Texas Instruments vs. Cerebras Systems: Which Chip Stock Has the Edge?

Texas Instruments reported Q2 2026 revenue growth of 23% year over year to $5.46 billion, with EPS rising 52% to $2.14 as gross margin expanded 350bps to 61.4%; Q3 midpoint guidance implies roughly 25% revenue growth and 62% EPS growth. Cerebras posted substantially faster core-revenue growth of 103% to $209.9 million and held $25.4 billion of RPO, but recorded a $450.5 million GAAP net loss, while core gross margin fell 590bps sequentially to 40.6%. The analysis favors TXN for its diversified end markets, profitability, $6.5 billion trailing-12-month cash generation, $5.8 billion in shareholder returns, and lower 10.48x forward P/S valuation versus CBRS at 21.40x.

Analysis

TXN is a cleaner way to express an industrial/auto semiconductor upcycle than the market may appreciate. The key incremental earnings driver is not merely revenue growth but factory-loading leverage: utilization gains at its owned 300mm fabs can sustain gross-margin recovery faster than revenue, while a broad catalog and direct-sales model reduce the pricing volatility seen at merchant digital peers. Over the next 1-3 months, the relevant catalyst is whether order strength broadens beyond data center into distribution replenishment; that would support upward FY27 EPS revisions and multiple resilience despite a still-elevated rate backdrop.

CBRS is effectively a leveraged bet on inference demand converting from contracted capacity into economically attractive delivered compute. Its backlog is not equivalent to funded, near-term revenue: customer deployment schedules, power interconnection, server availability, and any renegotiation of concentrated commitments determine both conversion timing and cash burn. The more important negative signal is declining unit economics during a demand surge; if rented capacity is necessary to serve workloads, incremental revenue can consume capital without demonstrating the eventual steady-state margin profile. This creates a plausible financing/dilution overhang within 6-18 months even if top-line estimates rise.

The non-obvious relative beneficiary of CBRS scaling is AMD: a viable non-NVDA inference architecture reinforces the market’s demand for diversified accelerator supply, but it does not necessarily displace NVIDIA near term because hyperscalers can run heterogeneous fleets. Conversely, CBRS's claimed avoidance of HBM and advanced-packaging bottlenecks limits read-through to HBM/CoWoS suppliers; investors should not extrapolate its growth to the broader AI hardware supply chain. Consensus may be too quick to treat TXN as a low-growth analog name and too willing to capitalize CBRS's RPO at cloud-like margins.

Falsification for TXN: a sequential deterioration in industrial bookings or guidance that implies utilization is plateauing, especially if gross margin fails to expand despite revenue growth. Falsification for the CBRS short thesis: two consecutive quarters of gross-margin expansion toward 50%+ alongside material RPO conversion, reduced customer concentration, and credible evidence that capacity additions are funded without equity issuance.

Key Decisions for Investors

• Initiate a 3-6 month long TXN / short CBRS pair, sized beta-neutral. The trade captures factory-utilization and cash-return durability versus execution-, concentration-, and funding-risk exposure; target a 15-20% relative return, with a 7% relative stop if CBRS demonstrates margin inflection and backlog conversion.
• Add TXN on post-results weakness only if management maintains industrial order momentum and gross-margin progression; use a 6-12 month horizon. Upside is driven by FY27 EPS estimate revisions rather than multiple expansion; reduce if distribution inventory metrics or automotive demand weaken.
• Avoid outright CBRS long exposure until the next earnings release provides cohort-level evidence on RPO conversion, customer prepayments, capex commitments, and normalized gross margin excluding rented systems. A revenue beat without improving unit economics is a sell-the-rally setup, not confirmation.
• For AI diversification, prefer a modest long AMD versus CBRS rather than treating CBRS as a direct NVDA short. AMD benefits if buyers seek alternative inference supply, while CBRS retains binary risks around delivery capacity and customer concentration; reassess after hyperscaler capex guidance.