Baidu (BIDU) Stock Today: AI Chips, Layoffs and a JPMorgan Upgrade – What December 2025 Really Changes for Investors

Published: December 6, 2025

Baidu, Inc. (NASDAQ: BIDU) has had a whiplash few weeks. The Chinese search and AI giant has reported its steepest revenue drop on record, announced large‑scale layoffs, unveiled its new ERNIE 5.0 model, and watched the stock first slump almost 10% in Hong Kong—then rip higher after Wall Street upgrades and fresh AI chip news. ^[1]

At the same time, analysts have been busily lifting price targets, and Baidu’s in‑house AI chip unit, Kunlunxin, is now preparing a Hong Kong IPO at a near‑$3 billion valuation. ^[2]

Here’s a deep dive into where Baidu stock stands today, what changed on and around December 6, 2025, and how current forecasts frame the risk–reward profile for BIDU over the next few years.

BIDU stock today: price, performance and the post‑upgrade rally

As of the U.S. close on December 5, 2025, Baidu’s ADRs traded at $125.66, up 5.85% on the day and essentially flat in after‑hours trading. ^[3]

Key trading stats:

Current price: $125.66
Previous close: $118.72
Intraday range: roughly $122.6 – $126.9
52‑week range: about $74.7 – $149.5 ^[4]
Year‑to‑date: GuruFocus estimates a >40% gain for 2025 so far, even after the ERNIE 5.0 sell‑off. ^[5]

The latest leg higher has been driven by a cluster of positive catalysts:

JPMorgan upgrade: JPMorgan upgraded Baidu to Overweight and hiked its price target from $110 to $188, helping send the shares up more than 7% on the day. ^[6]
Other banks joining in: Bank of America, Benchmark and others also raised targets (e.g., $151, $158), while Barclays lifted its target to $100 from $81, even while staying cautious. ^[7]
Renewed AI enthusiasm: The upgrade wave comes against a backdrop of 50%+ growth in Baidu’s AI‑related revenues and intensifying focus on chips, cloud, and autonomous driving. ^[8]

Short version: despite awful headlines about layoffs and losses, the market is now pricing Baidu as a recovering AI platform rather than a fading search‑ad dinosaur.

Q3 2025 earnings: worst sales fall on record, but AI business accelerates

Baidu’s latest results are the starting point for any honest BIDU discussion.

Headline numbers

For Q3 2025, Baidu reported:

Total revenue:RMB 31.17–31.2 billion (~$4.37–4.38B), down 7% year‑on‑year, its steepest revenue drop to date. ^[9]
Online advertising revenue:RMB 15.3 billion, down 18% YoY as China’s shaky macro backdrop and competitive pressure hammered ad budgets. ^[10]
AI cloud revenue:RMB 6.2 billion, up 21% YoY, a rare bright spot. ^[11]
AI‑powered businesses overall: first time disclosed at nearly RMB 10 billion, with >50% YoY growth. ^[12]
Adjusted net profit:RMB 3.77 billion, down ~36% YoY. ^[13]
Bottom line:net loss of about RMB 11.2 billion, mainly from a large asset write‑down. ^[14]

One more important datapoint: Baidu says it has invested over RMB 100 billion into AI since launching ERNIE in early 2023. ^[15]

In other words, profitability is getting squeezed from both sides: a structurally weakening ad business on one side, and very heavy AI investment on the other.

Layoffs: painful cost cutting with AI and cloud ring‑fenced

Toward the end of November, multiple reports confirmed that Baidu has started large‑scale layoffs that could continue through year‑end: ^[16]

Cuts span several business units, with some teams reportedly seeing headcount reductions as high as 40%. ^[17]
The mobile ecosystem group—closely tied to the traditional advertising and content business—appears to be bearing the brunt. ^[18]
Roles tied to AI and cloud computing are largely protected; Baidu is explicitly reallocating resources toward its AI infrastructure bets. ^[19]

Analysts read this as a classic “transition squeeze”: Baidu is still highly dependent on a declining ad model, but must continue plowing billions into AI, cloud and chips to remain relevant. The layoffs don’t necessarily mean the AI story is failing—but they are a clear signal that the old profit pool is shrinking faster than hoped.

Kunlunxin IPO: Baidu’s AI chip arm steps into the spotlight

The freshest hard news for December 5–6 is Reuters’ scoop that Kunlunxin, Baidu’s AI chip subsidiary, is preparing a Hong Kong IPO. ^[20]

Key details:

Kunlunxin was recently valued at ≈RMB 21 billion (about $3 billion) after raising over RMB 2 billion from a group of investors including a China Mobile‑backed fund. ^[21]
The unit expects 2025 revenue above RMB 3.5 billion, and more than half of that is projected to come from external customers, not just Baidu’s own workloads. ^[22]
Its most advanced chip, the P800, is already being deployed in state‑backed data‑center projects, positioning Kunlunxin as a domestic alternative to Nvidia for AI data‑center compute in China. ^[23]
Two next‑generation chips are lined up:
- M100 (inference‑optimized) targeting early 2026,
- M300 (training + inference) targeting early 2027. ^[24]
The IPO itself is tentatively targeted for 2027, with a Hong Kong listing application potentially filed as early as Q1 2026. ^[25]

This matters for BIDU shareholders because:

Vertical integration: Using homegrown chips reduces dependence on constrained Nvidia hardware under U.S. export controls and can lower long‑term AI inference/training costs. ^[26]
Sum‑of‑the‑parts upside: A separately valued Kunlunxin creates more transparent value for Baidu’s AI stack, much as Nvidia’s dominance props up its own valuation.
Execution risk: Custom silicon is capital‑intensive and can go out of date fast. If AI model architectures evolve rapidly, Baidu could find itself locked into expensive hardware that no longer fits the workload mix. ^[27]

Simply Wall St’s narrative on December 5 explicitly frames Baidu’s in‑house chip push as central to the long‑term thesis, while cautioning that profits and free cash flow are under pressure until these bets start paying off. ^[28]

ERNIE 5.0: technically impressive, market‑reaction brutal

At Baidu World 2025, the company unveiled ERNIE 5.0, calling it a “natively omni‑modal” foundation model that jointly handles text, images, audio and video, and powers a new wave of AI applications for consumers and enterprises. ^[29]

From a product and benchmark angle, ERNIE 5.0 looks competitive. But the market had seen this movie before and demanded more than another big model launch:

Hong Kong‑listed Baidu shares fell as much as 9.8% on November 14 after the ERNIE 5.0 event, their biggest drop in about seven months. ^[30]
GuruFocus notes that the stock still shows >40% YTD gains, but the reaction signaled investors weren’t wowed by incremental improvements. ^[31]
Commentary from AI observers like Implicator.ai emphasized that China’s AI market has shifted into a brutal price war, where technical excellence alone isn’t enough—cost structure, distribution and ecosystem lock‑in matter more. ^[32]

Meanwhile, Baidu has rolled ERNIE‑powered answers out aggressively:

eMarketer estimates that about 64% of Baidu’s mobile search result pages now surface AI‑generated answers, and roughly 70% of all Baidu results include some AI response. ^[33]

That’s great for user experience, but the monetization lag is real:

Baidu’s digital ad business—which still accounts for over half of total revenue—fell 18% YoY in Q3 and has been shrinking by 15–18% for multiple quarters.
eMarketer forecasts Baidu’s search ad revenues to fall 5.2% in 2025 and another 4.7% in 2026, even as peers like Tencent, PDD and ByteDance post 20–30% ad growth. ^[34]

In plain language: Baidu is rewriting its core product around AI faster than it has figured out how to make money from it.

Robotaxis and global expansion: Apollo Go as the long‑dated call option

If ERNIE is the software bet and Kunlunxin the silicon bet, Apollo Go is Baidu’s moonshot on wheels.

Recent data points:

In Q3 2025, Apollo Go’s robotaxi service (Luobo Kuaipao in China) delivered 3.1 million rides, up 212% YoY. ^[35]
By November, Baidu reported over 17 million cumulative robotaxi rides globally. ^[36]
Earlier in 2025, Apollo Go expanded to 16+ cities, including pilots in the UAE and new European deployments via partnerships (e.g., PostBus in Switzerland). ^[37]

Regulators are slowly catching up:

Baidu has secured permits for fully driverless operations in several Chinese cities and received approvals to run autonomous robotaxi services in Abu Dhabi. ^[38]

Financially, robotaxis are still a rounding error compared with search and cloud. Strategically, though, they matter a lot: they turn Baidu’s AI stack (models + chips + cloud + maps) into a vertically integrated transportation platform. That’s the kind of story Wall Street likes to project 10 years of optionality onto—provided the company survives the middle part.

What Wall Street is saying now: price targets and ratings

Analyst sentiment on BIDU has shifted noticeably more bullish since the Q3 print and subsequent chip/AI newsflow.

Consensus ratings and 12‑month price targets

Different aggregators paint a broadly similar picture:

StockAnalysis.com:
- 13 covering analysts
- Consensus rating: Strong Buy
- Average target:$144.23 (about 15% upside)
- Target range: $95–$188 ^[39]
MarketBeat:
- 24 analysts over the last 12 months
- Consensus rating: Moderate Buy (17 Buy / 6 Hold / 2 Strong Buy / 1 Sell)
- Average target:$145.95 (≈16% upside from $125.66)
- Target range: $92–$188 ^[40]
Zacks (snippet data):
- 19 analysts
- Average target: around $153.02, with a low near $100 and high near $190, implying somewhat more upside than the other consensus sets. ^[41]
GuruFocus aggregated target: about $140–141, with a cautiously optimistic tilt and a recommendation score around 2.1 on a scale where 1 is Sell and 3 is Buy. ^[42]

Underneath the headline targets, you see the individual calls that moved the stock this week:

JPMorgan: Hold → Buy, target $110 → $188
BofA Securities: Strong Buy, target $100 → $151
Benchmark: Strong Buy, target $115 → $158
Morgan Stanley: Hold, trimming target $140 → $130, signaling valuation concern despite recognizing AI upside. ^[43]

Long‑term growth expectations

On earnings and revenue, the consensus is surprisingly modest:

StockAnalysis’ forecasts (in CNY) show:
- 2025 revenue:RMB 131.9B, slightly down 0.9% from 2024.
- 2026 revenue:RMB 138.5B, +5.0% YoY.
- 2025 EPS: down ~16.8%, followed by ~7.6% EPS growth in 2026. ^[44]

Simply Wall St’s forward model (for the over‑the‑counter BAIDF line) is more aggressive on the bottom line:

Earnings growth: ~31% per year
Revenue growth: ~5.5% per year
EPS growth: ~32.5% per year
ROE in 3 years: ~6.1% ^[45]

Another Simply Wall St narrative pegs Baidu’s fair value around $151–152 per ADR, implying roughly 20–28% upside from recent prices, assuming Baidu can deliver around 4% annual revenue growth and over CN¥22B in earnings by 2028 despite the current slump. ^[46]

Morningstar, meanwhile, assigns Baidu a 4‑star rating and a fair value estimate of about $146, describing the stock as “moderately undervalued” after Q3. ^[47]

So the pattern is clear:

Nobody thinks Baidu is about to triple overnight.
Most serious models cluster in the mid‑$140s to low‑$150s as a “normal” 12‑ to 36‑month fair value range, assuming AI businesses grow and ads merely stagnate instead of collapsing.

Structural risks: ads, competition and geopolitics

Baidu is not just an AI fairy tale. There are concrete, ugly risks investors have to keep in mind.

1. Structural ad decline

We’ve already seen:

Online ad revenue down 18% YoY in Q3. ^[48]
eMarketer expects search ad revenue declines in both 2025 and 2026. ^[49]

At the same time, Chinese advertisers are reallocating budget toward ByteDance (Douyin), Tencent, PDD and others, many of which are still growing ad revenue at 20–30% YoY. ^[50]

That suggests Baidu’s weakness is structural, not just macro.

2. AI competition and price wars

Alibaba’s Qwen and the DeepSeek models are widely reported by Chinese developers to be at least as strong as ERNIE on many tasks, while also benefiting from powerful distribution through ecommerce and social platforms. ^[51]
An ongoing AI price war has forced Baidu and peers to slash model pricing, compressing near‑term margins in cloud and API businesses. ^[52]

In other words, Baidu is fighting simultaneously on product quality, ecosystem reach, and raw pricing. That’s expensive.

3. Balance sheet stress and layoffs

Q3’s RMB 11B net loss, combined with elevated AI capex and major layoffs, highlights the risk that Baidu’s transition burns a lot of cash before AI contributions materially replace lost ad profits. ^[53]

The company still has a strong net cash position, but investors need to watch free cash flow as closely as headline AI announcements.

4. Regulatory and geopolitical overhang

Reuters reports that the U.S. Pentagon has moved to add Baidu to a list of Chinese companies deemed to support the country’s military, alongside Alibaba and BYD. ^[54]

This doesn’t automatically translate into sanctions or delisting, but it:

Raises the risk of future restrictions on U.S. institutional ownership or government contracts.
Reinforces the broader U.S.–China tech decoupling context that already constrains Baidu’s access to high‑end Nvidia chips, making the Kunlunxin project not just strategic but existential.

For global investors, these are not abstract issues—they are live re‑rating risks.

So… is Baidu stock a buy after the December rally?

Nothing here is investment advice, but we can map the bull and bear cases as they look today, after the JPMorgan upgrade, Kunlunxin IPO news and Q3 fallout.

Bull case in one breath

Core AI metrics are strong: AI revenue up 50%+, AI cloud up 21%, robotaxi rides up 212% YoY, and ERNIE‑powered search deeply integrated into the core app. ^[55]
Vertical stack is taking shape: ERNIE (models) + Qianfan (platform) + Kunlunxin (chips) + Apollo Go (robotaxis) form a uniquely integrated Chinese AI ecosystem that’s hard to replicate. ^[56]
Valuation still reasonable: Most fair‑value estimates cluster 15–25% above current prices, with consensus targets in the mid‑$140s and several fundamental models (Morningstar, Simply Wall St) in the $146–$152 zone. ^[57]
Wall Street stance: The analyst mix has shifted from cautious to clearly positive—JPMorgan’s $188 target is aggressive, but it’s backed by a wider upgrade wave across major houses. ^[58]

For believers, Baidu is a rare big‑cap AI name that still trades at a mid‑teens earnings multiple on normalized estimates rather than a nosebleed SaaS valuation. ^[59]

Bear case in one breath

Old engine is stalling: Search‑anchored ads are shrinking at double‑digit rates, and leading industry analysts expect multiple years of revenue contraction in that segment. ^[60]
New engine isn’t fully built: AI cloud and chips are fast‑growing, but they’re capital‑intensive, highly competitive, and subject to price wars that could cap margins for years. ^[61]
Execution strain is visible: Record revenue decline, multi‑billion‑yuan write‑downs, and large layoffs across core operations are not what a smooth AI transition looks like. ^[62]
Policy and geopolitical risk: Being on the Pentagon’s radar, operating in a heavily regulated Chinese internet sector, and relying on domestic chips in a global arms race are all non‑trivial risk factors that don’t show up neatly in DCFs. ^[63]

If you think the ad decline keeps accelerating or that China’s AI landscape consolidates around other players, Baidu can easily become a value trap rather than a comeback story.

Key takeaways for December 6, 2025

BIDU has bounced hard to around $126, largely on the back of JPMorgan’s upgrade to $188, broader target hikes and growing AI optimism. ^[64]
Fundamentally, Q3 was rough: record revenue decline, an 18% collapse in ad sales, and an 11‑billion‑yuan net loss driven by write‑downs—even as AI businesses grow quickly. ^[65]
Kunlunxin’s planned Hong Kong IPO formalizes Baidu’s bet on becoming a major domestic AI chip supplier, potentially unlocking valuation upside but also increasing execution and capital‑allocation risk. ^[66]
Consensus views BIDU as undervalued but not a screaming bargain—targets in the mid‑$140s to low‑$150s assume that AI and cloud growth gradually outweigh ad erosion and that margins stabilize. ^[67]
The central question for 2026–2028 is simple: can Baidu turn its formidable AI stack—ERNIE, cloud, chips, and robotaxis—into cash flows large enough and fast enough to offset a structurally declining ad business in a highly competitive, politically noisy environment?

The stock is now positioned as a high‑beta, China‑specific AI turnaround: potentially rewarding if the execution goes right, but absolutely not low‑risk.