
Analysts Still Buy Google: Model Lag Doesn't Matter, Turnaround Hinges on TPU Chip Sales
Author: JR Research
Translated by: TechFlow
TechFlow Summary:Since summer, Alphabet's stock has weakened, with its share of frontier large language models falling behind OpenAI and Anthropic, while Meta's newly launched Muse agent adds competitive pressure. The author believes Google's true strategic focus has shifted to Google Cloud and TPUs: cost-effectiveness advantages, chip sales, and cloud infrastructure options could push revenue to approximately $730 billion by FY2028. Maintains a Buy rating, viewing the current valuation near historical averages as an entry window before a TPU-driven turnaround accelerates.
Key Takeaways
- Alphabet's summer pullback deepened, AI model rankings lag behind OpenAI and Anthropic, and Meta's newly released Muse agent further intensifies competition.
- GOOGL's strategic focus now lies on Google Cloud and in-house TPU hardware, which offers superior cost-performance and could open avenues for revenue diversification beyond advertising.
- TPU sales and infrastructure options are viewed as key growth engines, with Wall Street forecasting FY2028 revenue to reach approximately $730 billion.
- The author maintains a Buy rating, arguing that the current valuation is closer to historical averages, presenting a suitable entry point ahead of a potential TPU-driven turnaround acceleration.

Google's 'Summer Slump' Isn't Over Yet
Investors in Alphabet (GOOG) (GOOGL) have endured quite a scorching summer.
Since peaking in May, the stock has barely regained momentum.
For the past three to four months, Google shareholders have been digesting whether the company can still secure a front-row seat after gradually loosening its grip on the frontier AI race amid drastic environmental changes.
OpenAI and Anthropic have already widened their lead, even more so given the increasing popularity of OpenAI's coding models—especially since the release of GPT 5.6 and its upgrade to GPT 6 Astra.
Google's AI Models Remain Distant Third

However, Anthropic remains firmly at the forefront of transforming agentic AI workflows.
This short-term lead seems unlikely to be overturned by Google, which still sits as a distant third among leading US AI model developers.
Google's predicament is further compounded by Meta (META), which just launched the personal AI agent Muse this week.
The author views this as the first step taken by Meta CEO Mark Zuckerberg toward his grand ambition of 'personal superintelligence.'
It is too early to conclude where the Muse project will ultimately go.
Nevertheless, Wall Street increasingly believes it will drive up engagement with Meta's AI models and unlock more robust monetization opportunities through freemium tiers.
This week, it rapidly climbed to the third spot in the Apple App Store's free apps chart.
In the previous update on Google, the author noted that the market is highly concerned about weakening free cash flow margins.
But management made it clear: Google Cloud is now the north star driving the inflection point of growth—not only serving Google's own AI models, but also opening massive growth options for deploying TPU-based infrastructure.
Google's TPUs Have Proven Their Cost-Effectiveness

This has been validated: Google TPUs are regaining attention for their high cost-performance ratio. According to SemiAnalysis's InferenceX benchmark, Ironwood (TPUv7) can deliver performance gains of up to ~50% per dollar spent.
This naturally elevates market expectations for inference-focused TPUv8 and training-focused TPUv8, unlocking specialized use-case opportunities across both training and inference workloads.
More notably, Google is no longer just challenging Nvidia on hardware cost-performance but is also attempting to undermine the software stack advantage of CUDA.
If the current pace of AI capital expenditure continues unchecked, the world is beginning to face the reality that major tech companies could fall into negative free cash flow margins.
Cooling Token Spend Could Benefit TPU Sales

Ramp's recent update corroborates this: among the top 1% of companies with the highest AI spend, per-employee expenditure dropped from ~$8,000 in July to just over $7,200 in August.
Therefore, Google's message is clear—there is more than one way to access Google's AI infrastructure.
Google Cloud CEO Thomas Kurian highlighted these pathways at a recent conference:
- Access via subscription to cloud services;
- Direct purchase of hardware and complete systems;
- Acquisition through the 'Neocloud' channel advanced in partnership with Blackstone.
TPUs Could Become Google's Largest Revenue Growth Opportunity

Options tied to chip sales will become the key catalyst pushing Google out of its advertising-dependent model and toward revenue diversification.
Based on updated Wall Street consensus estimates, FY2028 revenue is expected to reach approximately $730 billion.
Against the backdrop of expanding AI compute supply, concerns over whether token pricing can hold may prompt investors to more carefully scrutinize Google's TPUs—particularly their reported cost advantages.
Amid this round of frenzied AI capex accumulation, free cash flow margins are being rapidly compressed, making the TPU narrative even more critical.
Should You Buy or Sell GOOGL?

The market remains relatively lukewarm on returning to Google stock: after surging to ~$400 in May, the shares have pulled back significantly.
This hesitation stems largely from the market waiting for Google to prove whether its AI coding capabilities can genuinely improve following the release of Gemini 3.8 Flash.
This may explain why Google's stock struggles to find momentum, yet has steadily approached levels near its June low, currently sitting just below $315.
The author anticipates bargain-hunting capital will enter more aggressively around these levels: Google's stock has also pulled back significantly to align closer with its long-term five-year average forward P/E of ~22.6x, currently trading just under 25x.
GOOGL isn't exactly a discount, but it is far less expensive than it was during the May/June highs—when the forward P/E briefly stretched well beyond 30x.
The author hasn't seen a flood of bottom-fishing capital return yet, believing investors prefer to first observe how TPU business growth opportunities materialize.
As more cautious investors give newly refreshed AI models time to regain traction, the present moment should be viewed as a timely buying window ahead of an accelerated turnaround.
RATING: MAINTAIN BUY.
Disclosure
Analyst Disclosure: The author/team holds beneficial long positions in GOOGL and META via stocks, options, or other derivatives. This article was written solely by the author expressing their own opinions. Aside from Seeking Alpha, the author received no compensation for this publication and has no business relationship with any companies mentioned herein.
Seeking Alpha Disclosure: Past performance does not guarantee future results. This article does not constitute investment advice for any specific investor. The views expressed above do not necessarily reflect Seeking Alpha's overall stance. Seeking Alpha is not a licensed broker-dealer, investment advisor, or investment bank; its contributors are third-party authors comprising both institutional and individual investors who may not be licensed or certified by any regulatory authority.













