SerpApi.Org
Alphabet (GOOG) (GOOGL)-owned Google is seemingly working on developing a chip for its own Gemini AI models, according to The Information. Codenamed “Frozen v2,” insiders believe that the chip could be about six to ten times more efficient than Google's tensor processing units, or TPUs, which it develops in collaboration with Broadcom (AVGO).
Expected to go live by 2028, the purpose of the chip would certainly be for Gemini to climb up the ranks in the battle of models, where none of its models are near the top 10 on the leaderboard.
A New Chip to Supercharge Gemini
The reports about Frozen v2 were timely, considering the delay of Gemini 3.5 Pro. And as one of the only companies (along with Amazon (AMZN)) to have full-stack capabilities and equally compelling consumer and enterprise business lines, it will be difficult to ignore Google in the AI war.
But how can Frozen v2 possibly help in this regard?
The chip would reportedly embed portions of the Gemini architecture directly into silicon. It would not permanently store the model weights, which was the less flexible idea behind an earlier Frozen design. Updated weights could still be loaded, while common calculations and data movement would be handled more efficiently.
That distinction matters.
A TPU must support numerous models and workloads, while Frozen v2 can sacrifice flexibility for efficiency on Gemini inference. At six times the efficiency, the power required for a given volume of tokens could theoretically fall by about 83%. At ten times, the reduction could reach 90%. Actual savings would be lower because memory, networking, cooling, buildings, and conventional processors would still carry costs.
However, the chip could nevertheless improve the user experience by making additional reasoning, longer context, faster responses, and heavier agent activity economically practical. Lower inference costs would also let Google serve more free users, offer more competitive Cloud pricing, and collect more interaction data for model improvement. This matters because Alphabet said in June 2026 that its model interfaces were processing about 19 billion tokens per minute, while total token usage across its products had reached 3.2 quadrillion per month. At that scale, even modest unit cost reductions become meaningful. Frozen v2 could also ease the capacity shortage that reportedly caused Google Cloud to turn away some business.
Alphabet: No Worries With the Financials
Alarms rose around Alphabet's high capex, debt raise, and equity financing. Yet, a look at the numbers should dispel those fears.
Alphabet reported solid financial performance in Q1 2026 with revenue advancing 22% year-over-year (YoY) to $109.9 billion. The cloud business, which carries higher margins, delivered particularly impressive results with 63% growth to $20 billion. Backlog for this segment more than doubled from the prior quarter to surpass $460 billion, with the bulk expected to convert to revenue over the next two years. The services segment, encompassing Google Search, increased 16% YoY to $89.6 billion.
Profitability also strengthened as earnings per share rose 82% to $5.11, comfortably exceeding the consensus estimate of $2.68. Over the past eight quarters, Alphabet has achieved earnings growth exceeding 30% YoY in seven of them, an impressive track record for a company of its size.
The firm raised its quarterly dividend to $0.22 per share, resulting in a current yield of 0.25%. Not a possibility, considering the company's focus on AI capex, yet the payout ratio of just 6.41% means that the scope for growth in dividends remains.
Cash generation remained healthy, with net cash from operating activities reaching $45.8 billion compared with $36.2 billion in the year-ago quarter. Free cash flow for the period stood at $10.2 billion, down sequentially from $24.5 billion in the fourth quarter of 2025. This metric holds particular importance given the company's substantial planned capital expenditures of 180 billion dollars to 190 billion dollars for the year. Alphabet closed the quarter with $38.1 billion in cash and no short-term debt on its balance sheet.
Despite these results and a somewhat subdued stock performance year-to-date (+12% YTD), Alphabet shares appear richly valued. The forward P/E ratio of 24.33 times, P/S multiple of 8.65 times, and P/CF multiple of 20.88 times all exceed the respective sector medians of 13.13 times, 1.19 times, and 7.75 times.
Alphabet is set to report its Q2 2026 earnings after the market closes on July 22. Analysts are expecting the company to report revenues of $116.9 billion and earnings of $2.90 per share, which would represent growth of 21.2% and 25.5% YoY, respectively.
Analyst Opinion of GOOGL Stock
Considering this, analysts have deemed GOOGL stock to be a consensus “Strong Buy,” with a mean target price of $434.37. This indicates an upside potential of about 23% from current levels. Out of 54 analysts covering the stock, 45 have a “Strong Buy” rating, four have a “Moderate Buy” rating, and five have a “Hold” rating.
On the date of publication, Pathikrit Bose did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.
More news from Barchart
・These 2 Analysts Just Upped Their AMD Stock Price Targets. Here’s Why.
・Intel Stock Just Got a Stunning Price Target Hike Ahead of Earnings. How to Play INTC From Here.
・How to Play RDW Stock as Redwire Boosts Manufacturing Capacity
