing cost and damaged credibility can persist. The prior stockholder litigation illustrates governance cost even after settlement.

10. Form a disciplined conclusion

Aeva could create value if FMCW velocity sensing and integration materially improve customer systems and can be manufactured at positive margin. It could retain value through patents, process learning and qualification. None is yet durable: revenue is prototype-scale, gross margin is negative and powerful OEMs can delay or substitute.

The balance sheet has no near-term debt wall and received substantial 2025 funding, but current burn nearly equals liquid assets. Common holders experienced major dilution and face another 22 million potential claims. Financial survival is stronger than operating proof.

Thesis-invalidating evidence includes lost production programs, persistent negative gross margin, manufacturing-yield failure, technology substitution, customer concentration causing material loss, cash burn above plan, or repeated equity issuance without proportional increases in probability-adjusted production value per share. Technology quality and valuation are separate questions.