Top Companies' AI Moves

Today's Stories

1. 1

   Arista Networks' balance sheet revealed a buildup of AI chip orders and deferred revenue months before its stock surged 76%, showing how close readers of financial details could have spotted the boom.

   Arista's purchase commitments—legally binding orders for future inventory, mostly high-end chips—jumped to $3.1 billion（約5000億円） by the end of Q4 2024 (up from $2.4 billion（約3800億円） the prior quarter), then hit $3.5 billion（約5600億円） in Q1 2025. Separately, the company's product deferred revenue (sales shipped but not yet officially booked due to acceptance clauses) grew by $320 million（約510億円） in Q3 2024, $150 million（約240億円） in Q4, and $219 million（約350億円） in Q1 2025. Management explicitly tied the chip orders to "purchases for chips related to new products and AI deployments." These two balance-sheet signals—rising chip commitments and a growing pile of unbooked revenue—telegraphed heavy AI demand long before earnings showed it. The deferred revenue specifically reflected large, complex AI projects underway with major customers that would eventually turn into recognized revenue. The CEO likened the pattern to a 2016 cloud boom, signaling a structural shift in customer demand rather than a temporary spike. For businesses tracking networking stocks, this shows that financial statements can reveal demand signals before they appear in official results.

   Arista is targeting $1.5 billion（約2400億円） in AI centers revenue for 2025 and has customers heading towards 50,000 GPU deployments. The company is managing this growth while holding its margins—a contrast to some legacy networking peers who are seeing profitability pressure as they pivot to hyperscaler AI architectures. One practical signal investors watch: companies that raise their own guidance in real time, tracked by Guidance Momentum rankings.

2. 2

   Three Publishers Account for 62% of Credit Card AI Citations, Raising Questions About Data Concentration in AI Training

   A new analysis found that three publishers supply 62% of the citations used to train AI models in the credit card industry, according to research highlighted in PR Newswire reporting on publisher contribution patterns to AI model datasets. This concentration suggests that AI systems in the financial sector may be drawing from a narrow set of sources, which could mean the models reflect a limited range of perspectives and information. For businesses and financial institutions relying on or building AI systems, this dependency on a small number of publishers may carry risks around data quality, bias, and reliability.

   The finding underscores ongoing debate over how AI training data is sourced and whether broader publisher participation is needed to create more representative AI models in financial services and other sectors.

3. 3

   Forward-looking leadership: How this telecom giant puts AI into action

   Forward-looking leadership: How this telecom giant puts AI into action

4. 4

   Thermo Fisher Scientific's leadership outlined how artificial intelligence and strategic partnerships are transforming the contract drug development sector.

   Dr Anil Kane from Thermo Fisher Scientific discussed at BIO 2026 how AI and strategic partnerships are reshaping contract drug development. The pharmaceutical and contract manufacturing sectors rely on partnerships and operational efficiency to bring drugs to market. AI and strategic alignment can streamline development timelines and reduce costs, which affects how companies plan outsourced development.

   The discussion centered on reshaping contract drug development—a core service area where Thermo Fisher operates. Further announcements about specific AI applications or partnership frameworks in this space would clarify the competitive implications for the broader contract manufacturing industry.

5. 5

   With $15M, BI becomes latest to buy into Immunai's AI platform

   With $15M, BI becomes latest to buy into Immunai's AI platform

6. 6

   Booking Holdings faces potential AI-driven disintermediation, but the company argues the threat is being overstated.

   Booking Holdings is under investor scrutiny over concerns that AI tools could bypass travel agencies and booking platforms, allowing customers to book directly with hotels and airlines instead. Booking Holdings' business model depends on being the intermediary between travelers and service providers; if AI reduces the need for platforms like Booking.com, the company's revenue and market position could face pressure.

   The article frames this as a strategic reassessment—Booking's ability to demonstrate that its platform remains essential despite AI advances will be key to maintaining investor confidence.