Bloomberg Tech: AI Spending; Nvidia Buys Hugging Face (Podcast)

The piece highlights the state of the AI boom, pointing to earnings read-throughs from Broadcom, Snowflake, and HPE and includes discussion of Nvidia’s $13B acquisition of Hugging Face. It also notes a deal between robotaxi startup Wayve and Uber to offer rides in London. Overall, it is promotional/roundup coverage with no specific new financial metrics or guidance changes disclosed.

Analysis

AI spend is still flowing to the highest-leverage layer of the stack, which is why NVDA and AVGO remain the cleanest beneficiaries while HPE is more of a cyclical read-through than a true compounder. The market should treat strong AI capex prints as a signal that accelerator and networking demand is still being pulled forward for the next 1-3 quarters, but software names with consumption-based models can lag if customers are still paying for infrastructure before proving ROI. That creates a subtle rotation risk: the boom can stay intact while multiple expansion shifts away from broad AI software exposure and toward chip vendors.

A hypothetical NVDA-Hugging Face combination is strategically more important than it looks because it pulls developer mindshare closer to the hardware owner and raises switching costs across the model deployment stack. Near term, that is bullish for NVDA’s ecosystem lock-in; over 6-18 months it could backfire if hyperscalers and frontier labs accelerate secondary sourcing to AMD, custom ASICs, or internal tooling to reduce dependence on one vendor. The real falsifier is not the announcement itself, but whether subsequent capex guides broaden beyond NVIDIA-centric spend.

The Wayve/Uber angle is a longer-duration option on autonomy monetization, not a 2026 earnings driver. If partner-led robotaxi supply scales in dense cities, UBER can monetize AV demand without owning the fleet, which is the best-case path to margin expansion; but if regulatory approval or safety economics stall, the market should strip out most of the AV optionality. Consensus is likely overestimating how fast this becomes revenue and underestimating how much value accrues only after repeated city-level approvals and utilization data.

Key Decisions for Investors

• Stay long NVDA versus SNOW for the next 1-3 months; thesis is that AI dollars are still buying infrastructure, not software ROI. Falsify if SNOW consumption growth reaccelerates or NVDA guide/ramp disappoints.
• Overweight AVGO on dips versus HPE over the next earnings cycle; AVGO has higher AI content and better operating leverage, while HPE is more exposed to server digestion and lower-quality backlog conversion.
• Set a watch alert on UBER rather than forcing a trade: only add risk if the London autonomy rollout shows explicit utilization, take-rate, and regulatory milestones over the next 6-12 months.
• If NVDA screens become too crowded, consider a small hedged position: long NVDA, short a basket of AI software proxies that need proof of monetization; the risk/reward works only if capex stays hardware-heavy for another quarter or two.