Voices
Voices
Ben Thompson Bullish
Google's AI bar is low -- it just needs Search not to die too fast.
Stratechery · Aug 18, 2026 · Full history
Quote and source
"Google just needs search to not die too quickly."
Ben Thompson Bullish
Thompson casts Search as See's Candies and AI as Google's BNSF: lower margin, far bigger absolute profit.
Stratechery · Aug 18, 2026 · Full history
Quote and source
will we look back on Google search like See's Candies? It feels like that's what's happening.
Ben Thompson Bullish
Thompson reads Google's equity issuance as confidence, not weakness: a smaller slice of a much bigger pie.
Stratechery · Aug 18, 2026 · Full history
Quote and source
Well, you have the smaller percentage of an astronomically larger pie. At the end of the day, no one's gonna be complaining.
Ben Thompson Bullish
Thompson names Google (with Amazon) as Nvidia's real long-term competitor, citing its equity-funding edge.
Stratechery · Aug 18, 2026 · Full history
Quote and source
Google can just issue equity. Shareholders don't love it, but their monetization capacity is much higher than Nvidia or Nvidia's customers are.
Ben Thompson Bearish
Thompson won't call Google a frontier lab: OpenAI and Anthropic are clearly there, Google is unclear.
Stratechery · Aug 18, 2026 · Full history
Quote and source
I think OpenAI and Anthropic are clearly on the frontier. Who knows what's happening with Google
Ben Thompson Bullish
Google is pitching Meta and large institutions on buying/hosting TPUs directly, not just renting via GCP.
Stratechery · Nov 25, 2025 · Full history
Quote and source
Ben's read: "shocking but not surprising" — incremental TPU sales let Google amortize TPU development costs over more units
Ben Thompson Bullish
Gemini 3 launch makes Google the biggest winner: full TPU-to-app stack, lowest cost, priced above GPT-5.1.
Stratechery · Nov 19, 2025 · Full history
Quote and source
Google = biggest winner: full TPU→model→app vertical integration (the structure Ben expects to win early in a new tech paradigm), massive cash flow, lowest cost structure among leading model makers via in-house TPUs.
Ben Thompson Bullish
Apple to pay Google ~$1B/yr for a custom Gemini model to power Siri, despite Anthropic scoring better.
Stratechery · Nov 10, 2025 · Full history
Quote and source
Siri white-labels Gemini: Apple finalizing a deal to pay Google ~$1B/year for a custom 1.2T-parameter Gemini model to power a rebuilt Siri
Ben Thompson Bullish
Thompson on Google's raised capex guide: the rational response is to wish it had invested even more, sooner.
Stratechery · Nov 3, 2025 · Full history
Quote and source
Ben's stance, explicit and bullish on Google's capex: the "rational response" to the capex raise isn't concern, it's to wish Google had invested more sooner, even accounting for a 41% rise in depreciation expense.
Ben Thompson Bullish
Thompson's most bullish Google piece: YouTube's AI auto-tagging makes 'every pixel monetizable.'
Stratechery · Sep 23, 2025 · Full history
Quote and source
he calls this making "every pixel monetizable."
Ben Thompson Bullish
pairing a frontier model (Gemini) with in-house silicon (TPUs) uniquely favors Google over Big Tech peers.
Stratechery · Sep 16, 2025 · Full history
Quote and source
This structurally favors Google, uniquely among Big Tech, because it pairs a competitive frontier model (Gemini) with its own custom silicon (TPUs) — an integrated model+chip strategy that neither Amazon (no leading model) nor Apple (no chip strategy at this scale) can currently match.
Brad Gerstner Neutral
You can never count Google out, Gerstner said, but flagged it has seen less AI traction lately.
Altimeter Capital · Jul 6, 2026 · Full history
Quote and source
You can never count Google out. They are the original pioneer here. They have plenty of compute. They have all the money in the world. I'm surprised that, frankly, they haven't seen more traction here as of late, but I think that they'll be in the hunt.
Brad Gerstner Bearish
Altimeter fully exited its ~$162.5M Alphabet stake in Q1 2026 after a 100% one-year run, near 27x trailing P/E.
Altimeter Capital · Jun 12, 2026 · Full history
Quote and source
Altimeter fully exited its 519,290-share Alphabet (GOOGL) position in Q1 2026, a stake worth roughly $162.5 million ... GOOGL traded near 27x trailing P/E after a 100% one-year run at time of sale.
Brad Gerstner Bullish
Alphabet was the clear winner of mega-cap earnings as search revenue grew 19%, Gerstner said.
Altimeter Capital · May 4, 2026 · Full history
Quote and source
Google, with its TPUs, its cloud services, is benefiting tremendously from what's going on in AI ... They grew search revenues 19% in the quarter, way higher than we forecast a couple years ago. I'm glad we had the mental flexibility to own Google along the way here.
Brad Gerstner Bearish
Altimeter's full Alphabet exit while buying AI infrastructure implies it sees Google's AI shift as slower or riskier.
Altimeter Capital · May 1, 2026 · Full history
Quote and source
Selling Alphabet entirely while aggressively buying AI infrastructure names is a pointed statement.
Brad Gerstner Bullish
Gerstner frames owning Google as a direct bet on Gemini and AI, alongside Amazon.
Altimeter Capital · Jan 28, 2026 · Full history
Quote and source
when I invest in Gemini, I'm investing in Google, I'm investing in Gemini and AI.
Brad Gerstner Bullish
Gemini is on a tear and Google's AI innovation rate has improved dramatically, Gerstner said, still holding it.
Altimeter Capital · Jan 6, 2026 · Full history
Quote and source
Gemini is on a tear. The model has improved dramatically. The rate of innovation has improved dramatically at Google ... We were happy to be shareholders. We continue to be shareholders in Google.
Brad Gerstner Neutral
Google's TPU won't match Nvidia's Reuben-family token efficiency despite the 'TPU kills GPU' narrative.
Altimeter Capital · Jan 6, 2026 · Full history
Quote and source
the TPU, while it has some advantages, is not going to produce the token efficiency that Reuben and the Reuben family of chips is going to produce.
Brad Gerstner Bullish
Broadcom partnering with Google on TPU7 is forcing Nvidia to get better, Gerstner said, still holding Google.
Altimeter Capital · Dec 2, 2025 · Full history
Quote and source
What Hock Tan has done partnering with Google to build out TPU7 is extraordinary. That's going to force Nvidia to be even better.
Dylan Patel Bullish
Google funded its AI buildout by halting buybacks, unlike peers piling on debt.
SemiAnalysis · Aug 25, 2026 · Full history
Quote and source
Google funded its buildout by halting buybacks without much market disruption, contrasted with the debt-funded phase other hyperscalers are entering.
Dylan Patel Bearish
SemiAnalysis: DeepMind is no longer a frontier lab; Gemini ranked 8th-9th.
SemiAnalysis · Aug 7, 2026 · Full history
Quote and source
"DeepMind is no longer a frontier lab."
Dylan Patel Bullish
SemiAnalysis projects GCP AI ARR of $73B plus $120B of TPU sales by end-2027, GCP revenue up >100% YoY now.
SemiAnalysis · Aug 7, 2026 · Full history
Quote and source
Gemini ARR ~$12B (Q2'26) against a projected $73B GCP third-party AI ARR + $120B TPU sales by end-2027
Dylan Patel Bullish
SemiAnalysis estimates Google's TPU/GCP buildout adds roughly $3 of EPS in 2027.
SemiAnalysis · Aug 7, 2026 · Full history
Quote and source
Estimated ~+$3 EPS to Google in 2027
Dylan Patel Bullish
TPUs are objectively strong hardware, running all of DeepMind and Anthropic's pretraining.
SemiAnalysis · Jun 30, 2026 · Full history
Quote and source
"TPUs are objectively strong hardware (they run all of DeepMind and Anthropic's pretraining)."
Dylan Patel Bearish
Google is getting squeezed on TSMC N3 capacity as Nvidia's share trends toward 70%+ by 2027.
SemiAnalysis · Mar 13, 2026 · Full history
Quote and source
Nvidia "trending toward 70%+" of N3 by 2027; "Google is getting squeezed" because it committed later.
Dylan Patel Bearish
Google's free cash flow goes to roughly zero in 2027 as capex absorbs every dollar earned.
SemiAnalysis · Feb 28, 2026 · Full history
Quote and source
There's no reason why Google will have any profit, uh, in 27 at all. Right? In terms of cash flow. They will just spend every dollar they make on AI infrastructure.
Dylan Patel Bullish
SemiAnalysis: Google's TPUv7 delivers extremely strong performance-per-TCO.
SemiAnalysis · Dec 4, 2025 · Full history
Quote and source
TPUv7 "extremely strong perf per TCO."
Dylan Patel Bullish
OpenAI already saved 30% on its Nvidia fleet cost purely from the threat of switching to Google TPUs.
SemiAnalysis · Dec 1, 2025 · Full history
Quote and source
"OpenAI hasn't even deployed TPUs yet, and they've already saved 30% on their entire lab-wide NVIDIA fleet."
Dylan Patel Bearish
Google's TPU engineer attrition to OpenAI hits the next-gen TPUv8, not the shipped TPUv7.
SemiAnalysis · Dec 1, 2025 · Full history
Quote and source
"chip timelines are so long, that didn't affect TPUv7, that's affecting TPUv8."
Dylan Patel Bullish
SemiAnalysis' ClusterMAX 2.0: Google tops the Silver tier and is on track for Gold or higher.
SemiAnalysis · Nov 6, 2025 · Full history
Quote and source
"Google aims to set the bar and command a pricing premium... on the fast track to Gold-tier or higher."
Dylan Patel Bullish
Patel says he turned meaningfully more bullish on Google over the prior two years as it began selling TPUs externally.
SemiAnalysis · Sep 30, 2025 · Full history
Quote and source
Turned "meaningfully more bullish on Google over the past two years as it began selling TPUs externally, improved its models, and increased infrastructure investment despite acknowledged internal dysfunction."
Gavin Baker Neutral
Google paid SpaceX ~$50B per gigawatt for compute, the richest hyperscaler rate Baker cited.
Atreides Management · Jun 11, 2026 · Full history
Quote and source
They just signed Anthropic at 22 to 23. They just signed Google at 50.
Gavin Baker Bearish
Google lost the Pareto-frontier lead on intelligence-per-cost as conservative TPU v8 choices ceded ground.
Atreides Management · May 20, 2026 · Full history
Quote and source
Google dominated the Pareto frontier (intelligence vs. cost) 9 months ago; now it's dominated by Anthropic and OpenAI, with Grok 4.3 also on the frontier ... and Gemini 3.1 hanging on, possibly subsidized out of pride. This is a consequence of Google losing their per-cost-token leadership as a result of very conservative TPU v8 design decisions to try to take share back from Broadcom and Nvidia, while others kept making aggressive choices.
Gavin Baker Bearish
Google made very conservative TPU v8 design choices while Nvidia and Trainium went aggressive.
Atreides Management · May 12, 2026 · Full history
Quote and source
Google, for a lot of reasons, made very conservative design choices with TPU v8. Nvidia and Trainium made very aggressive design choices ... I would never bet against Google, I would never bet against Broadcom. But I do think Trainium is super underestimated right now.
Gavin Baker Bullish
Google's TPU cost edge let it 'suck the economic oxygen' from the AI ecosystem, but Baker called it temporary.
Atreides Management · Dec 9, 2025 · Full history
Quote and source
suck the economic oxygen out of the AI ecosystem
Gavin Baker Bullish
Nvidia's biggest competitor is Google, whose Gemini traffic may already top OpenAI's, running on TPUs.
Atreides Management · Oct 30, 2025 · Full history
Quote and source
Nvidia's biggest competitor, it's not AMD, it's not Broadcom, it's certainly not Marvell, it's not Intel, it's Google ... I suspect on an actual traffic basis, Google is bigger than OpenAI, Anthropic, anyone today, and that business is going to run on TPUs.
Gavin Baker Bullish
Google's 5-billion-user Chrome moat makes it hard to bet against in the AI-browser fight.
Atreides Management · Oct 30, 2025 · Full history
Quote and source
there's something called Chrome that has 5 billion users, and if you're Google, you can just go look at what happened with Google Buzz, they are very cautious ... I do think it's tough to bet against the companies with large existing user bases today.
Gavin Baker Bullish
Google and xAI have the most unique proprietary data among AI labs, a durable post-reasoning edge.
Atreides Management · Oct 22, 2025 · Full history
Quote and source
to varying degrees, they all have with probably Google and XAI having the most unique data.
Gavin Baker Bullish
Google was able to attract the AI talent needed to make its scale a sustaining, not just disruptive, edge.
Atreides Management · Oct 22, 2025 · Full history
Quote and source
the core ingredients to success are compute and data. And these companies have the most compute and the most data ... Google was able to do that.