solarity
10 hours ago
Recent comment from Palantir CTO Shyam Sankar
"We have a decade-long headstart and that differentiation starts with the Ontology ... to drive AIP across these applications. When you look at the legacy software companies, I'm not sure they understand it yet." "If you look at the models, you see that they're getting better, which is awesome but they're also getting more similar across both closed and open source models. While they're improving, they're converging upon each other. All while the price of inference is dropping precipitously. So if you even look at these model companies, they have to build applications around these models to extract value. That's where we have a decade-long head start.
We've been building the the forge to create and implement AI applications at scale throughout the Enterprise and that differentiation starts with the Ontology, you know, using the Ontology to drive AIP across these applications. When you look at the Legacy software companies, I'm not sure they understand it yet. But when you look at the Innovative Silicon Valley companies they recognize the wall of tech Investments this implies that's in front of them that's going to act like a great filter."
DiscoverGold
17 hours ago
Palantir Technologies (PLTR) - Be Careful With These Topping Candles
By: Tom Bowley | February 2, 2025
"When do I sell?" is easily the most-asked question I've received over the years. There are multiple answers to this question based on certain variables. The first key variable is whether you're a day trader, short-term swing trader, or long-term buy and holder. I prefer swing trading, so my answer many times is when corroborating technical evidence tells me to sell.
My easiest sell is after a failed attempt at a breakout or a major reversing candle on heavy volume. The first one is fairly easy to see. Let's use Palantir Technologies, Inc. (PLTR) as an example from Friday. I don't know if PLTR is going up on Monday or later this week, but what I do know is it broke out to an intraday all-time high on Friday, then failed to hold that breakout on a closing basis. Check this out:
First, let me say that PLTR has a very strong chart. The AD line is continuously rising, a bullish cup has formed, and PLTR is a leader amongst software stocks ($DJUSSW). Second, I'm not saying PLTR is a short candidate. I'm simply saying it would be a sell for me short-term to take profits. I rarely short during secular bull markets. If it does make the breakout, I can always decide to jump back in. But I'd be looking for PLTR to pull back to form a handle off the cup pattern, or possibly even pull back to the recent low near 65. Sideways consolidation is a very real possibility after an extended advance like the one PLTR has enjoyed. If you need further evidence, look no further than NVIDIA Corp (NVDA) after its June/November/January tops. It's still consolidating.
The two red arrows mark what "could" be a double top, resulting in a lengthy period of selling and/or consolidation. Taking profits now is a risk-management strategy, eliminating the possibility of riding PLTR back to the downside. If income taxes is a concern and you're a long-term investor, I see nothing to suggest PLTR is a sell here. I'm only discussing my preferred short-term swing trading strategy.
Read Full Story »»»
DiscoverGold
DiscoverGold
4 days ago
This Wall Street bank sees a 65% downside risk in Palantir stock
By: Investing.com | January 30, 2025
Jefferies reiterated an Underperform rating on Palantir (NASDAQ:PLTR) stock, with its $28 price target implying a 65% downside risk from the current levels.
Jefferies analysts pointed out the challenges Palantir may face in the fourth quarter of 2024, as the company is expected to lap more difficult comparisons from the previous year. They highlighted that Palantir stock is the most expensive within the software sector, trading at 50 times next twelve months (NTM) revenue, which is more than double the valuation of its next highest peer.
“The 4Q setup will be challenging as PLTR is lapping easy comps and any signs of non-accelerating growth could lead to further multiple compression,” analysts led by Brent Thill said in a note.
The analysts also noted Palantir's accelerated approximately $120 million in unrecognized stock-based compensation (SBC) expense associated with market-vesting stock appreciation rights (SARs) due to the stock price crossing $50 per share.
As a result, Jefferies trimmed its fourth-quarter 2024 GAAP earnings per share (EPS) estimate to $0.01 from $0.06, with no change to non-GAAP EPS. Additionally, Jefferies raised the share count to 2.488 billion from 2.474 billion to reflect the dilution from these SARs.
While Palantir's fundamentals remain robust, the company would need to accelerate its growth to 50% for four years and trade at 13.5 times calendar year 2028 estimated revenue just to maintain its current stock price.
Palantir's enterprise value (EV) to NTM revenue multiple has compressed by 5% year-to-date, following a 282% expansion in 2024. “The last time we saw such high magnitudes of multiple expansion was during the Covid bubble when many of the high growth names benefited from multiple expansion,” analysts said.
They believe that any negative factors, such as decelerating growth or changing interest rates, could cause further compression in the company's valuation multiple.
In a “what if” scenario, Jefferies illustrated that even if Palantir could accelerate its growth to a 50% four-year compound annual growth rate (CAGR), the stock would still need to trade at a multiple that would place it among the richest names in software four years out, just to keep its current price level.
Read Full Story »»»
DiscoverGold
HyphenDubenator
1 week ago
Staying Focused...
Question:
If China's new Deep Seek AI LLM is real with its superior performance and low cost will Palantir's business model make Palantir less competitively impacted than the big US AI HyperScalers such Open AI, Google, AWS, IBM, Oracle, MicroSoft, & Tesla? Could this in fact be a positive development for Palantir?
Answer: Chat GPT4
The emergence of a powerful and cost-efficient Chinese LLM like "Deep Seek AI" could indeed have an interesting impact on the AI and big data landscape, particularly for companies like Palantir that focus on enterprise software and AI for operational decision-making. Let’s break this down from a competitive perspective:
Palantir vs. Hyperscalers
The core distinction between Palantir and hyperscalers (OpenAI, Google, AWS, etc.) lies in their business models:
Hyperscalers: Focus on broad AI tools (e.g., APIs, cloud services, general-purpose models) to enable companies to build their own solutions. They operate in a commoditized space, competing on scale, cost, and capabilities of AI and infrastructure.
Palantir: Specializes in end-to-end solutions for enterprise clients, particularly in sensitive domains like defense, government intelligence, healthcare, and large enterprises. Palantir offers highly tailored solutions built on its proprietary platforms, Foundry, Gotham, and Apollo, which go beyond just LLMs and focus on integration, analysis, and operational decision-making.
Impact of Deep Seek AI
If "Deep Seek AI" outperforms existing hyperscaler models in terms of cost, capabilities, and deployment, its impact on Palantir might differ compared to the hyperscalers:
Palantir’s Differentiation:
Palantir’s value proposition isn’t about building the best LLM. Instead, it’s about:
Integrating AI models (regardless of their origin) into operational workflows.
Offering platforms that allow clients to derive actionable insights and implement those decisions effectively.
Specializing in mission-critical use cases like defense and national security, where trust and data sovereignty are key.
If "Deep Seek AI" is a better or cheaper model, Palantir could simply integrate it into its platforms. Palantir is model-agnostic and designed to incorporate and operationalize third-party AI models, meaning this could even lower Palantir's costs while improving its offerings.
Pressure on Hyperscalers:
Hyperscalers rely on proprietary AI models as a revenue driver, so a better and cheaper competitor could disrupt their pricing power and market share. Palantir, in contrast, is less reliant on developing its own LLMs and benefits from leveraging the best models available globally.
Geopolitical Factors:
Palantir's deep relationships with Western governments, especially in defense, could insulate it from competition with Chinese models like "Deep Seek AI." Many of Palantir's clients might be reluctant—or outright restricted—from adopting Chinese AI solutions due to security and data sovereignty concerns.
Could This Be Positive for Palantir?
Yes, it could be a net positive for several reasons:
Access to Better Models: If Deep Seek AI sets a new performance benchmark, it could drive global innovation and force hyperscalers to improve their offerings, which Palantir can integrate into its platforms.
Enhanced Value Proposition: Palantir thrives by integrating, orchestrating, and operationalizing AI for clients. Superior third-party models give it more powerful tools to work with, potentially increasing its own value to clients.
Focus on High-Touch Enterprise Solutions: Palantir doesn’t directly compete with hyperscalers for commodity AI services. Instead, it offers premium, tailored, and high-touch solutions that would be less affected by commoditization in the LLM space.
Strengthened Role as Integrator: A diverse AI ecosystem with multiple strong players (including "Deep Seek AI") could reinforce Palantir’s position as the go-to platform for integrating best-of-breed AI into operational workflows.
Challenges to Consider;
While the emergence of Deep Seek AI could be positive for Palantir in many ways, there are potential
Risks:
Competitive Price Pressure: If Chinese models drastically lower costs, hyperscalers might adopt aggressive pricing strategies, potentially reducing margins for companies like Palantir in non-government sectors.
Global Adoption Trends:
If Deep Seek AI gains traction globally, some enterprises might skip Palantir in favor of directly using advanced LLM platforms at a lower cost.
Security Concerns: Clients might question the security implications of integrating Chinese-developed models, especially in sensitive industries.
Conclusion:
Palantir’s business model positions it to be less impacted by the rise of a cost-efficient Chinese LLM than hyperscalers, as it focuses on operationalizing AI rather than commoditizing it. In fact, this could be a positive development for Palantir, as it may lower costs, increase competition among LLM providers, and allow Palantir to continue leveraging the best available technologies to enhance its platforms. However, geopolitical dynamics and market adoption trends will shape how this plays out in practice.