JR Research, in a piece translated by TechFlowPost, said Alphabet still merits a buy rating even as Google remains behind OpenAI and Anthropic in the race for frontier AI models. The argument rests less on model leadership and more on Google Cloud and TPU commercialization, which the author sees as a path to broader revenue growth beyond advertising.
Alphabet shares have struggled since May as investors reassess Google’s AI standing
The article said Alphabet investors have gone through a difficult summer. Since peaking in May, the stock has struggled to regain momentum over the last three to four months as shareholders weigh whether Google can still hold a leading seat in AI after the competitive landscape shifted sharply.
According to the author, OpenAI and Anthropic have widened the gap. OpenAI’s coding models have gained traction, especially after the release of GPT 5.6 and the upgrade to GPT 6 Astra. Google’s AI models, the piece said, remain a distant third.
Anthropic, meanwhile, was described as staying at the front of the shift toward agentic AI workflows, with no clear sign that Google can reverse that position in the near term. Pressure increased again this week after Meta launched its personal AI agent, Muse. The author called it the first step in CEO Mark Zuckerberg’s broader push toward "personal superintelligence." Where Muse ultimately leads is still unclear, but Wall Street is increasingly treating it as a product that could raise engagement with Meta’s AI models and support monetization through a freemium structure. The article said the app had already climbed to No. 3 among free apps in Apple’s iOS store this week.
The investment case centers on Google Cloud and TPU deployment
The piece said investors have also been worried about weakening free cash flow margins. Management, however, has made its position clear: Google Cloud is now the company’s main growth pivot. In the author’s view, that business does more than support Google’s own AI models. It also creates a large commercial opening for TPU-based infrastructure deployment.
That shift is central to the thesis. The author does not dismiss Google’s weakness in frontier models, but the article places more weight on whether the company can turn its cloud footprint and custom silicon into a monetizable business line than on whether it can immediately reclaim model leadership.
TPU value proposition has gained attention on cost performance
The article said Google’s TPUs have already demonstrated a cost-performance edge. Based on SemiAnalysis’ InferenceX benchmark, Ironwood, identified as TPUv7, can deliver as much as a 50% improvement in performance per dollar.
That has lifted expectations for future generations. The author said investors are now looking more closely at inference-focused TPUv8 and training-focused TPUv8, with potential room for specialized use cases on both sides of the workload mix.
The challenge, in the author’s framing, is not limited to Nvidia’s hardware position. Google is also trying to erode the advantage built by CUDA’s software stack. If AI capital spending keeps expanding at the current pace, the market may have to confront the possibility that large technology companies end up with deeply compressed free cash flow margins, or even negative ones. In that setting, lower-cost infrastructure alternatives would carry more weight.
Cooling AI spend could strengthen Google’s case with enterprise buyers
The article also cited a recent update from Ramp. Among the top 1% of companies with the highest AI spending, per-employee spending fell from about $8,000 in July to just above $7,200 in August.
For the author, that matters because it suggests enterprise buyers are becoming more sensitive to AI costs. Google’s message, the piece said, is that access to its AI infrastructure comes through more than one route. Google Cloud CEO Thomas Kurian recently outlined three of them:
- access through cloud service subscriptions;
- direct purchases of hardware and full systems;
- and a "neocloud" channel developed in partnership with Blackstone.
That setup suggests Google is not treating TPU only as internal infrastructure for its own training and inference workloads. It is also trying to build a business around external sales and distribution.
Wall Street sees FY2028 revenue near $730 billion
Chip sales and infrastructure distribution are presented in the article as a key way for Alphabet to diversify beyond advertising. Based on updated Wall Street consensus estimates, the author said Alphabet could generate roughly $730 billion in revenue in FY2028.
The piece added that as AI compute supply expands, investors are also questioning whether token pricing can hold. That could force a closer look at Google’s TPU economics, especially given previously reported cost advantages. In the author’s view, the more aggressively the industry keeps piling into AI capex and the faster free cash flow margins compress, the more important the TPU story becomes.
Valuation has moved closer to historical norms, and the buy rating stays in place
The article said investors remain hesitant to rotate back into Google. After reaching about $400 in May, the stock has pulled back sharply. The hesitation, according to the author, reflects a wait-and-see stance on whether Google’s AI coding capabilities will improve materially after the release of Gemini 3.8 Flash.
The stock is now moving closer to its June low, at just under $315, the piece said. The author expects dip buyers to become more active around that level. On valuation, Alphabet has retraced toward a level closer to its five-year average forward price-to-earnings multiple of about 22.6x, with the current figure just below 25x. That does not make the stock cheap, the author said, but it is no longer as expensive as it was in May and June, when the forward multiple climbed well above 30x.
The author added that there is still no sign of a strong wave of bottom-fishing capital rushing back in. Investors appear more inclined to wait and see whether TPU growth opportunities materialize. Even so, JR Research views the current setup as a timely entry window ahead of any faster TPU-driven turnaround and maintained a buy rating on GOOGL.
Disclosure
The author disclosed that they or their team hold beneficial long positions in GOOGL and META through stock, options, or other derivatives. The piece was written by the author and reflects personal views. Other than from Seeking Alpha, the author said they did not receive compensation for the article and have no business relationship with the companies mentioned.
Seeking Alpha said past performance is not indicative of future results and that the article does not constitute investment advice for any specific investor. It also said the views expressed do not necessarily represent the platform’s overall position. Seeking Alpha is not a licensed securities dealer, broker, U.S. investment adviser, or investment bank, and its analysts are third-party contributors that include both professional and individual investors, who may or may not hold licenses or certifications from institutions or regulators.
