Deepseek 2.0 moment did not arrive: chip stocks stabilize, US stock market awaits earnings season
- Core Viewpoint: Kimi K3 did not trigger a sell-off in chip stocks like DeepSeek 1.0, as its narrative shifted from "computing power is sufficient" to "computing power is scarce." Google plans to launch an AI chip with model architecture embedded, the Frozen v2, further strengthening hardware demand. However, with the earnings season approaching, whether capital expenditure growth can continue to exceed expectations remains a key test for the market.
- Key Elements:
- The K3 model, with its 2.8 trillion parameters and extremely low inference cost, became a sensation, but it also put pressure on the inference infrastructure of Moonshot AI, triggering demand for computing power expansion and shifting the narrative to "efficiency unlocks demand."
- The "low cost, high computing efficiency" narrative of DeepSeek 1.0 had previously impacted the chip investment thesis; however, K3 proved that model efficiency improvements actually stimulated the expansion of chip demand.
- Google is developing an AI server chip codenamed Frozen v2, which will embed part of the Gemini model's architecture directly into the silicon. Its energy efficiency ratio is expected to be 6 to 10 times that of the seventh-generation TPU Ironwood.
- Frozen v2 is scheduled for deployment before 2028, sending a signal: AI giants are shifting their competition towards customized, specialized chips, creating more segmented semiconductor demand, which is a long-term positive for the industry chain.
- The upcoming week's earnings season is a critical test. The market will focus on whether the capital expenditures of Google, Meta, and Microsoft can be sustained, and whether SK Hynix can prove the profitability of memory chips.
- The market's threshold for exceeding expectations on capital expenditure is already high. Any marginal slowdown (such as downward guidance revisions) could trigger a new wave of sell-offs, placing the chip sector in a window of high uncertainty.
On July 21, after the U.S. stock market closed, the chip sector, which had been shaken by the sudden emergence of Kimi K3, finally caught a breather.
The "DeepSeek 2.0 moment" that the market had feared did not materialize further. Although SK Hynix still closed down 1.86%, previously high-flying stocks like SanDisk and Micron Technology turned upward, showing signs of stabilization and recovery. After more than a week of intense sell-offs, the entire semiconductor sector seemed to find a temporary equilibrium.
Why, this time, did Kimi K3 not replicate the panic script of DeepSeek 1.0?
1. K3’s Narrative Has Shifted: Not "Enough Computing Power," but "Computing Power Scarcity"
The reason DeepSeek 1.0 caused such a huge stir in March-April last year was its core narrative: "Low cost, high efficiency, no computing power shortage." It trained a model approaching top-tier levels using very few GPU resources, directly challenging the investment logic that "AI performance must rely on massive chip deployment."
But the story of Kimi K3 is entirely different.
While K3 astounded the industry with its 2.8 trillion parameters and extremely low single-inference cost, another equally notable consequence emerged post-launch: computing power scarcity. The overwhelming popularity of K3 exceeded expectations, putting immense pressure on Moonshot AI's inference infrastructure, making it an urgent priority to expand computing capacity.
What does this mean? It means K3 isn't proving that "we don't need that many chips"; rather, it's proving that "no matter how efficient models become, the growth rate of demand will always outpace supply." This narrative of "computing power scarcity" is precisely the strongest support for the chip sector—it reassures investors that demand for AI chips will not vanish due to improved model efficiency, but instead could expand continuously due to the explosion of application scenarios.
DeepSeek 1.0's logic was "efficiency replaces scale"; K3's logic is "efficiency unleashes demand." The impact of these two narratives on chip stocks is diametrically opposed. This is the fundamental reason why a "DeepSeek 2.0" scenario did not unfold.
2. Google's Trump Card: "Baking" Gemini into the Chip
At a critical moment when the chip sector was searching for direction, Google dropped a bombshell that could potentially change the rules of the game.
Alphabet is developing a new AI server chip, codenamed Frozen v2. The design approach for this chip is extremely aggressive: it directly etches part of the Gemini model's architecture onto the silicon itself.
This is not the traditional "optimizing software to fit hardware"; it's "baking the model blueprint into the chip." By reducing data movement between compute units and memory, it drastically cuts the power consumption and latency of each inference.
Google engineers estimate that the energy efficiency of Frozen v2 will be staggering: the number of tokens processed per unit of power could be 6 to 10 times higher than that of its current state-of-the-art Ironwood TPU. For context, Ironwood is already Google's seventh-generation TPU, and it achieved only a doubling of performance per watt compared to its predecessor. The generational leap of Frozen v2 far surpasses any previous chip upgrade by Google.
Following the news, Alphabet's stock price briefly rose from around $350 to nearly $359 during trading. The market is clearly reassessing Google's long-term competitiveness in the AI infrastructure field. If Frozen v2 can achieve commercial deployment before 2028, Google would possess one of the most efficient large model inference infrastructures globally, making it difficult for competitors to replicate its cost advantages in AI services.
But for the entire chip sector, the significance of Frozen v2 extends far beyond "Google's own business."
It sends a critical signal: AI giants are not slowing down their hardware investments due to improved model efficiency. On the contrary, they are pushing the competition to a more fundamental, more customized level—application-specific integrated circuits (ASICs). From NVIDIA's general-purpose GPUs to Google's specialized TPUs, and now to Frozen v2 with its model architecture etched directly into silicon, the AI hardware competition is shifting from "who has more cards" to "whose cards are smarter, more efficient, and more specialized."
This shift means that demand for AI chips will not shrink just because "models are getting smaller and cheaper." Quite the opposite: the trend toward specialization and customization will create more diverse and segmented chip demands—which is a long-term positive for the entire semiconductor supply chain.
3. The Earnings Season Test: The Coming Week Will Determine the Direction
Although the chip sector has stabilized for now, the test is far from over.
Over the next week or so, key players in the AI value chain will successively release their quarterly results. This earnings season is exceptionally significant—the market isn't just looking at the numbers; it's also watching for answers to several critical questions that will determine whether the AI narrative can continue:
For Google, Meta, and Microsoft: Will capital expenditures keep burning? Can the revenue generated from cloud businesses and AI services cover the increasingly heavy costs of depreciation, leasing, and electricity? If not, cracks will appear in the "AI monetization" narrative, and the inflection point for capital expenditure growth might arrive sooner than anticipated.
For SK Hynix: Can the massive spending by cloud providers ultimately translate into pricing power for memory chips, market share expansion, and profit growth? Hynix needs to prove in its earnings report that it is a part of the AI value chain that genuinely makes money, not just an "intermediary skimming a margin."
The real focus of this earnings season isn't simply whether capital expenditures are "high," but whether they can "continue to exceed expectations." After consecutive quarters of "surprises," the market's threshold for expecting overperformance has been raised significantly. Any marginal slowdown—whether a downward revision in guidance, softening language, or key data falling short of estimates—could act as a trigger for a new wave of sell-offs.
4. In Conclusion: In the Eye of the Earnings Season Storm, Options are a Crucial Anchor
The chip sector is navigating a window of high uncertainty. The bulls argue that K3's computing power scarcity and Google's Frozen v2 prove that AI hardware demand is far from peaking. The bears contend that capital expenditure growth is about to peak, valuations are already stretched, and earnings will struggle to keep beating expectations.
In this environment of information divergence, betting on a single direction carries risks far exceeding potential rewards.
The BIT options functionality provides a perfectly suited tool for this kind of market environment characterized by "unclear direction and high volatility":
- Holding chip stocks + buying put options: Insures your position before earnings, capping downside risk
- Buying single-direction call/put options: Bet on post-earnings direction with a fraction of the capital needed for the underlying stock, with maximum loss limited to the premium paid
- Buying options in both directions: Not sure if earnings will bring surprises or shocks? Bet on both sides simultaneously; profit as long as the volatility is big enough.
The winds of earnings season are approaching. In a market without clear direction, only those with options can afford to be composed.
Risk Disclosure: The market conditions, valuation estimates, and product descriptions mentioned in this article are for reference only and do not constitute investment advice. Trading in U.S. stocks and their derivatives involves market volatility, leverage, and liquidity risks. Short selling stocks may carry the risk of unlimited losses. Options trading carries the risk of total loss of premium. Past performance is not indicative of future returns. Investors should make prudent decisions based on their own risk tolerance and consult professional investment advisors when necessary.
