Baidu and Xiaomi are both at a critical juncture of shifting gears.

Baidu and Xiaomi have just released their Q2 2026 financial reports.

One saw its net profit drop by 68%, while the other posted a 42.6% decline in adjusted net profit.

If you only look at the income statement, neither company is having an easy time.

But the really interesting part is, after profits start to slide, what will they trade for their future?

The two companies have given completely different answers.

I. Don't rush to focus on profits first

The 68% profit drop of Baidu this time is quite staggering.

But when you break down the income statement, you will find that things are not that simple.

In the same period last year, Baidu's investment gains and losses reached 4.86 billion yuan; this year the figure is only 180 million yuan.

The missing 4.6 billion yuan basically explains the steepest downward curve on the income statement.

This part mainly comes from the fair value change of investment assets and exchange rate impact. The money has not vanished out of thin air from the company's account, but is more of a pullback of the book gains accumulated in the past.

Therefore, if you only judge Baidu's operating performance by the 68% drop in net profit, it is easy to mix up two separate accounts.

The actual money earned from core business operations is not in such a bad shape.

Baidu's operating profit in this quarter is about 3 billion yuan, compared with 3.3 billion yuan in the same period last year; its operating margin remains at around 10%.

Of course, Baidu's operation is indeed under pressure.

Online marketing revenue reached 13.1 billion yuan, down 19% year on year; overall revenue dropped 4% year on year, and its traditional business is still in a shrinking phase.

So Baidu's problem is neither a falsified income statement, nor a trouble-free operation.

A more accurate description is: the old business is gradually shrinking, while the new business has not yet fully taken over.

This is the real issue Baidu needs to address.

Xiaomi is a completely different case.

Its profit decline mostly comes from its core operations itself.

In the second quarter, its revenue hit 108.9 billion yuan, down 6.1% year on year; adjusted net profit stood at 6.22 billion yuan, down 42.6% year on year.

Its smartphone business generated 42.1 billion yuan in revenue, with a gross margin of only 8.5%.

The most troublesome variable in this cycle is storage.

In Q1 2026, the quarterly contract price increase of DRAM reached 93%-98%, and the growth rate of NAND Flash hit 55%-60%. The proportion of storage cost in the smartphone BOM has also risen rapidly from about 10%-15% in the past to more than 30%.

Xiaomi has suffered a particularly obvious impact.

The reason is not complicated. Entry-level and mid-range products account for a relatively high proportion of Xiaomi's smartphones, and these products have limited profit margins in the first place, making it very difficult to fully absorb the sudden extra cost.

But there is another interesting figure here.

The ASP (Average Selling Price) of Xiaomi's smartphones is still rising.

In Q4 2025, Q1 2026 and Q2 2026, the ASP of Xiaomi's smartphones was about 1176 yuan, 1310 yuan and 1351 yuan respectively.

The price is getting higher, but the gross profit has not gone up accordingly.

This shows that Xiaomi is going through a very typical hardware dilemma: raising prices can hold revenue, but may not necessarily preserve profits.

Profits of both companies are declining.

But from this point onward, they are facing two completely different types of pressure.

Baidu needs to find a new profit engine, while Xiaomi needs to build a new profit structure.

II. Baidu is replacing its profit engine

The most noteworthy figure of Baidu right now is its AI revenue.

In the second quarter, Baidu's AI business revenue increased by about 25% year on year, accounting for nearly half of its general business revenue.

This figure is quite impressive.

But when you break it down, you will find it represents both hope and pressure.

The rising proportion of AI revenue has two reasons.

On one hand, the AI business itself is growing; on the other hand, the traditional business is shrinking.

Revenue from traditional business fell 23% year on year, and overall revenue dropped 4%.

This means that a higher proportion of AI business does not equal that Baidu has completed its transformation. It is more like the old engine is decelerating, and the new engine has just started to run.

The problem is that the new engine has not yet reached the profit efficiency level of the old one.

The most profitable business Baidu used to have was search advertising.

Traffic comes in, ads are sold, and the marginal cost is very low.

Now Baidu is increasingly stepping into AI cloud, computing power infrastructure, and large model services.

This is a completely different type of business.

Servers need to be purchased, computing power needs to be built, data centers need to be operated, and model training costs a lot of money.

This is also why Baidu's gross margin dropped from 43.9% to 39% in the second quarter.

The operating margin can still maintain 10% thanks to cost control; the decline in gross margin more directly reflects the ongoing change in its business structure.

What Baidu is facing now is actually a very difficult gear shift process.

The old business is still generating cash, but its growth space is getting more and more limited; the new business has huge room for development, but requires massive capital investment first.

Li Yanhong is fully aware of the problem.

After the financial report release, Baidu still emphasized that it will push the Ernie Large Model to reach the top tier of foundational models as soon as possible.

This is surely important.

But a bigger problem also arises -- after the model enters the top tier, what will Baidu rely on to make profits?

This is the real question that needs to be answered in the next stage.

Because model capabilities eventually need to be embedded in products. But for Baidu today, none of its Agent products are competent enough to take on the core task, and most of them are just old wine in new bottles, which is a hard flaw that cannot be fixed in the short term.

III. Xiaomi is bearing cost pressure

If Baidu is replacing its engine, what Xiaomi is doing now is more like reinforcing its old engine.

Smartphone remains Xiaomi's largest business.

But this business is getting more and more difficult to run.

The global smartphone market is already highly mature, and the gap between leading brands is getting smaller and smaller.

The most lethal weapon Xiaomi used to have was efficiency: offering higher configurations at lower prices.

The problem is that when costs of storage, chips, screens, manufacturing and other links rise together, the cost-performance model will hit a very realistic ceiling.

You can raise prices. But after the price hike, will users still be willing to buy your products?

You can also keep prices unchanged. Then you have to bear the extra cost all by yourself.

In the second quarter, the ASP of Xiaomi's smartphones has reached 1351 yuan, but the gross margin is still only 8.5%.

This set of figures speaks volumes.

Xiaomi is working hard to push its smartphones to higher price ranges, but the process of brand upgrading takes time.

As a result, large home appliances have started to take on more responsibilities.

In the second quarter, Xiaomi's IoT and lifestyle consumer products generated 31.3 billion yuan in revenue, with a gross margin of 20.1%, more than twice that of the smartphone business.

This business is becoming increasingly important.

Air conditioners, refrigerators, TVs, washing machines, and a large number of smart home products give Xiaomi the opportunity to extend its previously built user base, channels and brand recognition into the home space.

This is also the biggest value of Xiaomi's ecosystem.

Buying a smartphone is an entry point for users.

After the smartphone enters the home, TVs, air conditioners, refrigerators, robot vacuums, door locks, smart speakers... all may continue to generate connections with the user.

But problems also exist.

The large home appliance sector is a very mature industry. The real estate cycle, consumer confidence, and the phase-out of national subsidies are all affecting the market.

If Xiaomi wants to expand its scale in this field, it also has to face competitors such as Haier, Midea and Gree that have been operating in this sector for decades.

Today's large home appliance business can help Xiaomi ease the pressure from its smartphone business, but it is not yet sufficient to independently support the next round of growth.

The real variable remains automobiles.

IV. The problems of the two companies are exactly the opposite

Xiaomi's automobiles seem to be more perceptible to the public than Baidu's AI. Consumers can see SU7 and YU7 running on the road every day.

In the second quarter of 2026, Xiaomi delivered 104,200 automobiles, up 28.2% year on year.

This performance is not bad.

But another set of figures is also quite notable -- the automobile and AI-related innovative businesses posted an operating loss of 2.6 billion yuan in the second quarter.

The loss in the first quarter was 3.1 billion yuan.

In other words, Xiaomi's automobile business has achieved considerable scale, but its business model has not been fully verified yet.

This actually conforms to the law of the automobile industry.

Automobile manufacturing requires huge R&D investment, as well as factories, supply chains, sales networks and after-sales systems. Only after the scale grows, can the fixed cost be amortized.

The problem Xiaomi is facing now is not whether the automobile business has opportunities.

The opportunities surely exist. The real question is, how much money does Xiaomi need to burn before the automobile business can start making real profits?

This is also why the extended-range electric vehicle model is worth paying attention to.

Xiaomi's previous two core products SU7 and YU7 have users concentrated in first-tier and second-tier cities, where the acceptance of pure electric vehicles is relatively high.

Entering the extended-range electric vehicle track now means Xiaomi is actively expanding its user boundary.

The 250,000-300,000 yuan extended-range SUV market is indeed large enough.

But Li Auto and AITO have already laid out in this segment.

What Xiaomi needs to prove next is whether its product capabilities can expand from the first batch of tech-savvy users to a broader group of family users.

This is a bit similar to the problem Baidu is facing.

Baidu needs to turn AI from model capabilities into products that users use every day.

Xiaomi needs to turn automobiles from tech products into a large-scale profitable business.

Both companies are waiting for a result: when the money invested today can be turned into profits tomorrow.

V. After profits decline, the gap begins to emerge

Looking back at these two financial reports, you will find a very interesting difference.

Baidu's danger lies in the fact that its old business is getting less and less valuable.

Xiaomi's danger lies in the fact that its old business is getting harder and harder to make profits.

But both companies have cards in their hands.

Baidu's cards are AI cloud, models, agents, Kunlun chips, as well as 283.1 billion yuan in cash and investments.

Kunlun Chip is particularly notable.

At present, many institutions in the market have carried out independent valuation for it. JPMorgan Chase has given an independent valuation range of 40-49 billion U.S. dollars, of which the part belonging to Baidu is about 27-34 billion U.S. dollars.

These figures cannot be regarded as the actual value yet.

After all, Kunlun Chip has not gone public, and its valuation will fluctuate along with market changes.

But this phenomenon itself is very noteworthy: the market has started to break down Baidu's assets for separate evaluation.

In the past, Baidu was treated as a single company. Search, cloud, autonomous driving, chips, AI were all included in one single valuation. Now, the market is increasingly willing to price these assets separately.

This means that Baidu's biggest opportunity in the future may not only lie in restoring profit growth, but also in gradually making the AI assets previously hidden inside the group more visible to the public.

Xiaomi's cards are simpler.

The smartphone business still has a huge user base, IoT has scale advantages, and the automobile business is seeing rapid sales growth.

Its problem has never been a lack of growth stories, but that there are too many stories, and each of them requires capital investment.

The smartphone business needs to preserve profits, the large home appliance business needs to expand scale, the automobile business requires continuous investment, and the AI business needs ongoing R&D.

The most important capability Xiaomi needs now is not to find another new growth story.

The core is to make these stories mutually supportive and form a positive cycle.

VI. The next stage is the real test

This is probably the most interesting part when we put the two financial reports together.

Both companies are experiencing profit decline.

But what the capital market really cares about is never just that they earned several billion yuan less this year.

It cares more about where the lost money has gone.

If after the money is burned, what you get is a worse business, then it is a real problem.

If the money is invested in new businesses and eventually forms a new growth curve, then it is a valuable investment.

Baidu is going through this process now.

Search advertising continues to decline, and the AI business has gradually become an important part of its revenue. While Fitch downgraded Baidu's rating, it also predicted that Baidu's AI business revenue will surpass its traditional business in 2026.

This statement is actually more noteworthy than the rating downgrade itself.

An institution that just downgraded your rating also judges that your new business may overtake the old business next year.

This shows that the market has not denied Baidu's future.

It just does not believe that Baidu has turned its future into today's profits yet.

The same goes for Xiaomi.

Smartphone profits are compressed by storage costs, and the automobile business continues to lose money.

But automobile sales are still growing, product lines are still expanding, and large home appliances are becoming a more important business.

What Xiaomi lacks most right now is not growth stories, but time.

Time will tell the market whether the automobile business is a core business that can make sustainable profits, or an expensive experiment that consumes cash for a long time.

Both companies are betting on the next stage, but the capital market will not only focus on stories in the end. It will check when the new business can become self-sustaining and generate profits on its own.

What will ultimately determine their fate is whether the new business can form a new profit source before the old profits disappear.

This is an increasingly common status for Chinese tech companies today:

Growth is still there, but profits have run off first. The real competition next is who can survive this most difficult period.

Words beyond the page:

Now the financial reports of more and more established large companies are starting to look less impressive.

But the reasons for the unsatisfactory results have begun to vary.

Some companies' old businesses are really failing; some have grown new businesses that have not yet had time to make profits; others actively accept a less impressive income statement today in exchange for growth in the next stage.

Therefore, when evaluating a company today, you cannot only ask how much profit it has made.