Conclusion on the AI Market: It's Not 'Demand' That Wins—What Cisco's Decline Teaches Us About Gross Margins
The topics this time may seem disparate at first glance (memory prices, Siri delays, AI for the wealthy, social media lawsuits, generative AI fundraising), but the common thread is that while AI is generating 'demand,' it has also shifted the 'cost structure and axis of competition' by one level. The market is beginning to re-evaluate the 'gravity of reality'—such as profit margins, pricing power, and regulatory risk—rather than just growth stories.
1. Memory prices hit home: Cisco Systems cannot be saved by 'AI demand' alone
As a giant in enterprise networking and data center equipment, Cisco has seen tailwinds in sales from AI-related demand, but rising memory prices have pressured profit margins, causing a significant reaction in its stock price. According to Bloomberg, Cisco reported that its adjusted gross margin for the quarter ending in April was approximately 66%, falling short of market expectations (approximately 68%). This is the market's 'pain point'.
The point of discussion in the program was also clear: analysts touched on the view of a 'margin erosion of about 200bp (2%),' and Cisco itself hinted that 'price pass-through is an option.' In other words, even if AI boosts sales, if component inflation continues, the 'share of profits' will shrink. Investors are looking more severely at 'where the bottlenecks are in the AI supply chain'.
1-1. The flip side of 'AI infrastructure construction': The winner is not 'demand' but 'pricing power'
Even if the expansion of AI infrastructure continues, in the short term, stock prices are likely to move along the path of 'component prices → gross margins → guidance.' Cisco has become a 'textbook example' of this.
2. 'Platform or feature?'—The re-ordering of software in the agent era
What was impressive in the program was this organization regarding AI agents:
'What matters is not the "feature" but the "platform"'
The view was that features are 'embedded' into higher-level platforms, while platforms become the 'foundation' that agents aggregate.
This perspective also applies to existing SaaS companies like Salesforce and Shopify. Does 'AI replace software,' or does 'AI make software richer (enrichment)'? The market is starting to bet on which layer will take the lead, rather than a 'total replacement'.
3. The meaning of Apple's Siri stumbling: Delays are not 'feature lags' but 'losses of competitive time'
As reported by Bloomberg, Apple is facing challenges with the new Siri (AI overhaul) in internal testing, and it is reported that some features may be gradually pushed back from the initially expected iOS 26.4 (March) to iOS 26.5 (May) or iOS 27 (September). The market's perception is that the highlight features related to 'personal context' in particular are expected to be delayed.
What is important here, as pointed out in the program, is not the 'iPhone sales themselves,' but the fact that Apple is 'losing time' against the speed of evolution of its competitors (various generative AI companies). For AI assistants, not only the level of completion but also the 'accumulated time for learning and improvement' is effective. Delays can directly become factors that widen the gap.
4. Tax agents shake the market: The 'model of destructive power' shown by Altruist's AI tools
Bloomberg reports that the background to the sell-off in stocks related to the wealthy, securities, and advisory services was triggered by Altruist expanding its AI tax planning function (Hazel). Including the company CEO's remark that 'tax is just the beginning,' the picture of 'agents absorbing tasks' has chilled investor sentiment.
The point is that Altruist has shown the potential to (1) drastically shorten work time, (2) use barriers to entry (infrastructure/data) as a weapon, and (3) pressure existing fee structures, rather than just 'replacing human advisors.' Barron's similarly organizes the flow of how AI disruption concerns led to a sharp drop in related stocks.
5. Is social media a 'digital casino'?: Regulatory risks shown by Meta Platforms' lawsuit
In the trial underway in Los Angeles over 'youth mental health and social media design,' it is reported that Instagram head Adam Mosseri testified, emphasizing the distinction that it is not 'addiction' in a clinical sense, but 'problematic use.' The focus is on how far the 'design responsibility' of tech companies will be questioned, including the prospect of Mark Zuckerberg's testimony.
6. Another main battlefield: Capital, computing resources, and autonomous driving
Finally, understanding the flow in the latter half of the program that 'AI is ultimately a game of capital and implementation' will solidify your understanding.
Anthropic: Massive fundraising continues, and Reuters and the FT have reported on a round of approximately $30 billion and a significant increase in valuation (the lineup of investors is also expanding). This is capital to accelerate not just 'research' but also data centers and commercialization.
Waymo: Officially announced raising $16 billion and indicated a policy to accelerate overseas expansion into cities like London and Tokyo. This is a symbol of investment as AI moves into the 'real world'.
Google: Bloomberg and others reported on the update to Gemini's reasoning mode (Deep Think). It emphasizes the bridge from research to practical application.
Conclusion
What this program depicted is the moment the perspective shifts from 'AI market = dream' to 'AI market = recalculation of costs, margins, and competitive advantage.' Memory prices confront us with the reality of hardware, Siri's latency confronts us with the value of time, tax agents shake up fee models, and social media lawsuits bring regulatory risks to the surface.
For investors, there is only one question—not who is on the 'receiving end of AI's benefits,' but who is on the 'side making the rules for the AI era.'
