Nvidia shatters earnings, Google’s woke AI falters, Groq unveils LPU breakthrough

Nvidia’s fourth‑quarter results blew past Wall Street forecasts, delivering $18.1 billion in revenue and a net income of $6.2 billion—both roughly double the company’s guidance, according to the firm’s earnings release cited by Forbes. The blowout lifted the stock nearly 10 percent in after‑hours trading, pushing the chipmaker back above the $1 trillion market‑cap threshold it crossed earlier this year. Analysts on the Bloomberg terminal flagged the numbers as “the strongest quarterly performance since the AI boom began,” noting that data‑center sales alone grew 45 percent year‑over‑year, a trend that mirrors the “dot‑com era” surge in demand for networking gear, as highlighted in a MarketWatch commentary linking Nvidia to the Cisco boom of the late 1990s.

The earnings surge has reignited debate over Nvidia’s long‑term valuation. A thread from investor Kobeissi on X (formerly Twitter) argues that the company’s terminal value could exceed $2 trillion if it sustains current growth rates, while a contrasting view from a Wall Street note—referenced in the All‑in podcast— warns that the “bear case” hinges on a potential slowdown in AI‑driven spend once enterprise budgets normalize. Nonetheless, the consensus among the podcast’s hosts is that Nvidia’s dominance in GPU‑based AI training and inference has created a “network effect” that is hard for competitors to dislodge, especially as the firm expands its custom‑silicon roadmap to address emerging workloads such as generative‑AI inference at the edge.

Google’s AI rollout, by contrast, has stumbled amid a cultural backlash that the All‑in podcast describes as a “woke AI disaster.” The company launched Gemini‑Advanced and a suite of Gemini‑for‑Workspace tools in early February, but internal tweets from employees—including a thread by Ben Thompson—suggest that the rollout was hampered by public criticism of the firm’s AI‑ethics stance. A post by Google’s own AI‑responsibility page emphasizes the company’s commitment to “human‑centered AI,” yet external commentary on X points to a perception that the policy framework is being used to curb the speed of product releases. The backlash has translated into a muted market reaction: Google’s parent Alphabet saw its shares dip 2 percent on the day the Gemini announcements were made, according to Bloomberg data cited in the podcast.

Amid the turbulence for the two AI giants, chip‑design startup Groq announced a “breakthrough” in its proprietary Learning Processing Unit (LPU) architecture. In a segment of the All‑in episode that runs from 49:37 to 1:17:17, Groq’s CEO explained that the LPU can deliver up to 2.5 × the throughput of a comparable GPU for inference‑only workloads, while consuming 30 percent less power. The company’s claim rests on a new “data‑flow” pipeline that eliminates the need for traditional memory‑fetch cycles, a design choice that the podcast hosts compare to the early advantages that ASICs held over general‑purpose CPUs in the 1990s. While Groq’s revenue remains modest—estimated at under $100 million in the latest private filing—the firm has secured contracts with several hyperscale cloud providers, suggesting that its LPU could become a niche alternative for latency‑critical AI services.

The juxtaposition of Nvidia’s record earnings, Google’s politically charged AI rollout, and Groq’s hardware innovation underscores a broader shift in the AI ecosystem: performance, governance, and specialization are now competing vectors for market share. Nvidia’s financial muscle gives it the bandwidth to double‑down on custom silicon and data‑center expansion, but it also invites scrutiny over whether its valuation is sustainable in a market that could see “AI fatigue” once the hype cycle matures. Google’s struggle to balance ethical guidelines with rapid product delivery may force the search‑engine titan to rethink its go‑to‑market strategy, especially as rivals like Anthropic and Microsoft accelerate their own Gemini‑style offerings. Meanwhile, Groq’s LPU breakthrough hints that the GPU monopoly is not unassailable; if the startup can scale production and prove real‑world performance gains, it could carve out a profitable niche in the increasingly heterogeneous AI hardware landscape.