Kaplan Fox & Kilsheimer LLP is investigating potential securities law violations by Cerebras Systems Inc. (NASDAQ: CBRS) after the artificial intelligence infrastructure company’s stock price dropped 19.61% in a single trading session. The investigation targets disclosures surrounding the company’s initial public offering and subsequent financial guidance, which may have misled investors about operational costs and margin sustainability.

The probe was announced on August 5, 2026, following a sharp decline in Cerebras’ share price from $226.72 on June 23, 2026, to $182.26 on June 24, 2026, a loss of $44.26 per share. This market reaction occurred immediately after Cerebras reported its first-quarter fiscal year 2026 results and provided outlook details for the second quarter of fiscal year 2026.

Cerebras, which designs and manufactures AI compute platforms including processors and data centers, conducted its initial public offering on or around May 14, 2026. The company sold 30 million shares of Class A common stock at an offer price of $185 per share.

Operational Cost Concerns

During the earnings call on June 23, 2026, Cerebras’ Chief Financial Officer disclosed that the company is temporarily renting its own systems back from an existing customer to accelerate service delivery for its contracted backlog. This strategy aims to make capacity available sooner while the company builds out its own data center infrastructure.

The CFO stated that the additional cost of renting third-party capacity will depress core cloud and other services margins temporarily from current levels. The company expects this impact to decrease margins by 10 to 15 points based on anticipated volumes before ramping back toward a target margin of 60% plus as it transitions away from rented systems.

What the Numbers Show

The disclosure reveals a significant dependency on third-party capacity to meet near-term demand, creating a temporary but substantial drag on profitability. While the company maintains a target margin of 60% plus, the immediate outlook suggests a double-digit percentage point reduction in core cloud margins. This divergence between long-term targets and short-term execution highlights the capital intensity of scaling AI infrastructure rapidly.

Investors who suffered losses or possess information relevant to the investigation are urged to contact Kaplan Fox & Kilsheimer LLP. The firm, founded in 1956, has recovered more than $10 billion for clients in securities, antitrust, and consumer protection actions nationwide.

Lovable and Cerebras Systems (NASDAQ: CBRS) announced a partnership on Aug 5, 2026, to integrate Cerebras' high-performance inference technology into Lovable's software creation platform. Based in Stockholm and Sunnyvale, Calif., respectively, the companies will run selected latency-sensitive workloads on dedicated Cerebras capacity. This move addresses the critical need for speed in AI-driven software development, where every instruction requires a round trip to a model. By reducing response times, the partnership aims to transform the user experience from a sequence of waiting periods into a seamless, interactive workflow for millions of developers and business users.

Partnership Details

Under the agreement, Lovable will utilize Cerebras' Wafer-Scale Engine to accelerate its operations. The two companies are jointly exploring product experiences that faster inference enables, with further technical results and availability details to be shared as the work progresses. This collaboration marks an expansion for Cerebras into new AI-native markets, demonstrating how ultra-fast inference can support interactive applications beyond traditional enterprise AI.

Technical Rationale

Cerebras built its Wafer-Scale Engine to remove speed constraints on AI capabilities. Unlike GPU-based systems that split model weights across many chips and incur networking taxes on every token, the Wafer-Scale Engine keeps an entire model's weights on a single wafer. This architecture delivers significantly greater memory bandwidth than the fastest GPU, generating tokens fast enough to make multi-step AI workflows feel instantaneous. Software creation involves a chain of decode-bound workloads—planning, scaffolding, writing components, and debugging—where output tokens must be produced sequentially. This is precisely the workload type the Wafer-Scale Engine is designed to accelerate, turning batch processes into conversational interactions.

Market Context

Since its launch in November 2024, more than 50 million projects have been built on Lovable, ranging from internal tools to entire companies. These projects are created by solopreneurs, small business owners, and teams at companies such as Adidas and Zendesk. Lovable is backed by leading investors including Menlo Ventures, CapitalG, and Accel. The platform operates with headquarters in Stockholm and teams in London, Boston, New York, and San Francisco.

Executive Commentary

Anton Osika, co-founder and CEO of Lovable, stated that the company was built on the idea that the person closest to a problem should be able to solve it. He noted that Cerebras helps make Lovable respond as fast as customers think, allowing users to build ideas without wondering if they are worth trying. Andrew Feldman, CEO and co-founder of Cerebras, emphasized that fast AI is more valuable than slow AI. He added that when AI responds in real-time, users do more with it, stay longer, and run higher-value workloads.

Regulatory Disclosures

Cerebras disclosed that it uses its investor relations page, X account (@cerebras), and LinkedIn page to disclose material non-public information in compliance with Regulation FD. Investors are advised to monitor these channels alongside press releases, Securities and Exchange Commission (SEC) filings, public conference calls, and webcasts. The release includes forward-looking statements regarding the partnership's implementation, anticipated benefits, and Cerebras' expansion into new markets. These statements are subject to risks including Cerebras' ability to sustain growth, access capital, and achieve profitability, as well as its dependence on significant customers such as OpenAI, Group 42 Holding Ltd, Mohamed bin Zayed University of Artificial Intelligence, and AWS.