Cerebras posts $238M profit while IonQ’s quantum revenue grows 287%

The split between Cerebras Systems and IonQ has become one of the clearest illustrations of a divide running through the technology industry: between companies that make money building AI infrastructure now, and companies scaling toward a fundamentally different kind of computing.

Cerebras built its business on a radical architectural choice. Its WSE-3 chip fills an entire 300mm semiconductor wafer in one continuous piece, eliminating the inter-chip communication bottlenecks that limit conventional GPU clusters. The platform serves enterprise, government, and high-performance computing customers across North America, Asia, and Europe. Cerebras reported $510 million in revenue for 2025, up 75.7% year-over-year, and generated $238 million in net income — a profit margin that is rare in AI hardware.

IonQ operates in a different physics. Its trapped-ion quantum computers use charged atomic particles suspended in electromagnetic fields as their computational units — qubits that, in theory, solve certain problem classes exponentially faster than any silicon device. The commercial reality is more incremental: IonQ generated roughly $130 million in 2025 revenue, a quarter of Cerebras’s total. But that figure grew at 202% for the year and accelerated to 287% year-over-year in the second quarter of 2026, the company’s fifth consecutive record quarter, delivered almost entirely through cloud partnerships with Amazon Web Services and Microsoft Azure.

Both companies’ financials carry substantive caveats. Cerebras’s profitability sits alongside negative free cash flow of $680 million for the twelve months through mid-2026, reflecting capital expenditure that operating earnings do not yet offset. Its competitive position faces sustained pressure from Nvidia‘s dominant hold on AI hardware and from custom chip programs at Google, Amazon, and Microsoft targeting the same enterprise workloads. IonQ burns $484 million in cash against $130 million in revenue; its price-to-sales multiple of 110.5x embeds assumptions about commercial quantum advantage that remain unproven at scale. Recent acquisitions including SkyWater Technology and Vector Atomic have drawn management attention and capital at the same moment the core technology needs both.

For the engineers and organizations actually purchasing computing capacity, neither product is simple to deploy. Cerebras hardware requires tight integration with specific AI frameworks; IonQ’s quantum cloud is productive today primarily for scientific simulation and optimization rather than general enterprise AI workloads — a constraint that concentrates IonQ’s growth in a narrow segment of buyers. Both companies operate in a regulatory environment complicated by US export controls on advanced semiconductor technology and national-security designations on quantum computing transfers.

IonQ is scheduled to report third-quarter 2026 results in early November. Cerebras has not confirmed its next earnings date, though analysts expect a report in the same window.

Tags: AI infrastructure, AI chips, quantum computing, Cerebras Systems, IonQ