Salesforce, Figma Decline After Earnings | Bloomberg Tech 9/4/2025

Key Concepts:

• AI Impact on Software: The current impact of AI on software company revenues and investor expectations.
• Generative AI: The adoption and integration of generative AI in various sectors.
• AI Infrastructure vs. Use Cases: The distinction between investing in AI infrastructure and AI use cases.
• Concentration Risk: The risk associated with concentrated investments in a few tech companies.
• Diversification: The strategy of spreading investments across different sectors and geographies to mitigate risk.
• AI Web Search: The development and integration of AI-powered web search tools.
• Agentic AI: The evolution of AI towards more complex, multi-step task execution.
• Data Center Buildout: The ongoing expansion of data center infrastructure to support AI and other technologies.
• AI Model Development: The development of advanced AI models with agentic capabilities.
• Export Restrictions: U.S. government restrictions on exporting technology to China.
• Open Source Models: AI models that are publicly available and can be modified and distributed.

1. Software Market and AI's Impact

• Salesforce Selloff: Salesforce's stock dropped because AI isn't paying off as quickly as expected, reflecting high market expectations.
• Discretionary Spending: Companies like Workday are struggling to see discretionary spending, impacting revenue for software companies except those focused solely on AI infrastructure.
• Figment's IPO: Figment, despite pivoting to generative AI, met but didn't exceed expectations post-IPO, indicating the need for improved fundamentals to justify its valuation.
• Investor Expectations: Investors are black-and-white, seeking immediate results from AI investments, while the integration and deployment of AI take time.
• Old vs. New Companies: Established companies like Salesforce have the cash flow to acquire smaller, innovative AI companies, creating a "whales and krill" scenario.
• Earnings Talk: Investors are focusing on whether companies meet or exceed high earnings expectations and if AI has a tangible contribution.
• Timeframe Understanding: Investors need to understand the timeframe for AI investments and whether companies are generating desired revenue, even if they don't beat the wildest expectations.
• Margin Analysis: Software companies generally have great margins relative to the S&P 500, making them cash-generating businesses.
• Switching Costs: Once a software system is implemented, it's difficult for companies to switch, providing stability.

2. Concentration Risk and Diversification

• Profit Estimates: Profit estimates for the MAG Seven (Magnificent Seven) and tech generally are much better than the S&P 500 aggregate, but concentration risk exists.
• Higher Valuations: Higher valuations and markets primed for production leave little wiggle room for pullbacks.
• Volatility Expectation: Investors should expect volatility and diversify into other industries, segments, parts of the globe, and income streams to balance portfolios.
• Hardware vs. Software: Hardware (e.g., Broadcom, up 30% year-to-date) is currently outperforming software (e.g., Salesforce, down 30% year-to-date).
• Data Centers: Building data centers takes time (5-6 years to get to the grid), creating various investment opportunities in hardware and reshoring activities.

3. Apple's AI Web Search and Siri Revamp

• AI Web Search Tool: Apple plans to launch an AI web search tool integrated into Siri, rivaling OpenAI and Google.
• Siri Overhaul: An overhauled version of Siri is expected around March, including tapping into personal data.
• Answer Engine: The AI web search tool will function as an answer engine, providing information from the open web.
• IOS 26.4: The AI web search tool will be launched as part of an update called iOS 26.4 and integrated into Safari and Spotlight.
• Google Search Revenue: Google Search is very important to Apple, generating roughly $20 billion per year.
• Siri Redesign: The Siri revamp includes a redesign, a health AI subscription service, and conversational features on home devices.
• Personal Data Access: Siri will be able to access personal data to fulfill queries, such as "Who did I meet with a month ago?"
• On-Screen Content: Siri will be able to fulfill queries based on on-screen content, such as "Tell me more about this picture."
• Health Plus Subscription: A health AI agent subscription service is planned for later in the year.
• Conversational Siri: A conversational Siri tied to a new robotic home device is expected the year after next.
• AI Model Partnerships: Apple has talked to OpenAI, Anthropic, and Google to rebuild Siri, with signs pointing to a partnership with Google's Gemini.
• Anthropic Cost: Anthropic's cost (north of $1.5 billion a year) is a concern for Apple, which feels it has little leverage.

4. Deepseek's AI Model Development

• Agentic Capabilities: Deepseek is developing an AI model with more advanced agentic capabilities to handle complex, multi-step tasks.
• Market Shift: The tech industry has shifted towards pushing out products with agentic capabilities.
• R2 Model: Deepseek's new model is expected to be released before the end of the year.
• Chinese Market: Deepseek, originally from China but now moving to Singapore, may find a bigger market within China and globally.

5. Revolut's Private Status

• Share Offering: Revolut is offering shares at a $75 billion valuation.
• Remaining Private: Revolut is working to remain private while retaining talent and pleasing VC backers.
• Employee Satisfaction: Keeping employees happy provides a competitive advantage.
• Paper Valuations: Paper valuations are not money in the bank until they crystallize.

6. HP's Performance and AI Strategy

• Fiscal Third Quarter Earnings: HP reported solid fiscal third-quarter earnings with record-breaking revenue and expanded profitability.
• Server Segment: The traditional server business is returning to historical levels, with AI making up more than 50% of orders.
• Profitability Improvement: AI server profitability is set to improve.
• Customer Types: HP participates in different segments (service providers, sovereign clouds, and enterprise) with different approaches.
• Enterprise Growth: Enterprise grew for the seventh consecutive quarter, doubling the number of logos and growing more than 250%.
• Traditional Server Business: HP is doing better than peers like Dell and Meta in the traditional server business due to a refresh cycle and strong platforms (HP Gen11 and 12).
• Storage Business: HP has transformed the storage industry by deploying a consistent architecture.
• Enterprise IT Spending Outlook: The pipeline is solid across network, server, and hybrid cloud segments.
• Modernization: Enterprises need to modernize and deploy AI, freeing up space, power, and cooling.
• Juniper Acquisition: The Juniper acquisition will create a complete network portfolio, making HP unique in building the best network in business.
• Cost Savings: HP expects $600 million in cost savings from the Juniper combination.
• Nvidia Partnership: HP has a strong partnership with Nvidia, co-engineering solutions like the AI Factor Enterprise.
• Generative AI Adoption: HP is moving from generative AI to gigantic AI, transforming business processes.
• Agentic AI Deployment: HP is deploying agentic AI across finance and marketing, with over 60 use cases deployed and 200 in testing.
• Cloud-Driven Company: HP's vision is to be a cloud-driven company, with the Juniper acquisition making it a network company.

7. Trump Administration and Tech

• Rose Garden Meeting: Tech leaders, including Mark Zuckerberg and Tim Cook, are set to gather at the White House for a discussion about AI and education.
• AI Investment: The Trump administration has made it clear they are investing in AI and domestic production.
• Chipmakers: The administration has been closely aligned with chip manufacturers.
• Export Restrictions: The Trump administration has torn up a Biden-era compromise allowing some chipmakers to maintain operations in China.
• Validated End User Authorizations (V.E.U.’s): The administration ended blanket waivers that allowed companies like Samsung and TSMC to avoid becoming collateral damage in the U.S. effort to rein in China’s AI ambitions.
• Data Leakage: Concerns about data leakage from factories in China and leveling the playing field with South Korean and Taiwanese companies led to the end of the waivers.

8. Mistral's Valuation and AI Startup Demand

• New Investment: Mistral is in talks to raise a new investment that would value the company at $14 billion.
• Open Source Models: Mistral's niche is developing open-source models.
• European Company: Mistral is a European company, potentially preferred by European users.
• Investor Appetite: There is an incredible amount of appetite for AI startups, particularly those focused on large language models and application layer AI.

9. Conclusion

The tech industry is currently navigating a complex landscape shaped by AI. While investor expectations for immediate returns from AI are high, the integration and deployment of AI take time. Diversification remains a key strategy to mitigate concentration risk in tech investments. Companies are actively exploring and implementing AI solutions, with a focus on agentic AI and transforming business processes. Government policies, such as export restrictions, also play a significant role in shaping the industry. The demand for AI startups remains strong, reflecting the long-term potential of AI technologies.