$SIVE $SIVEF THE 2025 ANNUAL REPORT IS NOTABLE FOR STRONG TOP-LINE MOMENTUM BUT ALSO MATERIAL ACCOUNTING, CONTROL, LIQUIDITY, AND EXECUTION RISK.
Sivers generated 306.6 MSEK of 2025 revenue, up 40%, driven primarily by Wireless, while adjusted EBITDA improved only marginally to -50.3 MSEK and reported EBITDA deteriorated to -93.4 MSEK. The investment case is therefore still less about current profitability and more about whether SATCOM, FWA, defense, AI photonics, and LIDAR programs convert into scalable production revenue during 2026–2027.
EXECUTIVE SUMMARY
Sivers appears to have moved from a largely development-stage semiconductor platform toward a more commercially validated but still financially fragile scale-up. The strategic positioning is attractive: Wireless is levered to SATCOM terminals, mmWave FWA, and defense electronics, while Photonics is increasingly tied to AI datacenter optical interconnects, external laser sources, co-packaged optics, and LIDAR. These are structurally high-growth markets with strategic relevance, particularly where sovereign communications, AI infrastructure power efficiency, and defense modernization are priorities. However, the company’s 2025 financials do not yet demonstrate operating leverage. Gross profit was -2.1 MSEK despite 306.6 MSEK of revenue, implying that revenue mix, project accounting, inventory adjustments, underabsorbed cost structure, and NRE economics remain problematic.
The most important positive signal is commercial momentum. Wireless revenue increased to 213.1 MSEK, Photonics revenue increased to 93.4 MSEK, and NRE revenue reached 178.7 MSEK. This confirms customer engagement, but it also highlights that revenue quality remains heavily development-program driven rather than volume-product driven. The first material production conversion appears expected from late 2026, with management pointing to https://t.co/J3qK1bN6x5, a Tier-1 FWA customer, Tachyon Networks, Doosan, ESA/IRIS2-related SATCOM opportunities, POET/O-Net AI optical opportunities, and a major LIDAR customer ramp beginning in Q4 2026 with meaningful revenue from 2027.
The most important negative signal is financial credibility. The annual report includes restatements, a change in income statement presentation, identified material weaknesses in internal control over financial reporting, a 2025 IT intrusion, and an auditor emphasis of matter on going concern. These are not cosmetic issues. They raise the cost of capital, reduce confidence in reported project margins and contract assets, and increase the risk that future reported revenue timing, capitalization, and impairment assumptions may require further revision.
FINANCIAL QUALITY
Revenue growth of 40% is the clearest headline improvement, but revenue quality is mixed. NRE revenue was 178.7 MSEK, representing roughly 58% of total revenue, while hardware revenue was 127.2 MSEK, or roughly 41%. For a semiconductor company seeking scale economics, this mix indicates that customer-funded development remains the dominant revenue source. NRE can be strategically valuable because it validates design wins and funds development, but it is typically less attractive than recurring production revenue unless it converts into high-volume shipments with sustainable gross margin.
Gross margin was effectively negative, with gross profit of -2.1 MSEK versus 4.2 MSEK in 2024. This is the central financial concern. The company grew revenue materially but failed to translate that growth into positive gross profit. Cost of goods sold was 308.6 MSEK, exceeding revenue. This suggests that current programs still carry unfavorable cost absorption, project costs, inventory valuation pressure, or low-margin development economics. In investment terms, the 2025 result does not yet support a thesis that scale alone will solve profitability without evidence of better production mix, pricing, manufacturing yield, and operating discipline.
EBITDA was -93.4 MSEK versus -68.3 MSEK in 2024, while adjusted EBITDA was -50.3 MSEK versus -51.3 MSEK. The adjusted result shows stabilization, not inflection. Reported EBITDA was burdened by 43.1 MSEK of one-time items, including 14.7 MSEK of strategic projects, 11.5 MSEK of restructuring, 8.5 MSEK of incentive-program costs, and 8.5 MSEK related to the IT intrusion. The recurring loss base remains meaningful even after adjustments.
Liquidity remains tight. Cash was 29.7 MSEK at year-end, with 13.8 MSEK of restricted cash. Operating cash flow was -38.5 MSEK and investment cash flow was -54.6 MSEK, producing cash burn before financing of -93.1 MSEK. The company funded this through equity issuance and debt, including 202.8 MSEK of 2025 gross equity proceeds and a 12 MUSD loan. Subsequent refinancing in February 2026 and the 125 MSEK directed issue approved in May 2026 are important liquidity bridges, but they also reinforce dependence on external capital.
BALANCE SHEET AND CAPITAL STRUCTURE
The balance sheet remains asset-heavy, with 1.0 BSEK of intangible assets, including 370.1 MSEK of goodwill, 478.5 MSEK of other intangibles, and 174.7 MSEK of capitalized development costs. Equity was 949.8 MSEK and reported solidity was 71%, which appears strong superficially. However, the high intangible weighting and negative free cash flow make tangible liquidity and capital-market access more relevant than accounting equity.
Impairment sensitivity is material. The Wireless cash-generating unit had 516 MSEK of headroom, while Photonics had 293 MSEK. The impairment model depends on aggressive multi-year growth assumptions, including average 5-year revenue growth of 28.2% for Wireless and 43% for Photonics, with pre-tax discount rates of 19.3% and 27.5%, respectively. These assumptions are not unreasonable for a semiconductor scale-up exposed to AI, SATCOM, and defense, but they leave valuation highly sensitive to execution slippage.
COMMERCIAL OUTLOOK
Wireless is the nearer-term strategic anchor. SATCOM traction with https://t.co/J3qK1bN6x5, ESA, Doosan, Intelsat, IRIS2-related terminal opportunities, and DIFI participation suggests meaningful relevance in electronically steerable terminals and sovereign satellite communications. The company’s claim of being the only EU-based supplier of silicon-based mmWave beamformer ICs for SATCOM is strategically important if accurate, particularly under European security and supply-chain sovereignty priorities.
FWA remains a more binary opportunity. The Tier-1 infrastructure customer is expected to launch a first-generation product by late 2026, and Tachyon placed a 2.8 MUSD production order for 28 GHz modules. This matters because FWA has historically been a difficult market for mmWave suppliers due to operator capex cyclicality, installation economics, and patchy global mmWave adoption. A credible Tier-1 launch would materially improve the investment case, but until customer identity, production volumes, pricing, and gross margin are visible, the FWA thesis remains promising but unproven.
Defense is becoming a more credible pillar. CHIPS Act-linked engagement, EW-STAR funding, a 6.4 MUSD year-2 award, FR3 beamforming transceiver work, and an 800,000 USD contract with a major U.S. defense contractor improve strategic relevance. Defense revenue can be attractive because it may support differentiated technology, long program duration, and non-dilutive funding. However, conversion from funded development to scalable production often takes years, and program timing can be volatile.
Photonics has the highest strategic upside but the least near-term financial proof. Partnerships with WIN Semiconductor, O-Net, POET, and LIGHTIUM position Sivers in external laser sources, DFB laser arrays, optical engines, TFLN integration, co-packaged optics, and pluggable transceivers for AI infrastructure. This is directionally well aligned with the industry’s transition toward 800G, 1.6T, and eventually 3.2T optical connectivity. The stated customer sampling activity and expected production commitments from 2027 onward are important, but the business still needs evidence of volume qualification, supply reliability, yield, and margin structure.
KEY RISKS
The principal risk is execution. Sivers must convert development programs into production revenue while simultaneously improving gross margin, strengthening controls, retaining engineering talent, managing supply-chain dependencies, and financing cash burn. This is a difficult combination for a small-cap semiconductor company competing against larger players with superior balance sheets, manufacturing leverage, and customer access.
Customer concentration is high. More than 50% of 2025 revenue came from 3 Wireless customers, with the 3 largest contributing 183.4 MSEK. This creates revenue volatility, negotiating leverage for customers, and elevated downside risk if any major program is delayed, descoped, or repriced.
Accounting and governance risk is elevated. The company corrected prior-period errors, changed its income statement presentation, disclosed material weaknesses in internal controls, and received an auditor emphasis of matter on going concern. These factors make reported financials less clean than preferred for a public semiconductor investment case and warrant a higher discount rate until remediation is demonstrably complete.
INVESTMENT VIEW
Sivers represents a high-upside, high-risk semiconductor scale-up rather than a conventionally de-risked growth company. The strategic assets appear relevant to several attractive markets, including SATCOM, defense, AI optical interconnect, and LIDAR. The 2025 report strengthens the argument that customer engagement and technology relevance are improving. However, it does not yet prove the economic model. Negative gross profit, negative adjusted EBITDA, cash burn, customer concentration, control deficiencies, and going-concern language materially constrain conviction.
The key underwriting question is whether 2026–2027 production ramps convert the company from NRE-heavy development revenue into repeatable product revenue with positive gross margin. Until that conversion is visible, the equity should be viewed as an option on commercial scale-up rather than a fundamentally validated operating compounder.
