Cerebras Makes a Strong Quarterly Debut, Morgan Stanley Raises Target Price to $273 and Maintains Overweight

Overview

Morgan Stanley views Cerebras' first post-IPO quarter as a "strong debut": revenue beat expectations, gross margin and operating margin improved even more sharply, and management's guidance for next quarter and the full year 2026 both came in above the prior model. The firm maintains Overweight and raises its target price from $250 to $273. The core investment thesis centers on demand for low-latency AI inference, the ramp-up of 750MW of contracted capacity, the buildout of Cerebras Cloud, and the potential incremental upside from the formal Amazon agreement.

Core views

The report's core view is that Cerebras has a differentiated architecture and a leading position in the fast-growing AI inference niche. The company has commercially deployed wafer-scale processors, which Morgan Stanley believes allow it to serve customers with strong demand for speed and high-value tokens. Although the market is still debating whether wafer-scale computing can scale once a model exceeds a single wafer, the company's performance in large-model demos such as Kimi K2.6 is seen as positive evidence of scalability. The most important near-term variable is whether the 750MW of contracted capacity can ramp on schedule; over the medium term, the focus is on whether the Amazon agreement, additional customers, and cloud capacity buildout can drive revenue above expectations.

Analysis framework

The report combines earnings comparisons, guidance revisions, scenario-based valuation, and peer multiple comparison. Morgan Stanley compares actual revenue, gross margin, operating margin, and EPS against its own model expectations, then updates its 2026-2028 revenue, gross margin, and EPS forecasts. The target price is based on roughly $6bn of adjusted revenue in 2028, a 13x revenue multiple, and a comparison with valuation levels among AI semiconductor peers.

Methodology notes

• Valuation FrameworkRevenue Multiple Valuation

  13x 2028 adjusted revenue

  The $273 target price is based on roughly $6bn of 2028 adjusted revenue and a 13x multiple; the multiple was raised from 12x to 13x, reflecting a recent valuation re-rating among AI semiconductor peers.

• Scenario AnalysisRisk Reward

  Bull, base, and bear scenarios

  The bull case is $400, the base case is $273, and the bear case is $90; the key differences are the speed of contracted-capacity ramp-up, hardware demand, cloud capacity utilization, and 2026-2028 revenue CAGR.

• Performance ComparisonMorgan Stanley ModelWare

  Company actuals versus Morgan Stanley estimates and consensus

  The report notes that, unless otherwise stated, metrics are based on the Morgan Stanley ModelWare framework; consensus data comes from Refinitiv Estimates.

• Theme AssessmentTechnology Diffusion

  Positive technology diffusion

  The risk-reward theme marks Technology Diffusion as Positive, reflecting the report's positive view on the commercial diffusion of Cerebras' rapid inference technology.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• CEREBRAS SYSTEMS INC / CBRS.US

  Core coverage name

  Strengths

  Clear low-latency inference positioning, differentiated wafer-scale architecture, 750MW of contracted capacity supports the revenue ramp, and both quarterly revenue and gross margin beat expectations.

  Weaknesses

  The business is still in an early scaling phase, cloud capacity buildout requires significant execution, and the market continues to debate wafer-scale scalability.

  Comparison

  The valuation reference is AI semiconductor peers, and the report believes a 13x 2028 adjusted revenue multiple is consistent with smaller-cap AI companies that have high AI exposure and faster-than-industry growth.

  Risks

  Capacity deployment delays, customer concentration, intensifying competition, weaker-than-expected hardware demand, and cloud infrastructure buildout risk.

• Amazon relationship

  Potential incremental customer and ecosystem partner

  Strengths

  The relationship has become a formal agreement and could provide Cerebras hardware support for decode workloads in the Trainium 3 ecosystem.

  Weaknesses

  Morgan Stanley's current model still treats the Amazon opportunity as a relatively small contribution, and actual revenue conversion remains to be seen.

  Comparison

  Compared with the 750MW contract, Amazon is not the biggest near-term driver in the model, but it could become a source of upside surprise.

  Risks

  Near-term compute capacity is already allocated to existing contracts, and the pace and scale of incremental demand realization remain uncertain.

• Cerebras Cloud

  Revenue ramp vehicle

  Strengths

  Once the contracted capacity comes online, it can absorb rapid inference demand, and if demand exceeds supply it has pricing and utilization flexibility.

  Weaknesses

  Cloud buildout is affected by progress on data centers, space, power, and other infrastructure.

  Comparison

  The report treats cloud service assumptions as more conservative than hardware purchase options and continues to model it as a cloud business.

  Risks

  Capacity deployment delays, utilization below expectations, and transition lease capacity dragging on margins.

Key data

• RatingOverweightRating unchanged.
• Target Price$273.00Raised from $250.00 to $273.00.
• Current Price$226.72Closing price on 2026-06-23.
• Core Revenueabout $191mnAbove Morgan Stanley's expectation of about $180mn; table disclosure shows Core Revenue of $191.3mn.
• GAAP Revenue$193.4mnUp about 94.3%-94.4% year over year, above Morgan Stanley's expectations.
• Next-Quarter Core Revenue Guidance$194mnAbove Morgan Stanley's expectation of $180mn, up about 88% year over year.
• Next-Quarter Non-GAAP Gross Margin Guidance37.0%Significantly above Morgan Stanley's expectation of 24.2%.
• 2026 Full-Year Revenue Guidance$860mnAbove Morgan Stanley's prior expectation of $831mn.
• Morgan Stanley 2026 Revenue Forecast$864mnUpdated 2026 core revenue forecast in the report.
• 2027 Revenue Forecast$2.71bnMorgan Stanley updated forecast.
• 2028 Revenue Forecast$6.46bnEssentially unchanged.
• 2026 Non-GAAP EPS Forecast$(0.78)Prior forecast was $(1.22), revised upward.
• Base-Case Revenue CAGR178%Assumed revenue compound annual growth rate for 2026-2028.
• Contracted Capacity750MWBase-case assumption is a ramp on schedule by 2028.

Impact & implications

This report is positive for Cerebras: the earnings and guidance validate the company's early post-IPO execution, better-than-expected gross margin eases concerns about margin pressure from rented G42 transition capacity, and the formal Amazon agreement improves visibility into potential customers and use cases. For investors, the main point of debate is no longer whether the company can meet a single quarter's numbers, but whether subsequent execution can continue to validate the size of the rapid-inference market, wafer-scale scalability, the pace of cloud capacity buildout, and customer concentration risk.

Risks

• Delays in capacity deployment and infrastructure build-out could slow the revenue ramp.
• High customer concentration means changes in key contracts or customer demand could materially affect forecasts.
• Rising competition in AI inference chips and cloud computing could compress growth or valuation multiples.
• Weaker-than-base-case hardware demand could hurt the coordinated growth of cloud capacity and hardware sales.
• Wafer-scale scalability remains a market debate point; if subsequent large-model validation is insufficient, the investment thesis could come under pressure.
• The free float is relatively thin, so the stock may experience large post-earnings volatility.

What to watch

• Whether the 750MW of contracted capacity can come online as planned by 2028.
• The progress of Cerebras Cloud's self-built capacity and its impact on gross margin improvement.
• Whether the formal Amazon agreement brings revenue contributions beyond the small assumptions in the current model.
• Whether rapid inference demand continues to exceed supply, especially in high-value, low-latency token scenarios.
• Whether full-year 2026 guidance for revenue, gross margin, and operating margin is raised again.
• Customer conversion and scaling cases following the Kimi K2.6 large-model demos.