PLTR overnight +29%, a 180x PE gamble

NEXT PICK · Market Insights

PLTR overnight +29%, a 180x PE gamble

Tuesday, August 4, 2026

At the close on August 4, PLTR was at $162.66, up 29.45% in a single day, with trading volume expanding to 4.2 times the daily average over the past three months—a company with a market cap of $389.95 billion has grown to nearly $89 billion in just one trading day.

The trigger was undoubtedly the post-market earnings report on August 3: quarterly revenue of $1.935 billion, up 93% year-on-year; U.S. commercial business revenue $764 million, up 149% year-over-year; net retention rate raised to 157%; and full-year revenue guidance sharply raised from $7.65 billion to $8.15 billion.

But what truly makes people uneasy is another set of numbers

This stock reached $207.52 last December, and this year has dropped all the way to around $116, with a maximum drawdown approaching 44%. Even including today's nearly 30% bullish candlestick, its returns over the past 12 months have been almost stagnant.

So, is today's bullish candlestick because the fundamentals have finally caught up with valuations, or has valuations once again surged ahead of fundamentals?

Latest price $162.66 ▲ +29.45%

Q2 Revenue $1.935 billion ▲ Year-on-year +93%

Implied market-to-sales ratio 47.8 times ▲ Based on the annual guidance caliber

It doesn't sell software; it reinstalls operating systems for governments and enterprises

Palantir was founded in 2003, initially serving the U.S. intelligence system, focusing on data linkage in counterterrorism investigations; More than twenty years later, it has expanded this capability to defense, federal civilian agencies, and large corporations.

Today, it is supported by four product lines

Gotham focuses on defense and intelligence, integrating multi-domain sensor data into a single operational decision-making chain; Foundry is the central operating system for enterprise data; Apollo is responsible for continuous software delivery and updates in virtually any environment; AIP is the true engine behind this round of growth.

AIP's positioning is crucial—it integrates open-source, self-hosted, and commercial large models, translating structured and unstructured data into objects that large models can understand, and then encapsulates organizational actions and processes into tools that both humans and AI agents can invok.

In other words, it doesn't compete with OpenAI or Anthropic for model development; it competes for orchestration rights above the model and beneath the business layer. Whoever controls this layer controls the gate where AI truly turns into money.

So why is this position so hard to replace? Just look at the contract names: DHS's investigative case management and analysis platform, USDA's SNAP secure network Starlight platform, HHS's DPMO and Aurora systems—these are not generic tool procurement, but customized infrastructure bonded to the agency's mission.

Replacing it means rebuilding the entire data pipeline, rewriting business logic, redoing security authentication, and bearing the risk of business interruption. Adding to this the high-level security qualifications required to serve intelligence systems—acquisition cycles measured in years—the moat is almost physical on the government side.

But in the commercial market, this wall is much thinner. Snowflake, Databricks, and Microsoft Fabric are all providing data platform capabilities, but PLTR's brand recognition still lags behind pure commercial customers. The irreplaceability of 9 points on the government side is discounted on the commercial side.

A 93% growth rate is real, and a price-to-sales ratio of 47.8 times is also real

First, settle the accounts. Based on a market value of $389.95 billion and a median full-year revenue guidance of $8.155 billion, the implied price-to-sales ratio is as high as 47.8 times. For reference, the historical median price-to-sales ratio for high-quality enterprise software companies typically falls between 10 and 20 times.

The price-to-earnings ratio is also not cheap

180.7 times, corresponding to earnings per share of $0.90 over the past twelve months. This reading comes after the company just delivered a quarterly net profit of $1.06 billion, diluted $0.41 per share, and exceeded Wall Street's $0.35 expectation.

The good news is that cash flow is no longer a black box. The company has set full-year adjusted free cash flow guidance at $4.5 billion to $4.7 billion and adjusted operating profit guidance at $4.889 billion to $4.897 billion—single-quarter GAAP operating profit of $912 million, margin 47%, adjusted operating profit of $1.19 billion, margin 62%.

In other words, this is not a company that trades losses for growth; its operating leverage has already been fully unleashed. But if you put $4.6 billion in free cash flow to a market value of $389.95 billion, the yield is just over 1.18%.

So why does the market offer a price-to-sales ratio of 47.8 times? The answer lies in a harsh arithmetic: even if revenue growth remains high at 40% and 30% respectively over the next two years, 2028 revenue will be around $14.9 billion. A 25x price-to-sales ratio corresponds to a market value of $372 billion—slightly lower than today.

If growth drops to 30% and 20%, 2028 revenue will be about $12.7 billion, and at a 15-times price-to-sales ratio, the implied market value would be $191 billion, with a downside of over 50%. Conversely, if growth rates remain at 50% and 40%, with 2028 revenue of 17.2 billion yuan, and a market value of 602 billion yuan at 35 times, the upside would only be just over 50%.

Given three scenarios weighted with probabilities of 30%, 40%, and 30%, the implied market value is about $383 billion—almost matching today's $389.95 billion. This result itself is the answer: Today's price has fully priced the baseline scenario.

What's even worse is the asymmetry

The downside for the pessimistic scenario exceeds 50%, the upside for the optimistic scenario is less than 55%, and the pessimistic scenario only needs to "drop from excellent to good" for growth to materialize.

This is a breakthrough, but what is breaking is the valuation ceiling or the fundamental floor

From the perspective of the industry cycle, PLTR is clearly in the breakout phase. The judgment is not narrative but several hard jumps: the full-year guidance was revised up by about $500 million in a single quarter, nearly $430 million above Bloomberg's consensus estimate of $7.73 billion, with the company itself admitting this was the largest guidance increase in its history.

Remaining performance obligations reached $4.9 billion, up 103% year-on-year; In a single quarter, 220 orders over one million USD were signed, with 73 of them exceeding ten million USD, both setting records; The total number of customers increased to 1,049, with 653 U.S. business customers, a 35% year-on-year increase.

Demand, order, and retention are all working together—this is a step forward, not a gradual progress. CEO Alex Karp summed it up to one thing—the demand for sovereign AI was fully unleashed, and customers wanted to control the value of data, models, and AI outputs themselves.

But the technical position tells a different story. The 50-day moving average is $130.72, the 200-day moving average is $152.89, with the short-term moving average still below the long-term moving average—this is the scar left by the 44% pullback over the past six months. Today's gap-up bullish candle only pushed the price back above both moving averages.

Within a day, prices jumped from below the trend to above, but the trend structure itself has not yet recovered. Whether the market accepts this breakout depends on whether the price can hold above the 200-day moving average in the coming weeks, not how long today's candlestick lasts.

Its competitors are not on the same shelf

The competitive landscape of PLTR needs to be viewed in two layers, and the most interesting layer is where competitors don't even make software.

The three companies named are Concentrix (CNXC), TaskUs (TASK), and Genpact (G)—all in the business process outsourcing industry. They are not competitors to PLTR products, but potential disruptors to PLTR agent AI capabilities.

The logic is very straightforward

BPO's business model is built on labor arbitrage, relying on manual execution of multi-step, highly repetitive workflows. AIP can encapsulate organizational processes into tools usable by AI agents. Once successful, the long-duration customer service interactions CNXC-relied on, TASK-dependent manual content review, and G-dependent finance and procurement process outsourcing will all face structural business volume deflation.

Does this count as already priced in? It's hard to judge—the price-to-sales ratio, enterprise value multiples, and other horizontal comparable data for these three companies are missing in the current dataset, making it impossible to quantify how much the market prices this disruption risk.

The second layer is the conventional level of competition

Snowflake、Databricks、Microsoft Fabric、Google Cloud Vertex AI。 PLTR differentiates itself in three ways—the qualification barrier for high-level government security certifications, sovereign positioning for running Nemotron open-source models in a sovereign environment in collaboration with NVIDIA, and end-to-end workflow orchestration from data to execution.

The downside is equally clear

Valuations are higher than all peers, with margins of safety cut to a very thin margin; Brand recognition in the commercial market is still not as strong as those few pure commercial players; As a pure software company, it does not occupy any scarce resource nodes in this hardware-constrained AI investment chain.

The ceiling isn't low, but it already stands halfway up the mountain

The government should first calculate first. The annual budget for federal IT in the U.S. is in the $100 to $120 billion range, with about 15% to 20% of spending on data analytics, AI, and cybersecurity, or $15 to $24 billion.

PLTR single-quarter US government revenue was $809 million, up 90% year-on-year. Including state and local governments as well as international government markets such as the UK and NATO allies, the total addressable market on the government side could expand to $30 billion to $40 billion; However, due to security certification and procurement threshold constraints, its truly competitive portion is about $15 to $20 billion, and its current penetration rate has reached 25 to 30 percent.

The space on the commercial side is where imagination lies. The global enterprise data, analytics, and AI platform market has addressable sizes ranging from $80 billion to $120 billion, while PLTR's full-year U.S. business revenue guidance is only $3.424 billion, with penetration rates in the single digits.

If AIP-driven penetration is realized within three years, annual government revenue could reach 7 to 9 billion RMB, commercial revenue 5 to 8 billion RMB, totaling 12 billion to 17 billion USD. Counting back from the $17 billion ceiling to today's $389.95 billion market cap, the implied price-to-sales ratio three years from now would still be around 23 times—more than half of what it is now, yet still sitting at the upper end of the high-end software valuation range.

It should be noted that the above federal IT budget size and penetration rate are based on industry common sense estimation and are not hard data that can be checked item by item.

What the bulls held was a report card that was almost flawless

This season's bullish logic doesn't need to rely on stories; almost every one can be found in the financial report with its corresponding figure:

▲ Bull Case

Bears don't deny earnings; what they deny is price

Almost no one who is bearish on this stock is questioning the 93% growth rate; they are questioning the price paid for that growth:

▼ Bear Case

In the next three quarters, every financial report will be a valuation hearing

At this price level, PLTR has no room left for "positive news to be realized," only the rhythm of "verification or disproven" remains:

•	Next quarterly report (Q3 earnings window): Whether the 149% growth rate of U.S. business can be sustained is the single most critical variable for the year
•	FY2026 full-year deliveries: $8.15 billion to $8.158 billion in revenue, $4.5 billion to $4.7 billion in adjusted free cash flow—any failure would trigger a repricing reset
•	AIP continues to see a surge in volume in federal civil agencies: HHS contracts have explicitly procured platform-as-a-service solutions embedded with agent AI capabilities, serving as a traceable implementation sample
•	Advancing sovereign AI collaboration with NVIDIA: The commercialization progress of Nemotron models running in sovereign environments determines how quickly international government markets open
•	Window for realizing BPO alternative logic (FY2027 direction): If agent AI achieves large-scale production deployment within two years, it will open up additional incremental markets
•	Wall Street's Continued Reset of Target Prices: The long-short target price range ranges from $156 to $245, with divergence itself a source of volatility in the coming weeks

From 207 to 116 and then to 162, completing one round trip a year

PLTR was directly listed in September 2020 at a reference price of $7.25, surged above $35 in 2021 amid the growth stock boom, and was slashed back to $6 to $8 during the 2022 rate upward cycle—even lower than the reference price.

In 2023, the generative AI wave surged, with the launch of AIP and the practical deployment of the Maven project, giving the market a concrete AI monetization narrative and triggering a new round of stock price upwards; From 2024 to 2025, the number of commercial clients accelerated as large government contracts were secured, pushing the stock price all the way to $207.52 in December 2025.

Then there is this year's textbook valuation compression 。 By mid-June, the yearly decline had reached 26%, and in early July, it briefly reached around $116, pulling back from its peak to nearly 44%. The key point is that the company has not failed in any of these declines—Q1 revenue was $1.633 billion, up 84.7% year-on-year, U.S. business grew 133%, and adjusted earnings per share were $0.33, beating expectations by 18%, marking the eighth consecutive time it has exceeded expectations.

What has fallen has never been earnings, but the market is willing to pay multiples for results. This is the most important background for understanding today's bullish candle: when valuations are first halved and then a record-breaking financial report is released, elasticity is naturally amplified to 29.45%.

The lessons left by history are equally clear

A sharp correction at extreme valuation levels is normal, not surprising. When it dropped from $207 that year, the company's fundamentals were equally impeccable.

⚠️ Risk Notice • Extreme valuation stretch: 47.8 times implied price-to-sales ratio; any slowdown in growth could trigger double-digit drawdowns • Single-day +29.45%, a surge in volume of 4.2 times, severe short-term overbought trading, concentrated profit-taking pressure • The weighted implied market capitalization of the three scenarios is now below the current market value, with the risk-reward ratio downward asymmetric • The 50-day moving average remains below the 200-day moving average, indicating that the medium-term trend structure has yet to recover • Management's communication style is unpredictable, and under extreme valuations, any remarks are amplified by the market

🟡 Neutral The growth was impeccable, but a price-to-sales ratio of 47.8 times already ended the victory three years ahead of schedule

💬 Discussion Don't chase at the peak of your emotions. Real-time buy and sell positions, post PLTR to the official account for viewing

Data source

•	Source: NextPick real-time snapshot + Palantir official earnings report and SEC announcement + SAM.gov contract records + mainstream financial media coverage, with key figures cross-checked twice.