Baidu's 'Apparent Decline,' Xiaomi's 'Real Struggle'

Both companies have seen significant drops in profits, but the nature of their challenges differs. On August 18, Baidu and Xiaomi unveiled their Q2 2026 financial results simultaneously. Baidu's net profit plummeted by 68% year-on-year, while Xiaomi's adjusted net profit declined by 42.6% year-on-year.

At first glance, these figures suggest that the two tech giants are facing similar 'profit halving' scenarios. The market's reaction was equally harsh: Baidu's stock price dipped in after-hours trading, and Xiaomi encountered a wave of skepticism.

However, a closer examination of the two reports reveals distinct narratives.

Approximately 4.6 billion yuan of Baidu's profit reduction stemmed from the fair value adjustment of investment assets and currency exchange rate impacts. These are non-cash losses that do not reflect operational cash outflows and do not undermine the core business.

Excluding these non-operational items, Baidu's operating profit stood at around 3 billion yuan, with an operating profit margin hovering around 10%.

In contrast, Xiaomi's profit decline was primarily driven by business factors: escalating storage costs eroded smartphone gross margins; reduced IoT subsidies dragged down sales revenue; automotive shipments increased by 28.2% year-on-year, yet losses persisted; and AI and other innovative ventures have yet to yield substantial revenue.

Faced with these triple cost pressures, Xiaomi's profitability was inevitably squeezed, with every yuan of lost profit feeling like a significant setback.

In essence, Baidu's paper losses magnified real operational challenges, while Xiaomi faced genuine profit erosion.

Many believe that the toughest times for Baidu and Xiaomi are behind them, with a second growth phase on the horizon.

But are these turning points genuine? A closer look at the financial reports reveals a crucial distinction: Which company is 'faking a fall' to invest in the future, and which is enduring cyclical pain without having replaced its growth engine?

Xiaomi's Real Struggle: Half Cyclical, Half Strategic Decision

The most notable figure in Xiaomi's financial report is its smartphone ASP (Average Selling Price): it surged by 25.9% year-on-year to 1,351 yuan, a record high; models priced above 3,000 yuan accounted for 32.1%, up 4.5 percentage points year-on-year.

The cost? Smartphone shipments dropped from 42.4 million units to 31.2 million units, a 26.5% year-on-year decline; gross margin also fell from 11.5% to 8.5%. Beyond the financial data, Xiaomi must also navigate external 'sentiments' such as shareholder doubts, supply chain pressures from reduced orders, and market concerns about market share.

Yet, after more than four years of pushing upmarket, Xiaomi's strategic choices are essential. Deciding what to sell and what not to sell requires independent judgment, not reliance on market conditions.

As Lu Weibing stated on the earnings call, 'We will no longer pursue smartphone volume in the short term.' This indicates that Xiaomi, facing cost pressures like rising memory prices, chose to actively reduce low-end smartphone shipments to prioritize profit protection.

Another set of data often overlooked is that, in Q2, internet services gross profit reached about 6.9 billion yuan, surpassing IoT's approximately 6.3 billion yuan for the first time, becoming Xiaomi's highest gross profit contributor. Advertising revenue was 7.2 billion yuan, up 4.8% year-on-year, with a gross margin of 76.8%, up 1.4 percentage points year-on-year.

When smartphones and IoT face pressures, this high-margin business serves as Xiaomi's profit stabilizer. Its value has been underestimated by the market.

In the past, Xiaomi's profits relied heavily on hardware businesses like smartphones; now, internet services lead gross profit contributions. The ecological value of 'People-Vehicle-Home' has become more prominent when hardware revenue is under pressure.

Now, consider two major investments: automotive and AI. Automotive deliveries reached 104,000 units, up 28.2% year-on-year; losses narrowed to 2.6 billion yuan from 3.1 billion yuan in Q1. For automotive, narrowing losses is more significant than revenue growth.

This suggests that Xiaomi's automotive business is beginning to achieve economies of scale, with costs being diluted. Meanwhile, the Pengcheng extended-range SUV, priced at 260,000 to 300,000 yuan and targeting family users with low overlap with SU7 users, fills gaps in model structure and gross margin.

Sedans build the brand, SUVs drive volume—two legs make for stable walking.

Xiaomi's AI strategy is also very Xiaomi: not rushing to monetize independently but embedding it into devices to feed the ecosystem. AI-related revenue from the MiMo large model is included in 'Other Related Businesses,' which totaled 1 billion yuan this quarter, with API calls and token schemes starting to contribute revenue.

While the revenue scale isn't disclosed separately, the commercialization path is clear. The financial report deliberately mentions that the humanoid robot achieved over 90% test success in Xiaomi's own factories and will soon be publicly displayed at the World Robot Conference—Xiaomi is signaling to the market: it has more cards to play.

In terms of valuation, Xiaomi's current P/E ratio is between 15 and 17 times, with net cash of about 176 billion yuan, or about 10.5 Hong Kong dollars per share; it repurchased 11.7 billion Hong Kong dollars worth of shares this year and launched a new 20 billion Hong Kong dollar repurchase plan.

Yet the market still values Xiaomi as a hardware company. If automotive gross margins continue to improve, MiMo revenue is disclosed separately, and overseas IoT scales, this standard will need revision. The market offers a hardware price, while the company's repurchases signal confidence—the gap between these two prices is the expectation gap in valuation logic.

But the turning point isn't solely up to Xiaomi. As Letter List analysis notes, its realization depends on three variables: Can Pengcheng carve a niche in the cooling extended-range market? Can Xuanjie O3 deliver as expected? Can AI remain in the SOTA (State-of-the-Art) tier?

Baidu's 'Apparent Decline': Two Ledgers Clash, Turning Point Arrives

Baidu's financial report must be read in reverse.

Online marketing, i.e., advertising revenue, was 13.1 billion yuan, down 19% year-on-year. This traditional growth engine is indeed slowing. But AI cloud revenue was 7.3 billion yuan, up 50% year-on-year; GPU cloud grew triple-digit for four consecutive quarters, hitting 283% this quarter.

Fitch downgraded Baidu's rating but simultaneously expects AI business revenue to surpass traditional business by 2026. This shows that institutions recognize Baidu's new growth engine's speed but remain cautious due to unrealized profits.

This is a clash between old and new ledgers, leading to seemingly contradictory external judgments.

One intriguing detail: J.P. Morgan gave Kunlunxin an independent valuation of 40-49 billion USD, with Baidu's portion valued at 27-34 billion USD. The market is starting to price Baidu as 'multiple assets' rather than 'one company.' This is a substantive revaluation of its technological reserves.

Smart money is also moving. According to NetEase Technology, citing 13F filings, Druckenmiller established a new position in Baidu ADRs in Q2—his first major Chinese stock holding since clearing Alibaba in 2023; Tepper also increased his stake during the same period.

Baidu's turning point is already evident in its business. In Q2, Baidu Core's AI-driven business revenue reached 12.5 billion yuan, up 25% year-on-year, accounting for half of general business revenue; in Q1, this proportion was 52%. While the proportion dipped slightly, the 25% year-on-year growth is real—the new engine isn't slowing.

Baidu's growth lies in vertical industries and Robotaxi. AI cloud continues to expand flagship cases in finance, gaming, embodied AI, and other fields, with vertical solutions offering higher stickiness and pricing power. Apollo Go now operates in 28 cities, with cumulative test mileage of 350 million km, achieving fully driverless commercialization in Dubai, partnering with Uber, and starting public road tests in London, Switzerland, and elsewhere.

Once the overseas model replicates, valuation shifts from 'domestic mobility service provider' to 'global autonomous driving technology exporter.' NetEase Technology observes that market questions have shifted: from 'When will AI generate revenue?' to 'Can this capability be replicated?' From office agents to GPU cloud, from self-developed chips to Robotaxi, the same path is being repeatedly validated: technology becomes products, products become clients.

Of course, Baidu is also experiencing growing pains. For example, gross margin fell from 43.9% to 39%, indicating heavier operations. AI cloud and GPU cloud require heavy upfront compute investment before returns, so margin declines are normal during the investment phase. AI application layer revenue was 2.5 billion yuan, with growth slowing to 3%—C-side monetization still needs time.

But fortunately, Baidu's core operating foundation remains intact, with effective cost control and no runaway transformation. Its future trajectory is far more certain than a year ago.

In my view, Baidu's current transformation resembles an Odyssean journey home: Ithaca hasn't disappeared, but the voyage must pass the Sirens' seas. Advertising revenue is the enticing song, but the mast is tied, and the direction remains unchanged.

For every unit of old revenue lost, new revenue replaces it—the speed of replacement determines the degree of pain. AI revenue exceeding half, Kunlunxin's independent valuation, and overseas Robotaxi deployments—these cards are already on the table.

For Robin Li, there's no turning back. The sunk investments, established relationships, and forming revenue structure all propel Baidu forward.

Two Types of Struggle, Different Turning Points

Viewing both financial reports together, the turning points take on different meanings.

Xiaomi's turning point lies in treating external wounds. Storage costs, IoT subsidy reductions, and automotive losses—all three cuts are business-related. These external wounds can 'heal' on their own: storage prices are cyclical, subsidy reductions are one-time, and automotive losses are narrowing.

Lei Jun's task is to stabilize morale and continue pursuing high-end and full-stack AI ecosystem strategies.

But before that, Xiaomi still relies heavily on 'industry cycles' to rebalance supply and demand. Its underlying growth engine hasn't fully changed—what will its strategy be when the next industry cycle arrives?

Baidu's turning point is that its business growth engine has switched more than halfway. AI's independent revenue-generating ability is rapidly filling the hole left by old businesses, with cloud, API, and token schemes all needing to monetize independently, forcing it to find its profit logic faster.

What Robin Li must prove is whether AI's new businesses can generate profit and whether the planned commercial blueprint can be strongly executed.

Capital markets are mercenary, using profits to define companies. A drop means 'it's over,' while a rise means 'it's back.' But money doesn't lie. Xiaomi's money flows into high-end choices, overseas IoT, automotive scale, and AI ecosystems—waiting for cost cycles to reverse and new businesses to scale. Baidu's money flows into AI infrastructure, Kunlunxin, vertical industries, and Robotaxi—waiting for new revenue to fill the old revenue gap.

For these two companies at pivotal moments, neither path is easy: one waits for the cycle to reverse, the other waits for its engine to upgrade. That's where the real drama lies.